Given everything else that’s happening related to corruption right now (much of it awful), perhaps it’s a mistake for me to be spending so much time thinking about fairly narrow doctrinal issues related to applications of the U.S. Foreign Corrupt Practices Act (FCPA). But my reflections on the recent court of appeals decision in US v. Hoskins (which held that a foreign national could not be charged as an accomplice or co-conspirator in an FCPA violation based on conduct occurring abroad) have gotten me thinking about—and questioning—what I had assumed was a well-settled and straightforward conclusion that the foreign official who takes a bribe from a person or entity covered by the FCPA cannot be charged with aiding and abetting, or conspiring to commit, that FCPA violation.

That conclusion—that bribe-taking foreign officials may not be charged as accomplices or co-conspirators in FCPA cases—was announced by a US court of appeals in 1991 in a case called United States v. Castle. In Castle, according to the allegations (which for present purposes I’ll assume to be true), two private US businessmen paid a $50,000 bribe to two Canadian government officials in order to win a contract to provide public buses to the provincial government. The US government charged the American citizens with violating the FCPA—which, if the facts are as alleged, they clearly did. The Canadian officials cannot directly violate the FCPA, which by its terms prohibits only covered entities from giving (or promising or offering) bribes to foreign public officials; the FCPA does not criminalize the act of taking a bribe. But in the Castle case, the US government tried to get around this problem by charging the Canadian officials with conspiracy to violate the FCPA, pursuant to the federal conspiracy statute, codified at 18 U.S.C. § 371. That section makes it a separate crime (“conspiracy”) for “two or more persons [to] conspire … to commit any [federal] offense,” as long as “one or more of such persons do any act to effect the object of the conspiracy.” According to the U.S. government’s theory of the case, once the Canadian officials agreed with the US businessmen to accept money in exchange for a public contract, they had all conspired to commit a federal crime, and once the US businessmen took action in furtherance of this conspiracy (by paying the money), all the parties, including the Canadian officials, were liable as co-conspirators. The US district judge rejected that theory, and the court of appeals affirmed, simply endorsing and reprinting (with one minor correction) the district judge’s ruling.

Since Castle, so far as I can tell, this principle that the US government can’t prosecute bribe-taking foreign officials as conspirators in an FCPA violation (or, similarly, as accomplices to an FCPA violation under another statute, 18 U.S.C. § 2(a)), seems to have become generally accepted, largely unchallenged by the US government, and treated as clearly correct as matter of legal doctrine. And it matters a great deal as a policy matter: If the Castle ruling had gone the other way, than the FCPA—complemented by the general conspiracy and complicity statutes—would give the US government a very powerful tool, for better or worse, to prosecute bribe-taking foreign government officials, at least those with sufficient ties to the US to establish personal jurisdiction (an important qualification I’ll return to later). I’d always assumed, without much reflection, that Castle was rightly decided. But after some digging into the case law, prompted largely by the more recent decision in Hoskins, and re-reading the Castle opinion, I think that Castle’s broad holding is doctrinally incorrect. If certain other conditions hold, a bribe-taking foreign official can be guilty as an accomplice to or co-conspirator in an FCPA violation, even though the foreign official could not directly violate the FCPA.

The starting point here is the well-established principle under US law that, as a general matter, a party may be guilty of conspiracy (under § 371) or complicity (under § 2(a)) even if that party did not commit all, or even any, of the elements of the underlying federal crime. Indeed, a party can be guilty of a conspiracy to commit a crime, or complicity in a federal crime, even if that party cannot, as a matter of law, commit the underlying crime (because the party is not a member of the specific class of individuals to whom the criminal statute applies). If that principle applied in cases like Castle, then there would be no question that the bribe-taking foreign officials were co-conspirators and/or accomplices.

