This document is in two parts, and the two parts are a single story. The first is the inventory of the celebratory season for the twenty-five years of Decree 231: the conferences, the price lists, the training credits, the certificates — the party that the market organised and that the State did not organise. The second is the analysis of the reform project that that season accompanied and applauded, and which on 4 August 2026, while this document was being finalised, became a government bill. The first part answers the question: who is celebrating what, and who pays and why. The second: what is in the regal parcel of the new clothes.
PART ONE — THE PARTY: THE CERTIFICATE OF ATTENDANCE
Inventory of the party and a reasoned guide to who collects and who pays.
I. The certificate
At the end of every training event on Legislative Decree 231 there arrives, by email, a personalised certificate of attendance. A sheet of paper. It certifies that a person, on a certain date, remained connected to a virtual classroom while other people discussed organisational models, that is, documents. It is paper attesting attendance at a discussion about paper, issued by an industry that lives on paper. In 2026 that sheet turns twenty-five, together with the decree that made it possible. This piece is the inventory of its birthday party.
The method is the usual one. Everything that follows belongs to one of three registers: established (events, dates, prices and documents verified from open sources, listed in the digest at the end), declared (what the industry says about itself, reported in quotation marks or as such), inferred (the conclusions, with the logical leap declared). Where the survey does not reach, it is written: no record found.
II. The frame: two dates and a committee
The dates, to begin with. Legislative Decree No. 231 was enacted on 8 June 2001 and entered into force on 4 July 2001. The twenty-fifth anniversary therefore falls twice, one month apart: a generosity of the calendar that the training industry has known how to appreciate, spreading the celebratory season over an entire academic year.
The engine of the party, however, is not the anniversary: it is the reform, that is, the gift parcel that required long and passionate labour. By decree of 7 February 2024 the Ministry of Justice established a Technical Committee (Tavolo tecnico) for the revision of the Decree, coordinated by the president of the Sixth Criminal Section of the Court of Cassation, Giorgio Fidelbo. Its work began on 12 June 2024 and concluded on 30 October 2025; the Committee heard in hearings Confindustria, Confcommercio, Assonime, the Association of members of supervisory bodies and the National 231 Observatory of the National Council of Chartered Accountants — that is to say, with a single partial exception, the representatives of the supply of, and the organised demand for, compliance. The proposal, twenty articles, was delivered to the Minister in December 2025 and made public in January 2026.
One fact must be noted: for the twenty-five years, the State organised no ceremony. There is no record of an institutional celebratory event by the Government or by Parliament. The State produced draft reform articles and a report; the party, all of it, was organised by the market. It is a division of labour that describes the subject better than many conferences.
III. The calendar of the party
The season opens early, like Christmas shop windows. Bologna, 17 October 2025: a conference on stocktaking and prospects of reform, hosted in the Oratorio dei Fiorentini by kind concession of a city bank, with the accompanying presentation of a collective volume on compliance published by Wolters Kluwer. The sequence — oratory, bank, publishing house — is already a small treatise on the sociology of the subject.
From January 2026, once the proposal was published, the machine runs at full capacity. Palermo, 20 February: the reform discussed in the lecture hall named after Borsellino, with the university, the professional bodies, Sicindustria, a scientific journal and an international auditing firm on the same organising committee. Milan, 5 March, Palazzo delle Stelline: first reflections on the draft, with the participation of the Committee’s coordinator. Milan again, 13 March: an entire day, from half past nine to half past six, on the evolving lines of the liability of entities. Turin, 18 March, hall 74 of the Palazzo di Giustizia: three training credits for lawyers, of which — this is the detail that deserves verbatim quotation — one in professional ethics. Listening to a conference on the reform of Decree 231 is worth, for the rules of the legal profession, one twenty-fifth of the annual professional ethics requirement.
April is the month of the local professional bodies. Lecce, 10 and 11 April: two full days on the analysis of twenty-five years of application. Padua, 21 April: international sanctions and the 231 Model. Milan, 23 April, Università Cattolica: “25 years of 231. Towards what reform?”. May raises the tempo: a web seminar on compliance as a “strategic lever” (8 May), the paid full day on artificial intelligence applied to the 231 system (13 May, we shall return to it at the price list), a real-estate conference on the reform moderated by an economic journalist (14 May), cybersecurity put to the digital test (19 May), and finally — 22 May — the major academic day: the Statale and Bocconi together, under a title that promises “new paradigms for the criminal law and procedure of entities”.
Then June, the month of the real birthday. On 4 June the University of Padua convenes its own study conference in the Archivio Antico of Palazzo del Bo, seven training credits, morning and afternoon sessions. The title deserves to be transcribed in full: “Gone with the Wind. Twenty-five years of the liability of entities for offences”. A quarter of a century of the punitive law of entities, and to give it a name the academy must borrow the title of an epic Hollywood blockbuster about the end of a world. On 8 June — the exact day of enactment — the dignity of labour and criminal intervention in the enterprise are discussed in Alghero. On 11 June ANAC opens its doors in Rome, free of charge and in streaming. On 16 June, in Milan, in the Congress Centre of the Fondazione Cariplo, the annual conference of the association of supervisory bodies: “The Supervisory Body between present and future”. In Teramo, the university gathers around the same table four members of the Technical Committee, coordinator included, with a Google Meet link and three credits from the local professional body: the reform explained by those who wrote it, on tour.
Because this is the constant trait of the season: the judge who coordinated the proposal is the most sought-after guest in Italy. From Milan to Teramo, the person who drafted the articles illustrates them to the audience that will have to apply them, sell them or teach them. There is nothing irregular in this, and much that is instructive: the material legislator as itinerant speaker is the form that Italian norm-production has chosen to give itself. The season does not close for the summer: the catalogue already includes an environmental master’s in streaming from late October to mid-December 2026, whose launch page celebrates the twenty-five years of the Decree and, in the same paragraph, refers the reader to the enrolment form.
IV. The price list
So far the free party, or almost: university conferences, professional bodies, associations. Then there is the paying party, and here the figures are public. The Milan day of 13 May on artificial intelligence applied to the 231 system — four hours of live streaming, from 9.30 to 13.30, three speakers from the same law firm — costs 650 euros plus VAT. With early booking, 520. One hundred and sixty-two euros and fifty cents an hour, VAT excluded, to attend remotely.
The brochure, however, is more interesting than the price. There is the “Young Excellence Project”: by enrolling a second staff member under thirty, a fifty per cent discount — the industry’s nursery must be cultivated in good time. There is the “Corporate Quota”: from the fifth enrolee of the same organisation, minus fifty again — crime prevention knows economies of scale. There is the item “Funded training”: the fee “may be fully reimbursed through vouchers promoted by the Joint Interprofessional Funds”, with “complete and free management” of the reporting borne by the organiser. Translated: the party is paid for with money the companies have already paid in, and the host offers to fill in the reimbursement claim himself. There is, finally, the certification that the event is “purchasable on the MePA”, the electronic marketplace of the public administration: even the State, which did not organise the ceremony, can buy a ticket to other people’s.
Around the conference circuit stands the heavy manufacturing: the master’s programmes. The specialisation course of Milan’s Statale on the liability of entities for offences costs 916 euros. A Turin executive master’s in compliance, entirely online from December 2025 to March 2026, reaches 2,440 euros including VAT — and that is the discounted price, payable in two instalments. A private Milanese school offers sixty-four hours with access to regional vouchers and ANPAL funds. A business school belonging to the main national financial daily notes that its master’s on the 231 Model is “activatable” with the interprofessional funds. A training body offers the “Master in 231” with a tariff of rising loyalty: enrolling two people gives access to the “Silver Formula”, three to the “Gold”, four or more to the “Platinum”. Crime prevention has a points programme. Another operator issues, with the same course, fourteen credits for lawyers and thirty-six hours of refresher training for safety officers: a single outlay, two training obligations discharged. And in Palermo the Decree has made the definitive species leap: there is a master’s degree course in “Compliance, business development and crime prevention”, with an attached laboratory on 231 models. The Decree no longer needs courses: it has itself become an academic qualification.
V. Paper upon paper
The anniversary is also celebrated in bookshops and on news-stands, where Decree 231 enjoys a privilege that no other Italian rule can claim: a journal all of its own. “La responsabilità amministrativa delle società e degli enti”, known to the market as Rivista 231, has come out every quarter since 2006 — twenty years of a periodical devoted to a single decree — and over time has equipped itself with a portal, a monthly circular for subscribers and videos “accredited for continuing professional training”: the subscription that incorporates the credits, reading that becomes a compliance duty. Issue no. 1 of 2026 duly came out in January, in time for the season.
On the books side, the Wolters Kluwer collective volume on compliance and the liability of entities, six academic editors, was presented in Bologna in October 2025, at the opening of the season. On the side of position papers, the choreography is complete: Assonime publishes in April 2025 a position paper with its own reform proposals; the Committee hears Assonime in a hearing; once the twenty articles are out, Assonime publishes in July 2026 a second position paper which judges the draft “broadly supportable in its essential lines” — a text, it declares, that “gathers the best settled holdings of the case law and practice that have emerged in twenty-five years in force”. The interlocutor heard in the hearing congratulating itself on the outcome of the hearing: the circle of the consultation closes by applauding itself. Confindustria has filed its own “prospects for reform”; the chartered accountants have a National 231 Observatory and have already produced, with their Foundation, the document that joins the 231 Model to the “ESG factors” — the coupling of the new wagon to the old locomotive.
Finally, the diffuse commentary, which is the connective tissue of the entire economy. Every stage of the reform — report, publication, position paper — is relayed within a few hours by online journals, update portals, sector newsletters and law-firm websites. The recurring format deserves a taxonomic mention: the article that diligently explains the reform proposal and ends, without a break, with the list of the 231 services of the firm that signs it — drafting of the model, training, appointments as supervisory body. Dissemination as shop window; the anniversary as a client-acquisition campaign. And the financial press does its part: in the spring an agency dispatch picked up by the Borsa Italiana website defines the 231 Model as an “indispensable tool”, lists its ancillary rewards — reduction of the INAIL premium, access to funds, legality rating, even attractiveness to “new talent” — and recalls, in the same text, Confindustria’s figure: only 36 per cent of small and medium-sized enterprises have adopted it. Promotional register and statistical register on the same page, without either noticing the other.
VI. Double-entry bookkeeping
Let us then try to write the two columns, keeping to the registers. In the receipts column stand, in order of appearance: the catalogue training bodies, with fees from 520 to 650 euros for half a day and from 900 to 2,400 for a master’s; the publishing houses, with the dedicated journal, the collective volumes, the subscription portals; the professionals of the ancillary economy — consultants who draft the models, trainers who explain them, members of supervisory bodies remunerated, according to the practice documented by the operators themselves, with a fixed annual fee on a typically three-year mandate; the law firms, for whom the celebratory season is at once professional refreshment and marketing; the business schools and the universities, from the specialisation course to the master’s degree; and the party’s logistics: congress centres, host foundations, streaming platforms.
In the outgoings column stands a single subject, with three disguises. The company pays out of its own pocket, when it pays the fee; it pays with money already its own, when the fee is “fully reimbursed” by the interprofessional funds fed by the contributions of the companies themselves; and it pays as a taxpayer, when the purchaser is a public administration on the MePA or when the training travels on public vouchers. In all three cases the flow has the same direction and only the name of the channel changes. The one who never appears in either column is the nominal beneficiary of the entire system: the prevention of crime, which issues no invoice and of which there is no record of its ever having issued a certificate of accomplished prevention.
How much is the party worth, in euros? Here the register must be declared bluntly: there is no record of any public survey of the overall turnover of the 231 industry — not of training, not of consultancy, not of supervisory bodies. It is a market that lives on other people’s transparency and does not publish its own. Any aggregate figure would therefore be an inference, and the most honest is also the simplest: let the reader take the established price list of a single morning in streaming, multiply it by a prudent number of connections, and then run through the digest at the end, counting the mornings. The logical leap lies entirely and only in that multiplication; the factors, at least, are documented.
There remains the underlying paradox, which no speaker has put on the programme: what is being celebrated as the keystone of corporate legality is a safeguard which, by the declaration of the main employers’ association, two small and medium-sized enterprises out of three do not have. The twenty-fifth-birthday party is the party of an organised minority — the one that writes the models, sells them, supervises them and teaches them — held in honour of a majority that did not turn up.
VII. The adversarial case
Fairness demands that the party be given its best lawyer. The strongest version of the defence goes like this: the Technical Committee (Tavolo tecnico) is serious work, conducted by front-rank judges, academics and lawyers; publicly debating a reform before it becomes law is exactly what a mature legal community must do; a good part of the calendar — the Statale and Bocconi, the Cattolica, Padua, Teramo, Palermo, the local bar associations, ANAC — is free of charge, scientifically grounded and open; and an anniversary is a legitimate occasion for taking stock, not a fault. All true. The distinction between scientific debate and the compliance industry is real, and anyone who erased it for the love of polemic would commit the very error this piece seeks to document in others: mixing up the registers.
The point at which the defence breaks down is a different one, and it is quantitative. In twenty-five years the field has produced a celebratory, training and publishing apparatus of dimensions that this register documents only in part; it has never produced, in parallel, a public measure of its own preventive effectiveness. Onto the conference platforms climb the drafters of the reform, the sellers of the models, the supervisors of the supervised; what never takes the stage is a verifiable figure stating how many offences the paper has prevented. And there is a second point, which must be marked as an inference: the reform itself, as its admirers describe it — simplification, rewarding mechanisms, adequacy, measures calibrated for small businesses — promises a new wave of updates to the models, and therefore new demand for consultancy and training. If the inference is correct, every reform of Decree 231 is also, objectively and with no need for intentions, the business plan of its own satellite industry. The breaking point of the counter-thesis is declared: the intention is not proven, and there is no need to prove it; the effect is enough.
VIII. Valediction
When the season is over — and the catalogue shows it will continue at least until December — what will remain of these twenty-five years are the conference proceedings, the position papers, the quarterly issues, the speakers’ slides and, above all, thousands of named certificates of attendance, duly sent out at the end of each training event to those who “duly connected to the virtual classroom”. It is the only document this industry produces on its own account and actually delivers: paper certifying that one has attended. The guest of honour, as at all previous editions, was not in the room. The prevention of offences does not register, does not log on, does not accrue credits. And it does not issue certificates of attendance.
Register of the celebratory season (open sources)
A service apparatus. Every entry has been verified against an open source as at 3 August 2026; the indication of the organiser is limited to the entities, with no names of individuals except institutional office-holders already mentioned in the text. The register is, by construction, incomplete: it records what the open sources return, not everything that happened.
7 FEBRUARY 2024. Decree of the Ministry of Justice: establishment of the Technical Committee for the revision of Legislative Decree 231/2001 within the Minister’s Private Office (Ufficio di Gabinetto). Work from 12 June 2024 to 30 October 2025; hearings of Confindustria, Confcommercio, Assonime, AODV231 and the CNDCEC’s National 231 Observatory.
