865 F.3d 767 (D.C. Cir. 2017)
Nicholas Slatten, Paul Slough, Evan Liberty and Dustin Heard (“defendants”) were contractors with Blackwater Worldwide Security (“Blackwater”), which in 2007 was providing security services to the United States State Department in Iraq. As a result of Baghdad shootings that injured or killed at least 31 Iraqi civilians, Slough, Liberty and Heard were convicted by a jury of voluntary manslaughter, attempted manslaughter and using and discharging a firearm in relation to a crime of violence (or aiding-and-abetting the commission of those crimes); Slatten was convicted of first-degree murder. They now challenge their convictions on jurisdictional, procedural and several substantive grounds.
For the following reasons, we hold that the Court has jurisdiction pursuant to the Military Extraterritorial Jurisdiction Act (“MEJA”), 18 U.S.C. §§ 3261 et seq., and that venue in the District of Columbia was proper. We further hold that the district court did not abuse its discretion in denying the defendants’ motion for a new trial based on post-trial statements of a government witness. Regarding the challenges to the sufficiency of the evidence, we hold that the evidence was sufficient as to all except one of Liberty’s attempted manslaughter convictions, and that the evidence was sufficient as to Slatten. We further hold that Slatten’s indictment charging first-degree murder did not constitute vindictive prosecution. The Court concludes, however, that statements made by a co-defendant shortly following the attack, statements asserting that he—not Slatten—fired the first shots on the day in question, were admissible. Accordingly, the Court concludes that the district court abused its discretion in denying Slatten’s motion to sever his trial from that of his co-defendants and therefore vacates his conviction and remands for a new trial. Moreover, the Court concludes that imposition of the mandatory thirty-year minimum under 18 U.S.C. § 924(c), as applied here, violates the Eighth Amendment prohibition against cruel and unusual punishment, a holding from which Judge Rogers dissents. The Court therefore remands for the resentencing of Slough, Liberty and Heard.
On September 16, 2007, a car bomb exploded in Baghdad near a United States diplomat who was under the protection of Blackwater, a private security firm under contract with the State Department. The defendants were members of Blackwater’s Raven 23 team, which was sent to provide secondary support in the effort to evacuate the diplomat. Rather than meeting the primary team at the pre-arranged checkpoint, Raven 23 shift leader Jimmy Watson ignored his orders and directed the team to Nisur Square, a traffic circle in downtown Baghdad that Watson intended to “lock down.” A car bomb had exploded in Nisur Square earlier that year, in response to which Iraqi security had been dramatically increased, with multiple checkpoints at the Square’s entrances for potential threats.
The Raven 23 convoy, which consisted of four armored vehicles, came to a stop at the south end of the Square, and together with Iraqi police they brought all traffic to a halt. Two or three minutes later, witnesses heard the “pops” of shots being fired, and a woman screaming for her son. The car that had been hit, a white Kia sedan, had been flagged days earlier by a Blackwater intelligence analyst as a type that might be used as a car bomb. According to the government, the Kia then rolled forward and lightly bumped the vehicle in front of it. The driver’s side of the Kia windshield had a hole in it and was splattered with blood.
Two nearby Iraqi police officers approached the Kia on either side, and they saw the driver’s face full of blood, with a bullet wound in the middle of his forehead. One turned back to the convoy, waving his hands to indicate the shooting should stop, while the other made similar gestures as he tried to open the driver’s door. At that point, the vehicle in front of the Kia moved away, causing the Kia to roll forward again. Heavy gunfire erupted from the Raven 23 convoy into the Kia, and the Iraqi officers took cover behind their nearby kiosk. Multiple grenades were fired at the Kia, causing it to catch fire. The Kia passenger was shot and killed.
Indiscriminate shooting from the convoy then continued past the Kia, to the south of the Square. Victims were hit as they sought cover or tried to escape, giving rise to the bulk of casualties that day. At some point a Raven 23 member radioed that they were taking incoming fire, but others could not locate any such threat. When the shooting died down, a radio call indicated one of the Raven 23 vehicles had been disabled and needed to be hooked up to another vehicle to be towed. During the hook-up, a member of the Raven 23 convoy saw an Iraqi shot in the stomach while his hands were up, by an unidentified Blackwater guard who had exited his vehicle. Once the hook-up was complete, the Raven 23 convoy began moving slowly around the circle and north out of the Square, where isolated shootings continued both to the west and north. By the time the convoy finally exited the Square, at least thirty-one Iraqi civilians had been killed or wounded.
