The Science of the Crimes of the Powerful

To understand what is missing here, it is worth looking where it exists. In the United States — and in the anglophone world that gravitates around them — there has existed for decades a science of corporate crime: not a doctrine of the defence, but a criminology that observes how large corporations offend, measures their harm in the field and derives models from it. Models that serve, in the end, one concrete purpose: to help those who judge to read the facts of a disaster or of a great corruption case. I follow it here along two threads that concern us closely — environmental crimes and international corruption — because it is there that that science has given its best; and I follow them up to the present day, because both subjects breathe with the breath of politics, and today — in the United States of Donald Trump's second term — that breath is held.

I. A mature science

The starting point has a date, 1939, and a sentence. From the platform of the American Sociological Society Edwin Sutherland coins white-collar crime and defines it: a crime "committed by a person of respectability and high social status in the course of his occupation".1 It is not a provocation: it is the breaking of a scientific prejudice, the idea that crime is the business of the poor. Sutherland shows that crime also dwells, and with far greater profit, on the upper floors; and that it is learned, like a trade, inside the organisation that practises it. Even then, in his list of white-collar crimes, there appeared commercial bribery and the bribery of public officials "in order to secure favorable contracts and legislation".1 The two threads of this essay are, in embryo, already there.

From that intuition the discipline becomes empirical. In 1980 Marshall Clinard and Peter Yeager publish the first great systematic study: five hundred and eighty-two large American corporations, more than one thousand five hundred federal proceedings in just two years; around 60 per cent of the firms with at least one legal action against them, and a recidivist minority — 8 per cent — responsible for more than half of the violations.2 It is the "tip of the iceberg", they write; and among the six categories they analyse there is already that of environmental violations. The methodological message is that corporate wrongdoing is not the anomaly of a few bad apples, but a recurrent phenomenon, concentrated by sector and by pressure for results — something to be counted, not deplored.

It then becomes organisational. Studying the Space Shuttle Challenger disaster, Diane Vaughan coins the formula that more than any other serves a jurist: the normalization of deviance. No one at NASA decided "to do evil"; a chain of individually innocuous decisions shifted the threshold of the acceptable step by step, until the risk became invisible from within.3 Disaster as a product of the structure, not as the fault of a single villain. It is the exact reverse of the rhetoric of the "isolated human error"; and it is, as we shall see, a key to organisational fault. Vaughan adds to it a second concept, structural secrecy: in large organisations knowledge is fragmented across levels and functions, and no one possesses the whole picture — so that the absence of an omniscient culprit is not proof of innocence, but a physiological effect of the structure.3

Onto this trunk are grafted the orientations that explain the why. The theories of strain and institutional anomie: a culture that places economic success above every counterweight makes wrongdoing a rational shortcut.4 The opportunity perspective: the white-collar offender commits crime by exploiting the legitimate access his profession gives him — the separation between offender and victim, the appearance of legality.5 The theory of the techniques of neutralization: the rationalisations that defuse the moral brakes before the act — "I didn't hurt anyone", "everybody does it", "I was doing it for the company".6 And Kagan and Scholz's model of the "amoral calculator": the firm that violates the rule when the expected sanction is lower than the cost of compliance — a calculation, not a pathology.7 James Coleman would hold these threads together in an integrated theory of white-collar crime, linking motivation (the culture of competition) to opportunity (the structure of the role).8

Then there is the wing that asks how to respond, again with empirical method: Ayres and Braithwaite's "responsive regulation", the pyramid that starts from persuasion and climbs to severe sanction only when the firm persists.9 And there is the verification, uncomfortable for many commonplaces, of the empirical studies on sentencing: the great Yale project on white-collar offenders showed, from the federal case files, that the offenders are for the most part not extremely rich but middle-class, and that judges, far from being lenient, often justify the sentence precisely by the social harm done.10

