My post last week discussed the recent U.S. Court of Appeals decision in United States v. Hoskins, which held that a foreign national cannot be charged with aiding and abetting a violation of the Foreign Corrupt Practices Act (FCPA), or with conspiracy to violate the FCPA, unless that foreign national either took some action connected to the violation within US territory, or else acted as an agent of a US domestic concern or an issuer of securities in the US. That’s a bit of a mouthful. To put this another way: The FCPA itself says that it applies extraterritorially to US nationals (including US firms), to non-US firms that issue securities on US markets, and to the officers, employees, directors, and to agents of firms in either of the preceding categories. The FCPA also applies to foreign individuals or firms (other than issuers) if but only if they engage in some part of the wrongful conduct while in US territory. The question is whether such foreign individuals (including non-issuer firms), who act outside of US territory, and so cannot be charged directly with violating the FCPA’s anti-bribery provisions, can nevertheless be charged with aiding and abetting, and/or conspiring with, some other actor’s FCPA violation. In Hoskins, the U.S. Court of Appeals for the Second Circuit said no: Not only can a foreign national (other than an issuer or an agent of a US domestic concern or issuer) not be charged with FCPA violations based on conduct abroad, but such a defendant’s conduct abroad also cannot support a charge of aiding, abetting, or conspiring in an FCPA violation.

Perhaps because appellate court decisions on legal issues related to the FCPA are so rare, Hoskins has attracted considerable commentary. Most of this discussion, including my post last week, focuses on summarizing the court’s holding, considering its implications for future cases, and assessing whether Hoskins’ limitation of complicity and conspiracy liability is likely to improve or worsen FCPA enforcement overall. However, I haven’t seen very much commentary on the question whether, as a matter of legal doctrine and legal interpretation, Hoskins was decided correctly—that is, whether it is consistent with precedent, statutory text, and generally-accepted jurisprudential principals. That’s entirely understandable—most of the initial wave of commentary is coming either from law firms that want to explain to their clients what this decision means for them, or from those interested more in the policy issues than in parsing the doctrine. Nevertheless, I do think it’s worth getting a conversation going about whether Hoskins’ reasoning is (legally and doctrinally) sound. I may not be the best person to do this, as I’m not a criminal law specialist, but I figured I might as well take a crack at it, if only in the hopes that doing so might prompt some of the real experts to weigh in.

After reading the case a few times, and delving into some of the earlier case law and other materials, it seems to me that Hoskins is a hard case. Really really hard. And I tentatively think that is was probably decided incorrectly. Or maybe “incorrectly” is too strong—instead, perhaps I should say that the Hoskins result is in tension with existing doctrine, and the result the court reaches, though defensible, requires an aggressive expansion of traditional doctrinal principles, one that the court doesn’t really acknowledge. For those readers out there who care more about the policy bottom-line than about the intricacies of legal doctrine, you may want to stop here. Law nerds, read on!

The starting point for the discussion is the U.S. law on complicity and conspiracy. The complicity statute, codified at 18 U.S.C § 2(a), states that anyone who “aids, abets, counsels, commands, induces or procures [the commission of a federal crime] is punishable as a principal [that is, as if he or she had committed the offense].” The conspiracy statute, codified at 18 U.S.C. § 371, creates a separate offense for conspiring with at least one other person to commit a federal crime, so long as at least one of the conspirators performs “any act to effect the object of the conspiracy.” (To keep things simple, I’ll refer to these together as “complicity or conspiracy liability,” though that’s a bit sloppy, as complicity is a separate theory of liability while conspiracy is a separate offense.) The important thing to understand about complicity and conspiracy liability is that they can apply even to defendants who did not and could not themselves commit the primary offense. Thus the getaway driver for a bank robbery can be guilty of aiding and abetting bank robbery, even though the getaway driver’s conduct doesn’t satisfy the elements of the federal bank robbery offense, and a defendant who helps a bankrupt party hide her assets can be guilty of conspiring to hide those assets, even though the primary offense can be committed only by the party in bankruptcy. Complicity and conspiracy liability are thus quite expansive—as long as someone committed the underlying federal crime, then someone else who helped with, instigated, or planned the crime can be charged with complicity and/or conspiracy, even if that latter party doesn’t belong to the limited class of individuals who are legally capable of committing the primary offense.