In the immediate aftermath of the shootings, the State Department conducted mandatory de-briefing interviews of the Raven 23 team. Because the testimony of certain witnesses before the grand jury relied on those statements, the district court dismissed the case as tainted as to all defendants. United States v. Slough, 677 F. Supp. 2d 112, 166 (D.D.C. 2009) (citing Kastigar v. United States, 406 U.S. 441 (1972)). This Court agreed that the oral and written statements that resulted from the de-briefings were compelled, and thus could not be used directly or indirectly by the government against the defendants who made them, but remanded the case for a more individualized analysis of the effect of the taint. United States v. Slough, 641 F.3d 544, 548, 554-55 (D.C. Cir. 2011).
On remand, the government used a new prosecutorial team and convened a new grand jury, which returned indictments against the defendants for voluntary manslaughter, attempted manslaughter and using and discharging a firearm in relation to a crime of violence. Slatten moved to dismiss the charges against him as time-barred, which this Court ultimately granted by writ of mandamus. In re Slatten, No. 14-3007 (D.C. Cir. Apr. 18, 2014). The government thereafter obtained an indictment charging Slatten with first-degree murder. The defendants were tried jointly in the summer of 2014, and after seven weeks of deliberation, the jury returned guilty verdicts on all counts except three. The district court sentenced Slatten to life imprisonment, and it sentenced Slough, Liberty and Heard to the mandatory term of imprisonment of thirty years for their convictions under 18 U.S.C. § 924(c), plus one day on all of the remaining counts.
We begin with the defendants’ challenges to the applicability of MEJA. The defendants argue that they are entitled to acquittal on all counts because MEJA does not authorize their prosecution. Alternatively, even if their actions do fit within MEJA’s scope, the defendants maintain that the jury was erroneously instructed regarding MEJA. On both claims, we disagree.
Historically, civilians accompanying American armed forces overseas were subject to military court-martial for crimes committed in a host country. See Reid v. Covert, 354 U.S. 1, 3-4 (1957) (plurality op.). In a pair of opinions, however, the United States Supreme Court put an end to that practice, deeming it unconstitutional because the courts-martial failed to provide civilians with certain constitutional rights guaranteed by the Fifth and Sixth Amendments. Id. at 5 (“[W]e reject the idea that when the United States acts against citizens abroad it can do so free of the Bill of Rights.”); Kinsella v. Singleton, 361 U.S. 234, 249 (1960) (civilian defendant “is protected by the specific provisions of Article III and the Fifth and Sixth Amendments and . . . her prosecution and conviction by court-martial [was] not constitutionally permissible”).
Thereafter, many crimes committed by civilians overseas fell into a jurisdictional vacuum as generally our country’s criminal statutes do not apply extraterritorially and, “[a]lthough host foreign nations [did] have jurisdiction to prosecute such acts committed within their nation, they frequently decline[d] to exercise jurisdiction when an American [was] the victim or when the crime involve[d] only property owned by Americans.” H.R. Rep. No. 106–778, Pt. 1, at 5 (2000); accord United States v. Arnt, 474 F.3d 1159, 1161 (9th Cir. 2007).
In 2000, the Congress began to address the “jurisdictional gap” by enacting MEJA. H.R. Rep. No. 106–778, at 5. In its original version, MEJA authorized the prosecution of extraterritorial crimes committed by civilians employed by the Department of Defense (DOD) or its contractors. See 18 U.S.C. § 3267(1)(A) (2000). Following a series of high-profile offenses committed by non-Defense Department contractors—including those committed by private contractors employed by the United States Interior Department at the Abu Ghraib prison in Baghdad, Iraq—the Congress expanded MEJA’s scope. See 150 CONG. REC. S6863 (daily ed. June 16, 2004). Indeed, then-United States Senator Jeff Sessions—the chief sponsor of the 2004 amendment—acknowledged that the amendment’s purpose was to address a jurisdictional gap through which “private contractors who may not have in every instance been directly associated with the Department of Defense . . . might not be prosecutable under [MEJA].” Id. Sessions noted that the gap “highlighted [the Congress’s] need to clarify and expand the coverage of the act” by giving “the Justice Department authority to prosecute civilian contractors employed not only by the Department of Defense but by any Federal agency that is supporting the American military mission overseas.” Id.
Senator Charles Schumer likewise noted that the proposed amendment addressed “a dangerous loophole in our criminal law that would have allowed civilian contractors who do the crime to escape doing the time.” Id. at S6864.
As amended, then, two key sections of MEJA work together to authorize the prosecution of qualifying offenses committed by a civilian overseas: Section 3261 and Section 3267. See 18 U.S.C. §§ 3261, 3267.
