United States v. Robert Haas

In the United States Court of Appeals For the Seventh Circuit ____________________ No. 20-3269 UNITED STATES OF AMERICA, Plaintiff-Appellee, v. ROBERT A. HAAS, Defendant-Appellant. ____________________ Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 19 CR 00486 — Edmond E. Chang, Judge. ____________________ ARGUED JANUARY 5, 2022 — DECIDED JUNE 21, 2022 ____________________ Before KANNE, 1 WOOD, and BRENNAN, Circuit Judges. WOOD, Circuit Judge. After posting death threats against former U.N. Ambassador Nikki Haley on her Instagram page, Robert Haas was (not surprisingly) visited by diplomatic security officers. He reiterated his violent message during that 1 Circuit Judge Kanne died on June 16, 2022, and did not participate in the decision of this case, which is being resolved under 28 U.S.C. §46 (d) by a quorum of the panel. 2 No. 20-3269 conversation, and then escalated matters by posting vile, antiSemitic statements on a Russian social-media website called VK.com. An FBI agent questioned him about the new round of threats, but Haas just doubled down with additional threats via texts and voicemails, this time against the agent. Criminal charges for transmitting threats in interstate commerce and for the threats against the FBI agent followed. A jury convicted him on all counts, and the district court sentenced him to 51 months in prison. Haas now appeals, raising four issues. Only one of those points, however—a multiplicity challenge to the indictment— was properly preserved in the district court. His contentions that the government’s evidence was insufficient to support the verdict, that the indictment was constructively amended, and that some counts were improperly grouped for sentencing, were all forfeited. We therefore review them only for plain error. Because we conclude that Haas failed to support his multiplicity argument and his other arguments do not undermine the district court’s judgment, we affirm. I Robert Haas believes that Jews are responsible for crimes against humanity, that Israel is “the biggest racist organization on the planet,” and that anyone who supports Israel or Jews is a traitor to the United States who deserves to die. In 2018 Haas posted a threatening comment along these lines on the public Instagram page of Ambassador Haley. The comment drew the attention of federal authorities, who visited Haas in his home to speak with him about it. Displeased with the visit and believing that federal authorities were in cahoots with the Jews, Haas turned to VK.com, a Russian social-media No. 20-3269 3 site, where he made several posts vowing to kill Jews and the officials that protect them. These posts caught the attention of FBI Task Force Officer Joseph Kostuchowski, who arranged to meet with Haas at his job site to discuss the posts. When Kostuchowski arrived, Haas refused to speak with him. After Kostuchowski gave up and left, he began receiving phone calls, text messages, and voicemails from Haas. In many of these communications, Haas conveyed his belief that Kostuchowski deserved to die for supporting Israel. In 2019, Haas was arrested. For the VK.com posts, Haas was charged with eight counts of transmitting threats in interstate commerce in violation of 18 U.S.C. § 875 (c). For the direct threats against Kostuchowski, Haas was charged initially with four counts of threatening a federal official under 18 U.S.C. § 115 (a)(1)(B). A fifth count was added later, based on Haas’s tirade to an Illinois State Police patrol sergeant who was transporting him from Ottawa (Illinois) to Chicago; Haas’s general message was that Kostuchowski should die for protecting Jews. Haas proceeded pro se during pretrial and trial proceedings, though he was represented by appointed counsel for certain posttrial motions. He moved to dismiss the first three section 115 counts as multiplicitous, but the district court denied the motion. At the close of the government’s evidence, Haas moved for judgment of acquittal on all counts, see FED. R. CRIM. P. 29, arguing that the government had not met its burden of proof. Although Haas’s oral motion focused specifi- cally on the element of intent, the court construed his motion as addressing all the elements of both section 115 and section 875(c). It denied this motion. 4 No. 20-3269 The case proceeded to a jury, which convicted Haas on all 13 counts. The district court sentenced Haas to 51 months in prison. After the verdict, Haas again moved for judgment of acquittal, this time with the assistance of appointed counsel, but the court again denied the motion. Haas now appeals. II Haas has raised four issues for our consideration. First, he reiterates his multiplicity argument regarding Counts 1–3. Second, he challenges the sufficiency of the government’s evidence on the interstate-commerce element of his section 875(c) convictions. Third, he argues that the government’s evidence and the court’s jury instructions constructively amended the indictment, in violation of the Grand Jury Clause of the Fifth Amendment. Finally, he challenges the district court’s separation of Counts 6–13 into two groups for purposes of the