There is a broad consensus that foreign bribery harms the citizens and governments of developing nations. But in most cases where enforcement agencies in a “supply side” jurisdiction (that is, the home jurisdiction of the companies that paid the bribes) reach a settlement with a company accused of bribing foreign officials, the settlement does not provide for any remedial payments to the government or citizens of the “demand side” country where the bribery took place. Given the inherent difficulties in setting right the harm corruption causes, this is hardly surprising. Nevertheless, scholars and activists have increasingly called for settlement agreements between supply side enforcers and bribe-paying companies to include requirements that the companies make such remediation to the victims of the foreign bribery scheme, and some prosecutorial agencies, like the U.S. Department of Justice (DOJ) and the U.K. Serious Fraud Office (SFO), have occasionally done something along these lines. They have done so, however, only intermittently, and as an exercise of prosecutorial discretion, without any overarching policy agenda or conceptual framework.
In a recent article, I proposed a framework that could achieve more consistent outcomes and be used as a benchmark for developing best practices. I do not focus on grand designs for a private right of action for the foreign victims of corruption, or on obligations under international law. Because the action is happening on the ground, through the exercise of prosecutorial discretion in negotiating settlements, that’s where I focus. In this post, I outline the factors that enforcement agencies should take into account when deciding whether to pursue remediation in any given case.
Before proceeding to the factors that enforcement agencies ought to consider, it is important to distinguish between two concepts of harm, and two associated concepts of victimhood—distinctions which, in turn, suggest a crucial distinction between two different forms of remediation in this context.
First, we should distinguish between direct and indirect harm. Direct harm exists where the harm suffered by the victim is directly attributable to the misconduct of the perpetrator. The deferred prosecution agreement (DPA) negotiated between the SFO and Standard Bank provides an illustrative example of direct harm in the foreign bribery context. Standard Bank’s subsidiary paid $6 million in bribes, using funds that would have otherwise been paid to the Tanzanian Government. The Tanzanian government thus suffered a direct harm as a result of Standard’s bribery: But for the bribery, that $6 million would have gone to the Tanzanian government. The DPA therefore required Standard Bank’s subsidiary to pay $6 million to the Tanzanian Government. Indirect harms, by contrast, are the diffuse adverse effects of foreign bribery, where the victims could be said to be the general populace.
This distinction between direct and indirect harm is useful in distinguishing different kinds of corruption victims. What I call “first-order” victims are those who have suffered direct harm, while those who have suffered second-order harm are “second-order” victims. These distinctions, in turn, provide us with a vocabulary for describing two different categories of remediation in foreign bribery enforcement actions: compensation and reparations. Compensation seeks to make a first-order victim whole to the greatest extent possible by providing an amount of remediation equal to the direct harm suffered by the victim. The Standard Bank DPA noted above is as an example of compensation. Reparations are intended to benefit second-order victims of corruption’s indirect harms. For example, the SFO and DOJ have included terms in DPAs and plea deals that require the defendant firm to provide funds earmarked for public infrastructure initiatives or charities in the countries where the bribery took place.
With this distinction in hand, we can consider the criteria that enforcement agencies ought to use when deciding whether to insist on compensation and/or reparations in settlement agreements.
With regard to compensation, enforcement agencies should take the following factors into account when deciding whether to provide compensation through the terms of a DPA:
- First, whether there is an identifiable first-order victim, or victims;
- Second, the extent to which the degree of loss to the victim is reasonably ascertainable;
- Third, whether there is a risk that paying compensation to the victim (especially if that victim is a foreign government) will lead to additional corruption—for instance, that the money paid might be repurposed for corrupt ends—and if so whether that risk can be sufficiently mitigated.
- Fourth, whether providing compensation would be impractical or unnecessary—an admittedly broad catch-all consideration that might include, for example, whether the victim has other available avenues for seeking relief, or is a wealthy government that has no inclination to pursue relief. (The breadth of this final consideration is necessary given that enforcement agencies need to be able to take the unique facts of each case into account.)