So that’s where the analysis begins. If it were also where the analysis ended, then it would be straightforward to conclude that a defendant like Mr. Hoskins can be convicted as an accomplice and/or conspirator in an FCPA violation. After all, if the allegations against him are true, Mr. Hoskins (a foreign national) instructed executives at a US firm in how to go about making bribe payments to foreign government officials, arranged the connections with the middlemen who would handle the bribe payments, and promised to help ensure that the payments were approved by the corporation. Mr. Hoskins did not himself violate the FCPA (indeed, as a foreign national who was outside the US the whole time, he could not have violated the FCPA). But so what? As just explained, complicity and conspiracy liability do not require that the defendant did, or legally could have, committed the underlying crime. In this case, assuming the allegations are true, an FCPA violation clearly occurred—a US firm and its executives paid bribes to a foreign government official to obtain business—and Mr. Hoskins clearly both conspired with these US actors, assisted them, and indeed instructed them in the violation. Under ordinary circumstances, that would suffice to establish complicity and conspiracy liability.

The case is harder, though, because of two potential limits on complicity and conspiracy liability: the presumption against extraterritorial applications of US law, and the so-called “affirmative legislative policy” exception to complicity and conspiracy liability. The Court of Appeals that decided Hoskins thought that each of those limits was independently sufficient to find that a foreign national acting outside the US can’t be guilty of aiding or conspiring in an FCPA violation. I think that’s wrong: Neither of these limitations on its own, at least as traditionally interpreted and applied, would support the court’s conclusion. However, it’s possible that, for reasons the court seems to intuit but doesn’t fully articulate, the combination of these factors might somehow justify the conclusion that a foreign national acting abroad cannot be found guilty of acting as an accomplice or co-conspirator in an FCPA violation. I don’t think I agree with that view either, and it would seem to require breaking some new doctrinal ground, but I acknowledge that it’s not necessarily inconsistent with existing doctrine. I’ll proceed by discussing each of these possible limits one at a time, and then consider their possible combination.

Let’s start with extraterritoriality. Although the United States has the legal authority to apply its law outside U.S. territory (subject to some important limitations), the background presumption is that federal statutes only apply in US territory unless there’s a clear indication that the statute was intended to have extraterritorial effect. Neither the complicity statute nor the conspiracy statute says explicitly that it applies abroad, so one might conclude that only those who act within US territory can be charged with complicity or conspiracy. That’s a logically coherent position, but it’s not the current law. Rather, complicity and conspiracy liability are extraterritorial if (but only if) the primary offense to which they attach has extraterritorial application. (This principle also applies, by the way, to other “ancillary” offenses, such those under the “accessory after the fact” and harboring statutes and the Racketeer Influenced and Corrupt Organizations Act.) In other words, for those federal crimes that are only crimes if committed in US territory (the default position), an accomplice or co-conspirator can only be found liable if he or she also acted in US territory, but if liability for the underlying crime is extraterritorial, then complicity and conspiracy liability also apply extraterritorially. So, for example, the foreign recipient of a cocaine shipment from a US party can be charged with conspiracy to violate a US law prohibiting the export of controlled substances, and a foreign bank can be charged with conspiracy to commit money laundering for its role in a money laundering scheme, even if the bank doesn’t take any action in furtherance of the conspiracy in US territory.

Now, with respect to certain parties—US domestic concerns, issuers of US securities, and their officers, employees, and agents—the FCPA is without question an extraterritorial statute. So it would seem to follow that the prohibition on aiding, abetting, instigating, or conspiring in the commission of such a violation by one of those parties also has extraterritorial application—even if the accomplice or co-conspirator could not and did not violate the FCPA. The Hoskins court’s discussion of this issue just seems confused: The court says over and over that complicity and conspiracy liability are extraterritorial only to the extent that the underlying primary offense is extraterritorial (true) and that the FCPA itself does not apply extraterritorially to foreign nationals like Mr. Hoskins (also true), and the court then concludes from this that accomplice and conspirator liability can’t apply to foreign nationals like Mr. Hoskins. But that conclusion doesn’t follow. Under conventional doctrinal principles, which the appeals court didn’t purport to alter, as long as the statute prohibits the extraterritorial conduct by some class of actors (which it does), then other actors (including those who could not legally be charged with violating the primary prohibition) may be charged with extraterritorial complicity or conspiracy with that violation.