Sentencing Guidelines, ch. 3, Part D. We begin with the multiplicity assertion, which Haas properly raised in the district court. The other three points were forfeited, but we will examine them for plain error. A Counts 1–3 involve the direct threats against Kostuchowski. Haas asserts that the use of three separate counts was improper. An indictment is multiplicitous when it charges a single offense as separate counts. At one time, there was a worry that multiplicity raised concerns under the Double Jeopardy Clause of the Fifth Amendment, but the Supreme Court put that to rest when it held that “[t]he Clause protects only against the imposition of multiple criminal punishments for the same offense, and then only when such occurs in successive proceedings.” Hudson v. United States, 522 No. 20-3269 5 U.S. 93 , 99 (1997) (emphasis and citations omitted). Nonetheless, Federal Rule of Criminal Procedure 12(b)(3)(B)(ii) recognizes the prohibition against multiplicity and permits a timely objection to such an indictment. Haas complied with this rule, and so we may assess this argument. Our review in this respect is de novo. To determine whether an indictment is multiplicitous, “we look to the applicable criminal statute to see what the allowable ‘unit’ of prosecution is—the minimum amount of activity for which criminal liability attaches.” United States v. Allender, 62 F.3d 909, 912 (7th Cir. 1995). The applicable statute here reads as follows: Whoever[] … threatens to assault, kidnap, or murder, a United States official, a United States judge, a Federal law enforcement officer, or an official whose killing would be a crime under such section, with intent to impede, intimidate, or interfere with such official, judge, or law enforcement officer while engaged in the performance of official duties, or with intent to retaliate against such official, judge, or law enforcement officer on account of the performance of official duties, shall be punished as provided in subsection (b). 18 U.S.C. § 115 (a)(1)(B). It further states: “A threat made in violation of this section shall be punished by a fine under this title or imprisonment for a term of not more than 10 years, or both, except that imprisonment for a threatened assault shall not exceed 6 years.” 18 U.S.C. § 115 (b)(4). The parties offer competing theories on the key question of the unit of prosecution. Haas focuses on the first quoted subsection, section 115(a)(1)(B), while the government 6 No. 20-3269 emphasizes the statute as a whole, including section 115(b)(4). Haas points out that section 115(a)(1)(B) identifies the possible victims, not individual threats. This suggests, he argues, that the unit of prosecution is the broader scheme of threatening one of the designated federal officials, not any particular threat or communication that was issued. Since the indictment separately charged three threats directed at only one victim, Kostuchowski, Haas concludes that it was multiplicitous. The government responds by urging us to take a broader look at the statute, and it suggests that when one does so, it readily appears that each threat is individually indictable. We agree with this approach. Section 115(b)(4) states that “[a] threat made in violation of this section shall be punished by a fine … or imprisonment … or both.” (Emphasis added.) “A threat” means just that: an individual threat. Haas retorts that the reference to “[a] threat” in section 115(b)(4) at best makes the statute ambiguous given the focus on the victim found in section 115(a)(1)(B). At a minimum, then, he urges that the rule of lenity demands that we adopt his interpretation. See Bell v. United States, 349 U.S. 81, 83 (1955). We do not buy his premise: section 115(b)(4) does not introduce ambiguity; it fills in useful information. The clarity offered by section 115(b)(4) also undermines the remainder of Haas’s arguments. For instance, Haas compares and contrasts section 115 to three other federal statutes. According to Haas, section 115(a)(1)(B) is like the bank-fraud statute, under which the unit of prosecution is the broad scheme of fraud. See 18 U.S.C. § 1344 ; United States v. Ajayi, 808 F.3d 1113, 1123 (7th Cir. 2015). Section 115(a)(1)(B) differs, he suggests, from the mail- and wire-fraud statutes, which No. 20-3269 7 make each individual communication a proper unit of prosecution. See Badders v. United States, 240 U.S. 391 , 393–94 (1916); Ashland Oil v. Arnett, 875 F.2d 1271 , 1278 (7th Cir. 1989). We are not persuaded. Unlike section 115, the bank-fraud statute explicitly refers to “a scheme or artifice” as the basis for liability. See 18 U.S.C. § 1344 . Nothing indicates that a single fraudulent transaction is properly described as a “scheme or artifice.” Compare 18 U.S.C. § 115 (b)(4) (“[a] threat … shall be punished”) with 18 U.S.C. § 1341 and 18 U.S.C. § 1343 (containing no explicit reference to “a fraudulent mailing” or “a fraudulent wire” as the basis for prosecution). Haas next argues that the government’s interpretation is inconsistent with the penalty structure of section 115, because it purportedly would lead to absurd results: a person who merely threatens a federal official (but does so numerous times) could serve more time in prison than a person who physically assaults a federal official (but does so only once). See 18 U.S.C. §§ 115 (b)(1), (b)(4). For support, Haas turns to a First Circuit decision that rested on this concern in the context of the federal murder-for-hire statute, 18 U.S.C. § 1958 (a), which Haas argues has a similar penalty structure. In United States v. Gordon, the court concluded that the murder-for-hire statute’s “graduated sentencing scheme … conveys … Congress’s apparent belief that the greater the harm to the victim, the harsher the punishment should be for the offender.” 