Because this, above all, is the discipline's point of view. When it speaks of social harm — all the way to zemiology, the proposal to study harm instead of the "crime" defined by the powerful — it adopts the gaze of those who suffer the harm: the workers, the sick, those who drink the poisoned water, the plundered peoples.11 Many of the worst forms of conduct are "lawful but awful": a science that looked only at what is criminally prohibited would not see them. That is why it works on the harm, not on the label; and that is why it has behind it a method that gathers data where data exist — the archives of the Environmental Protection Agency, the federal pre-sentence reports, the databases on corruption enforcement.12 It is an institutionalised field — a dedicated journal, handbooks, a scientific division of the American Society of Criminology, multi-year empirical programmes13 — and along its two most mature strands it shows its strength better than anywhere else.

II. First thread: environmental crimes

The criminology that studies harm to the environment has a name and a date of birth: green criminology, an expression proposed by Michael Lynch in 1990 as a development of critical criminology and its political economy.14 In a generation it has become a field with handbooks, journals and an abundant literature — from Rob White to Nigel South, from Piers Beirne to Avi Brisman.15

Its first move is theoretical and sharp: to shift the focus from the crime to the harm. The legalist approach defines environmental crime as what the criminal law prohibits; the harm-based approach — today prevalent — studies conduct "harmful to human beings, environments and non-human animals, regardless of its legality".16 It is a methodological choice heavy with consequences, because many of the worst environmental harms are, at least in part, lawful: authorised, tolerated, under-regulated. Green criminology also widens the notion of victim — to ecosystems and non-human animals: Piers Beirne has proposed a "non-speciesist" criminology and the concept of theriocide, while Rob White articulates three forms of justice: environmental, ecological and species justice.17

The underlying explanation is one of political economy, and it is called the treadmill of production, theorised by Allan Schnaiberg and brought into criminology by Stretesky, Long and Lynch.18 The capitalist economy is structurally expansive: to grow, it must extract ever more resources and discharge ever more waste; environmental harm is not an accident along the way but a systematic product, a cost offloaded onto the community while the profit remains private. The environment is treated as an externality; and an externality, in criminological language, is a harm that someone pays in the place of the one who produces it. Onto this basis is grafted the debate on ecocide, for which a panel of jurists proposed in 2021 a definition — "unlawful or wanton acts committed with knowledge that there is a substantial likelihood of severe and either widespread or long-term damage to the environment" — destined for a possible fifth crime under the Rome Statute.19

The point of view, once again, is that of those who suffer. Not by chance does green criminology converse with the movement for environmental justice, born in the United States from the Warren County protests (North Carolina, 1982) against a toxic-waste landfill in an African-American community: from there the notion of "environmental racism" and the studies of Robert Bullard.20 Green criminology has translated it into research: Stretesky and Lynch have shown that poor minorities live closer to chemically hazardous plants; and that the particulate matter from coal-fired power stations causes more deaths in the United States than homicides do — a diffuse "corporate violence", measured like any other cause of mortality.21

A case that explains everything: the story of PFAS

No case better illustrates the grammar of corporate environmental crime: the long story of the perfluoroalkyl substances — PFAS, and in particular PFOA, "C8". It is proof positive that harm can be at once lawful for decades and known from the very beginning by those who produce it.

It all begins by chance, on 6 April 1938, when a DuPont chemist, Roy Plunkett, discovers polytetrafluoroethylene: Teflon.22 From 1951 DuPont, at its Washington Works plant in Parkersburg, West Virginia, uses PFOA purchased from 3M to produce it.23 Knowledge of the risk arrives almost at once: as early as 1961 DuPont's in-house toxicology laboratory establishes that C8 is toxic to the liver of rats and recommends that it be "handled with extreme care", avoiding contact with the skin.24 In the 1970s the substance is found in the blood of the workers — at extremely high concentrations — and 3M concludes internally that PFOA and PFOS "should be regarded as toxic"; but does not report this to the federal agency.24 In 1984 DuPont knows that C8 is by now in the public drinking water of the area; it sets an internal guard threshold and discovers that it is exceeding it by around a hundred times.25 It is the normalization of deviance in slow motion, on the scale of half a century: every threshold exceeded becomes the new normality, and silence toward the regulator is the organisational choice that keeps the system standing.

