A few weeks ago, I had the good fortune to be able to attend an event at the University of Buenos Aires (co-sponsored by the New York University Law School), that focused, among other things, on a new draft bill, currently under consideration in the Argentinian legislature, that would impose criminal liability on corporations and other legal persons for corruption-related offenses. I’m largely unfamiliar with Argentina’s legal system, so I was very much an outside observer for this discussion, but there were a couple of things about the draft bill that struck me as interesting and worthy of attention from the wider anticorruption community. (Apologies for not providing a link: I’m working off a hardcopy of an unofficial English translation of the draft bill, which I can’t find on the web.)
A lot of the provisions in the bill are fairly standard, though in many respects the bill is quite aggressive. For example, Article 3 makes parent companies jointly and severally liable for sanctions imposed on their subsidiaries (without any requirement to show that the subsidiary was an agent of the parent), while Article 4 imposes successor (criminal) liability in all cases of merger, acquisition, or other corporate transformation. In both these respects, the draft Argentinian bill imposes more sweeping corporate criminal liability than does U.S. law. Also, like U.S. law, the Argentinian bill (in Article 2) would make corporations criminally liable for the actions of its officers, employees, and agents.
But what most caught my attention were the draft bill’s provisions on sanctions:
Perhaps the most striking thing, at least to me, was the fact that fines to be imposed on guilty corporations are expressed not in absolute dollar/peso terms, but rather as a percentage of the firm’s gross annual income. According to Article 8 of the draft, corporations (or other legal persons) criminally convicted of corruption offenses must pay between 0.5% and 20% of the corporation’s annual gross income (though that can be lowered a bit – more on that in a moment), in addition to confiscation of all proceeds of the crime (as specified in Article 23).
Just to get a sense of what that approach to penalty calculation would mean in monetary terms, consider Siemens, which had 2016 worldwide revenues of a little below $80 billion. If Siemens got caught misbehaving again, and the penalty scheme were as it appears in the Argentinian draft bill, the minimum penalty would be around $400 million, and the maximum penalty would be around $16 billion. These are very sizable penalties – Siemens’ actual penalties in the blockbuster 2008 bribery case, were around $1.6 billion total, divided roughly evenly between U.S. and German enforcement authorities. By contrast, consider Odebrecht. Its 2016 gross revenues were a little over $1 billion. So if this new law were in place and Argentina were to find Odebrecht criminally liable for paying bribes in Argentina, the penalty range would be between about $5 million and about $200 million. For comparison, in the real FCPA case against Odebrecht recently brought by the U.S. Department of Justice, the fine imposed on Odebrecht was $2.6 billion. Articles 9-11 of the Argentinian draft bill, quite sensibly, makes the selection of a specific penalty within these very broad ranges turn on factors like the seriousness of the crime, the extent of cooperation, and so forth. But because firms differ so much in their annual revenues, the penalties can vary drastically – note the minimum monetary penalty that could be imposed on a firm with revenues the size of Siemens’ is about twice the size of the the maximum monetary penalty that could be imposed on a firm like Odebrecht.
I have no idea whether this approach to monetary sanctions is standard under Argentinian law in other areas, or whether other countries also explicitly tie corporate sanctions to corporate revenues (or some other indicator of corporate wealth). Not having seen or thought about anything like this before, I’m not quite sure what I think of it. On the one hand, the Siemens-Odebrecht comparison above makes me uneasy. I have this intuition, which I imagine others might share, that a criminal penalty should be based on the magnitude of the crime, not the size or wealth of the criminal. And simple economic logic might support that intuition: Ideally, we want to set the expected penalty (that is, the magnitude of the penalty discounted by the probability of getting caught and punished) so that it offsets the expected gain from misconduct. As long as the potential criminal is risk-neutral (probably a fair assumption to make for sufficiently large corporations), then the wealth of the defendant is irrelevant (except insofar as the defendant’s total assets may impose an upper bound on the size of the penalty that can be imposed).
Then again, there’s something about the Argentinian approach that seems appealing (putting aside the question of whether the particular percentages are calibrated correctly). I can’t put my finger on exactly what it is. The goal of corporate criminal sanctions is (or should be) primarily instrumental—to deter misconduct and encourage appropriate investment in controls, while at the same time not chilling desirable conduct or prompting inefficient over-investment in compliance. Sanctions on corporations hit the bottom line, but the size of a sanction that will sting may depend–or so it seems–on the corporation’s size and wealth—notwithstanding the economic logic sketched above. I’m not sure if there’s an existing literature on whether/when penalties should depend on a defendant’s wealth – if there is, and any readers know about it, I’d appreciate some references! In any event, maybe part of my intuition here is that the agency problem between a firm’s shareholders and its managers means that the latter may have too great a capacity to “hide” penalties that are small relative to the firm’s overall revenues, since the impact on any individual shareholder’s returns will be small enough that it will be ignored entirely. At the same time, though, if that’s true then the gains from any individual instance of bribery will also be small enough that they may not be noticed by shareholders. Clearly, I haven’t thought through this with anything resembling rigor, but it’s an interesting enough approach that I wanted to highlight it here, and perhaps I’ll return to this issue in a future post.