875 F.3d 26, 33 (1st Cir. 2017). Accordingly, the First Circuit reasoned that “[t]he government’s proposed unit of prosecution,” which would have defined each individual act in furtherance of a murder plot as a separate unit, not the broader plot, “would frustrate this congressional aim.” Id. “[A] person who made ten telephone calls in service of a 8 No. 20-3269 failed plot that caused no injury to anyone [would be exposed] to a much steeper maximum sentence than a person who, as a result of a single telephone call, caused substantial personal injury to a victim.” Id. Because it believed this would be an irrational result, the First Circuit held that “[t]he unit of prosecution advocated by the defendant (which focuses on the number of plots) is much more consistent with the victimcentric sentencing scheme formulated by Congress.” Id. Haas’s penalty-structure argument ignores a key difference between section 115 and the murder-for-hire statute at issue in Gordon: the murder-for-hire statute contains no provision analogous to section 115(b)(4), which clarifies the unit of prosecution. In the absence of such a provision, the First Circuit’s reasoning may have been tenable. Here, that reasoning is foreclosed by the plain text of the statute. The risk of absurd outcomes is also not inevitable. Defendants charged under section 115 enjoy several protections against counterintuitive and disproportionate sentences, including the option of concurrent sentences and the rules calling for grouping of multiple counts. Finally, Haas argues that “[t]he only federal court of appeals to have considered the issue,” the D.C. Circuit, “supports [his] interpretation of the unit of prosecution under 18 U.S.C. § 115 (a)(1)(B).” See United States v. Klat, 156 F.3d 1258 (D.C. Cir. 1998). But Klat did not deal with multiplicity; it dealt with its converse, duplicity. Id. at 1266 ; see U.S. Dep’t of Justice Archives, Criminal Resource Manual, § 812, Duplicity and Multiplicity Issues (“Duplicity occurs when more than one offense is alleged in the same charge ... [c]onversely, multiplicity occurs when two charges allege the same offense.”). The Klat decision states that multiple threats may be charged No. 20-3269 9 in the same count (if they effectively make up a single scheme) without risking duplicity. Klat, 156 F.3d at 1266 . It does not hold that (or even consider whether) they must be charged in the same count to avoid multiplicity. Klat is therefore unilluminating. In sum, we are convinced that the allowable unit of prosecution under 18 U.S.C. § 115 is the individual threat, not some broader scheme or plot to threaten. Section 115(b)(4) says as much. This leaves the question whether an individual threat occurs in every separate text or email, or if a single threat might be communicated in a couple of steps. If a defendant sends six threatening texts in quick succession, may or must each text be charged as a separate count, or should all six texts be treated as a single, ongoing threat? Context will govern the answer to that question. As we now explain, the indictment here draws reasonable lines. This indictment stated that Haas’s relevant communications conveyed three separate threats. The threat described in Count 1 read as follows: “[t]ake off the gun and badge like you said you would pussy. Come meet true evil. � I got something for you terrorists.” It was sent on May 8, 2019, while the texts that form the bases for Counts 2 and 3 were sent the following day, May 9. While we do not rule out the possibility that messages sent on separate dates may constitute a single threat, we are satisfied that is not the case here. Besides being separated by date from the other two, the threat contained in the May 8 text in no way depended on additional information or communications for its completion. Thus, it was properly charged as a discrete threat. Counts 2 and 3 rest on texts that were closer in time, but we are satisfied that they, too, were separate threats. Count 2 10 No. 20-3269 was based on a text that Kostuchowski received on May 9 at 10:37 a.m. It said: “I thought you wanted to chat you monkey jew. You coward old man. You know I get what I want and I think you deserve death.” Count 3 stemmed from an 11:04 a.m. text that said: “I’m not afraid to walk out of my door in the morning. You should be, however, considering you support Jewish terrorism, and your anti-American bullshit is going to get you killed.” Haas argues that because the texts are similar in content and were