17 OCTOBER 2025 — BOLOGNA. Conference “Legislative Decree 231/2001: between stocktaking and prospects for reform” (DPEI, under the patronage of the Franco Bricola Association), Oratorio dei Fiorentini, as guests of Banca di Bologna; presentation of the Wolters Kluwer 2024 collective volume on compliance and the liability of collective entities for criminal offences. In person only.
DECEMBER 2025. Delivery to the Minister of Justice of the reform proposal (twenty articles) drawn up by the Technical Committee.
12 DECEMBER 2025 — MILAN. Conference “Organisational fault and criminal fault, between theory and practice” (listed on the AODV231 events noticeboard).
14–22 JANUARY 2026. Publication of the Technical Committee’s final report and draft articles (Giurisprudenza Penale on the 14th, Sistema Penale on the 22nd); immediate relaunch in the main online criminal-law journals and on the professional-update portals.
JANUARY 2026. Release of issue no. 1/2026 of the Rivista 231 (“La responsabilità amministrativa delle società e degli enti”), a quarterly entirely devoted to the decree since 2006.
10 FEBRUARY 2026 — ROME. Presentation of the Corruption Perceptions Index (CPI) 2025 (on the AODV231 noticeboard among the ecosystem’s events).
20 FEBRUARY 2026 — PALERMO. “Reform of Decree 231 between liability, rewarding mechanisms and organisational models”, Aula Borsellino, University of Palermo — Compliance Laboratory of the master’s degree course in Compliance, business development and crime prevention, with Sicindustria, the COA and the ODCEC of Palermo, Sistema penale and PwC; full recording circulated online.
5 MARCH 2026 — MILAN. “The future of Legislative Decree 231/2001: first reflections on the draft reform”, Palazzo delle Stelline (AODV231), with a speech by the coordinator of the Technical Committee; summary published on the association’s website.
13 MARCH 2026. AODV231 study day “Lines of development in the liability of collective entities for criminal offences”, 9.30 am to 6.30 pm.
18 MARCH 2026 — TURIN. “The liability of entities for criminal offences: prospects for reform of Legislative Decree 231/01”, COA of Turin, room 74 of the Palazzo di Giustizia; in person and online; three training credits, one of them in professional ethics.
25 MARCH 2026 — MILAN. Conference on the reforms on waste (the “Terra dei fuochi” decree, the EPR regime for textiles), within the 231 orbit (AODV231 noticeboard).
10–11 APRIL 2026 — LECCE. Fifth “Vittorio Aymone” Conference of the Lecce Bar Association: “Legislative Decree 231/2001 and the liability of legal persons. An analysis of 25 years of application and prospects for reform”. Two days.
21 APRIL 2026 — PADUA. “International sanctions and corporate compliance: new responsibilities for businesses within the 231 Model”.
23 APRIL 2026 — MILAN. “25 years of Decree 231. Towards which reform?”, “Federico Stella” Advanced School on Criminal Justice, Università Cattolica del Sacro Cuore.
8 MAY 2026. Web seminar “Compliance and Risk Governance: from regulatory obligation to strategic lever for professionals and companies”.
13 MAY 2026 — STREAMING. “Artificial Intelligence applied to the 231 system” (Paradigma), 9.30 am to 1.30 pm; fee 650 euros + VAT, early booking 520; 50% reduction for a second registrant under 30 (“Progetto Giovani Eccellenze”) and from the fifth registrant onwards (“Quota Corporate”); full reimbursement possible through the Joint Interprofessional Funds (Fondi Paritetici Interprofessionali) with free handling of the paperwork; event purchasable on the MePA; named certificate of attendance; three speakers from the same firm.
14 MAY 2026 — MILAN. ASPESI conference (with Relatus) on the reform of Decree 231, Centro Culturale di Milano, with moderation entrusted to a financial journalist.
19 MAY 2026 — MILAN. “Secure communications, liability and cybersecurity: Legislative Decree 231/2001 put to the digital test”.
22 MAY 2026 — MILAN. “New paradigms for the criminal law and criminal procedure of entities. Reflections on the proposed reform of Legislative Decree 231/2001”, study day of the law departments of the Università Statale and the Università Bocconi; relaunched in the sector’s online journals.
4 JUNE 2026 — PADUA. “Gone with the Wind. Twenty-five years of the liability of entities for criminal offences”, Archivio Antico of Palazzo del Bo, University of Padua with the COA and the ODCEC of Padua; morning and afternoon sessions; seven training credits.
8 JUNE 2026 — ALGHERO. On the exact day of the twenty-fifth anniversary of the decree’s enactment: a conference on the dignity of work, the protection of rights in business activity and the prospects of criminal-law intervention (reported by the online legal press).
11 JUNE 2026 — ROME. ANAC conference (“first experiences compared”), attendance free of charge on request, live streaming.
16 JUNE 2026 — MILAN. AODV231 annual conference, Centro Congressi Fondazione Cariplo: “The Supervisory Body between present and future”, with streaming; subject: the role, composition and prerogatives of the Supervisory Body in the light of the proposed reform.
SPRING 2026 — TERAMO. Conference of the University of Teramo “25 years of Legislative Decree no. 231 of 2001: stocktaking and prospects for reform”, with four members of the Technical Committee (coordinator included); remote connection via Google Meet; three credits from the COA of Teramo. (Exact day: no record found in the sources collected.)
2 APRIL 2025 / JULY 2026 — ROME. Assonime: position paper with reform proposals (April 2025) and position paper no. 8/2026 commenting on the Committee’s draft, judged “broadly supportable”; relaunches on professional portals and online journals.
28 OCTOBER – 16 DECEMBER 2026 — STREAMING. HSE Master’s course (33rd edition) run by a private environmental-training operator, launched with a celebratory article for the twenty-five years of the decree and the fifteen of Article 25-undecies.
4 AUGUST 2026 — ROME. Council of Ministers no. 185: approval of the government bill on the “revision of the rules on the administrative liability of legal persons” (see Part Two, section 36). Entry added after the closing of the census, which remains referred to 3 August.
The price list (established prices)
€650 + VAT. Half a day in streaming (4 hours) on AI and the 231 system; early booking €520; −50% for a second under-30 resource; −50% from the fifth registrant; full reimbursement possible via the Joint Interprofessional Funds; purchasable on the MePA.
€916. Advanced training course of the Università Statale di Milano “Legislative Decree no. 231/2001: the liability of entities for criminal offences and corporate compliance”, academic year 2025/2026 (fee inclusive of stamp duty).
€2,440 VAT INCLUDED. Executive master’s course in compliance (Turin, online, December 2025 – March 2026), discounted price in two instalments; discounts for AODV members and multiple registrations.
AT LIST PRICE. “Modello 231” master’s course from the business school of the leading national financial publishing group, activatable with interprofessional funds; a private Milanese master’s course of 64 hours with ANPAL and regional vouchers; a “Master in 231” with Silver loyalty formulas (−20% for 2 registrants), Gold (−30% for 3), Platinum (−35% from 4), 20 CNF credits; a course with dual recognition: 14 legal training credits and 36 RSPP/ASPP hours; an SDA Bocconi course on compliance, 231 models and sustainability (−20% from the third registrant); an online master’s course from the Wolters Kluwer/IPSOA group; a master’s degree in Compliance in Palermo. Prices not always published: where the list price is not on record, it is not invented.
The library of the party
THE DEDICATED JOURNAL. “La responsabilità amministrativa delle società e degli enti” (Rivista 231): a quarterly entirely devoted to Legislative Decree 231/2001 since 2006; a portal, a monthly circular for subscribers, accredited videos for continuing professional development. There is no record, as at the date, of the announcement of a celebratory monographic issue for the twenty-fifth anniversary: this is noted as a census datum, not as a ruling-out.
THE COLLECTIVE VOLUME. “Compliance – responsabilità da reato degli enti collettivi”, Wolters Kluwer 2024, six academic editors; presented in Bologna on 17 October 2025 at the opening of the season.
THE POSITION PAPERS. Assonime (April 2025 and no. 8/2026); Confindustria, “Prospects for reform of the ‘administrative’ liability of entities”; CNDCEC–FNC, “The 231 Model and ESG factors”; Assonime, “The Supervisory Body in company practice twenty years on from Legislative Decree 231/2001”.
THE DIFFUSE COMMENTARY. Relaunches and comments in online criminal-law journals and on professional portals (including Sistema penale, Giurisprudenza penale, Diritto Bancario, Altalex, IPSOA Quotidiano, compliance portals); celebratory articles by local criminal bar chambers and training operators; contributions by law firms with a promotional tail on their 231 services; a launch picked up by the Borsa Italiana website (April 2026) on the 231 Model as an “indispensable instrument”, with the Confindustria figure of 36% adoption among SMEs.
Methodological note to Part One
Census conducted on open sources (institutional, university, association, publishing and organisers’ websites) as at 3 August 2026. Registers: the events, dates and prices reported are established on the sources indicated; the self-definitions (“strategic lever”, “indispensable instrument”, “broadly supportable”, “Giovani Eccellenze”, “Formula Platinum”) are declared and reported as such; the aggregate assessments of the turnover and of the reform’s satellite-industry effect are inferences, with the logical leap declared in the text. No record found of: a state institutional ceremony for the twenty-fifth anniversary; a public survey of the turnover of the 231 industry; a celebratory monographic issue of the sector journal (as at the census date). No unlawful conduct is attributed to anyone: everything described is lawful, public and publicised — that is precisely the point. The register can be updated: later versions will incorporate the autumn tail of the season.
PART TWO — THE GIFT: THE FIDELBO PROJECT, FROM THE COMMITTEE TO THE COUNCIL OF MINISTERS
Structure and critique of the reform, updated to the government bill of 4 August 2026.
Part One took the census of the party; this second part examines the gift. On 4 August 2026 — the census of the season had closed a day earlier — the Council of Ministers approved the government bill that is the declared offspring of the Fidelbo Committee's proposal: the reform that the industry surveyed in Part One awaited, discussed at conferences and, in the ways documented by the register, applauded. The pages that follow analyse it in the only full version readable today — the Committee's report and draft articles — and record in the final post scriptum what of the 4 August approval can already be verified. Party and gift must be read together: it is the only way to understand whose party it is, and to whom the gift is delivered.
The question
Before any detail, the question through which this text reads the project, because it is the blog's question: does a reform of the liability of entities for criminal offences serve to reduce corporate crime, with the right safeguards in the right place — or does it serve to shield the company still further, overestimating guarantism and the need for it in a system where the punitive power to be balanced is already close to zero? Guarantism is a set of scales: on one side it weighs a power that punishes, on the other the rights of those it may strike. When the pan of power is empty — and the figures on the application of Decree 231, documented in the first two parts of this investigation, say it is close to empty — adding weights to the other pan is not guarantism: it is something else, and at the end of this text I shall try to call it by its name. The answer, as will be seen, is not the easy one in either direction.
And one figure, before beginning, must be kept on the table: forty million euros as the minimum of the maximum penalty for 231 violations of a wholly particular kind. It entered Decree 231 on 24 January 2026, almost without witnesses, and the document in this series "What Lies Around the Corner for Decree 231?" is entirely devoted to it. This analysis closes there, because that figure is the measure against which every page that follows will have to be reweighed.
A note on method
This analysis is based on the final report of the Technical Committee (Tavolo tecnico) established at the Ministry of Justice for the revision of Legislative Decree No. 231 of 8 June 2001, coordinated by Giorgio Fidelbo, section president of the Court of Cassation, and on the proposed draft articles that accompany it. The document bears on its title page the date November 2025, was delivered to the Minister on 22 December 2025 and was made public on 14 January 2026. On 4 August 2026 the Council of Ministers approved a government bill revising the rules which is the declared offspring of that project; as at the date of this analysis the approved text has not been published, and the analysis therefore remains conducted on the Committee's report and draft articles — section 36 records the development, what in the two texts already appears to correspond according to official and press sources, and what will have to be verified when the government bill becomes readable.
The analysis is conducted on the full text of the document — final report and draft articles — in the version published by the journal Cassazione penale. Every quotation in quotation marks has been verified against the primary source; the few exceptions are flagged. A philological caveat: at several points the draft articles carry drafting options in square brackets — the deadline for exercising the delegated power, for example, is "[eight months]" —: a sign that some choices were left, deliberately, to the political decision-maker. For the genesis and the positions of the interested parties, qualified secondary sources are used, always indicated: Assonime's Position Paper 8/2026, Assolombarda's summary, the commentaries of the Osservatorio 231 of the Union of Criminal Chambers and of Impresa e Diritto.
I have taken account of the diagnosis documented in the two parts of "The Paper Gold of Decree 231" already published: this analysis presupposes it and uses it as a yardstick, because it is the only yardstick available — the reform project, as will be seen, offers none. On international comparison I have compiled a separate dossier, not used here except for the legal obligations Italy has already undertaken: those are not comparison, they are law in force.
I. THE DOCUMENT, THE AUTHORS, THE REQUESTERS
1. What exists, and where it stands
Two public texts exist and, since 4 August 2026, a third text that is not yet public. The two readable ones: a final report of some seventy pages and a proposal of draft articles in twenty articles — which Assonime, in its position paper, sums up in "five guiding lines". The third is the government bill for the "revision of the rules on the administrative liability of legal persons" approved by the Council of Ministers No. 185, on Tuesday 4 August 2026: the legislative form awaited for seven months. It will be presented to the Houses of Parliament — as at the date of this analysis there is no record of a bill number, and the approved text has not been published: of what went into it, and what was left out, we know only what the Government has said and the press has reported.
The chronology, then, is this. The Committee is established by decree of the Head of Cabinet of the Ministry of Justice on 7 February 2024, begins its work on 12 June 2024, and closes it on 30 October 2025. The report bears the date November 2025. The text is delivered to Minister Nordio on 22 December 2025 and made public on 14 January 2026 by Giurisprudenza Penale, then on 22 January by Sistema Penale: two legal journals, not the Ministry's website, on which there is no record of any act of publication. Then seven months without legislative initiative; Confindustria's invitation of 26 May; and the government bill of 4 August, approved on the eve of Parliament's summer recess.
That fact is already a piece of political information, and it must be read together with what happened on 26 May 2026 at Confindustria's assembly. President Orsini, in the official address, devotes to Decree 231 the harshest passage of the justice chapter: the law is "the symbol of the inability to translate into concrete fact the reforms shared only in words"; "over time, however, Decree 231 has turned into an almost exclusively punitive instrument, approaching forms of strict liability"; and he closes: "that reform is no longer merely necessary, it is absolutely urgent. With the contribution of all political parties, let us do it. Let us do it now". The Minister of Justice, the same day, replies: "we have accepted the invitation to amend the 231 rules, on which moreover we have long been working already. This does not at all mean limiting, or indeed leaving unpunished, entrepreneurs who break the law. The contrary"; and, after complaining that a model that is "certified" may not be recognised "in court", he concludes that the Government is "already well on the way to granting these legitimate requests".