The story becomes public by the most unpredictable route: a farmer's cows. Between 1998 and 1999 Wilbur Tennant, whose animals were dying next to a DuPont landfill, turns to an Ohio corporate lawyer, Robert Bilott, who crosses from the defence to the prosecution.26 Discovery brings to light decades of internal documents; the class action Leach v. DuPont is born, which in the early 2000s closes with a settlement and — above all — with a scientific clause without precedent: a panel of independent epidemiologists, the C8 Science Panel, studies for years a population of nearly seventy thousand people and in 2011-2012 establishes the "probable link" between exposure to C8 and six diseases, including kidney cancer and testicular cancer.27 In 2017 DuPont and its heir Chemours close around three thousand five hundred individual lawsuits with a settlement of $670.7 million.28 3M, which was the producer of the PFOA, had meanwhile begun the phase-out in 2000; in 2018 it settles with the State of Minnesota for $850 million.29

The true measure of the harm, however, arrives with the drinking water of half the country. In the great multidistrict litigation gathered in South Carolina, in 2023 3M agrees to pay up to around $10.3 billion to the public water systems for PFAS remediation, and the DuPont/Chemours/Corteva group a further $1.18 billion.30 They are among the largest environmental settlements in American history — and, note, they remain settlements: one pays to close the case, without a full finding of liability.

On the regulatory front, the State's race to catch up with the harm is exceedingly slow. After years of merely non-binding "reference values" — 400/200 parts per trillion in 2009, 70 in 2016, levels close to zero in 2022 — only on 10 April 2024 does the EPA set the first binding federal limits for PFOA and PFOS in drinking water: 4 parts per trillion, with an ideal goal of zero.31 And, again in 2024, it designates the two substances as "hazardous" for the purposes of the remediation statute (CERCLA), opening the way to the recovery of costs from those who polluted.32 Half a century after 1961.

This story hands the jurist a precious template, which green criminology generalises under the name of the manufacture of doubt: David Michaels and, on the historical side, Naomi Oreskes and Erik Conway have documented how entire industries — from tobacco to asbestos to PFAS — deliberately produced scientific uncertainty in order to delay regulation.33 The template has three uses in court: it distinguishes genuine uncertainty from manufactured uncertainty; if the firm knew the risk from its own internal studies, it defeats the defence of unforeseeability (subjective foreseeability is proved by its own documents); and concealment, being a deliberate organisational choice, reinforces organisational fault and illuminates the causal link. The PFAS story — knowledge since 1961, silence toward the regulator, "lawful" harm for decades — is that template in its pure state. (It hardly needs saying how much it resonates with our stories at home: PFAS are also the substance of the largest environmental disaster in Italy.)

III. Second thread: international corruption

The second thread springs from the same stock. Corruption has been inside the concept of white-collar crime from the origin — Sutherland was already listing it in 1940 — but it is recent criminology that has studied it as an organisational phenomenon, not merely as a question of bad morals or as a problem for economics.34 The great political-economy tradition of Susan Rose-Ackerman, which explains corruption through incentives and the failures of the State, must be kept distinct here: most useful, but of a different cut from a criminology that looks at the firm as offender and at the organisation as the machine of the crime.35

The reference work on the criminological side is by Nicholas Lord, who has shown how anticorruption regulation, even in countries equipped with severe laws, tends toward a "default position of accommodation": one negotiates, one settles, and "reaching a settlement is usually the most that is achieved".36 With Michael Levi, Lord then distinguished the finances for corruption — how the money of the bribes is generated and moved — from the finances from corruption, the concealment of the proceeds.37 And Liz Campbell has brought into focus the most advanced frontier of liability: the offence of "failure to prevent" introduced by the British Bribery Act of 2010, which punishes the company for not having prevented bribery committed in its interest, unless it proves that it had adopted "adequate procedures".38