sent only 27 minutes apart with no intervening communication, they constitute a single, ongoing threat. But there is little other than offensive content to link them. While they were closer in time to each other than to the first text, they still were separated by nearly half an hour—we can hardly characterize them as a string of successive texts. And while their content may have been similar, neither text was dependent on the other, in the sense of finishing a sentence or completing a thought. As before, each is a completed threat in its own right. We thus conclude that they were appropriately charged as separate counts. We emphasize that this is a fact-intensive inquiry. We are not saying that any time threatening communications are separated by at least 27 minutes, they may appropriately be charged as separate threats. In some cases, communications separated by longer periods may best be understood as a single, ongoing threat. But on the record before us, we are satis- fied that the May 9 texts were properly charged separately. B Haas’s next challenge is to the sufficiency of the government’s evidence. He argues that the government failed to No. 20-3269 11 prove that his threats were transmitted “in interstate commerce,” as required by 18 U.S.C. § 875 (c), which underlies the charges in Counts 6–13. Given the cross-border nature of the Internet, the government contends that it proved the interstate-commerce element of the section 875(c) charges simply by showing that Haas transmitted his threats over the Internet (specifically, to VK.com). Haas contests the notion that use of the Internet alone can satisfy the interstate-commerce element of a federal statute worded as section 875(c) is. As a threshold matter, the parties dispute whether Haas properly preserved this issue for our review. We agree with the government that he did not. Haas moved for judgment of acquittal twice, once at trial and once in a posttrial motion. Both times, Haas made no mention of the interstate-commerce element even as he explicitly argued about others, such as intent. When he first moved for judgment of acquittal orally at trial, Haas was pro se, which might entitle him to some latitude. See United States v. Lewis, 817 F.3d 1054, 1055 (7th Cir. 2016). But when he moved for judgment of acquittal in a posttrial motion, he was represented by counsel. And while the written posttrial motion carefully argued the sufficiency of the evidence for other elements, counsel either neglected or chose not to address interstate commerce. Since Haas failed to raise the argument despite multiple opportunities to do so, including while represented by counsel, he forfeited the issue, and we review only for plain error. See United States v. Flores, 929 F.3d 443, 447 (7th Cir. 2019). Plain-error review requires that the defendant show (1) an error, (2) that is plain, (3) that affects substantial rights, and (4) that “had a serious effect on ‘the fairness, integrity, or public 12 No. 20-3269 reputation of judicial proceedings.’” Greer v. United States, 141 S. Ct. 2090 , 2096–97 (2021). This is a tall order. Haas begins with a truism: the particular wording of the interstate-commerce element of a statute establishes what the government must prove. Statutes that contain language such as “in interstate commerce” require proof that state lines were crossed; by contrast, statutes with language such as “affecting commerce” or “any facility of interstate commerce” require proof only that the criminal activity involved an instrumentality or channel of interstate commerce. See Circuit City Stores, Inc. v. Adams, 532 U.S. 105, 115 (2001); United States v. Schaefer, 501 F.3d 1197, 1201 (10th Cir. 2007), overruled on other grounds by United States v. Sturm, 672 F.3d 891 (10th Cir. 2012). Finally, Congress sometimes exercises its broad power to regulate even local activities that have a substantial effect on interstate commerce. See Gonzales v. Raich, 545 U.S. 1, 17 (2005). Congress’s choice of language in any given statute is thus critical. How it articulates the interstate-commerce element of a statute tells us what that statute will reach. See Circuit City Stores, Inc., 532 U.S. at 115 (“Congress uses different modifiers to the word ‘commerce’ in the design and enactment of its statutes. The phrase ‘affecting commerce’ indicates Congress’ intent to regulate to the outer limits of its authority under the Commerce Clause. … Unlike those phrases, however, the general words ‘in commerce’ … are understood to have a more limited reach.”). None of this is controversial or even novel. The question is how we are to apply these established principles to the Internet. Haas argues that the phrase “in interstate commerce” requires a showing that the relevant communication No. 20-3269 13 physically traveled from a server in one state to a server in another. The government argues that such a showing is not necessary. Given the inherently interstate nature of the Internet, the government believes it needed to show only that the Internet was used. A circuit split has developed around this point. The First, Second, Third, and Fifth Circuits have taken the