I record two things about this exchange, without comment. The first: the only empirical assertion — Decree 231 as an "almost exclusively punitive instrument" — is the one that the available data contradict most sharply, because an instrument almost never applied is not almost exclusively punitive, it is almost exclusively unapplied. The second: in the lexicon of both men, the reform is a request with an owner, and the problem to be corrected is that a model built "with many costs and sacrifices" may not be recognised in court. It is the definition of the problem proper to those who buy and sell the model; neither of the two speakers that day represented those harmed by corporate crime.
2. Who wrote it
The Committee is composed of twenty people, including the coordinator. Giorgio Fidelbo is president of the sixth criminal section of the Court of Cassation and one of the judges who have done most to build the Court of Cassation's case law on the liability of entities. The other members, according to note 1 of the report, are Annetta, Caggiano, Catalano, Ceresa-Gastaldo, Corasaniti, Del Coco, Di Geronimo, Giarratana, Gullo, Manacorda, Milani, Mongillo, Parrotta, Piergallini, Quattrociocchi, Raimondo, Ravazzin, Selvaggi, Stifano.
It is a composition of high technical calibre and of clear criminal-law predominance: professors of criminal law, corporate criminal defence lawyers, judges. Piergallini and Mongillo are two of the three or four Italian authors who have written the pages on organisational fault that get cited. On legal competence there is nothing to object to.
What must be observed is the absence of non-legal competence. In a Committee of twenty people called upon to rewrite the rules on organisational deviance there is no sociologist of organisations, no scholar of organisational behaviour, no economist of enforcement, no statistician. There appear exclusively the disciplines that the second part of this investigation identified as structurally blind to the organisational fact. The Committee that must establish when an organisation is guilty of having organised itself badly is composed entirely of people whose training does not include the study of organisations. This is not a reproach to anyone: it is a fact of composition, and it alone explains many of the product's choices.
There is no record of publication of the establishing decree, of the criteria for selecting the members, or of minutes of the meetings.
3. Who was heard, and what they had asked for
The report gives account of hearings of the "associations that have always been particularly sensitive to the demands for reform", which "have offered interesting contributions, including in writing". Those heard were: Confindustria, Confcommercio, Assonime, the Association of members of supervisory bodies (AODV231) and the Osservatorio nazionale 231 of the National Council of Chartered Accountants. Five parties: three associations representing the companies to which the rule is addressed, the trade association of those who sit on supervisory bodies, the observatory of a professional order that makes its living from Decree 231. There is no record of hearings of prosecution offices, trade unions, victims' associations, ANAC, INAIL, the Labour Inspectorate, or anti-corruption organisations; in the readable part of the report there appears no reference to the OECD, whose 1997 Convention was the reason Decree 231 was written.
Of the written contributions, one is public and permits the most instructive comparison in this whole affair: the position paper Confindustria published on 11 April 2025, "Prospettive di riforma della responsabilità 'amministrativa' degli enti". Nineteen pages that diagnose the "unease and disorientation" of companies and ask, in order: organisational fault as a constituent element of the offence, with the burden of proof on the prosecution; the adoption of the codes of conduct of trade associations as a parameter for the judgment of adequacy, with an obligation on the judge to give specific reasons when disregarding a model that conforms to them; a probation-style supervision (messa alla prova) of the entity with extinction of the offence, on the model of the American agreements; the alignment of the statute of limitations with that of the natural person, with extension of the bar on proceeding (improcedibilità) under Art. 344-bis; the exclusion of micro-enterprises from the Decree; subjective parameters — financial solidity, reputational damage — as preconditions for precautionary measures; the pruning of the catalogue of offences; the entity's right to silence.
I set this list aside: at the end of the analysis I shall compare it item by item with the text of the project. I anticipate only the method of the comparison: it is not enough to count how many requests were granted; one must look at which were rejected, because it is by the refusals that the independence of the drafter is measured.
One methodological point stands: of those hearings there is no record of any published account — not the minutes, not the complete list of the documents filed. In the absence of records, no one can verify which proposal was advanced by whom and which was accepted. The text presents itself as neutral technical elaboration, while the only accessible documentary trail leads to five parties who all sit on the same side of the table. And the other side is silent outside too: as at 5 August 2026 — the day after the approval in the Council of Ministers — there is no record of a single critical stance on the project from the judiciary's associations, trade unions, victims' associations or anti-corruption organisations — the public criticisms from the labour side are all concentrated on the parallel Sisto worksite on workplace safety. The Fidelbo project crosses the public debate without a declared adversary. For those who write rules it is the most comfortable of conditions, and the most dangerous.
4. The field in which the project lands
The project is not the only thing that is moving, and this matters for understanding it.
On the parliamentary front there is a competing legislative proposal, A.C. 2632, filed in the Chamber of Deputies on 29 September 2025 with Morrone as first signatory and drawn up by the research group of the Department of Law of the University of Bergamo coordinated by Anna Lorenzetti, with the contribution of AIGA. It moves in a more radical direction: it excludes from the application of the Decree “small-sized entities”, identified by cumulative criteria — total assets of up to 220,000 euros, revenues of up to 440,000 euros, an average of ten employees — to which is added a qualitative requirement of the absence of managerial autonomy of the administrative body; entities exercising direction and coordination remain subject to the Decree. It is the only text that is currently before Parliament.
On the ministerial front there is a second commission, chaired by Deputy Minister Sisto, established by ministerial decree of 27 March 2024 on prevention and safety in the workplace, which concluded its work on 22 July 2025 and whose report was delivered on 12 May 2026: another document in this series is devoted to it. It proposes a new Article 590-septies of the Criminal Code which limits to gross negligence the employer’s punishability for negligent homicide and negligent injury committed in breach of accident-prevention rules, on one condition: that the entity has adopted an organisational model compliant with Article 30 of Legislative Decree 81/2008. The benefit does not operate for the fundamental breaches — failure to appoint the head of the prevention service or the competent doctor, failure to draw up the risk assessment document, failure to supply protective equipment, failure to provide training. I shall return to this condition, because it is the point at which the two building sites touch and neither of the two declares it. Against this proposal — not against the Fidelbo project — the CGIL has spoken out (“it lightens companies’ responsibilities”), as have the Five Star Movement (“totally inadmissible” in a country with three workplace deaths a day) and part of the press, which read into it a “legal shield”.
On the European front, and this is the point that weighs most, the picture changed while the Committee was at work, and in the direction opposite to its own.
Legislative Decree No. 211 of 30 December 2025, in force since 24 January 2026, implemented Directive (EU) 2024/1226 on restrictive measures and introduced into Decree 231 a new Article 25-octies.2 which, for the first time in twenty-five years, abandons the quota system and pegs the entity’s monetary penalty to a percentage of global turnover: from 1 to 5 per cent for the most serious offences, with a subsidiary cap of 40 million euros where turnover cannot be determined. Twenty-five years of the presumed untouchability of the quota system were closed by an implementing decree, without fanfare. The maximum of the Decree’s ordinary statutory penalty range remains 1,549,000 euros: the ratio between the two ceilings is roughly one to twenty-six.
Legislative Decree No. 81 of 21 April 2026, in force since 2 June 2026, implemented Directive (EU) 2024/1203 on the criminal protection of the environment, broadening the catalogue of predicate offences in Article 25-undecies; but — the datum has been verified against the text in the Official Gazette (Gazzetta Ufficiale) — it adjusted the penalties while keeping to quotas: environmental disaster rises to 600-1,200 quotas, that is, to a maximum of 1.86 million euros, against the forty million (or 5 per cent of worldwide turnover) that the Directive indicates as the minimum threshold for the maximum. I shall return to this choice, because it is a counter-proof that carries weight.
Directive (EU) 2026/1021 on the fight against corruption, adopted on 29 April 2026 and in force since 1 June 2026, requires Member States, by 1 June 2028, to provide for public and private corruption monetary penalties on legal persons whose maximum is not lower than 5 per cent of total worldwide turnover or 40 million euros, and to introduce among the non-monetary penalties exclusion from public procurement and public funding, the revocation of licences, judicial supervision, the imposition of compliance programmes, publication of the decision and, in the most serious cases, the dissolution of the entity.
Holding these three dates and these figures together is already half the critique. The Committee concluded its work on 30 October 2025, deferring the entire matter of penalties to a future enabling law, with the guiding criterion of “graduality and proportionality”. Two months later, the Italian legislature, under European obligation, raised the ceiling to forty million. Six months later, the European Union made that ceiling compulsory for corruption too, that is, for the offence that lies at the origin of the 2001 law. On the point that the OECD has for years indicated as the most critical, the project arrives already overtaken by events. And the 231 market is silent, perhaps stunned by the EU’s sudden overtaking on the left, at least on the environment which, indeed, in Italy has been elegantly neutralised by a rebellious and conservative transposition of the Directive. An infringement procedure may yet follow. A dedicated part of the series addresses the subject.
II. THE STRUCTURE OF THE PROJECT
5. The architecture
The draft articles comprise twenty articles. The introduction to the report declares its programme in discursive form: having preserved the “philosophy” of the 2001 decree — a preventive vocation, judicial assessment within the criminal trial with the guarantees of the accused — the reform elevates organisational fault to a constitutive element of the offence, introduces “new rewarding mechanisms, aimed at favouring the entity’s return to legality”, enhances “the function of best practices as orientative and ‘conformative’ instruments of the judicial evaluation”, completes on the procedural side “the process of assimilation to the position of the accused”, and defers to a delegation the revision of the catalogue of offences and of the penalties. However one summarises it, one observation remains: none of the declared directions concerns the effectiveness of the judicial assessment in practice, the victims or the measurement of results. With one exception, which the introduction does not announce and which lies hidden in the body of Article 13: it is the true surprise of the project, and it will be found at paragraph 13.
The map: Articles 1–9 rewrite the substantive part (attribution, model, suitability, autonomy, extinction, small entities, confiscation, the statute of limitations); Articles 10–12 adjust three predicate offences; Article 13 reforms the procedure in ten moves; Articles 14–16 cover enforcement, sector-specific grounds of extinction and the register; Article 17 contains the delegation; Articles 18–20 coordination, the procedure for the delegation and transitional provisions. It should be noted that the text still carries options in square brackets — the deadline for the delegation, standing for the enforcement set-off — and, on the statute of limitations, two alternative texts: it is not an armour-plated set of draft articles, it is a menu with declared variants.
6. Objective attribution in negligent offences (Article 1)
The new paragraph 1-bis of Article 5: “In the case of negligent offences committed by the persons indicated in paragraph 1, the entity’s interest or advantage exists when a saving in expenditure or an increase in production deriving from non-compliance with the provisions governing the conduct of the activity is pursued or obtained, to an appreciable extent”.
The declared purpose is to codify the landing point of the ThyssenKrupp Joint Chambers, which refer interest and advantage to the conduct and not to the event. The report, however, says more: it declares that it wishes to “settle any conflicts in the case law” and expressly censures two orientations — the one that gave weight to “the systematic nature of the precautionary breaches, as evocative of a corporate policy aimed at saving” (“a selective filter with too wide a mesh”), and the one that excluded the entity’s liability in the case of unconscious negligence on the agent’s part (“an unjustified privilege for the entity that did not trouble itself to confront careless and slovenly practices”). The codification therefore cuts in both directions; but the two new filters it introduces — the quantitative one of “to an appreciable extent” and the typological one of the only two relevant benefits, saving in expenditure and increase in production — operate in one only. I shall return to this.
7. Organisational fault (Article 2)
It is the heart of the project. The new Article 6, paragraph 1: “The entity is liable for the offences committed in its interest or to its advantage by the persons indicated in Article 5, paragraph 1, when their commission was determined or facilitated by the failure to adopt, or the ineffective implementation of, the organisation and management model referred to in Article 7, suitable for preventing offences of the kind that occurred”. A single formula for senior officers and subordinates; organisational fault as a constitutive element, to be proved by the prosecution; the reversal of the burden of proof gone; neither fraudulent circumvention nor the reference to the omitted or insufficient oversight of the Supervisory Body reproduced.
Paragraph 2 contains a provision of balance that must be acknowledged: “Where liability is excluded for lack of organisational fault, the confiscation of the profit that the entity has drawn from the offence is in any event ordered, including in the form by equivalent value”. The blameless entity does not pay, but does not keep the gain from another’s offence.
One point must be noted with precision, because in the current reading it gets lost. The two limits that the report attaches to attribution — the offence that is “concretely uncontainable (unavoidable)”, which severs the causal link, and the offence that is “concretely unforeseeable”, which excludes fault — are not in the text of the article: they are in the explanatory report, as systematic consequences which “follow from it”. The provision hands the interpreter the formula “determined or facilitated”; the boundaries of attribution will be drawn by the case law, with the report as a guide. The same holds for the opposite issue: the report discusses at length the “corporate policy oriented towards wrongdoing” — the intent of the societas glimpsed by the Impregilo-bis judgment — and chooses not to legislate on it, “owing to the difficulties that appear to stand in the way of a correct translation into legislation”, remitting it here too to the case law. On both sides, the perimeter of liability remains judicial.
8. The structure of the model and the Supervisory Body (Article 3)
The new Article 7 codifies the structure of the model in nine requirements: mapping of the at-risk activities and of their intensity; protocols with segregation of functions and safeguards; management of financial resources; a Supervisory Body “endowed with autonomous powers of initiative and control as well as adequate financial resources, whose members possess requirements of professionalism”; information flows towards the Body; training obligations; a description of the internal control system; internal reporting channels and a prohibition on retaliation “taking account” of Legislative Decree 24/2023 on whistleblowing; a disciplinary system. At the opening, a principle of proportion: the model must be commensurate with legal form, nature, activity, delegated powers, size, organisational configuration and intensity of risk.
On the Supervisory Body, three choices. Professionalism and resources enter the law. The option of entrusting oversight to the board of statutory auditors — paragraph 4-bis of the current Article 6, a “hasty innovation of 2012, vitiated by an evident functional and preventive incoherence” — is not reproduced. And independence remains outside the law, with explicit reasoning: it has “by now penetrated into practice” and enshrining it in legislation “could foment mistaken interpretations”, such as holding that an internal member compromises it.
Paragraph 3 reproduces the current mechanism of the guidelines: drawn up by the representative associations, communicated to the Ministry of Justice, which “may formulate observations within thirty days”. It is the same procedure that the report, a few pages earlier, declares to have become “bureaucratised” to the point of weakening its function: the text reproduces it identically. Paragraph 4 entrusts to the Ministry, “after consulting the most representative associations of the entities”, the drawing up of simplified procedures for SMEs.