The second possible limitation on the broad scope of complicity or conspiracy liability is the so-called “affirmative legislative policy” exception. The idea here is that even though it’s usually the case that someone who assists in the planning or commission of a crime can be charged as an accomplice or conspirator, in some cases it’s clear that doing so would undermine a deliberate legislative choice to impose (certain) penalties only on some participants in a crime, not on others. This exception is extremely narrow, and has traditionally been invoked in only two kinds of cases (that I was able to find):

• First, there are certain criminal offenses where the consent (in the sense of agreement or passive acquiescence) of another party is usually (though not necessarily always) involved, but where the statute in question, by its terms, only criminalizes the conduct of the instigating party. Courts have reasoned that in cases like this, treating the acquiescing party as an accomplice or co-conspirator—as would seem to be allowed under the plain terms of the complicity and conspiracy statutes—would thwart the legislative judgment that those acquiescing parties are more like victims, and should not be punished. The leading Supreme Court cases on this point involve an old statute called the Mann Act, which made it a felony to transport any woman or girl across state lines for prostitution or “any other immoral purpose.” The Supreme Court reasoned that it would be improper to charge the woman or girl, who went along with a man’s arrangements to transport her across state lines for sexual purposes, with aiding and abetting a Mann Act violation. However—and this is crucial for the FCPA context—the Court was very clear that this limitation only applies when the woman, though nominally “consenting,” is not the active instigator and planner of the travel; if she is, then she can be held liable (for violating the Mann Act directly, and presumably also as an accomplice or co-conspirator). We see an application of that same basic principle—and limitation on that principle—in the context of the application of a statute called the Hobbs Act to the bribery of public officials. The Hobbs Act prohibits a state or local public official from abusing her official power to extort money or property from any person. Suppose a private party pays a bribe to a public official in exchange for some favorable exercise of the latter’s power. Can the bribe payer be charged with aiding and abetting a Hobbs Act violation (or with conspiracy to violate the Hobbs Act)? If the public official made the bribe demand, and the private party merely acquiesced, then the answer is no, on a logic similar to that applied in the Mann Act cases: Since Congress only criminalized the taking of the money by the public official, it would seem contrary to the clear design of the statute to punish the (passively acquiescing or consenting) bribe payer, as Congress appeared to consider the latter the victim. But—again drawing on the analogy to the Mann Act—if the private party in some sense instigated or encouraged the transaction, then the private party can be charged as an accomplice or co-conspirator in a Hobbs Act violation, even though the private party could not, as a matter of law, be charged with violating the Hobbs Act. (See also here, here, here, and here.) Though the Hoskins court leans heavily on this limitation, especially the old Supreme Court case on the Mann Act, this exemption doesn’t appear to apply in Mr. Hoskins’ case. The sort of complicity and conspiracy with which Mr. Hoskins was charged are not the sorts of things that are typically involved in any FCPA violation. (In that sense this case is quite different from an earlier case, United States v. Castle, which held that a foreign official who takes a bribe cannot be guilty of abetting or conspiring in an FCPA violation.) Moreover, Mr. Hoskins (again assuming the allegations are true) did much more than passively acquiesce to the FCPA violation. The idea that we shouldn’t interpret the complicity and conspiracy statutes so broadly as to punish those whom Congress appears to have considered to be victims rather than offenders would therefore seem to have no application in the Hoskins.
• Second, the “affirmative legislative policy” exception has also been invoked in the context of statutes that create special, and extremely severe, penalties for those who lead a criminal enterprise (such as organized crime “kingpins”). If courts were to apply the complicity statute as written, it would create the odd result that everyone in a criminal organization could be charged with “aiding and abetting” the head of the enterprise, and therefore all of them could be punished as if they themselves were “kingpins.” Courts have therefore understandably reasoned that when a statute seems to identify a subgroup in a criminal enterprise and single out that subgroup for special, different treatment (mainly harsher penalties), this evinces a clear legislative policy that other members of the criminal enterprise should not be subject to these different rules—even though, in contrast to the Mann Act and Hobbs Act cases, one can’t really say that the other employees of the criminal enterprise are passively acquiescing “victims.” (There’s a disagreement among various courts of appeals, which to my knowledge has not yet been resolved, about whether this principle applies only to employees of the criminal enterprise, or also to other accomplices and co-conspirators with the kingpin, such as the corrupt police who provide protection. See here and here.) This narrow exception again has no applicability to Hoskins.