Because indirect harm is diffuse and difficult to quantify, the framework for reparations does not require the same degree of precision with respect to identifying victims and ascertaining the level of harm. Instead, when deciding whether to include reparations in a DPA or similar settlement, the enforcement agencies should consider the following factors as preconditions for including reparations:
- First, even though reparations are distinct from compensation payments, there ought to be some kind of nexus between the nature of the indirect harm and the form of the reparations. For example, if bribes paid to secure a monopoly over a state’s public health sector led to defective services in that sector, reparations payments should be earmarked for hospitals, medical supplies, or the health sector generally. To date, enforcement agencies that have sometimes included reparations agreements in settlements have not employed a nexus requirement. But if enforcement agencies are committed to providing reparations in a consistent and principled manner, then they ought to adopt some variant of a nexus requirement as a precondition to reparations, or at least should provide an explicit and plausible justification for declining to do so.
- Second, as with compensation payments, enforcement agencies considering the inclusion of reparations in settlement agreements must take into account the risk of repeat corruption. But assessing and mitigating this risk is more complex in reparations cases, because the agency administering reparations has to satisfy itself that funds dispersed through a publicly accessible apparatus will not become repurposed for corrupt ends. (The U.S. and U.K. exhibit different approaches to managing this risk. Settlements negotiated by the DOJ tend to impose obligations on the parties to the agreement, stipulating how reparations monies can and cannot be spent. The SFO adopts similar terms, but leaves the administration of reparations schemes to other U.K. government agencies.) Agencies might consider whether a portion of the funds used for remediation should be set aside to establish transparency and oversight measures to ensure the proper spending of reparations.
- Third, as with compensation, enforcement agencies must also consider whether reparations are appropriate in light of the totality of the circumstances.
Scholars and policymakers should consider how the international foreign bribery regime, which focuses on the supply side of corruption, might be wielded to assist those in the developing world most harmed by foreign corrupt practices—and supply side enforcement agencies have increasingly used DPAs in foreign bribery cases to advance this remedial objective. But this practice ought to operate according to a set of clearly defined principles and best practices, rather than proceeding in an ad hoc and inconsistent fashion.
GAB | The Global Anticorruption Blog 
Thanks for a thought-provoking post. I am not so sure the difference between direct and indirect harm, however, is as clear as you suggest. And although you elide it with your compensation versus reparations distinction, I am not sure that solves it.
The U.S. Mandatory Victim Restitution Act requires a court upon sentencing a defendant to award damages to “a person directly and proximately harmed” by the offense. In United States v. Simmons, 544 F. App’x 21 (2d Cir. 2013), defendant Simmons pled guilty to defrauding financial institutions by falsely claiming he was purchasing a condominium. One consequence of the fraud was that several condo units ended up in bankruptcy as he never went through with the purchases. In addition to the harms the banks suffered, the court held the company that managed the units was directly harmed as well because it incurred expense in maintaining the properties while they were in bankruptcy.
Now compare the Seventh Circuit decision in U.S. v. Shepard, 269 F.3d 884 (7th Cir. 2001). There a hospital employee drained the bank account of one of the hospital’s elderly patients. The was sued by the patient’s estate, settled and then recovery of the settlement at defendant’s sentencing. The court held it was not “directly harmed” by defendant’s crime.
I have spent a good deal of time sorting through the cases looking for some principle or even rough guide to when a harm is “direct” and when it is not. A search motivated by the hidden debt case in Mozambique. Large bribes were paid government officials in return for which they borrowed some $2.1 billion for projects of little or no value. When the scheme was revealed, the economy dipped, and millions were thrown into poverty. Was the harm they suffered a “direct” or “indirect” consequence of the hidden debt? Or as some continental writers would put it, was it a “direct” effect or a “ripple” effect?
The closest to a principle I have found distinguishing direct from indirect harm is commentary on tort law. Most tort cases arise from circumstances where the damage was unintentional – accidents caused by malfunctioning factory machinery, motor vehicle collisions, defective consumer goods. Practical considerations and a sense of justice have led courts to develop rules of duty, causation, and damages to avoid imposing a crushing liability on a defendant. One word they hang their hat on is “direct.” If they think liability is not warranted, perhaps because they think it would retard economic development (industrial accidents) or because it is out of proportion to the degree of fault (car crashes), they say claimant’s damages are “indirect.” Other terms for evading liability are in England a finding that the damage is “too remote” or “not foreseeable.” American courts will sometimes say the defendant didn’t owe the injured party a “duty” or the damage was not “proximately” caused by the defendant’s conduct.