The explanations are the very ones of the criminology of the corporation, brought down onto corruption: the anomie of the pressure for results, the opportunity structure of those who deal with governments, the neutralisations ("it's the cost of doing business there"), and above all the role of the intermediaries. In its 2014 report the OECD found that in three international bribery cases out of four an intermediary was involved — agents, consultants, vehicle companies — the diaphragm that puts distance between the firm and the bribe.39 It is exactly the opportunity structure that criminology describes: legitimate access, separation between offender and victim, the appearance of legality.

The critical heart of the thread is enforcement and its limits. The formal instruments are imposing: the United States Foreign Corrupt Practices Act of 1977, born of the post-Watergate scandal of illicit foreign payments; the OECD Anti-Bribery Convention of 1997, with its Working Group on Bribery which evaluates the countries in rotation.40 And yet empirical criminology shows a systematic gap between the law on the books and the law in action. Brandon Garrett, in Too Big to Jail, documented the rise of the agreements that suspend or exclude trial (the deferred and non-prosecution agreements): the company settles, pays, and its executives almost never stand trial.41 Mike Koehler speaks openly of a "façade of enforcement"; Lord and King have shown how prosecution, more and more, is "deferred" or in effect exempted.42 And Transparency International's data record, year after year, that the great majority of world exports come from countries which, in practice, do not prosecute foreign bribery.43

Here too the point of view is that of the victims — but of distant and silent victims: the peoples of the plundered countries. The literature on grand corruption defines it as the abuse of power at high level that benefits the few at the expense of the many, up to the violation of human rights; and a line of scholars — from Ndiva Kofele-Kale, who spoke of "patrimonicide", to Anne Peters, on the nexus between corruption and human rights — now inscribes it within the international law of human rights.44 There even exists an architecture for returning the spoils: the UN Convention against Corruption, with the principle of asset recovery, and the StAR initiative of the World Bank and UNODC.45 Not by chance the most exposed sector is the extractive one, oil and gas: the 2014 OECD report places there the highest share of all cases — the point where the enormous rent meets the direct relationship with governments, and the moment of the award of licences becomes the crossroads of the bribe.46

The textbook cases say it better than any theory. Siemens, in 2008, settled with American and German authorities for more than one and a half billion dollars, over a structured system of slush funds and "cash desks" for the systematic payment of bribes — not the initiative of a disloyal executive, but a corporate procedure.47 Nigeria's Bonny Island, the TSKJ consortium — with an Italian name inside it too, Snamprogetti — brought KBR/Halliburton to a $579 million settlement in 2009, over hundreds of millions in "consultancies" partly destined to bribe Nigerian officials.48 Odebrecht, the Brazilian construction colossus, was described by the American Department of Justice in 2016 as the largest foreign-bribery case in history, with a dedicated "bribery department" inside the company.49 And BAE Systems, over the enormous Saudi arms contract Al-Yamamah, offers the limit case that carries us to the next thread: the investigation by the British Serious Fraud Office was dropped in December 2006 on grounds of "national security".50 Not the company escaping the State: the State halting the investigation into its own strategic company.

IV. The knot: the State that lets it happen

Here the two threads tie into a knot. American criminology has a name for this too: state-corporate crime, the crime that arises at the intersection of government and business. Ronald Kramer and Raymond Michalowski coined it at the end of the 1980s, and defined it as "illegal or socially injurious actions that occur when one or more institutions of political governance pursue a goal in direct cooperation with one or more institutions of economic production and distribution".51 Corporate crime is no longer a private matter that the State represses from outside: it is something that State and business produce together.