position that the government asks us to adopt here, that is, that the government can satisfy the “in interstate commerce” element of a statute simply by showing that the Internet was used. See United States v. Lewis, 554 F.3d 208 , 214–15 (1st Cir. 2009) (addressing 18 U.S.C. § 2252 (a)(2), which at the time contained the language “in interstate … commerce” but has since been amended to say “in or affecting … commerce” and “any means or facility of … commerce”); United States v. Harris, 548 F. App’x 679, 682 (2d Cir. 2013) (same); United States v. MacEwan, 445 F.3d 237, 243 (3d Cir. 2006) (same but with respect to 18 U.S.C. § 2252A(a)(2)(B)); United States v. Runyan, 290 F.3d 223, 239 (5th Cir. 2002) (same but with respect to 18 U.S.C. § 2251 ). The Ninth and Tenth Circuits, on the other hand, have sided with Haas: they hold that the government must prove that the online communication crossed state lines, not simply that it was made on the Internet. See United States v. Wright, 625 F.3d 583 , 590–95 (9th Cir. 2010) (addressing 18 U.S.C. § 2252A(1), which at the time contained the language “in interstate … commerce” but has also been amended since then); Schaefer, 501 F.3d at 1200–02 (same but with respect to 18 U.S.C. §§ 2252 (a)(2), (a)(4)(B)). How the established paradigm applies to the Internet is certainly an interesting question, but this case does not 14 No. 20-3269 require us to choose sides. Because we are reviewing only for plain error, we need not adopt a holding as sweeping as the one prevailing in the majority of circuits to have considered the issue. Nor need we insist on something as technologically reductive as proof that a communication traveled from a server in one state to a server in another, a dubious requirement that fails to take full account of the realities of the Internet and its functioning. 2 Under a plain-error standard of review, the government’s convictions survive either way. The existence of a circuit split is some evidence that the court would not have “plainly” erred if it had concluded that the government needed to show only that the Internet was used. And under that understanding, the government’s evidence that Haas used the Internet to transmit a post from Illinois to Russia would be more than sufficient to uphold the jury’s verdict. In addition, while it is true that in proving interstate commerce the government focused mostly on the transmission to Russia, the record also contained evidence that Haas’s posts were viewed by a human-rights organization in Los Angeles, California. Based on this evidence, a jury could have concluded that Haas’s threats crossed not only international, but also interstate lines, if that is in fact what the government needed to prove. Haas protests that the California evidence should be discounted because the California organization did not view the posts until months after they were posted, and so he could not have “knowingly transmit[ted] a threat to [California] on the 2 See “How Does the Internet Work?”, https://web.stanford.edu/class/msande91si/www-spr04/readings/week1/InternetWhitepaper.htm. No. 20-3269 15 dates alleged in the indictment.” “[A] threat communicated on a particular date,” he argues, does not “establish[] endless liability if it is subsequently viewed by a third party months later.” Haas’s point might be better taken if the California organization had viewed only later re-posts by a different user. In that case, whether Haas himself transmitted the posts to California on the dates in question might be relevant. But the evidence here was that the California organization viewed Haas’s original posts on the platform on which they were originally posted. Regardless of when this viewing occurred, the jury could have concluded that the posts crossed state lines the day that Haas posted them and remained available in California until they were seen. The evidence was also sufficient to support a finding that Haas knew about the interstate and international dimensions of his posts. The fact that he used a Russian platform did not require the jury to find that he knowingly transmitted his posts only to Russia. Haas wrote his posts in English, not Russian. And although VK.com is based in Russia, the jury heard evidence that it is like Facebook, which is based in the United States yet used around the world. Finally, Haas himself testi- fied that his intended audience was U.S. law enforcement of- ficials, whom he wished to discourage from investigating him—it was not Russians. Based on this evidence, the jury could have concluded that by posting on VK.com, Haas knowingly transmitted his posts not only to Russia but also to other states, including California, where they were in fact viewed. At minimum, the district court did not plainly err when it left the jury’s verdict undisturbed on that basis. 