9. The suitability judgment and the foreign entity (Article 4)
The new Article 7-bis, paragraph 1: the judge, in assessing the suitability of the model, “must take specific account of conformity with the guidelines drawn up by the representative associations of the entities or with the simplified procedures […], with the standards accredited by the technical-scientific community as well as with good practices, provided that they prove adequate to prevent the offence”.
Two textual clarifications that matter. The final clause — “provided that they prove adequate” — leaves to the judge the assessment of the adequacy of the standard itself: the constraint is less rigid than it is made out to be. And the burden of enhanced reasoning for the judge who departs from them — “in the manner of a dissenting opinion, attended by a burden of specific reasoning”, in the words of the report — is in the report, not in the letter of the provision. What the provision imposes is the “taking specific account”: a duty of point-by-point engagement with sector standards, from which the enhanced reasoning will follow by way of interpretation. The ex ante perspective of the judgment, likewise, is entrusted to the report.
Paragraph 2 governs the entity with its principal seat abroad for an offence committed in Italy: the judge assesses whether the safeguards adopted under the law of the State of origin are in practice equivalent to the requirements of Article 7 — without demanding exact correspondence, for example, to an Italian-style Supervisory Body, if compliance functions and internal and external controls guarantee its functional equivalence.
10. The autonomy of liability and the grounds of extinction (Article 5)
The new Article 8, paragraph 1: the entity’s liability exists even when the perpetrator “has not been identified or lacks capacity or is not punishable for a lack of culpability determined by organisational deficiencies of the entity”. The third hypothesis is new: the organisation that has rendered its own agent blameless — the confused procedures that induce error, the impossibility of compliance induced by disorder — answers alone. The report defines it a phenomenon “anything but marginal in complex organisations”.
Paragraph 2 confirms the decoupling of the grounds of extinction with two derogations: amnesty (already in force) and, a novelty, the withdrawal of the complaint. Paragraph 3: “If the offence is not punished under Article 131-bis of the Criminal Code, the non-punishability extends to the entity when the administrative offence dependent on it proves occasional” — the report’s “conditional bilateral decoupling”, designed to deny the benefit to the entity in which the trifling breach is repeated at the hands of different persons. Oblation, reparatory conduct under Article 162-ter and probation-style supervision (messa alla prova) of the natural person remain incapable of being extended to the entity.
11. Extinction of the offence through reorganisation (Article 6, new Article 8-bis)
The project’s new institution, now in the text. The entity which, before the offence, “has adopted and implemented the organisation and management model” may ask the judge, “within ninety days of notification of the notice of conclusion of the preliminary investigation”, for a period within which to eliminate the deficiencies found by the public prosecutor. The request must contain the reorganisation proposal, the offer of any compensation, an indication of the activities to eliminate the harmful or dangerous consequences and the making available of the profit. The judge assesses in chambers under Article 127 of the Code of Criminal Procedure, with an optional expert, and may indicate further measures; if he grants the request, he suspends the proceedings “as well as any precautionary measures that may have been ordered”, sets a deadline and determines a security under Article 49, paragraph 2. On a positive outcome he declares the extinction by judgment, “ordering the confiscation of the profit made available”; in the event of non-performance he immediately revokes the suspension. The rejection is not open to challenge; the request may be resubmitted up to the opening of the trial. Final paragraph: “In the case of an administrative offence dependent on a felony the extinction may not be declared more than twice” — for contraventions, no limit. The report describes the institution as “a probation-style supervision (messa alla prova) designed for the entity” and “a sort of reinstatement of time limits”.
The options discarded: the United States model of self-reporting and agreements, rejected with three textual reasons — the “ferocious internecine struggles fought by way of denunciations”, the middle managers as “sole scapegoats”, the “celebration of the criminal law of the natural person” —; and the filter of the “slightness” of the deficiency, held to be unascertainable in complex organisations. Access to the institution may also be sought again at the stage of opposition to a penalty decree (Article 64, new paragraph 3-bis).
12. Entities without structure and SMEs (Article 7, new Article 12-bis)
The text: “When the entity’s legal personality is not concretely distinguishable from that of the person who committed the offence, the court disapplies or reduces the financial penalty applicable to the entity, taking account of the punishments imposed by the same judgment on the perpetrator of the offence”, including with the conversion criteria of Article 135 of the Criminal Code; in the case of separate proceedings, taking account of the punishment already imposed with final effect. The subjective coincidence “exists” when the offence is committed by a senior officer “who is the sole owner or holds a largely majority shareholding in the entity, and there is no appreciable organisational structure, including by reason of the small number of employees and collaborators”. Excluded are entities that exercise or are subject to direction and coordination. Article 14 adds the new Article 74-bis for the deduction before the enforcement judge, when the conviction of the natural person arrives later in separate proceedings.
The Committee rejected the size-based exemption with four arguments: the difficulty of defining exhaustively those to be exempted; the criminological evidence (“small businesses, generally, do not in themselves present lower crime risks: in some sectors — for example workplace safety, labour exploitation, tax fraud — the criminal risks appear even higher”); the collision “with European and international obligations”; and the quantitative fact — 95 per cent of Italian businesses have fewer than nine workers, so that the exemption “would in fact empty the decree’s scope of application”.
13. The proceedings (Article 13): the part nobody had asked for, and the others
Article 13 rewrites ten junctures of the proceedings. It is worth dividing them by direction of travel.
In the direction of effectiveness — and this is the part that none of those heard had asked for — stand three interventions. Registration (Article 55): “The public prosecutor must immediately enter in the register referred to in Article 335 of the Code of Criminal Procedure the notice of the administrative offence”; the heading changes from “annotation” to “registration”. The report gives as its reason the only piece of applicative data it cites in the entire document, and it is a figure that deserves to be transcribed: “even in the face of criminal proceedings commenced for offences which, by law, would entail the liability of the collective person, in fewer than 10% of cases the annotation under Article 55 is made, with the disconcerting result of the substantial circumvention of the law”. And it continues: the removal of inaction from the court’s oversight “has legitimised a sort of autarky, a practice of internal management of the offences in respect of which each prosecution office (and even each individual deputy prosecutor) makes choices of mere expediency that disregard the obligations specifically laid down by the law”. Dismissal (Article 58): no longer a direct decree of the public prosecutor with mere hierarchical oversight, but a request to the judge for preliminary investigations when the evidence “does not allow a reasonable prediction of conviction to be formulated”, with a reasoned decree of the judge and reference to Articles 409, 410-bis and 411 of the Code of Criminal Procedure. The charge (Articles 59 and 61): the accusation must set out “in clear and precise form” also “the organisational deficiencies that determined or facilitated the commission of the offence”, the interest pursued and the advantage obtained — the end of “implicit” charges traced over the indictment of the natural person; and direct summons to trial enters among the acts of charging, avoiding artificial separations.
On this block, moreover, the case law of the Court of Cassation got there first: judgment no. 143 of 2026 of the Sixth Section — with Di Geronimo as reporting judge, who is a member of the Committee — held that a public prosecutor who proceeds for a predicate offence and has evidence concerning the entity “is obliged to carry out the relevant assessment”, because the choice “is not discretionary”. The Court of Cassation and the reform project say the same thing, in the same half-year, from a pen that partly coincides.
One absence in this block is, however, deliberate, and the report justifies it in a way that must be quoted: in the dismissal there is no reference to Article 410 of the Code of Criminal Procedure, “it being inconceivable here to institute a procedure of opposition to the request for dismissal by the person harmed by the offence (who does not appear, even indirectly, entitled to react against inaction towards the collective person) or by an undefinable person harmed by the administrative offence (a figure that has no citizenship, as case law and scholarship concordantly hold, in the system of the administrative liability of legal entities)”. The court’s oversight of inaction is there; the voice of the victim, by declared choice, is not.
In the direction of safeguards stand the rewritten Article 44 — incompatible as witnesses are the defendant charged with the predicate offence, the legal representative in office, the person designated under Article 39 and “the person who represented the entity at the time of the commission of the offence”, all heard with the safeguards of a defendant in connected proceedings: the entity’s right to silence becomes full — and the retouching of Article 45, paragraph 1: in the precautionary request “the public prosecutor shall specifically indicate the organisational deficiencies found”.
In the direction of the negotiated exit stand the negotiated settlement (patteggiamento) and seizure. The new Article 63: the entity and the public prosecutor may request the application of the financial and disqualification penalties “reduced by up to one third, as well as the confiscation under Article 19 in the measure deemed to correspond to the price or profit of the offence” — the quantification of the profit enters the perimeter of the agreement —; excluded are only the offences under Article 444, paragraph 1-bis, of the Code of Criminal Procedure and the cases of permanent disqualification; the entity’s settlement is uncoupled from that of the natural person. And, paragraph 4: the settled offence “is extinguished” when, after five years (disqualification of up to one year) or two years (financial penalty only), the entity does not commit an offence of the same kind. Article 8 of the proposal closes the circle by inserting into Article 19 confiscation “also” with the judgment applying the penalty upon request.
The new Article 53 on preventive seizure: “Where there are serious indicia of the existence of the entity’s liability for an administrative offence dependent on a criminal offence, the court may order the seizure of the things whose confiscation is permitted”. The seriousness of the indicia — “including organisational fault”, the report specifies — becomes a precondition of the seizure, aligned with the disqualification measures. And paragraph 2: “The seizure shall be revoked by the court if the entity offers security adequate to guarantee the execution of the confiscation”. The special rules on the management of seized businesses are repealed, reabsorbed into Article 104-bis of the implementing provisions of the Code of Criminal Procedure.
Finally, the register of records (Article 16 of the proposal). Judgments of extinction under Article 8-bis and orders of non-punishability under Article 8, paragraph 3, are entered in the register — the court sees them, and they serve for the twice-only limit — but they are excluded from the certificate requested by the entity; and the entity that makes substitute declarations under Articles 46 and 47 of Presidential Decree 445/2000 “is not required to indicate” those measures. In calls for tenders and in questionnaires, the entity that has passed through extinction declares, legitimately, nothing.
14. The statute of limitations (Article 9): two texts
Today: a single term of five years, and the bringing of the action definitively interrupts its running — the proceedings against the entity, once begun, are never time-barred. The report denounces the asymmetry: after the Cartabia reform the trial of the natural person has “certain and contained timeframes”, that of the entity “may drag on sine die”.
Text A equalises: the penalties are time-barred “upon expiry of the terms laid down for the offences entailing the entity’s liability”, with the Criminal Code rules on commencement, interruptions and suspensions; a limitation period “always waivable by the entity”; and a new Article 37-bis extends to appeal proceedings the maximum duration terms of Article 344-bis of the Code of Criminal Procedure “as well as the other provisions contained therein insofar as applicable”.
Text B preserves autonomy: a single term of six years; interruption by effect of the “application of disqualifying precautionary measures” — the adoption, no longer the request — and of the charge; definitive cessation of the running with the first-instance judgment, resuming in the event of annulment with regression; and the same graft of the bar on proceeding (improcedibilità) under Article 344-bis. The extension of the bar on proceeding is the only point on which the report records the agreement of all the members.
15. The sectoral exits (Article 15)
In environmental matters, the new Article 318-septies.1 of the consolidated act (“Prescriptions addressed to the entity”): when the summary offence is committed in the interest or to the advantage of the entity and is attributable to organisational deficiencies, the supervisory authority or the criminal investigation police prescribes to the entity the elimination of the deficiency, with corrective measures “technically certified by the competent specialised body”; the offence is extinguished, like the summary offence, upon compliance within the term and payment under Article 318-quater. In tax matters, the new Article 13-bis.1 of Legislative Decree 74/2000 (“Non-punishability of the entity”): the voluntary correction (ravvedimento operoso) that renders the natural person not punishable extinguishes the corresponding offences of the entity, “where the entity has eliminated any organisational deficiencies that determined or facilitated the commission of the offence”, with the possibility of a supplementary term.
16. Predicate offences, penalties, delegation (Articles 10–12, 17–18)
Three immediate amendments. Article 10 delimits criminal association: the entity is liable for the simple association offence “only if the offences that are the purpose of the association are included among those for which its liability is provided” — a reception of the restrictive case law, against the use of Article 416 as a picklock to circumvent the exhaustiveness of the catalogue. Article 11 intervenes on negligent homicide with breach of accident-prevention rules: the fixed penalty of one thousand quotas — the only non-adjustable penalty in the Decree — is replaced by a range with a minimum of three hundred and fifty quotas and a maximum that remains one thousand; the report reasons on grounds of proportionality and adjustment to the concrete case. Article 12 adds self-laundering and the use of cultural goods to Article 25-duodevicies.
The delegation is Article 17, with six guiding criteria. Letter a): redefine the criteria for determining quotas “taking account of the entity’s size and its economic, financial and asset conditions” and revise the sentencing ranges for “proportionality and greater coherence” — and here the report admits, verbatim, that the current amounts may prove “excessively burdensome for smaller businesses or, conversely, devoid of any effective deterrent impact on entities of greater economic significance”, and cites as a constraint Directives (EU) 2024/1203 and 2024/1226, which “lay down binding indications on the maximum amount of the applicable financial penalties, commensurate with the entity’s global turnover”. Letter b): revise the disqualification penalties “ensuring that they are applicable to the most serious categories of offence”. Letter c): “limit the current catalogue of predicate offences […] to those most pertinent to corporate crime”. Letter d): close the three “open” catalogues (terrorism, payment instruments, smuggling), today entrusted to indeterminate cross-references. Letter e): extend the catalogue to offences against public safety — those which the 2000 enabling act already provided for and which the 2001 decree left out, with the consequence, recalled by the report itself, of “serious disastrous events occurring in industrial contexts or in public transport” for which “criminal prosecution could concern only natural persons”. Letter f): introduce, for environmental offences, an obligation of recovery and restoration on the entity, in implementation of Directive 2024/1203.
Article 18, among the coordinating provisions, also rewrites Article 30, paragraph 5, of the consolidated act on safety: the presumption of conformity of the models is hooked to the UNI ISO 45001:2023 standard, the phrase “upon first application” is deleted, and the periodic updating of the standards is entrusted to the Permanent Consultative Commission. What in 2008 was born as transitional becomes permanent: I shall return to the weight of this choice in the critique.
III. THE CRITIQUE
I begin with what in the project is acceptable, because criticism that does not do so serves no purpose and dismantles itself.