The decisive distinction is between two forms. There is state-initiated corporate crime, when the firm deviates "at the direction of, or with the tacit approval of, the government"; and there is state-facilitated corporate crime, which is ours: it occurs "when government regulatory institutions fail to restrain deviant business activities, whether through direct collusion or because they share goals whose attainment would be hampered by aggressive regulation".52 The State, here, does not order: it omits. It does not see, does not inspect, does not enforce; it lowers the standards, wears down the supervisory agencies, shares with the firm the end — growth, the national champion — to which the criminal law becomes an obstacle.

The framework is at once organisational and political-economic. Kramer and Michalowski propose analysing each case on three levels — institutional, organisational, interactional — crossed with three catalysts: the push toward goals, the structure of opportunities, the operativeness (or inertia) of the controls.53 The structural mechanism that makes it all possible has a name that comes from economics: regulatory capture. As early as 1971 George Stigler showed that an agency born for the public interest tends to be captured by the industry it is supposed to oversee.54 A captured regulator is precisely the instrument of the facilitator State: by under-regulating and not enforcing, it renders lawful — or in practice unpunished — what it ought to prevent. To measure its degree, Kauzlarich and others have proposed a complicity continuum, from action to omission, from the explicit to the implicit.55

The cases compose an almost didactic typology. The Challenger is the model of the state-initiated form: NASA, under political pressure over the launch calendar, presses the contractor to authorise the cold-weather launch despite the engineers' alarm.56 The fire at Imperial Food Products in Hamlet, North Carolina, in 1991 — twenty-five workers dead behind barred emergency exits, in a plant never inspected in eleven years — is the manifesto model of the state-facilitated form: not the State that orders, but the State that, in a deregulatory climate, does not inspect.57 Flight ValuJet 592 (1996) shows the agency, the FAA, divided between two conflicting mandates: to regulate safety and at the same time to promote the industry.58 And Deepwater Horizon, BP's disaster in the Gulf of Mexico in 2010, is state-facilitated environmental crime in its pure state: the regulator, the Minerals Management Service, combined within itself the collection of royalties, the promotion of drilling and the supervision of safety — a conflict so structural that, after the disaster, the agency was broken up into three separate bodies.59 The same pattern was seen, in financial form, in the crisis of 2008: deregulation as a criminogenic environment, the control fraud of those who defraud from inside the firm, and the almost total absence of prosecutions at the top, in stark contrast with the more than one thousand convictions that followed the savings-and-loan crisis.60

The theoretical frame is explicit. David Whyte speaks of regimes of permission: corporate harm is co-produced by the State, which actively creates its conditions.61 It is the criminological translation of the "national champion" — the private firm to which the State guarantees a dominant position and protection in the name of the "national interest". Penny Green and Tony Ward inscribe it within the broader state crime as "organisational deviance involving the violation of human rights";62 and a line that runs from Frank Pearce to Steve Tombs and David Whyte draws the lesson from it: the selectivity of the criminal justice system, when it meets power, turns over into impunity — not by accident, but by structure.63 It is at this level that the two threads reveal themselves to be the same thread: the corruption that protects the national champion (the BAE investigation halted "for national security") and the disaster made possible by the captured regulator (Deepwater Horizon) are two faces of the same figure.

V. The political wind: the measured retreat (2025-2026)

There is one last thing, and it is the most topical. If the repression of corporate crime is, as we have seen, a function of the will of the State more than of the texts of the laws, then it follows the political wind — and today, in the United States, the wind has turned. The laws remain standing; what changes, and can be measured, is the decision to apply them. It is the living proof of this essay's thesis: the retreat of 2025-2026 almost never passes through the legislature, but through acts of the executive — orders, circulars, staff cuts, budgets.