16 No. 20-3269 At oral argument, counsel for Haas offered a slight variation of the argument against the California evidence. Counsel suggested that, as a technical matter, Haas transmitted his posts only to the website in Russia, which in turn (counsel postulated) retransmitted them to other locations, such as California. In other words, even though the California organization viewed Haas’s original posts on their original platform, it technically only viewed re-posts, because the foreign platform had to retransmit them from Russia to the rest of the world. We are skeptical of the technical and legal soundness of this claim. But even accepting it as technically accurate, it does nothing for Haas. First, the jury heard no evidence to that effect. Second, even if it had, it remained at liberty to conclude—based on the language of choice, the seemingly global nature of the platform, and the intended audience—that Haas knowingly used VK.com to facilitate an interstate transmission, just as one would use the Post Office, UPS, or FedEx to mail a letter rather than personally driving it across state lines. In short, Haas’s argument cannot survive plain-error review. C Haas’s next argument is that the district court and the government constructively amended the indictment, meaning that he was tried and convicted of a charge different from the charge brought by the grand jury. A constructive amendment violates the Fifth Amendment’s guarantee that “[n]o person shall be held to answer for a … crime, unless on a presentment or indictment of a Grand Jury.” U.S. Const. amend. V; see also United States v. Turner, 836 F.3d 849, 863 (7th Cir. 2016). Haas concedes that he forfeited this argument, and so we again review only for plain error. No. 20-3269 17 “A constructive amendment to an indictment occurs when either the government (usually during its presentation of evidence and/or its argument), the court (usually through its instructions to the jury), or both, broadens the possible bases for conviction beyond those presented by the grand jury.” United States v. Cusimano, 148 F.3d 824, 829 (7th Cir. 1998) (internal quotation marks omitted). That said, “not all variations in proof that contradict or supplement verbiage in the indictment,” Turner, 836 F.3d at 863 , or “variations between an indictment and the jury instructions,” United States v. Johnson, 827 F. App’x 586 , 590 (7th Cir. 2020), amount to a constructive amendment. “Rather, the offense ‘charged in the indictment must be materially different or substantially altered at trial, [so that] it is impossible to know whether the grand jury would have indicted for the crime actually proved.’” Turner, 836 F.3d at 863 . Haas argues that both the court (through its jury instructions) and the government (through its presentation of evidence) constructively amended the indictment. As to the former, he argues that while “he was charged only with threats in interstate commerce … the jury instructions permitted a finding of guilt based on evidence of threats in foreign commerce.” As to the latter, he argues that the government’s evidence only proved foreign commerce, not interstate commerce. Thus, the indictment was constructively amended to allow the jury to convict him based on evidence of foreign commerce, even though he was charged only with transmitting threats in interstate commerce. Haas’s contention that the jury instructions and the government’s evidence diverged from the precise charge in the indictment is not wholly without support. The jury was 18 No. 20-3269 instructed that the interstate-commerce element was met if the jury found that Haas’s threats crossed “from inside a State to outside a State.” While the government insists that this language defines “interstate commerce,” we agree with Haas that the instruction is most reasonably understood as defining the phrase “interstate or foreign commerce.” A direct flight from Michigan to Russia can avoid crossing state lines, but it would satisfy the jury instructions. Thus, the instructions permitted the jury to convict Haas based on evidence of foreign commerce, even though the indictment charged him only with transmitting threats in interstate commerce. And as previously discussed, to prove the interstate-commerce element, the government rested heavily on the fact that Haas’s communications were transmitted from Illinois to Russia (although, as discussed, the record did contain some evidence supportive of interstate transmission between Illinois and California). In any event, while the court’s instructions and the government’s evidence may have diverged from the strict language of the indictment, the standard of review is critical here. The Internet is an instrumentality of interstate commerce. For the reasons we already have reviewed, we are satisfied that this jury could have found movement of these messages from or through one state to another. Illinois does not share a common border with Canada (or any other foreign country), and so on the most mechanistic understanding of activities “in” commerce (think of the Pony Express model), some interstate movement is necessary before a message originating in Illinois could wind up in any foreign country, Russia included. Even a transmission bounced from Illinois to a geostationary satellite and down to a server in Russia might cross the airspace of a different U.S. state. But on plain-error review, that kind of speculation cannot carry the day. We No. 20-3269 19 have no doubt