The project contains at least ten choices that must be acknowledged. The elimination of the reversal of the burden of proof is clean and coherent — within the premise the project assumes, namely that the entity is entitled to the constitutional status of culpability; whether that premise is truly an obligation or a choice is the question that section 19 opens, because nobody poses it. The compulsory registration of the entity and the judicialisation of dismissal attack — with words of unusual harshness towards the prosecution offices — the true applicative disgrace of the law, its charging at an office’s discretion. The confiscation of the profit even in the absence of organisational fault prevents the blameless entity from retaining the gains of the offence. The non-reproduction of paragraph 4-bis closes the fiction of the board of statutory auditors supervising itself. The refusal of the size-based exemption for small businesses, against an explicit and powerful request, is the most courageous page of the report. The refusal of subjective parameters in precautionary measures — asset solidity as a shield from seizure — is a second denial that must be put on the record. The autonomous liability of the entity when the organisation has misled the material perpetrator is a strong intuition. The equalisation of the entity’s procedural safeguards with those of the defendant — silence, incompatibility as a witness — is systematic coherence, not a gift. The obligation to charge the entity, “in clear and precise form”, with the specific organisational deficiencies — the end of charges photocopied from the indictment of the natural person — is a safeguard that also serves the fact-finding. And in the delegation, alongside the lightenings, stand two real extensions: the offences against public safety, awaited since 2000, and the obligation of environmental restoration on the entity.
That said, the project has a fundamental problem that none of its merits offsets, and it is from there that I start.
17. A reform without a diagnosis
In the entire report, only two figures about the real Italy appear. The first is demographic: 95 per cent of businesses have fewer than nine employees (ISTAT), and it is used — rightly — to reject the exemption of small firms. The second is the only piece of applied data in the whole document, and it is a confession: when criminal proceedings concern a predicate offence, the entity's offence is recorded "in less than 10% of cases" — "with the disconcerting result of the substantial circumvention of the law". Everything else is missing: how many judgments, how many convictions, how many dismissals, how many times a model has been found adequate, how long a trial of an entity lasts, which penalties have actually been imposed, whether the models prevent anything at all. The only figure the report cites is the rate of inaction of the prosecutors' offices; and it uses it to found the only part of the project that none of those heard had asked for. On the preventive effectiveness of the paper that all the rest of the project prizes, not one number.
This is not a defect of style: it is a defect of method that vitiates every conclusion downstream. A reform is a causal hypothesis — if I change rule X, behaviour Y will change in this way — and a causal hypothesis without a measurement of the starting state is neither verifiable nor falsifiable. The Committee diagnosed the uncertainty of the judgment of adequacy as the law's principal ailment. It may be right. But that diagnosis is not drawn from a series of conflicting decisions: it is drawn from the perception reported at the hearings by five associations, four of which represent parties who buy or sell models.
The point is best seen in the negative. If the uncertainty of the judgment of adequacy were the central problem of Decree 231, one would expect to find, over twenty-five years, a mass of rulings assessing models with divergent outcomes. That mass does not exist. The available figures — those of the university observatory on Italy's most active court, documented in the previous parts — say that of 343 entities judged in Milan between 2016 and 2022 only 64 had a model adopted before the fact: eighty-one per cent arrived at trial without one; and acquittals on the merits founded on the model's adequacy, in the entire history of the law, the known case law records them: there are two. The Committee devoted the most refined intervention of the project — Article 7-bis, the duty to take specific account of best practices — to making predictable an assessment that in practice almost never takes place. It has carefully regulated the valve of a system that never comes under pressure.
It will be said that the perception of risk matters anyway, because it is on that perception that companies base their decisions. That is true, and it is precisely the point: if companies perceive as intolerable a judicial risk that the figures place close to zero, the problem is not the risk, it is the perception and those who produce and sell it. A reform that takes that perception as a given and shapes the rule to it does not reduce the uncertainty: it certifies its market value.
18. The one-way consultation
Five parties heard, all from the side of those to whom the rules are addressed. The objection is not that companies should not have been heard: it is that no one else was. Decree 231 governs a conflict. On one side there is the entity that wants to reduce its own exposure; on the other there are the workers who die in accidents, the polluted territories, the savers, the honest competitors excluded from public contracts, and the State that must establish the facts. The Committee listened to one side only. There is no need to hypothesise bad faith: the physiology of every asymmetric consultation is enough. Those who speak write.
Then there is the question of the records. There is no record of any published account of the hearings; there is no record that the establishing decree was published; there is no record that the criteria for choosing the twenty members were declared; no public consultation was held on the text. A reform that asks companies for traceable decision-making procedures, documented information flows and verifiable disciplinary systems was drawn up through a procedure of which no accessible trace exists.
And there is, finally, the silence of the other side. On the Fidelbo project there is no record, seven months after publication, of a single critical stance by the associations of the judiciary, trade unions, victims' associations, anti-corruption organisations. The only documented dissenting voice is that of the 231 Observatory of the Unione delle Camere penali — which criticises it from the opposite side, judging it "moderately protective of defence rights" and complaining that it does not touch precautionary measures or the catalogue. A project that rewrites corporate liability and does not find a single public opponent on the victims' side is not a project that has convinced everyone: it is a project that one side of the conflict has not read. That moment has now arrived: since 4 August the project has been a government bill, and the silence has not yet been broken.
19. The burden of proof and the question nobody asks: to whom does Art. 27 belong?
On this point the chorus is unanimous. The report presents the transfer of the burden of proof to the prosecution as a constitutionally required adjustment; the commentaries repeat it; even the critics of the project concede it at the outset, as one concedes the obvious. It is worth pausing instead, because the obvious, here, is not obvious at all — and the question nobody asks is the simplest of all: for whom was Article 27 of the Constitution written?
The answer of the Constitutional Court, when it has given it in earnest, is well known. In the twin judgments of 1988 — no. 364 and no. 1085 — culpability became a constitutional principle for reasons that all have a human holder: only those who could have acted otherwise are punished, because punishing the blameless means treating a person as an instrument; and punishment must be capable of re-educating, which presupposes someone to whom something can be reproached — the third paragraph speaks of the “convicted person”, of the “sense of humanity”, of re-education. The guarantee is written for the body, which prison can lock away; for the conscience, which reproach can reach; for liberty and dignity, which punishment can destroy. Only for human beings is imputation without fault odious, because only a human being can be degraded by it. The legal person has no body to imprison, no conscience to reproach, no life to re-educate: it has assets and an organisation. It is a person because a rule says it is — by fictio iuris — and the penalty that strikes it strikes a balance sheet, not a biography.
The story of how Art. 27 entered this field, moreover, ought to be told in full every time, because it is the story of two symmetrical distortions. For half a century the societas delinquere non potest rested precisely on Art. 27: the entity has no capacity for culpability nor can it be re-educated, therefore it cannot be punished — and as early as 1970 Bricola had put “the cost” of that principle on the ledger. In 2001 the legislature circumvented it with two moves: the administrative label, so as not to have to reckon with the Constitution; and the construction — the ministerial report says so almost in these very words, and the document in this series “What Lies Around the Corner for Decree 231?” has documented it on the text — of an “own act” and an “own fault” of the entity, invented as devices of imputation to shield the new liability from the objection of unconstitutionality. The Joint Chambers, in the ThyssenKrupp case, completed the work: the system is legitimate because the entity answers for its own culpable act. Note the direction: the entity’s culpability was born as an argument for being able to punish it. Twenty-five years later, the same construction has changed masters: if the entity has culpability, then presumptions are odious, the burden falls on the prosecution, doubt acquits. The anthropomorphic costume, sewn to open the door to punishment, now closes it from the inside. Two distortions, one beneficiary.
And when the Constitutional Court was called upon to say whether the statute of Art. 27 extends to punitive non-criminal sanctions — it happened in 2019, judgment 112, points 8.2.1 and 8.2.3 —, it answered no: punitive sanctions attract legality, proportionality, fair trial; not the full complement of criminal culpability. Decree 231 is, by the label its own legislature chose, administrative liability. The project therefore demands for it, in the name of the Constitution, a statute that the Constitutional Court does not even recognise as due — and it demands it in favour of the only inhabitant of the legal order that lacks all the anthropological preconditions of the guarantee.
There is a simple way of checking whether the entity’s culpability is truly a necessity of legal civilisation: look at the legal civilisations. In the United States the criminal liability of corporations has been strict and vicarious since 1909 — New York Central v. United States: the agent’s act, in the exercise of his functions and for the benefit of the entity, is imputed to the entity, with no organisational fault to be proved; the diligent organisation is not exempt: compliance operates afterwards, as a mitigating factor in sentencing under the Sentencing Guidelines and as a criterion for prosecution. In France, Art. 121-2 of the code pénal imputes to the entity the act committed “on its behalf” by its organs or representatives, by rebound. And the United Kingdom — the homeland of due process — wrote in 2010, in section 7 of the Bribery Act, exactly the mechanism the Italian project declares intolerable: the company answers for the corrupt act committed by those who operate for it unless it proves, itself, that it had adopted adequate procedures; and far from repenting of it, it has extended it — to the facilitation of tax evasion in 2017, to fraud with the Act of 2023, in force since 1 September 2025. As for the European Convention, the presumption of innocence applies to entities too, but from Salabiaku onwards Strasbourg has admitted reasonable presumptions, and there is no record that it has ever censured either vicarious regimes or defences with the burden on the company. The comparative picture says this: the great liberal legal orders treat entity liability as the price of organisation, not as moral reproach; they place prevention in the discounts, not in punishability; and they are moving towards placing the burden on the company, while Italy moves away from it, declaring the opposite direction obligatory.
The check can also be made without leaving the Italian legal order. Art. 2049 of the Civil Code has always made masters and principals answer without fault for the acts of their servants, and nobody finds it odious. Antitrust and data protection sanction the company for the acts of its employees on substantially strict criteria, up to ten and four per cent of turnover, and no court has ever demanded that the authority prove organisational fault. The civil Joint Chambers, in 2001, fixed the criterion that governs the allocation of proof: proximity to the evidence — the burden follows the party that has at its disposal the means to discharge it. Well then: who possesses the evidence of a company’s organisation, if not the company? Loading the prosecution with the proof of what lies, in its entirety, in the archives, the information systems and the practices of the accused is the exact reversal of that criterion — practised in the one sector where the informational asymmetry is greatest and where, as will be seen in a moment, the prosecution is given no means to bridge it.
The three serious objections must, in fairness, be stated. The first: the penalty on the entity falls on blameless third parties, shareholders and workers. It is true — and it is true of every civil judgment, of every bankruptcy, of every antitrust penalty: that problem has a constitutional name, and it is proportion (Art. 3), not culpability (Art. 27); it is governed through the measure and through remedial sanctions, not through the an of liability. The second: without culpability the criminal reproach loses legitimacy. Perhaps so — then let a choice be made: either the matter is truly criminal, and one has the honesty of the label with everything that follows from it, on the side of the penalties too; or it is administrative-punitive, as the legislature christened it, and the guarantees the Court assigns to it apply — no more. What cannot be done is what the project does: keeping the administrative label for the penalties and claiming the criminal statute for the guarantees. The third, the most serious, comes from economic analysis: pure strict liability can disincentivise self-monitoring and self-reporting, because those who investigate themselves produce evidence against themselves — it is the paradox pointed out by Arlen. But it is an argument about the how much and the path, not about the an: it is solved with rewarding mechanisms — robust discounts for serious compliance, for cooperation, for disclosure, on the model of the Guidelines and of 8-bis itself — not with immunity. Organisational fault can live perfectly well as a sentencing criterion and as a leniency route; it is its elevation to a constitutive element, with the burden on a disarmed prosecution, that turns an incentive into an exemption.
This, then, is the conclusion on the quaestio, and it must be formulated with precision: not that reversing the burden is constitutionally prohibited — it is that its elimination is not constitutionally required. Between the American vicarious model, the English burden-defence and the constitutive element of the project runs a space that the Constitution, as the Court has read it, leaves entirely to politics; and within that space the project’s choice is the most protective of the entity ever adopted by a comparable legal order. Presenting it as a duty has a precise function: to remove it from judgment on results. A political choice is judged by what it produces; a constitutional obligation is not. The true guarantees — those the Constitution reserves for the human being — are not at issue and remain intact: in 231 proceedings the flesh-and-blood defendants retain Art. 27 in full, as is right. The project does not implement the Constitution: it uses it — and it uses it in favour of the only subject of the legal order that exists because a law says it exists.
There remains the operational problem, which the constitutional premiss made invisible and which can now be looked at for what it is: a choice of system, with calculable consequences. The Committee wrote the rule and ignored the system in which it will have to operate. Proving that an organisational deficiency “determined or facilitated” an offence within a complex organisation means reconstructing who decided what, which incentives rewarded which behaviour, which signals had been ignored and at what level, which practice had eroded which procedure. It is an organisational assessment, not a legal one. It requires tools, time, expert consultancy, the ability to read internal control systems, access to corporate data.
The project transfers this burden to the public prosecutor and gives him not one thing more to discharge it: no new inspection powers, no obligation on the entity to preserve or hand over documents, no administrative authority to carry out upstream fact-finding, no specialised office. And at the same time it demands more of him and gives him less time. More on three fronts: registration becomes mandatory and immediate; dismissal goes through the judge; and the charge — including the precautionary one, by effect of the retouch to Art. 45 — must state “in clear and precise form” the specific organisational deficiencies, which means that the organisational assessment must be brought forward to the moment of the charge. Less time, because a parameterised or six-year statute of limitations and the bar on proceeding (improcedibilità) are on their way. A duty to proceed, a higher standard of proof, a charge to be reasoned at once, no new means, shorter time limits: anyone who has seen a prosecutor’s office knows how this equation is solved. It is solved with requests for dismissal for want of a “reasonable prospect of conviction” — reasoned, vetted by the judge, and inevitable.
The foreseeable result is not a trial with more guarantees: it is a trial that does not take place, but with one more stamp. The Committee does not say so, does not estimate it, provides for no compensation: it reformed the evidentiary rule as if the trial were a neutral place where the parties command comparable means. It is not, and in the field of corporate crime it is less so than anywhere else.
I add the severest remark, and it concerns internal coherence. The Committee declares that organisational fault is an autonomous, “normative” category, with “quite peculiar criteria of assessment”, and on this basis rejects the assimilation to individual fault carried out by the Court of Cassation. Then, however, when it comes to delimiting imputation, it takes from individual fault the two limits that most narrow it — foreseeability and avoidability — and applies them to the entity. The autonomy of the category holds when it serves to say that the entity is reproached for its own act; it does not hold when it would serve to say that the criteria for assessing the entity are different from those for the person. It is an argumentative asymmetry, and it points in the same direction as almost all the others.
20. Foreseeability and unavoidability: two exit doors written in pencil
The two limits the report attaches to imputation — the “concretely uncontainable” offence, which breaks the causal link, and the “concretely unforeseeable” offence, which excludes fault — deserve a remark of their own, starting from their location: they are not in the text of Art. 6, they are in the explanatory report. They are not clauses, they are instructions to the interpreter. Which makes them at once less solemn and more insidious: they will enter the system without the legislature’s scrutiny, with the elasticity of case-law maxims.
The limit of unavoidability reproduces, shifted onto the entity, the classic defence of the worker’s aberrant conduct, the argument with which the employer’s liability in workplace accidents has been dismantled for decades. Entrusted to a counterfactual judgment on a model nobody knows how to measure, it becomes a permanently available argument: it can always be maintained that a senior manager determined to offend would have circumvented any safeguard.