Another notable feature of the sanctions provisions of the draft bill is that, according to Article 11, if the corporation implements an appropriate integrity program (as described in Articles 29 and 30 of the draft law), and if the corporation also self-discloses the offense to the authorities and fully cooperates, then the court has the discretion not to impose any fine at all (or any other penalty), notwithstanding the other provisions of the law that seem to set a minimum penalty of 0.5% of the firm’s annual gross revenues. So, while the Argentinian draft law does not include the “compliance defense” that many in the business community, defense bar, and academic community have pushed for, it does include at least a somewhat more specific, concrete provision – in the law itself – regarding what sort of integrity program the corporation needs to adopt, and what the benefits of adopting such a program would be. Also worth noting here is the fact that the benefit of having adopted such a compliance program–the reduction in the minimum penalty form 0.5% of total revenue to zero–applies only if the corporation also self-discloses and cooperates. It’s not clear to me whether advocates of the compliance defense in other contexts think that the defense should be available only under these conditions, but it seems at least on the surface like a sensible approach.
One thing about the bill that seems slightly less sensible. There’s a set of provisions in the bill (Articles 20-27) that allows the public prosecutor’s office to enter into a cooperation agreement (apparently the rough equivalent of a U.S.-style deferred prosecution agreement); that seems sensible enough. Article 22 of the draft law requires that the monetary penalty imposed under such an agreement shall not be below 0.1% of the corporation’s annual gross income. So, a cooperation agreement drops the minimum criminal fine from 0.5% of gross revenue to 0.1% of gross revenue, which seems reasonable. But – and this is the weird part – there’s no provision in Article 22 for reducing the penalty still further (including to zero) if the corporation has an integrity program, self-disclosed, and cooperated, even though according to Article 8, as noted above, a court would have the discretion not to impose any fine at all (or could impose a fine lower that 0.1% of annual revenues) if those conditions were met.
The document I’m looking at is just a preliminary proposal, so I’m assuming that this is a drafting error – but maybe not? Is there a sensible argument for limiting the minimum penalty a prosecutor could agree to impose on a corporation, even if a court would be authorized to impose an even lower penalty, or no penalty at all? Perhaps if one were concerned about “sweetheart deals” between the prosecutor and the corporation, something like this might make sense? Maybe we might want more judicial scrutiny – in the form of a trial – before we say that this corporation committed a criminal offense but only deserves a fine below 0.1% of its annual revenue? Again, I’m not sure, but I’m throwing this out there to see what other people think.
I gather the Argentinian legislature is in the process of considering and debating this bill right now. I have no idea how long the process is likely to take or what the bill’s chances are of moving forward and becoming law in one form or another. But the proposal itself contains enough interesting and possibly innovative material that it deserves attention even from those with no direct interest or stake in what happens in Argentina specifically.
The current EU anti-money laundering directive also includes a percentage approach when it comes to administrative fines for breaches of the Directive.
“maximum administrative pecuniary sanctions of at least EUR 5 000 000 or 10 % of
the total annual turnover according to the latest available accounts approved by the management body”.
Art. 59 3(a)
It’s a bit complicated because as a directive it provides guidelines for member states, rather than hard and fast rules, and countries can opt for different approaches (I think). I’ve no idea whether this has been put into practice, but it’s an interesting precedent.
Thanks for bringing this to my attention! Since I wrote the post, I’ve learned that basing corporate penalties on some measure of firm value (revenues, sales, etc.), is actually quite common, at least outside the United States. I gather, for example, that EU competition law (what in the U.S. we’d call antitrust law) sets the penalty ranges in a manner similar to what you describe for AML rules — based on “turnover.” And as Diego Moretti points out in his comment below, the Brazilian Anticorruption law sets the penalty range in a manner quite similar to the Argentinian draft law (indeed, it’s so similar that I suspect that this portion of the Argentinian bill might have been based on the Brazilian law).
So I’ll apologize for being such a parochial American that I wasn’t already aware of this fairly common practice. But I’ll also suggest that, as I note in the post, this approach might be deserving of a bit more critical scrutiny and reflection, because it seems contrary to what at least a first-cut utilitarian/economic analysis would indicate is the right way to set penalties. I’d be surprised if there isn’t already an academic research literature on this, but I haven’t had time to look into it yet. Do you happen to know what the conventional justifications are (in the EU or elsewhere) for setting penalty ranges based on the corporation’s revenues, rather than simply setting them based on the magnitude of the harm caused?