So if the traditional version of the “affirmative legislative policy” exception don’t apply in the Hoskins case, how and why did the court conclude that this principle did preclude accomplice or conspirator liability for a foreign national like Mr. Hoskins? So far as I can tell, the court engaged in a novel and aggressive expansion of this traditionally narrow exception, reasoning that because the legislative history of the FCPA indicates a desire to limit liability of foreign nationals either to those who act in U.S. territory or those who act as agents for a US firm or US issuer, there was a sufficiently clear “affirmative legislative policy” to exclude foreign nationals who don’t fall into those categories from extraterritorial liability for complicity or conspiracy in an FCPA violation by some other actor who is subject to extraterritorial FCPA liability. (In addition to the legislative history, the court advanced some half-hearted and unpersuasive suggestions that the text and structure of the FCPA itself evinced a congressional intent to exclude foreign nationals from extraterritorial complicity or conspiracy liability. In particular, the court noted that the statute enumerated the particular individuals who could be held liable for FCPA violations. But as has already been discussed, under conventional principles of complicity and conspiracy liability a statute’s listing of the specific classes of individuals who can commit an offense does not implicitly limit the scope of complicity or conspiracy liability to those classes.)

On this point I suppose reasonable minds can disagree, but I didn’t find the court’s argument here very convincing. Not only does it go well beyond the traditional applications of the so-called “affirmative legislative policy” exception, but the fragments of legislative history the court cites don’t strike me as indicating a clear legislative decision that parties like Mr. Hoskins can’t be charged as accomplices or co-conspirators. On the Senate side, the court notes that an early draft of the FCPA only applied to companies, with the assumption that complicity and conspiracy liability would be used to go after individuals, but subsequently the FCPA was amended to explicitly list certain classes of individuals who could be directly liable for foreign bribery. Yet the legislative history quoted by the court doesn’t indicate that this express listing of certain classes of individuals was meant to displace complicity or conspiracy liability for others, especially given the background rule, already discussed, that the specification in the statute that certain individuals can be liable for primary violations does not imply that individuals not falling into those classes can’t be charged as accomplices or co-conspirators. On the House side, the court notes that an SEC official testified, in general terms, that applying the FCPA to a corporate agent who never enters the US could create “jurisdictional problems.” But of course the FCPA does apply (extraterritorially) to foreign nationals who are agents, and in any event it’s not clear what these “jurisdictional problems” would be, how Congress understood the concern, or whether those problems would be implicated by applying accomplice and conspirator liability extraterritorially (when the underlying crime is itself extraterritorial). It’s true that the legislative history demonstrates an understandable concern about not extending primary FCPA liability to foreigners acting outside US territory. Fine. But that’s not the question—it’s whether there’s a clear legislative policy to exempt foreigners acting outside US territory from complicity or conspiracy liability, when the primary offense is committed by a US person (or US issuer). And I’m not sure how I see, in Congress’s decision to exempt non-agent foreign nationals acting abroad from primary FCPA liability, an affirmative legislative policy that would be undermined if such foreign nationals could be charged as accomplices or co-conspirators (especially given background due process rules that would shield from liability any foreigner without any connection to the United States).

So I don’t think that either the presumption against extraterritoriality or the affirmative legislative policy exception is enough, by itself, to defeat accomplice or conspirator liability for a foreign national who takes action outside the United States to assist, plan, or support an FCPA violation by a party who is clearly covered by the FCPA’s extraterritorial jurisdiction. But perhaps it’s something about the combination of the extraterritorial aspect of the case and the indications that perhaps Congress made a policy judgment that it would be problematic to impose FCPA liability on a certain class of individuals that together justifies the result in Hoskins. That is, perhaps the usual presumption that an extraterritorial offense implies an extraterritorial application of the complicity and conspiracy statutes (in connection with that offense) can be defeated by evidence—weaker than would ordinarily be required to satisfy the “affirmative legislative policy” exception—that Congress considered and rejected applying the underlying statute itself extraterritorially to the relevant class of individuals, and that invoking complicity or conspiracy liability would thwart that congressional judgment. Maybe. But I think developing and defending this position would require much more analysis than the Hoskins court provided, and I don’t find it terribly persuasive on its face, especially since (as Judge Lynch notes in his concurring opinion) the actual outcome in Hoskins seems to lead to a result that it’s hard to imagine Congress would have wanted if it had considered parties like Mr. Hoskins.

My tentative conclusion, therefore, is that Hoskins was probably wrongly decided, should be reversed if reviewed by the Supreme Court, and should not be followed by other courts of appeals (and perhaps also disregarded by the US government in other jurisdictions). I suspect that mine will be very much a minority position among FCPA commentators, so I welcome corrections, objections, and general pushback.