18 U.S.C. § 3261 provides:
(a) Whoever engages in conduct outside the United States that would constitute an offense punishable by imprisonment for more than 1 year if the conduct had been engaged in within the special maritime and territorial jurisdiction of the United States—
(1) while employed by or accompanying the Armed Forces outside the United States
...
shall be punished as provided for that offense.
18 U.S.C. § 3267 sets out alternative definitions of “employed by the Armed Forces outside the United States” depending on the defendant’s employment status. Section 3267(1)(A)(iii)(II) applies to the defendants and provides as follows:
(1) The term “employed by the Armed Forces outside the United States” means--
(A) employed as . . .
(iii) an employee of a contractor (or subcontractor at any tier) of . . .
(II) any . . . Federal agency . . . to the extent such employment relates to supporting the mission of the Department of Defense overseas . . . .
When Section 3267(1)(A)(iii)(II) applies, we believe there are two preliminary questions posed by MEJA’s text: 1) whether the defendant’s criminal conduct occurred “while employed by” a non-DOD contractor; and 2) whether his employment (not his conduct) “relates to supporting” the DOD overseas mission. See 18 U.S.C. §§ 3261, 3267. The latter question, however, is subject to an additional restriction. Section 3267(1)(A)(iii)(II)’s “to the extent” clause operates as a temporal limitation applicable only to non-DOD contractors. See id. That is, because MEJA authorizes the prosecution of only those crimes a defendant commits “while” employed by a non-DOD contractor and “to the extent” such employment relates to a DOD mission, it applies only if the defendant’s employment at the time of the offense relates to supporting a DOD mission. See id. (emphasis added).
Although the United States Supreme Court has yet to address Section 3267(1)(A)(iii)(II)’s “relates to” language, it has interpreted similar language broadly. For example, in Smith v. United States, the Supreme Court concluded that “[t]he phrase ‘in relation to’ is expansive,” noting that “[a]ccording to Webster’s, ‘in relation to’ means ‘with reference to’ or ‘as regards.’” 508 U.S. 223, 237-38 (1993). Likewise, in District of Columbia v. Greater Washington Board of Trade, the Supreme Court interpreted “relate to,” as used in the Employee Retirement Income Security Act of 1974, to include any law that “has a connection with or reference to” a covered benefit plan, thereby “giv[ing] effect to the ‘deliberately expansive’ language chosen by Congress.” 506 U.S. 125, 129 (1992) (emphasis added) (some internal quotation marks omitted) (quoting Pilot Life Ins. Co. v. Dedeaux, 481 U.S. 41, 46 (1987); Shaw v. Delta Air Lines, Inc., 463 U.S. 85, 97 (1983)); accord Morales v. Trans World Airlines, Inc., 504 U.S. 374, 383 (1992) (“For purposes of the present case, the key phrase, obviously, is ‘relating to.’ The ordinary meaning of these words is a broad one—‘to stand in some relation; to have bearing or concern; to pertain; refer; to bring into association with or connection with[.]’” (citing BLACK’S LAW DICTIONARY 1158 (5th ed. 1979)). Circuit precedent, too, employs a broad interpretation. We have noted that the “ordinary meaning” of “relating to” is a “broad one,” see Friedman v. Sebelius, 686 F.3d 813, 820 (D.C. Cir. 2012) (internal quotation marks omitted) (quoting Morales, 504 U.S. at 383), and that “a statutory provision containing the phrase therefore has ‘broad scope,’” id. (quoting Metro. Life Ins. Co. v. Mass., 471 U.S. 724, 739 (1985)).
Having addressed both MEJA’s required elements and expansive scope, we next consider whether the evidence was sufficient to support jurisdiction under MEJA.4 The district court denied the defendants’ motion for judgment of acquittal on this ground and the Court must affirm so long as any reasonable factfinder could conclude that the evidence, viewed most favorably to the government, satisfied each element beyond a reasonable doubt. United States v. Kayode, 254 F.3d 204, 212 (D.C. Cir. 2001); see Jackson v. Virginia, 443 U.S. 307, 319 (1979).
We begin with two unchallenged elements. It is undisputed that all of the charges against all four defendants are within MEJA’s scope as it relates to included offenses. See supra 6-7; 18 U.S.C. § 3261(a). Moreover, all four defendants were employed by Blackwater Security Consulting, LLC, a contractor of the United States Department of State. JA 3743, 3760, 3776, 3794, 1228-29. At the time of the Nisur Square attack, they were therefore “employee[s] of a contractor (or subcontractor at any tier) of . . . [a] Federal agency.” See 18 U.S.C. § 3267(1)(A)(iii)(II). The remaining question is whether that employment, at the time of the attack, related to supporting DOD’s mission.