Corruption is a deliberate, intentional, harmful act, and doctrines applicable to unintentional acts therefore have no place. Students of tort law find the idea is reflected in common law court decisions. One is the imposition of strict liability for certain actions. Virtually any harm such actions cause is compensable. The classic example is the use of explosives to demolish a building.
Where a defendant’s actions don’t give rise to strict liability in tort but are nonetheless deliberate and harmful, a court will stretch to find that the defendant’s actions “caused” the harm or that the harm was foreseeable or “direct.” The closest that idea has ever come to being stated explicitly in American law, at least that I have found, appears in section 501 of the Restatement (Second) of Torts —
“The fact that the actor’s misconduct is in reckless disregard of another’s safety rather than merely negligent is a matter to be taken into account in determining whether a jury may reasonably find that the actor’s conduct bears a sufficient causal relation to another’s harm to make the actor liable therefore.”
In Revenue and Customs Commissioners v Total Network SL, Lord Scott made a similar point. What distinguishes the tort of lawful means conspiracy from an unlawful means conspiracy is the egregiousness of the conduct.
“The difference … demonstrates the essential flexibility of the action on the case [the origin of both torts]. It is not all conduct foreseeably likely to cause, and that does cause, economic harm to another that is tortious. . . . . The circumstances must be such as to make the conduct sufficiently reprehensible to justify imposing on those who have brought about the harm liability in damages for having done so.”
André Tunc, a leading authority on French law, finds the same reciprocal relationship between causation or directness and culpability in French tort cases. The more culpable the defendant, the less of a causal connection between the act and a resulting injury a French claimant is required to obtain compensation.
By the way, the link to your article isn’t working. Anxious to read it.
Great comment. Thank you for reading. Apologies about the article not working, try here: https://cjil.uchicago.edu/publication/remediation-foreign-bribery-settlements-foundations-new-approach
I appreciate that the doctrine surrounding direct vs indirect harm in criminal restitution and tort law cases is at times unclear and seemingly inconsistent. It’s important not to get too bogged down in the label courts affix to a particular legal remedy, and to remember that prosecutorial discretion is the driving force here rather than jurisprudence. Prosecutors, using the tools outlined in the Article, are able to make a judgment call as to whether a particular type of harm is direct or indirect. The direct harm analysis is appropriate when dealing with identifiable victims, while the indirect harm analysis is appropriate when attempting to deliver remediation to large numbers of victims/an entire populace. Transplanting private and criminal law jurisprudence would lead to the imposition of impracticably rigorous standards on enforcement agencies – which they have neither the time, resources, nor inclination to adhere to. I opted to provide a straight forward definition of harm, which enforcement agencies can apply at their discretion. When you actually look at the type of harm that agencies are trying to remediate through DPAs, you see that the direct/indirect analysis is quite straight forward.
Admittedly, some alterations might be needed if I were minded with formulating a statutory cause of action for the victims of foreign bribery – that would indeed require a more rigorous account of causation – but that’s not what I am trying to do. I am simply trying to suggest ways to improve the practices of enforcement agencies.
Additionally, I wouldn’t discard tort law as an inapposite point of analogy altogether. Tort law theory encompasses more than just negligence. The corrective justice theory I refer to in developing my theory of compensation, while a creature of private law scholarship, is by no means limited to unintentionally caused harm. I also disagree that the mental culpability of a corrupt wrongdoer distinguishes foreign bribery remediation from tort law entirely as well (at least for the sake of thinking about remediation). Save for a few exceptional legal doctrines, the bulk of both criminal and private law jurisprudence regarding legal remedies focuses on the harm suffered by a victim, with relatively little regard given to the mindset of the defendant.