I’m afraid I don’t, sorry. I can definitely see your point about it perhaps not being the most effective way of penalising companies. However, I do think that often financial penalties are built in as the “cost of doing business”. Obviously this is somewhat different for the larger fines, but I’m not sure how much a few hundred million dollars of fines make for big companies. When I was at Global Witness, we were pushing for much greater personal accountability for senior managers. https://www.globalwitness.org/en/campaigns/corruption-and-money-laundering/banks-and-dirty-money/
Understood, but of course the upper bound on fines has the opposite effect from the one you’re pushing. If a firm is considering engaging in corruption that would $10 million in social costs, and knows it has only a 5% chance of getting caught, then to deter we’d want the penalty to be (at least) $200 million. But if the firm’s annual revenues are $500 million, then the maximum fine that could be imposed under a statute like the Argentinian draft law would be $100 million. So the fact that we want the fines to be big enough to matter doesn’t necessarily imply that we want to set maximum fines based on firm revenues, rather than egregiousness of the conduct.
And at the lower end, even those who want very strict enforcement might blanch if we’re required by statute to impose very large fines for very small violations.
I have two thoughts on setting sanctions based on the firm’s income. On the one hand, this would avoid setting sanctions that would bankrupt the company. On the other hand, it provides yet another incentive for firms to engage in fraudulent accounting–not to hide the penalties from stockholders, but rather to reduce the visible income.
The combination of giving up the illicit income plus a portion of the firm’s income is interesting, but I, too, wonder if the percentage scheme is appropriate. For example, if the illicit gains were obtained over a period of years, does it make sense to cap the other part of the sanction to a percentage of one year’s income?
The reduction in sanctions for firms with strong integrity programs that also self-report is increasingly standard and seems to work well. However, I’d like to see a provision for firms that develop a pattern of self-reporting.
I’d like to know if the bill contains an article on debarment and, if so, if the debarment would be effective at all levels of government. A particularly problematic issue here in Mexico (a federation) is that firms debarred by one state win contracts with other state and municipal governments.
All good points.
Limiting myself for the moment to the question you raise at the end: The draft bill not only authorizes debarment, but goes much further. Article 8 says that in addition to fines and disgorgement, a convicted firm may be subject (in the words of my unofficial English translation) to:
a) full or partial suspension of its activities, which in no case shall exceed ten (10) years;
b) suspension of the use of patents and trademarks, which in no case shall exceed ten (10) years;
c) full or partial publication of the conviction judgment entered against the legal person, at its expense, for two (2) days in two (2) newspapers of national circulation;
d) loss or suspension of any benefits or State subsidies granted to the legal person;
e) suspension of access to benefits or State subsidies, or of the participation in public tender or bidding processes or any other State-related activities, which in no case shall exceed ten (10) years;
f) dissolution and liquidation of the legal person; this sanction may only apply if the legal person has been created for the sole purpose of committing the crime or if the commission of crimes constitutes its main activity.
This is–for lack of a better word–awesome! It is the equivalent of a prison sentence (or in the case of part f, a death sentence) for firms–just what is needed to complement the punishment of individuals within the firm when the corruption is systemic. Thanks!
Providing more incentives for self-disclosure and self -cleasing efforts was one of key agenda of 2017 B20 Responsible Business Conduct and Anti-Corruption Cross Thematic Group’s policy recommendations to G20. Looking forward to observing various kinds of developments around the world in the future.
Click to access b20-ctg-rbac-policy-paper.pdf
The Brazil’s Anticorruption Law (Law n. 12.846/13, Article 6) sets forth administrative fines that can also range from 0.1% to 20% of the company’s annual gross revenue.*
Provisions based on annual gross revenue target the balance between at least two interests: (i) that the company will not go bankrupt because of the fine and (ii) that the company is sufficiently dissuaded from committing a second offense. The fine is adjusted to the company’s financial situation.
*However, if the annual gross revenue cannot be determined, fines shall range between BRL 6,000 (approx. USD 1,800) and BRL 60,000,000 (approx. USD 18,000,000).
Very interesting. The similarities are similar enough that, as I mentioned in my response to Robert Palmer’s post above, I wonder whether the provision in the Argentinian draft bill was based on the Brazilian anticorruption law?
I cannot make a statement about the origin of that specific provision. However, the Argentinian Executive branch explicitly mentions the Brazilian Law 12.846/13 (Clean Company Act) in the letter that introduces the draft bill to the Argentinian Congress (end of p. 9): https://www.argentina.gob.ar/sites/default/files/oa-responsabilidad_penal_personas_juridicas_-_proyecto_pen.pdf
It is quite probable, although Brazilian anti-corruption law does not deal with criminal, but administrative and civil liability. But there is an attempt by the law to reconcile the calculation based on the company’s revenue and the amount of damage caused by establishing that the fine between 0.1% and 20% of gross sales in the last year will never be less than the advantage achieved.