The Castle court rejected that conclusion on the authority of a 1932 Supreme Court case called Gebardi v. United States, which dealt with a statute called the Mann Act. The Mann Act made it a federal crime to transport a woman or girl across state lines for prostitution or other immoral purposes. In Gebardi, an unmarried man and woman made plans to cross state lines for purposes of engaging in sexual relations (illegal at the time because they were not married). The man, who apparently made all the arrangements and purchased the train tickets, was charged with violating the Mann Act, and his conviction was not at issue. The government, though, had also charged the woman with conspiracy to violate the Mann Act. The Supreme Court held that this was improper, even in though the woman was a (consensual) party to the offense, and would therefore seem to fall within the scope of the conspiracy statute. The Court reasoned that finding any woman who consented to a Mann Act violation to be a co-conspirator would contravene Congress’s evident purpose in criminalizing the transportation of the woman, but not criminalizing the woman’s conduct. The Court’s reasoning seems to be that Congress must have thought that a woman—even one who “consents” in the sense of agreeing or acquiescing to her transportation—is more like a victim than a co-conspirator, and should not be punished. As the Court put it, “we perceive in the failure of the Mann Act to condemn the woman’s participation in those transportations which are effected with her mere consent, evidence of an affirmative legislative policy to leave her acquiescence unpunished.”

The Castle court thought that Gebardi barred the conclusion that bribe-taking foreign officials were co-conspirators in FCPA violations. As the court reasoned, “Congress intended in both the FCPA and the Mann Act to deter and punish certain activities which necessarily involved the agreement of at least two people, but Congress chose in both statutes to punish only one party to the agreement.” Furthermore, the Castle court noted, the fact that the FCPA criminalizes only bribe-giving, and conspicuously fails to criminalize bribe-taking, is in notable contrast to other federal anti-bribery statutes that address bribery of federal public officials (e.g. here, here, and here), as those other statutes prohibit both bribe-giving and bribe-taking.

That analysis, which I used to find persuasive, is deeply flawed, at least as a doctrinal matter. (I’ll put aside for the moment whether it’s good policy.) It’s true that Gebardi held that, notwithstanding the broad language of the conspiracy statute, when Congress chooses to punish only one party to a crime that typically involves the consent or acquiescence of another party, the passive or acquiescing party, whom it appears Congress considered more of a victim, should not be punished as an accomplice or co-conspirator simply by reason of his or her consent. However, the Gebardi Court was careful to distinguish that situation from the alternative scenario (discussed in an earlier Supreme Court case called United States v. Holte), that a woman who was “the active or moving spirit in conceiving or carrying out [her] transportation” could violate the Mann Act directly, and presumably could also be charged as a conspirator. Moreover, precisely that distinction—between a party who passively consents to a criminal scheme and one who is the “active or moving spirit”—has been deployed in the context of an important federal anti-bribery statute that explicitly criminalizes only one side of a bribe transaction: the Hobbs Act, codified at 18 U.S.C. § 1951, which prohibits “extortion,” defined in part as “obtaining of property from another, with his consent, … under color of official right.” That somewhat convoluted language has been interpreted by the Supreme Court as the “rough equivalent of what we would now describe as ‘taking a bribe.’” (Because more specific federal statutes deal with bribery of federal officials, the Hobbs Act is used principally against state and local officials.)

The Hobbs Act, then, is like the FCPA in that it only explicitly criminalizes one side of a bribe transaction: While the FCPA criminalizes paying a bribe but not taking a bribe, the Hobbs Act criminalizes taking a bribe but does not (explicitly) criminalize paying a bribe. But can the private party who pays a bribe to a public official be guilty of conspiracy (with the public official) to violate the Hobbs Act? The question has arisen repeatedly, and the Supreme Court and multiple courts of appeal have given a consistent answer: Yes, but only if the bribe-payer is more actively involved in suggesting or encouraging the bribe transaction—not if the bribe-payer is, as the Supreme Court put it in its 2016 decision in Ocasio v. United States, “merely complying with an official demand.” As Ocasio explained, “Holte and Gebardi make perfectly clear that a person may be convicted of conspiring to commit a substantive offense that he or she cannot personally commit. They also show that when that person’s consent or acquiescence is inherent in the underlying substantive offense, something more than bare consent or acquiescence may be needed to prove that the person was a conspirator.” In so holding, Ocasio was not breaking any new ground: A long string of court of appeals opinions has held that a bribe payer may be guilty of conspiracy to violate the Hobbs Act (or as an accomplice to such violation)—even though the relevant provision of the Hobbs Act applies directly only to the bribe-taking public official—as long as the bribe-payer “exhibits conduct more active than mere acquiescence.” (See also here, here, and here.) A similar logic has been applied in the context of other anticorruption statutes that, like the Hobbs Act, only explicitly criminalize one side of an unlawful transaction.