The limit of unforeseeability is more insidious. Proving the foreseeability of a criminal risk in a given organisation requires historical data, sector benchmarks, literature on organisational deviance: exactly what does not exist in Italy, because nobody collects it. The assessment is handed a criterion that refers to a body of knowledge the system does not produce.
A third door, this one written in ink, sits in Art. 5: in negligent offences the entity’s interest exists only if the saving or the increase in production is pursued “to an appreciable degree”. Who determines when a saving is appreciable? In workplace accidents, the omitted precaution is often worth a few thousand euros — the postponed maintenance, the skipped training course, the device not bought. For a large group, none of this is “appreciable” in relation to its balance sheets; the dead man, however, remains. The filter designed to bar automatic imputations may become, for the largest companies, a threshold of irrelevance: the bigger the organisation, the less appreciable its lethal savings.
Combined with the reference to best practices, these limits produce an effect that must be stated without circumlocution: the entity that has bought the model conforming to the guidelines, and which has therefore by definition covered the risks “foreseeable” according to the sector standard, has an almost automatic defence for everything that standard does not contemplate. The paper model, which today counts for little in court because the judge assesses it freely, would tomorrow count for much more. For completeness, the reverse must be added: the report also leaves to the case law the opposite criterion, the “company policy oriented towards wrongdoing” of the Impregilo-bis judgment, which could have aggravated the entity’s position and which the Committee chooses not to legislate, out of respect for the principle of precision. The symmetry is declared; but a limit pro reo entrusted to the maxims flourishes, an unwritten criterion contra reum dies. Anyone with courtroom experience knows it.
21. The Supervisory Body: the guardian is professionalised and the whistle is repealed
Under the current framework, letter d) of Article 6, paragraph 1, provides that the exemption does not apply where there has been omitted or insufficient supervision by the body. It is the only provision in the entire decree that attaches legal significance to the actual conduct of the Supervisory Body: it imposes no obligations on it, it does not sanction it, but it ensures that its silence weighs on the entity, and therefore that the entity has an interest in a body that actually works.
The draft does not reproduce it. The stated reason is “to enhance the entire system of controls”. In exchange, the body receives the reference to the professionalism of its members and the guarantee of financial resources.
Let us line up the result. The body emerges from the reform with more requirements to certify, with resources guaranteed by law, and without the only provision that made its inertia legally relevant. It does not receive independence, which the report declines to codify because it has “already penetrated into practice” — an argument that proves too much: what is already in practice can be written down at no cost; what one does not want to write down is usually not found in practice. It does not receive any obligation towards the outside: no duty to report to the authorities, which the legislator had given it in 2007 in anti-money-laundering matters and took away from it in 2017. It does not receive preventive powers, and therefore continues to have no position of guarantee. It does not receive liability.
A figure is professionalised and financed, and the only point at which its silence produced a consequence is taken away from it. In twenty-five years, with tens of thousands of bodies in office, there is no record of a single public case in which a supervisory body discovered an offence and brought it to light — this is the finding documented in the second part; there are, however, judgments certifying its inertia, and in the Monte dei Paschi trial that inertia served to found the bank’s fault. Under the draft, that same inertia would no longer found anything.
It must be acknowledged that the failure to reproduce paragraph 4-bis eliminates the worst possible confusion, that between controller and statutory auditor: the report calls it an “evident functional incoherence”, and it is significant that Assonime — that is, the front of the addressees — asks for its reinstatement in the name of the “greater discretion of the company”. On this point the Committee is right and the requesters wrong. But an improvement in architecture does not compensate for the removal of the only lever of accountability.
There is, in honesty, a serious defence of the operation: the body is not a sheriff, it is a systems auditor; loading it with reporting obligations would turn it into an internal informer and would dry up the information flows on which it lives; and a principle of civilisation demands that no one be punished for omission who has no preventive powers. It is a true argument, and it explains why the body is made this way. It does not explain why it must be paid this way — for a trade in which nothing is owed and no outcome is measured — nor why the reform professionalises it and finances it by law without asking of it in return a single verifiable result.
22. *Best practices*: the technical standard written by the regulated
The reference to sector guidelines, to the “standards accredited by the technical-scientific community” and to good practices is the draft’s most skilful choice and the most difficult to contest in words, because it presents itself as technical. A textual clarification, first: the proposed provision requires the judge to “take specific account” of conformity with the standards, “provided that” — the guidelines and the practices — “prove adequate to prevent the offence”: the final clause leaves to the judge the scrutiny of the adequacy of the standard itself, and it is a real tempering. The burden of aggravated reasoning for the judge who departs from them — “in the manner of a dissenting opinion, accompanied by a burden of specific reasoning” — sits in the report, not in the letter of the provision: it is the authentic interpretation that will accompany the text into the courtrooms. That said, the substance must be contested on the facts.
First: who writes the guidelines. They are written by the trade associations of the entities addressed by the provision — the same ones heard by the Committee. The draft attributes conformative effect on the criminal judgment to a normative product drawn up by the regulated, without public quality control over its content and without adversarial participation in its formation. When the technical standard steers the penalty, the technical standard is normative power; and normative power is not delegated to the regulated party without counterweights. That this was exactly the design is shown by the comparison with the requests: Confindustria had asked, verbatim, for the sector codes as a “reference parameter” and for the obligation of specific reasoning for the dissenting judge. It is the request granted with the greatest precision of all.
Second: the parallel with medical negligence does not hold. The guidelines under the Gelli-Bianco law are produced by scientific societies accredited under a public procedure, entered in a national system, founded on experimental evidence. The 231 guidelines are founded on no evidence: no one has ever measured whether a compliant model prevents more offences than a non-compliant one, because the data does not exist. The mechanism of evidence-based medicine is imported into a field that has no evidence at all.
Third: the draft contradicts itself, and this can now be demonstrated on the text. The report admits that the drafting of the guidelines “has suffered from a ‘bureaucratisation’ which has, in a significant way, weakened their orientating function”; the new Article 7, paragraph 3, reproduces the current mechanism identically — drafting by the associations, communication to the Ministry, optional “observations” within thirty days. The instrument given conformative effect on the criminal judgment remains formed exactly as before. The reform does not correct its formation; it increases its effects.
Fourth: the precedent already exists, and it is silent. Article 30, paragraph 5, of Legislative Decree 81/2008 has since 2008 presumed the conformity of models adhering to the UNI-INAIL guidelines or to the technical standard referred to — “at the stage of first application”, said the provision. Eighteen years have passed without that presumption producing either measurable safety or case law. The draft not only extends the mechanism to the entire judgment of suitability: it perfects it in the direction opposite to the one experience would suggest. Article 18, paragraph 2, rewrites paragraph 5, anchoring the presumption to the UNI ISO 45001:2023 standard — a private standard, available for a fee, conformity with which is attested by private certifiers — and deletes the phrase “at the stage of first application”: the presumption born as transitional becomes a permanent structure of the system, with the updating of the standards entrusted to the Consultative Commission. In the same six months, Decree-Law 159/2025 was carrying out on its own account the same substitution of standard: legislator and Committee converge in handing the legal parameter over to a commercial product.
Fifth: the asymmetry of the burden of reasoning. In the design of the report, the judge who departs from the practices must positively demonstrate the “profiles of unsuitability” of the standard, like a dissenter writing his own opinion; no corresponding burden falls on the entity, none on the judge who conforms to the standard. A one-way argumentative burden does not reduce discretion: it steers it, and steers it against the party that has fewer technical means to contest the standard.
Sixth: the certifiers. The draft gives weight to the “standards accredited by the technical-scientific community” and to certified systems, and does not regulate who certifies: requirements, independence from the consultant who drafted the model, liability for complaisant certification. Italy is the first country in the world for ISO 37001 certifications. The draft turns the stamp into an element of the criminal judgment and does not write a line about who affixes it.
23. Article 8-bis: the reward of the American agreement without the price of the American agreement
The Committee expressly rejects the United States model of agreements, with three reasons that do not withstand examination. The first — the “ferocious internecine struggles fought by way of denunciations” — is an argument of custom that proves too much: it would apply identically against whistleblowing, which is the law of the land and which the same draft incorporates into the model. The second — in the United States “it is never the top executives who pay, but the middle managers” — is true, and is documented by the most serious literature; but it is an argument against the scapegoat, not against the agreement, and the scapegoat is precisely what the Italian system produces without having the agreements. The third — the agreement as a “celebration of the criminal law of the natural person” — is untenable: the deferred agreement is by construction an instrument directed against the entity.
But the point is not the weakness of the three reasons. The point is that, after adducing them to reject the American model, the Committee builds an institution that takes the benefit of that model and leaves out all the counterweights. The benefit is early exit from the trial. The counterweights that exist there and not here: the admission of the facts in a public document; the publication of the agreement; the independent monitor who verifies implementation for years; the reopening of the proceedings if the plan is not implemented after closure. In 8-bis there is no publicity, there is no register, there is no monitor, there is no verification after the judgment, there is no admission. There is a hearing in chambers, an adversarial exchange declaredly “collaborative”, an extendable time limit, a security deposit, a judgment that extinguishes.
And there is — this is the discovery that changes the weight of the institution — the line on the register of records: judgments of extinction are not mentioned in the certificates requested by the entity, and the entity is exempted from indicating them in its self-declarations. Translated into the practice of public procurement: the entity that has passed through 8-bis returns to the market without a trace. Where the American model publishes the agreement and entrusts it to a monitor, the Italian model declares it by judgment and then erases it from the certificate. This is not an executive detail: it is the difference between an institution of negotiated justice and a criminal-records cleaning service.
The three defects already visible before remain, and the full text adds a fourth. The distributive one: only those who already had the model gain access to the institution, that is, the large company — not the eighty-one per cent that arrives at trial without one; the rewarding precondition is the possession of a document whose quality no one measures. The serial one: the limit of two uses applies only to serious offences (delitti); for contraventions — the terrain of serial corporate offences, environment and safety — the institution can be used without limit, and the draft, which in Article 5 recognises the problem of the repeated trivial violation and attaches to it the condition of non-occasionality, forgets it here. The mute one: the injured party has no procedural role — compensation is a precondition of extinction, but the person who suffered the harm has no voice on the plan and no power to object. This is no oversight: it is the same thesis that the report states with regard to dismissal, and which I shall take up again among the absences. Legally defensible; written into a reform that multiplies the entity’s exits from the trial, it sounds like a manifesto.
The fourth defect is the parallel track. Alongside 8-bis, the new negotiated settlement (patteggiamento): detached from the fate of the natural person, with penalties reduced by up to one third, with the quantification of the profit to be confiscated brought inside the agreement (“in the amount deemed corresponding”), and — paragraph 4 — with the extinction of the settled offence after two years (financial penalty only) or five (short disqualification) without reoffending of the same kind. Added to the register of records, the system of exits becomes a complete fan: the organised entity exits 8-bis with a judgment of no case to proceed and a clean certificate; the entity that cannot gain access to it settles, serves, and after two years the offence is extinguished. In neither of the two paths is a trial held; in neither of the two does an opposable trace remain in the certificates requested by the entity or in its self-declarations.
24. The central contradiction
The draft’s two major interventions rest on two opposite judgments of the same subject. Article 7-bis starts from the premise that the judge does not know how to evaluate a model: for this reason it binds him to the sector practices and imposes aggravated reasoning on him when he departs from them. Article 8-bis entrusts to the same judge a task incomparably more difficult: to predict, in a hearing in chambers, whether a future reorganisation plan will eliminate the shortcomings that produced the offence. Evaluating an existing document is an assessment; evaluating the prognosis of an unimplemented plan is a wager. And here the judge is bound to no practice, is not required to give reinforced reasoning, has no standard, has no mandatory expert, and his rejection cannot be appealed.
The draft distrusts the judge when he judges and gives him unlimited trust when he rewards. I know of no reading of this asymmetry that is not the obvious one: the constraint was placed where it served to reduce the entity’s exposure, and not where it would have increased it.
25. Precautionary measures: the bogeyman, its extinction, its legacy
Disqualifying precautionary measures were, in the 2001 design, the system's true deterrent: the only thing that could stop a company before judgment. And for a brief season they truly were. Between 2003 and 2007 the system showed its teeth: the first judicial receivership in Rome in 2003; the ban on contracting with the public administration imposed on Siemens in 2004, which went round the world; the disqualification of Impregilo, Fisia and FIBE from the Campanian waste cycle in 2007, with seizures of some seven hundred and fifty million. Anyone who sold models in the years that followed was selling, in the final analysis, the memory of that season.
Then the bogeyman went out. There is — and this is in itself a fact — no national statistic on precautionary measures against entities: neither the Ministry, nor the Committee's report, nor the Milan observatory, which surveys only the trial stage, counts their applications. But the indicators converge: the Court of Cassation's case law on Article 45 amounts to a handful of rulings in twenty years; procedural scholarship describes them as very rarely applied; and the most recent snapshot was taken by the Court of Cassation itself in judgment 143 of 2026 — the public prosecutor who seeks the disqualifying measure against the director, even a front man, and does not even register the entity. The perceived threat outlived the real threat by twenty years: the market of fear, documented in the second part, lives precisely in that gap.
What does the project do with this? Four things, pointing in different directions, and they must be weighed one by one.
On the preconditions for disqualifying precautionary measures it does not intervene: Article 45 stays as it is, and Confindustria's request to make precautionary measures conditional on subjective parameters — financial solidity, reputational damage — has been rejected. It is a refusal that weighs in favour of the Committee, because that request amounted to writing that the large company is not to be touched.
On preventive seizure it intervenes twice, and in the same direction. First: to seize with a view to confiscation, there will need to be "serious indicia of the existence of the entity's liability", "including organisational fault", the report specifies. In the abstract this is consistency: if confiscation is a penalty, its precautionary anticipation cannot rest on the confiscability of the asset alone. In practice it means that to freeze the profit of the offence in the first weeks of an investigation the public prosecutor will already need to hold the most difficult proof in the entire system — that organisational failing which, with an investigation barely opened, he has no instrument to demonstrate; and the retouch to Article 45 demands the same for disqualifying measures, requiring that the precautionary request itself already indicate "specifically the organisational deficiencies found". Second, and this is the paragraph that on a first reading of the project had escaped everyone: "The seizure shall be revoked by the court if the entity offers security adequate to guarantee the execution of the confiscation". Revoked, not revocable: upon the deposit of adequate security, the constraint on the assets falls away. For the well-capitalised entity — that is, for the only one worth seizing from — the seizure converts into a balance-sheet item. The seven hundred and fifty million frozen against Impregilo and FIBE in 2007, under this double regime, would probably never have been seen: first the evidentiary standard would have delayed them, then the security would have released them. Timely seizure had remained the only weapon of the precautionary system actually used; the project makes it conditional on the proof that the rest of the project makes more difficult, and makes it redeemable in any event.