As to Simmons and Shepard, my direct/indirect dichotomy is wholly irrelevant to these two cases. What I am concerned with is formulating an approach that agencies can readily and consistently adopt in foreign bribery enforcement actions being resolved via consensual means. Enforcement agencies are not bound by the rigours of MVRA doctrine in this context. However, if for the sake of example, I were to attempt to apply the logic of my article, I would suspect that an agency would find that the victims in these cases would have suffered direct harm. Any enforcement agency minded to pursue remediation would have no difficulty applying a basic “but for” analysis. Accordingly, compensation would be the appropriate remedy in both cases. Indeed, the identity of the victims in these cases is known, and their loss is ascertainable.
Moving to the Mozambique case you mentioned, that would clearly be indirect harm – and it is also a more appropriate case study for the application of my conceptions of harm. The causal analysis would resemble a “ripple effect”, with a diffuse number of victims, and harm would be indirect.
Very interesting post, Sam. Given the concern you highlight that reparations payments may be misappropriated or used for corrupt ends, I wonder if an alternative would be to direct the payments to good government or anti-corruption organizations in the victim country. Of course using the money in this way would be more attenuated than if the funds went to the victimized sector (e.g., public health) — still, it’s better than nothing.
I agree completely Sam. I suggest something similar in the article for those circumstances where it is not possible or practicable to deliver reparations to the sector of society that suffered harm.
Pingback: Episode 242 – the Lion or Lamb edition | The Compliance Podcast Network
Pingback: This Week in FCPA-Episode 242, week ending March 5, 2021 – the Lion or Lamb edition - Compliance ReportCompliance Report
Sam, this is a fascinating post. Some thoughts –
First, on the framework you suggest for first order harms, I wonder if it risks creating the wrong incentives. Corrupt governments plausibly get to take a bribe and then also secure compensation. One of the limiting factors you suggest – the risk of additional corruption – is hard to accurately measure. A corrupt government can easily project blame onto a few officials for instance.
Second, your compensation / restitution distinction brings to mind the classic contract damages categories. If we have a case of bribery in which all of the criteria you lay out are satisfied, is your definition of compensation damages limited to just making a party whole or putting them in the position they’d have been if there had been no bribery? (expectation damages) While this is a broader definition, it is also perhaps a fairer outcome.
Hi Disha.
Thanks for your comments.
Regarding the first point, whether there is a risk of repeat corruption (and more importantly, whether that risk can be managed) would be determined by an enforcement agency. Presumably, the agency’s investigation would produce evidence enabling it to determine whether the bribe recipient had been a lone operator or not – rather than simply taking the word of the relevant government. Furthermore, World Bank, FATF, and Transparency International rankings could also inform such a determination if there was no evidence from the investigation. Ultimately, the agency needs to convince itself that there is no risk of repeat corruption in dispersing remediation – and I don’t think the agency would be reliant on representations made by the government in doing that.
With that said, the Article talks about managing the risk of repeat corruption rather than simply detecting it. The presence of risk itself is not dispositive here – the key question is whether that risk can be effectively managed. Risk can be managed even if there is a broader culture of corruption by tying the remediation to accountability measures (such as the use of rigorous external audit – as was done in the James Giffen case). The issue is obviously fact-intensive and fraught and requires a fair degree of nuance depending on whether one is dealing with compensation or reparations.
Regarding the second point, the distinction arises from private law generally – my versions of compensation and reparations are an offshoot of Weinribian conceptions of tort law damages, while restitution as I employ the term is an offshoot of criminal law restitution and private law unjust enrichment. Compensation in contract does operate off of the “pre-harm state” principle but it is a much narrower concept than the tort law theory I employ. And the account of compensation offered is loss-based, not expectation-based. A broader concept, such as expectation damages, is far too amorphous for an enforcement agency to approach through DPA negotiations.
Thanks for this post Sam! It was rather delightful and thought provoking. Your proposal for supply side enforcers to also provide compensation to the government or citizen of the demand side is rather fascinating. However, for indirect harms I wonder why we need a framework. Why not just have a strict liability process wherein any company caught bribing an official or private company in a foreign jurisdiction also earmarks funds for public infrastructure initiatives in such country. This way, it serves as a stronger deterrent against foreign companies paying bribes to government officials or private companies in a developing country irrespective of their rationale for doing so.