The Castle opinion is flatly inconsistent with this aspect of the doctrine. A fair application of Gebardi and Holte, as well as the line of Hobbs Act cases culminating in Ocasio, would allow the government to charge a foreign official who receives a bribe proscribed by the FCPA with conspiracy and/or complicity in an FCPA violation, but only if the government can prove that the government official did something more then merely consent to receive the bribe. If there is “conduct more active than mere acquiescence,” and certainly if the government official is the “active or moving spirit” behind the bribe transaction, then under binding Supreme Court authority, as well as the persuasive authority of a lengthy line of opinions on the Hobbs Act and similar statutes, a bribe-taking public official can be guilty of conspiracy to violate the FCPA and/ or with aiding and abetting such a violation, even though it is incontrovertible that such a government official cannot directly violate the FCPA. Of course, the Castle court can’t be faulted for failing to anticipate subsequent cases, like the Supreme Court’s decision in Ocasio, though it’s fair to criticize Castle both for misreading Gebardi itself and for neglecting the prior line of clearly analogous appellate opinions concerning the Hobbs Act. At the very least, Castle’s doctrinal underpinnings have been eroded to the point where there’s a strong argument to be made that it’s no longer good law. (For a commentary reaching a similar conclusion about how Ocasio may undermine Castle‘s reasoning, see here. My take is largely similar, though I don’t think Ocasio actually broke much new ground; rather, that case seems to have endorsed a longstanding view that most courts had already endorsed, but that Castle ignored or implicitly rejected.)

Now, there are two important qualifications to the above conclusion.

• First, even if the above analysis is correct, it doesn’t mean that the US will be able charge all foreign officials who actively solicit bribes in FCPA cases as accomplices or co-conspirators. Some officials, particularly heads of state, may be protected by foreign sovereign immunity doctrines. And even for those who aren’t, the U.S. would still have to establish personal jurisdiction over the defendant, consistent with the Due Process Clause of the US Constitution. This will typically entail establishing that the foreign government official has sufficient “systemic and continuous” contacts with the US to establish general jurisdiction, or that the FCPA violation in question arose from the defendant foreign official’s conduct within US territory (say, the payment or agreement to make a payment was concluded in US territory).
• Second, even if the Castle court’s analysis is flawed, it’s possible that its holding might be justified on other grounds. This may bring us back to Hoskins, which relied in part on the reasoning of cases like Gebardi, but emphasized in particular the alleged affirmative legislative policy under the FCPA to limit extraterritorial complicity or conspiracy liability. I explained in my last post why I don’t think Hoskins’ analysis is correct on this point, but someone who finds Hoskins persuasive with respect to the limits on extraterritorial applications of complicity and conspiracy liability in FCPA cases might endorse the outcome in Castle even if they find my critique of Castle’s reasoning persuasive. That said, it’s perhaps worth pointing out that the Castle rule and the Hoskins rule don’t always lead to the same conclusion. Consider a foreign public official who takes a bribe from a US firm while in US territory. The Hoskins holding would not (by itself) bar a charge against the official for conspiracy to violate the FCPA, since such a charge would not be extraterritorial, but the Castle rule would bar such a charge. Similarly, a non-US co-conspirator whose relevant acts are all abroad but who is not the bribe-taking official (for example, someone like Mr. Hoskins), would be protected by the Hoskins rule but unaffected by the Castle rule.

I’m not sure, as a policy matter, whether it would be a good idea for the US government to be able to charge bribe-taking foreign officials (over whom the US can properly assert personal jurisdiction) with conspiracy or complicity in FCPA violations. That seems to me a hard question, and one that merits further thought. But as a legal matter, the idea that such officials cannot be charged as FCPA accomplices or conspirators—a view that many people (me included, until very recently) take to be both settled and settled correctly—is almost certainly wrong, in the sense of being inconsistent with the relevant case law. At the appropriate time, perhaps the US government should revisit this question and consider testing the issue by bringing charges in an appropriate case.