On the way out it does intervene: the granting of the request under Article 8-bis suspends any precautionary measure that has been ordered, with security. The precautionary measure thus becomes reversible on the initiative of an entity that promises reorganisation: the bogeyman, wherever it still materialised, has an off switch.
On the future of definitive disqualifications, finally, the delegation: revision "in order to limit their applicability to the most serious categories of offence". Extrema ratio, in the commentators' glosses.
The balance of the precautionary field is therefore this: no tightening at any point; the most used instrument (seizure) made conditional on a proof that is almost impossible at the stage when it is needed; the most feared instrument (disqualification) left formally intact but fitted with a rewarding off switch and set, by delegation, on the road to exceptionality. Anyone wishing to argue that the reform "deflates the bogeyman" would be saying something imprecise: the bogeyman had deflated on its own, through non-application. The reform does something different and more definitive: it transforms non-application in fact into non-applicability in law. It writes practice into the rule.
26. The only courageous provision is also the loneliest
There remains the block that runs in the opposite direction to everything else, and deserves a separate discussion because it is the proof that the Committee knew perfectly well where the real problem lay.
The mandatory and simultaneous registration of the entity, and dismissal removed from the public prosecutor and entrusted to the control of the court, strike at the applicative pathology that the previous parts have documented from the market side: Decree 231 as a local speciality, charged by a few metropolitan prosecutor's offices and ignored elsewhere. The project certifies it with the figure of the annotation missing nine times out of ten and calls it by the words already transcribed at paragraph 13 — the "autarky", the choices "of mere expediency" of each individual office. And the Court of Cassation, with judgment 143 of 2026, said the same thing in the form of a principle: proceeding against the entity is not a discretionary choice, and when the risk of repetition lies with the organisation the right precautionary measure is directed at the organisation, not at the front man of the moment.
This block must be defended without reservation: it is the only part of the project that increases effectiveness instead of reducing it, and it is technically well built. But its fragility must also be stated, and it is twofold. The first: the obligation to register without the means to investigate produces, downstream, reasoned dismissals — the obligation changes the accounting of inaction, not necessarily the inaction. The second is political: this is the orphan provision of the project. None of those heard asked for it; the Observatory of the Criminal Chambers has already flagged it as a risk, for the extra-procedural consequences that the status of entity under investigation carries with it; and companies, once they grasp its scope — every report of a predicate offence will mean, by law, a registered entity — will call for its head. When the project enters the legislative process, every other part of it will have an organised defender; this one will have none. The package of guarantees will pass, the orphan provision will be lost along the way, and the project without that provision is another project — the one this analysis describes in the preceding sections, without the mitigating factor any longer.
27. The statute of limitations: the only clear subtraction
Today the proceeding against the entity, once the prosecution has been brought, has no lapse deadline: a 2001 choice that the Court of Cassation has always held not unreasonable, because the entity's offence is an offence of organisation, its assessment requires time, and the entity does not have an expectation of oblivion commensurate with the individual's. The project introduces lapse for the first time: a statute of limitations pegged to the offence's statutory sentencing range (Text A) or a single term of six years (Text B), definitive interruption moved to the judgment at first instance, the bar on proceeding (improcedibilità) on appeal and in cassation. And it does so in a system in which the proceeding against the entity is already rare, already long, already concentrated in a few offices — and in which, by effect of the same project, the prosecution will from now on have to discharge a heavier burden of proof.
The argument is symmetry with the natural person. But the symmetry is invoked in one direction only: from the rules on the person one takes the term that extinguishes, not the rigour that punishes — that half is deferred to delegation. And there is a detail worth more than any argument: the report records the internal dissent on how to calculate the terms, but unanimity on the extension of the bar on proceeding. Out of twenty members, the only point on which all agree is the one that guarantees the entity an exit from the proceeding through the passage of time.
28. Penalties: the blank delegation, already expired
The OECD has been noting for years that Italian financial penalties on entities are inadequate to their purpose: the maximum of the ordinary range is 1,549,000 euros — one thousand quotas of 1,549 euros — and mitigating factors reduce it further. For a multinational it is an operating cost. The report, in fairness, recognises the problem to the letter — the current amounts "devoid of any effective dissuasive impact" on the large players, the European requirement of global turnover recalled in the delegation, as seen at paragraph 16. But recognising is not providing: no range is adjusted now; the European requirement is cited, not implemented; and the guiding criteria — "proportionality and greater coherence", consideration of "the entity's size and its economic, financial and asset conditions" — are bidirectional by construction, in a climate in which the documented demand is for relief. For disqualifications, the criterion is "to ensure that they are applicable to the most serious categories of offence": in the lexicon of the report it also means filling irrational gaps (false corporate reporting without disqualifications, laundering of cultural goods without disqualifications), but the declared direction remains selection upwards.
Then there is, among the immediate amendments, a small and revealing tell-tale: Article 11 replaces, for negligent homicide with breach of accident-prevention rules, the fixed penalty of one thousand quotas with a range from three hundred and fifty to one thousand. The report justifies it on proportionality grounds — it was the only non-graduable penalty in the Decree, and a defect of rigidity there truly was. But there were two ways to cure it: raise around it, or open downwards. It opened downwards: for the death of a worker, the minimum penalty on the entity falls by sixty-five per cent, in a decree whose ordinary ceiling the OECD already judges inadequate. It is the only immediate amendment to penalties in the entire draft articles. As for what has already happened to penalties outside the project, while the Committee deferred — the figure placed on the table at the opening — it belongs to the close of this analysis, and there it stays.
On the catalogue, the criterion of the delegation — the offences "most pertinent to corporate crime" — is a formula that decides everything and says nothing. With one exception to be acknowledged: the delegation would finally include offences against public safety — massacre, arson, railway disaster — which the enabling law of 2000 already provided for and which the 2001 decree left out. Anyone who remembers that for the Viareggio railway massacre no entity could answer under Decree 231 understands what that gap has cost. But here too: it is delegation, not law; and the same delegation that includes public safety thins out the rest with a blank criterion.
29. Smaller entities: where the Committee was serious, and where it was not
The rejection of the size-based exemption deserves full recognition: sound arguments, sustained against an explicit demand — Confindustria was asking for the exclusion of micro-enterprises — and against the competing parliamentary proposal, which writes that exclusion in black and white. The offset in Article 12-bis starts from a real problem — substantive double jeopardy (bis in idem) in the single-person entity — and is built with prudence: stringent requirements of subjective coincidence, confiscation and disqualifications intact, exclusion of groups.
Two reservations remain. The micro-enterprise is the segment where fatal accidents, irregular labour and environmental illegality are concentrated: any valve opened there must be watched, and the formula "absence of an appreciable organisational structure" is elastic enough to stretch beyond the single-person company. And the simplified procedures for SMEs, entrusted to the Ministry "after consultation with the comparatively most representative associations", are once again the technical rule written with the regulated — this time for ninety-five per cent of the productive system.
30. The register of requests
I close the critique with the comparison announced: Confindustria's list of April 2025 set beside the text of the project.
Organisational fault as a constitutive element, burden on the prosecution: asked, granted. Sector codes as a parameter of the judgment, with specific reasons required from the dissenting judge: asked, granted — it is Article 7-bis, almost to the letter. Probation-style supervision (messa alla prova) with extinction of the offence: asked, granted — with, in addition, the cleansing of the certificate, which not even the request dared. Aligned statute of limitations and the bar on proceeding: asked, granted in both variants. The entity's right to silence: asked, granted. Thinning of the catalogue: asked, delegated. Exclusion of micro-enterprises: asked, refused. Subjective parameters for precautionary measures: asked, refused. Certification with automatic exculpatory effect: a historic request, not granted in form — granted in the attenuated substance of 7-bis.
Five full concessions, one delegation, two refusals, one conversion. The refusals fall on the two most brazen requests, the ones no serious jurist could have signed; the concessions cover almost all the rest. It is an outcome more refined than passive transposition: the Committee rejected what would have made the project indefensible and accepted what makes it presentable. And one must record, to its credit, what it added off-list: the mandatory-registration block, which nobody had asked for, and public safety in the delegation, which nobody had asked for. But the measure of the reform is not given by the count: it is given by the perimeter. The questions outside the list and outside the additions — the inadequate penalties to be adjusted now, the data to be produced, the missing authority, the victims — were not discussed, or were discussed and deferred. Those who wrote the list wrote, in large part, the agenda. This, not the individual article, is the sense in which the project belongs to its requesters.
IV. WHAT IS NOT THERE
A reform project is judged also by what it does not touch. The absences that follow have been verified against the text of the report and the draft articles.
There is not a single provision that produces knowledge. No obligation to collect and publish data on proceedings against entities. No public database of decisions. No statistics on the outcome of suitability assessments. No register of the models adopted, no census of supervisory bodies. No effectiveness indicator, no periodic review clause, no ex post impact assessment. A reform that declares itself oriented towards prevention and does not provide a single instrument for knowing whether prevention works repeats, in terms of method, the error of 2001 — with the aggravating factor that in 2001 nobody could know it, and today they can.
There is no authority. No public body charged with developing or validating the standards, with verifying the models, with carrying out fact-finding upstream of the trial, with publishing guidance. The only public body to which the project entrusts a new task is the Ministry of Justice, and the task is to develop simplified procedures for SMEs in consultation with the trade associations. The system remains entirely trial-based: entrusted to a judge who intervenes rarely, late and without instruments of measurement.
There is no obligation to adopt the model. The model remains an optional burden, and on that optionality the project builds an ever more elaborate structure of rewards: those who have one gain access to extinction, to the shield of the parallel workstream on workplace safety, to the judge's constrained consideration; those who do not, do not. The rewards multiply for conduct which the law continues not to require.
There is no regulation of certification. The project gives judicial weight to the “standards accredited by the technical-scientific community” and to certified systems, and says nothing about who certifies: requirements, independence from the consultant who drafted the model, liability for complaisant certification. There is, more generally, no regulation of conflicts of interest: nothing on the incompatibility between those who draft the model and those who sit on the supervisory body, nothing on the declaration of interests of those who develop the guidelines to which the judge will have to conform.
There is nothing on corporate groups, and this time the silence is declared: “the Technical Committee cannot but associate itself with the ‘silence’ kept by the legislator of 2001”, the report writes. Twenty-five years of laborious case law on the attribution of the offence committed in the interest of the parent company are handed over to another twenty-five. The silence is not neutral: it favours articulated structures, that is, the largest ones.
There is nothing on corporate transformations — mergers, demergers, transfers of business, Articles 28 to 33 — in an economy in which corporate reorganisation is the ordinary instrument for managing legal risk: the draft articles, read in full, do not touch them.
There is no role for the victim, and here it is not a matter of silence: it is a matter of a thesis, the one already transcribed at paragraph 13 — the victim not “entitled to react” to inaction, “a figure that has no citizenship” in the system. In dismissal, no opposition. In extinction under Article 8-bis, compensation is an objective precondition, but the victim has no voice on the plan and no power to object. In the withdrawal of the complaint extended to the entity, the victim appears as the author of an act that closes, not as a party that argues. Three new mechanisms, three times no role — and a theory that justifies it.
The OECD, in the entire document, appears only once: as a historical note, to recall that the 1997 Convention was the reason for the decree. The periodic evaluations of its Working Group on Bribery on Italy, the repeated recommendations precisely on penalties and the statute of limitations, largely unimplemented, are never mentioned or discussed. A report that rewrites that law without engaging with the body that measures its effectiveness says something about the perimeter within which the Committee felt it had to move.
V. JUDGMENT
31. What this project really is
It is a technically refined and politically oriented project, written by excellent jurists who have elegantly solved the problems that were put to them — and the perimeter of the problems put to them coincided, with two exceptions, with the list of those who asked.
The balance sheet must be drawn up in three columns, because that is more honest than the caricature in either direction.
The provisions that increase the effectiveness or the balance of the system: the mandatory registration of the entity and dismissal under the control of the judge; the charge that must set out the specific organisational deficiencies; autonomous liability where the organisation has made the perpetrator unpunishable; confiscation of the profit even without fault on the part of the entity, including by equivalent value; the conditional decoupling of particular triviality, which denies the benefit to the repeat-offending entity; the exhaustive delimitation of the association offence and the closure of the open catalogues; in the delegation, public safety awaited since 2000 and the obligation of environmental restoration.
The provisions that reduce the entity's exposure: the burden of proof on the prosecution without a single new means of discharging it; the limits of unforeseeability and unavoidability entrusted to the report; the judge bound by the practices written by the regulated; extinction through reorganisation reserved for those who already possessed the model, without publicity, without a monitor, without a victim, with a cleaned certificate; the autonomous negotiated settlement (patteggiamento) with a discount of up to one third, confiscation negotiated in its amount and deferred extinction; the “appreciability” filter in negligence offences; seizure made conditional on proof of organisational fault at the stage in which that proof is not available, and in any event revoked against security; the statute of limitations, parameterised or six-year, and the bar on proceeding (improcedibilità) — the Committee's only point of unanimity; the minimum penalty for workplace death reduced by sixty-five per cent; the delegation that lightens catalogue, penalties and disqualifications with bidirectional criteria, in a one-way climate.
What is missing: the data, the authority, the regulation of standards and certifiers, the groups, the victims, the OECD, any instrument for knowing in ten years' time whether the reform has worked.
The first column is real and must be defended. But it carries little weight, for two reasons internal to the project: autonomous liability and mandatory registration require, in order to operate, precisely that capacity for organisational fact-finding which the second column makes more difficult and the third column does not build; and the first column is politically orphaned — none of the parties who asked for the reform asked for those provisions, and no one will defend them.
32. The answer to the question
I return to the question posed at the outset: an instrument for reducing corporate crime with proper safeguards, or a further shield with an overestimated garantismo?
The blunt answer is: as it stands, the project is the second thing, built with the materials of the first. But the formula needs explaining, because the word “shield” is at once too crude and too kind.
The project does not reduce corporate crime, and cannot reduce it, because it contains nothing that acts on the real mechanisms of organisational deviance: it produces no knowledge, builds no external controls, does not touch the incentives inside organisations, gives no one — not the judge, not the public prosecutor, not an authority — the capacity to look inside a company sooner or better than today. All the project knows how to do is redistribute procedural positions around a fact-finding process that remains blind. Prevention, in the project, coincides entirely with paper: bought beforehand (the model), displayed during (7-bis), promised afterwards (8-bis), certified at the end (a cleaned record).
As for garantismo: the true safeguards the project introduces — the burden of proof, the right to silence, the judge's control over dismissal — are just, and must be defended without embarrassment. Garantismo is overestimated not in the safeguards, but in the premise: that there exists, to be balanced, an exorbitant punitive power. The numbers say the opposite: a rare trial, concentrated in a few prosecutors' offices, with penalties at the level of an operating cost, precautionary measures almost never applied and a single season of severity twenty years ago. Shielding an entity from a trial that almost never comes is not balancing a power: it is making sure it never comes. The scales of garantismo, here, have one empty pan — and the project adds weights to the other.