that if Haas had raised this point in the district court, the government easily could have either supplemented its evidence of interstate commerce or obtained a superseding indictment charging foreign commerce. Thus, even if we were to find that Haas satisfies the first three elements of plain-error review (and we have not said that, to be clear), leaving this result alone does not impugn the integrity of the judicial process. D Haas’s final challenge is to the way the court grouped the counts related to the VK.com posts for sentencing purposes. He again concedes that he has forfeited this issue and that our review is for plain error. The sentencing guidelines provide that “[w]hen a defendant has been convicted of more than one count, the court shall … [g]roup the counts resulting in conviction into distinct Groups of Closely Related Counts (“Groups”) by applying the rules specified in § 3D1.2.” U.S.S.G. § 3D1.1. Section 3D1.2 states that “[a]ll counts involving substantially the same harm shall be grouped together into a single Group.” U.S.S.G. § 3D1.2. It then itemizes the situations where “[c]ounts involve substantially the same harm within the meaning of this rule.” U.S.S.G. § 3D1.2(a)–(d). For purposes of Haas’s argument, it suffices to note that “multiple counts involving making a threatening or harassing communication to the same victim are grouped together” while multiple counts involving different victims are not. U.S.S.G. § 2A6.1 cmt. n.3 (emphasis added). The district court separated Haas’s 13 counts into three groups for purposes of calculating the advisory guidelines range. The first group captured Counts 1–5, which involved 20 No. 20-3269 the threats that Haas made directly to Kostuchowski. It divided the threats that Haas posted on VK.com, which made up Counts 6–13, into two groups. Haas does not challenge the Kostuchowski group, but he argues that the district court erred in separating Counts 6–13 into two groups. In his view, they effectively involved the same victim (society as a whole) and thus should have been grouped together under U.S.S.G. § 3D1.2. The district court, accepting the grouping proposed in the Presentence Investigation Report, grouped Counts 7, 8, 10, 11, and 13 as counts involving federal employees, whom it regarded in the aggregate as one victim. It grouped the remaining Counts—6, 9, and 12—as relating to society in general as the victim. Haas argues that the federal employees and “society” should have been treated as one victim. He relies on application note two of section 3D1.2, which defines “victim” as: Generally, the[] … one person who is directly and most seriously affected by the offense and is therefore identi- fiable as the victim. For offenses in which there are no identifiable victims (e.g., drug or immigration offenses, where society at large is the victim), the “victim” for purposes of subsections (a) and (b) is the societal interest that is harmed. In such cases, the counts are grouped together when the societal interests that are harmed are closely related. U.S.S.G. § 3D1.2 cmt. n.2. Haas argues that all of Counts 6–13 (not just Counts 6, 9, and 12) involved nonspecific threats directed to no identifiable victim (other than society itself) and, therefore, they should have been grouped together per application note two. No. 20-3269 21 Haas’s argument is not without merit. We grant that it is dif- ficult to draw a neat line between those threats classified as threats against “federal employees” (e.g., “Lots of #Feds #FederalOfficers #StateDepartment #Aipac #JEWS will be killed when Americans see this Jewish trash plan.”) and the rest (e.g., “I don’t care if it’s a cop, prosecutor, judge, politician or elite. You try to stop me from telling the truth I will cut every throat in your home. Try me!”). On the other hand, note two calls for the careful exercise of discretion by the district judge, who has the unenviable task of discerning the societal interests at issue and their relation to the harms. In that setting, we are loathe to find plain error. The note offers some guidance at the extremes: on the one hand, there is the typical case in which a single victim is identifiable, while on the other, there are cases involving “drug or immigration offenses, where society at large is the victim.” U.S.S.G. § 3D1.2 cmt. n.2. It leaves room for the atypical case, such as this one, where the victim is neither society at large nor a single identifiable individual, but rather a victim category such as “federal employees.” Indeed, the examples that the application note lists when discussing the “offenses in which there are no identifiable victims”—drug and immigration offenses—suggest that it has in mind something qualitatively different from the situation presented by Haas’s charges. After all, drug and immigration offenses by their very nature and design target society at large, even if the underlying conduct also has a serious effect on drug users, businesses, and particular communities. In this context-dependent situation, we find no plain error in the district court’s grouping decisions. 22 No. 20-3269 III We AFFIRM Haas’s convictions and sentence.