On international corruption the turning point has a date. On 10 February 2025 the president signs an executive order that suspends the application of the Foreign Corrupt Practices Act: a pause of one hundred and eighty days on new actions, re-examination of all pending matters, revision of the guidelines.64 It is, note well, an executive withdrawal: the law is not repealed, but it ceases to be applied. The new directives of the Department of Justice, of June 2025, reorient enforcement toward different priorities — cartels, national security, harm to American companies — and centralise its decisions.65 The result is measurable: in 2025 the Department of Justice's FCPA actions collapse to a handful — around seven, against a historical average close to thirty-six a year — the SEC opens none and dismantles its own anticorruption unit, and around half of the pending investigations are closed.66 The OECD and Transparency International openly denounce the retreat of the country that had been the standard-bearer of anticorruption; in 2026 only a selective resumption can be glimpsed, with a few settlements and a few individual convictions.67

On the environment the dynamic is identical, and it touches precisely the PFAS. In May 2025 the new EPA announces that it will keep the limits of 4 parts per trillion for PFOA and PFOS, but extends the compliance deadline from 2029 to 2031 and rescinds or reconsiders the limits just set for the other four PFAS: the known harm remains, the protection slips.68 Around it, a broader retreat in the enforcement of environmental laws: dozens of deregulatory actions announced in a single block, circulars forbidding inspectors to concern themselves with environmental justice, a cut in the agency's staff of the order of one employee in three, and, in February 2026, the revocation of the historic endangerment finding of 2009, the legal foundation of the regulation of greenhouse gases.69 The most eloquent measure is supplied by an independent watchdog: in the first year of the new administration the Department of Justice's new civil environmental lawsuits fall by more than seventy per cent, settlements collapse, a third of the federal environmental lawyers leave.70 The agency claims a "record year" of collections, but the figure is inflated by a single large case: behind the headline, the enforcement of the law is in retreat.

It is the experimental confirmation of everything else. Corruption and pollution do not suddenly become lawful: they remain prohibited on paper. But the State chooses not to see them — and in doing so it passes, before our eyes, from the figure of the bulwark to that of the facilitator. The criminology we have followed does not merely predict it: it counts it.

VI. Templates for the judge

It is worth, at this point, drawing the threads together on the side that matters to a jurist: what, of all this, is of use in a courtroom. The answer is that an empirical criminology of corporate crime offers not opinions, but templates for reading the facts — and the facts, in corporate trials, are the hardest thing.

The normalization of deviance offers a template for causation and organisational fault: it explains how a catastrophe can be the product of accumulated defective micro-decisions, and takes away from the defence the shortcut of the "isolated error". Structural secrecy explains why the fragmentation of knowledge across levels does not amount to the innocence of the entity. The manufacture of doubt, with its documentary moulds — long-standing knowledge, concealment, disclaimers, the PFAS story — gives a grid for assessing the subjective foreseeability of the risk and dismantling the unforeseeability of convenience. The amoral calculator helps to read the motive: wrongdoing as an economic choice when the expected sanction is too low. Regulatory capture explains why an administrative authorisation, or the absence of sanctions, does not mean absence of harm or of fault — indeed, they may be its symptom. And the techniques of neutralization give a key for reading emails, minutes and defence briefs, distinguishing genuine unawareness from strategic rationalisation.

They are instruments of interpretation, not shortcuts to conviction. They make thinkable — and therefore ascertainable — facts that the dogmatics of individual causation alone tends to dissolve. It is the difference between a judge who has a theory of the facts at his disposal and a judge left alone before a disaster that "no one could have foreseen".

VII. The rest, and why it matters

Much remains, and I say so for the honesty of the map: American criminology of corporate crime is other things too — cultural criminology, social learning theory, the vast discussion of compliance and its limits. They are important branches, but here they have stayed in the background, because the design was another: to show that, where that science is alive, it observes, measures and names the crime of the powerful from the side of those who suffer it; that its two most mature strands — the environment and international corruption — converge at the point that concerns us most, the State that protects the company instead of the citizens; and that, precisely because all of this depends on political will, it can be swept away in a year, as is happening.

This is not erudition. It is the toolbox that we lack, and it decides whether a disaster or a great corruption case remain "facts that do not subsist" or become, at last, readable. Because the first way in which economic power becomes unpunished is not to bribe a judge: it is to ensure that no one — in the universities and in the courts — any longer has the tools, and the will, to say what happened.