There is a simple way of saying all this, and it is the language of the previous parts. The Decree 231 market lives on a product never measured, sold on the fear of a trial that almost never comes. This project is, objectively, the greatest act of promotion ever written for that market: it makes the ante factum model a condition of extinction, of the employer's criminal shield in the parallel workstream, of the judge's constrained reasoning; it entrusts its content to the guidelines of those who sell it; and it erases from the certificate the only trace its failure would leave. Those who until now had not bought the paper will from now on have to buy it; those who sold it will sell it with the law in their hands. The reform does not shield the company from a power: it converts the entity's liability into a procedure administered by the product the entity buys. It is not declared impunity; it is organised impunity, with the forms of law and the terminology of prevention.
There remains the part that does not fall within this judgment, and it must be stated with the same clarity: the block formed by mandatory registration and controlled dismissal, with the report's very harsh words on the “autarky” of prosecutors' offices, is a provision of equality before the law, and it is the only part of the project which, if it survived, could increase proceedings instead of reducing them. The forecast made above — that it will be the first to fall in the parliamentary passage, because it does not have a single organised defender — is the implicit wager of the whole project. If it falls, the shield remains. If it holds, what will remain is an obligation to proceed entrusted to a prosecution without instruments, within shorter time limits, towards reasoned dismissals. In neither scenario does corporate crime decrease. In both, the paper is worth more.
33. More is needed
If one wanted a reform that addresses the problem instead of its perception, the minimum would be this. Measure: a statutory obligation to collect and publish annually the data on registrations, dismissals, judgments, penalties and the outcomes of suitability assessments, with a public database of decisions and a five-yearly review clause based on those data — it is the intervention that costs least, and nobody proposes it. Give the prosecution the instruments that the burden of proof presupposes: structured documentary acquisition, organisational expert advice, specialised offices and, as a choice of system, an authority with powers of inspection and direction, on the model of what exists elsewhere for the prevention of corruption. Without this, shifting the burden means switching off the fact-finding.
Then the technical standard and its custodians. Guidelines formed through adversarial procedure, validated by a public body, updated on the evidence; certifiers regulated, with requirements of independence and liability. For the supervisory body, at least one duty towards the outside and one consequence for inertia — or its abolition, with the function distributed across the system of controls, which would be a legitimate choice; what does not hold is professionalising it, funding it and asking nothing of it. And if the rewarding mechanism is wanted, take its price too: publicity of the decision, admission of the facts, independent multi-year verification, reopening in the event of non-compliance, a limit on use extended to contraventions, a voice for the victim, no cleaning of the certificate.
Finally, three choices of timing. Address penalties now, not in the delegation: the turnover-proportional parameter has been in the decree since January 2026 and will be mandatory for corruption from June 2028; postponing means arriving at transposition with an organic reform just approved and already in need of rewriting. Defend the orphan provision: mandatory registration and control over dismissal are the part of the project to be saved at all costs, and the only one for which nobody will fight. And do not separate the workstreams: the reform of Decree 231 and that of workplace safety are linked by a point that neither text declares — the model under Article 30 as a condition of the employer's shield, sector practices as the judge's constraint on the entity. Put together, they make the purchased document the key to both liability regimes. If both passed in their current version, the combined result on workplace accidents would not be a reform of safeguards: it would be an insurance excess, and the price of the excess would be written in the price list of those who sell models.
34. The political risk, in one line
The project is a menu, and the legislative process is a customer who chooses. The lightening items — the statute of limitations, the bar on proceeding, the burden of proof, 7-bis, 8-bis, offsetting — each have an organised defender and a documented record of having been requested; the burdensome items — mandatory registration, controlled dismissal, confiscation without fault, the refusal of the size-based exemption — have none. When such a text enters Parliament after being announced to a Confindustria assembly as the answer to “legitimate requests”, the forecast requires no prophetic gifts: the former will pass, the latter will fall, and the result will be worse than the text criticised here — which by then will have served as the trailblazer, with the signatures of twenty excellent jurists to guarantee the operation.
35. Forty Million
At the opening I asked that one figure be kept on the table. The document in this series “What Lies Around the Corner for Decree 231?” explained it in full; it is time to set it alongside this project.
Since 24 January 2026 — the datum is verified against the Official Gazette (Gazzetta Ufficiale) — the forty-million cap is not a European prospect: it is Italian law in force. The new Article 25-octies.2, introduced into Decree 231 by Legislative Decree 211/2025 for the violation of restrictive measures, gears the entity’s penalty to 1-5 per cent of global turnover, with a subsidiary range of three to forty million where turnover cannot be determined. For a group with a turnover of one billion, the frame runs from ten to fifty million; for one of ten billion, from one hundred to five hundred. Alongside it, in the same decree, the quota system stops at one and a half million — and the law’s internal arithmetic has become indefensible: for the same entity, violating an embargo can cost up to twenty-six times more than killing a worker. No dogmatic distinction stands up to this proportion, which is a comparison between two articles of the same text.
This is not an episode: it is a direction, with a deadline. The 2026 anti-corruption directive imposes the same model — five per cent or forty million — by 1 June 2028 for corruption, the offence around which Decree 231 was born. And there is already counter-evidence of how Italy moves when it believes it has room for manoeuvre, subsequent to the delivery of the project: Legislative Decree 81/2026, transposing the environmental directive that demanded the same maximum, kept the quotas — a maximum of 1.86 million against the forty required. Faced with the same obligation, in the same half-year, the legislator wrote forty million where it could not avoid doing so and one million eight hundred thousand where it believed it could afford to. It is the practical demonstration of what the criterion of “graduality and proportionality” can become in the hands of the delegated legislator: the organic reform had the evident task of curing the inconsistency; it delegated it, and the first application of the implicit delegation is already on the record.
Now set this number alongside the project, and every preceding page changes weight. The burden of proof without instruments, the exits from the trial multiplied, the statute of limitations with the bar on proceeding (improcedibilità), the “gradual and proportionate” delegation: each of these choices meant one thing when the maximum was one and a half million — a dispute over detail around a penalty at the level of an operating cost — and means another when five per cent of a turnover is on the horizon. A quibble is a quibble only as long as the stakes are small. The forty million tear quibbling to pieces: they turn every exit door from the trial into an exit door from a penalty which, for the first time in twenty-five years, would be capable of wounding those who wound. A law that has never frightened anyone is about to begin doing so; the project that lightens its assessment reaches maturity exactly now. There is no need to postulate a design, and this analysis does not postulate one: it is enough to observe the order of the works. But anyone wishing to neutralise the moment at which Decree 231 becomes dangerous would not need to write anything different from what has been written.
And all of this is happening in silence. The most radical innovation in penalties in the history of the Decree has not produced a single article in the general press, not a line from the business associations that have filed position papers on every comma of the reform: the precedent that changes the nature of the law entered without witnesses, and none of those who saw it had any interest in pointing it out. That document extended the check to the specialist circle that makes its living from Decree 231: not even there has the comparison with the price of a worker’s life ever been written. From here to June 2028 the real game will not be played at conferences: it will be played in the delegation, in the transposition decrees, in the frames written into the last paragraph, where nobody looks — in January it already happened that the revolution entered without a newspaper headline, in May it already happened that a European obligation of forty million came out reduced to less than two. There it will be decided whether the precedent of January 2026 remains a foreign body to be reabsorbed or becomes the measure of all penalties; whether Decree 231 will arrive at its first appointment with a real penalty carrying an assessment that works, or the one designed here. The project never names this appointment. This analysis ends by naming it: forty million. The rest is commentary.
36. Postscript: 4 August
As these pages were being closed, the forecast of section 34 began to come true, and with a punctuality that gives no satisfaction whatsoever. On Tuesday 4 August 2026, at 18.04, the Council of Ministers, at its meeting no. 185, approved the government bill for the “revision of the rules on the administrative liability of legal persons”. Seventy days after the Confindustria assembly at which president Orsini asked “let us do it now” and the Minister replied that he was “already well along”, the project has its normative dress. It must be said with precision what has happened and what has not: the text is approved by the Government and will be presented to the Chambers; as at 5 August there is no record of a bill number, Parliament is in its summer recess, and the text of the government bill has not been published. Parliamentary scrutiny will begin in the autumn.
Since the text cannot be read, what is known of its contents is what the Government has said and the press has reported — and this section records it while declaring the register. The official words first, because they are themselves a document. The President of the Council, in statements carried by the news agencies: the Council of Ministers has approved the reform of corporate liability, “a zero-cost measure that cuts the administrative burdens borne by companies”. Deputy Minister Sisto, in the Ministry’s news bulletin: the punitive part “is flanked” by a “highly important preventive part”; “we have simplified some mechanisms”. These are the most authoritative sources possible on the nature of the operation, and they say what these pages have maintained while having to argue for it: the reform of corporate criminal liability is presented by its own author as a measure to reduce administrative burdens. The frame of the protected interests, of the victims, of the workplace deaths for which the Decree has Article 25-septies, appears in no official statement of 4 August of which there is any trace.
As for the content, the qualified reports — Il Sole 24 Ore, whose headline says the models are “suitable if they comply with the associations’ guidelines and good practices”, and the specialist agencies — indicate as survivors the devices analysed in these pages: organisational fault as a constitutive element with the burden of proof on the prosecution (sections 7 and 19); the parameter of guidelines and good practices in the judgment of suitability (section 22); what the reports call the “probation-style supervision” (messa alla prova) of the entity, that is, the reparative-extinctive device of Article 8-bis (sections 11 and 23); the simplified procedures for small and medium-sized enterprises (sections 12 and 29); and, on objective attribution in negligence offences, the criterion of a “saving in expenditure or increase in production” of an “appreciable” measure (section 6). If the correspondence is confirmed by the text, the analysis above holds for the government bill as it held for the project. The three choices that matter most remain unknown: whether the penalty delegation of Article 17 has entered the government bill, and with what criteria; which of the two options on the statute of limitations has been chosen; what has become of the catalogue. The first check to be made, the day the text becomes public, is one alone: to look for the word “turnover”.
Two final records. The first: on the day of the approval and on the following day there is no record of a single critical stance — not from the associated judiciary, not from the trade unions, not from the victims’ associations; the most severe comment available remains the civil-libertarian one of the 231 Observatory of the Union of Criminal Chambers, and even the most favourable, Assonime, had warned that the judgment on the model “in synergy” with the overall system risks reproducing the very logic it declares it wishes to overcome. The second: this section is written on press sources and official statements, and it is the only one in the document that does not rest on a readable normative text. It will be rewritten on the text of the government bill on the day of its publication, and any gap between what is reported here and what is written there will be accounted for.
Note on sources
Principal document: the Final Report and proposed draft articles of the Technical Committee (Tavolo tecnico) for the revision of Legislative Decree 231/2001, coordinator G. Fidelbo, title page dated November 2025, delivered to the Minister on 22 December 2025, made public on 14 January 2026 (Giurisprudenza Penale) and on 22 January 2026 (Sistema Penale); the text — report and draft articles — is consulted in the version published by the journal Cassazione penale, and every quotation in quotation marks is verified against that text. The draft articles carry drafting options in square brackets not yet resolved.
Other sources: Confindustria, “Prospettive di riforma della responsabilità ‘amministrativa’ degli enti”, position paper, 11 April 2025; the report of president Orsini to the Confindustria assembly, 26 May 2026, and the reply of Minister Nordio (GNews, LaPresse, Radiocor); Assonime, Position Paper 8/2026 (2 July 2026); Assolombarda, summary of the project; Osservatorio 231 UCPI (in Mondo Professionisti); Impresa e Diritto, 26 January 2026; A.C. 2632 (Chamber of Deputies, 29 September 2025) with explanatory report; the final report of the Study Commission for prevention and safety in the workplace (chaired by Sisto), 30 April 2026, delivered on 12 May 2026; Court of Cassation, Criminal Section VI, no. 143/2026; Legislative Decree 211/2025 (Article 25-octies.2); Legislative Decree 81/2026; Directive (EU) 2026/1021; Decree-Law 159/2025, converted into Law 198/2025; for the application data, Parts I and II of “The Paper Gold of Decree 231” and the sources cited there (Observatory on 231 case law of the Court of Milan, Università Statale).
On the constitutional quaestio of section 19: Constitutional Court judgments nos. 364 and 1085 of 1988; Constitutional Court judgment no. 112/2019, points 8.2.1 and 8.2.3; the ministerial report on Legislative Decree 231/2001 (on the construction of the entity’s “own act” and “own fault”); Court of Cassation, Joint Chambers, no. 38343/2014; F. Bricola, Il costo del principio “societas delinquere non potest” nell’attuale dimensione del fenomeno societario, in Riv. it. dir. proc. pen., 1970; New York Central & Hudson River R.R. v. United States, 212 U.S. 481 (1909); U.S. Sentencing Guidelines, chapter 8; Article 121-2 of the French code pénal; Bribery Act 2010, section 7; Criminal Finances Act 2017; Economic Crime and Corporate Transparency Act 2023, with the offence of failure to prevent fraud in force from 1 September 2025; European Court of Human Rights, Salabiaku v. France, 1988; Article 2049 of the Civil Code; Court of Cassation, civil, Joint Chambers, no. 13533/2001 (proximity of the evidence); for the economic-analysis objection, J. Arlen, The Potentially Perverse Effects of Corporate Criminal Liability (1994). The statement that the Strasbourg Court has censured neither vicarious regimes nor defences placing the burden on the company is formulated in the negative (“no record found”) and holds within the limits of the survey conducted.
On the approval of 4 August 2026 (section 36): the press release of the Council of Ministers no. 185 and the agenda (ANSA, 31 July 2026); the statements of the President of the Council carried by the agencies (Italpress, 4 August 2026) and of Deputy Minister Sisto in the news bulletin of the Ministry of Justice (gNews, 4 August 2026); Il Sole 24 Ore NT+ (A. Iorio, 4 August 2026); PublicPolicy (3 August 2026); the series of analyses by M. Vallone in the Giornale delle Partite Iva (9, 16, 23 and 31 July 2026). The indications as to the content of the government bill remain matters reported by the press until the text is published, and are declared as such in section 36.
The following remain reported from secondary sources, and are flagged as such in the text: the “moderately civil-libertarian” judgment of the UCPI Observatory; the circumstance that the subjective preconditions for precautionary measures requested by Confindustria were rejected in the course of the works (Assolombarda); the accounts of the Confindustria assembly. The survey of the silence of the general press and of the business associations on Article 25-octies.2 was conducted on 31 July 2026 with targeted searches, including by domain, repeated on 5 August 2026, and holds within the limits of the search engines’ indexing. Any report of error on these or other points will be verified and, if founded, incorporated with due account given.
whitecollarcrimes
a collective byline · guarantor: Avv. Luca Santa Maria · 5 August 2026