The government sufficiently established the DOD’s overseas mission. By 2007, “the mission of the Department of Defense overseas”—specifically, in Iraq—went beyond military operations against the insurgency. Id. Witnesses testified that the Defense Department mission was to rebuild the war-torn country, including the fostering of economic and political stability. United States Army Colonel Michael Tarsa testified that the military’s goal was to “stimulat[e] local governance” by “identifying local leaders [and] trying to organize them.” JA 1374. Tarsa also recounted that the military sought to improve the Iraqis’ “quality of life” by “restoring essential services, sewer, water, electricity [and] trash removal” and by “foster[ing] economic development,” all with the hope that such restoration would “dissuad[e] people from joining the insurgency.” JA 1373-77. Tarsa’s testimony was echoed by United States Marine Corps Officer Shelby Lasater, who testified that, as the United States’ presence in Iraq continued, the mission became “to rebuild the country and set up a government.” JA 1478-79. Then-Deputy Secretary of Defense Gordon England affirmed that the Defense Department “strategy” was to “help the Iraqi people build a new Iraq with constitutional representative government that respects civil rights and has security forces sufficient to maintain domestic order and keep Iraq from becoming a safe haven for terrorists.” JA 2949.
The government also produced abundant evidence that the defendants’ Blackwater employment supported the Department of Defense’s expanded mission at the time of the Nisur Square attack. Paralleling the testimony of Tarsa, Lasater and England, Blackwater guard Matthew Murphy testified that Blackwater’s “clients . . . the State Department [were] trying to bring along the country, . . . trying to mentor the Iraqi government and . . . get them up and running.” JA 1044. England also testified that the “U.S. Government had to rely on all of its departments and agencies in order to achieve the mission in Iraq.” JA 2950. The State Department was an important part of the rebuilding effort the Defense Department was engaged in; its diplomats were helping the Iraqis restore their country. Blackwater employed the defendants to provide security for the diplomats whose work plainly supported the DOD mission. The defendants’ employment, then, “relate[d] to”—that is, had a “connection with or reference to,” see Greater Wash. Bd. of Trade, 506 U.S. at 129 (internal quotation marks omitted)—supporting the Defense Department’s rebuilding mission.
In addition, the defendants’ contracts required them to complete unspecified “security-related duties requested by Blackwater or [the State Department] in support of the Engagement.” JA 3761. This necessarily requires consideration of the types of duties that Blackwater or the State Department in fact requested in order to determine whether they “relate[] to supporting the mission of the Department of Defense.” 18 U.S.C. § 3267(1)(A)(iii)(II). The evidence showed that, consistent with this contract provision, Blackwater employees were assigned to assist distressed military units during firefights, train Army security escorts and provide escorts to Provincial Reconstruction Teams when Army escorts were unavailable. JA 1622-23, 1762-64, 2956. Although it may be true that the defendants did not themselves participate in these assignments, this evidence nevertheless illustrated for the jury the types of “security-related duties” within the scope of the defendants’ employment. JA 3761.
The defendants’ employment “relate[d] to supporting the [DOD overseas] mission” in another way; it allowed military personnel previously responsible for providing State Department security to concentrate exclusively on their rebuilding mission. See 18 U.S.C. § 3267(1)(A)(iii)(II). Tarsa affirmed that the Defense Department was “able to reduce the amount of [its] platoons . . . . dedicated for Department of State security convoy missions” as “Blackwater took the majority of those tasks.” JA 1381. The platoons were then able to return to, inter alia, “the continued development of the Iraqi security forces.” JA 1382. United States Army Lieutenant Peter Decareau and England corroborated Tarsa’s testimony. JA 2581 (testimony of Army Lieutenant Peter Decareau) (agreeing that “from roughly February 2007 going forward, [Decareau’s] company and platoons within it did not need to provide [State Department] escort service missions anymore,” allowing his platoon “to focus on what [he] described as civil affairs and . . . night operation missions”); JA 2952 (testimony of Deputy Secretary Gordon England) (before Blackwater’s arrival, State Department “was draining personnel from the DOD mission”). Again, then, the defendants’ employment, which increased the manpower available to the military by replacing military personnel previously assigned to guard State Department personnel, had some “bearing or concern” regarding—that is, “relate[d] to”—supporting the Defense Department mission. See Morales, 504 U.S. at 383 (internal quotation marks omitted) (citing BLACK’S LAW DICTIONARY 1158 (5th ed. 1979)). Providing security to State Department personnel who themselves acted jointly with the Defense Department to aid the Iraqi people and whose protection would have continued to require military personnel but for the defendants’ employment necessarily “relate[d] to” supporting the Defense Department’s mission.
The defendants attempt to narrow MEJA’s scope by reading the “to the extent” language of 18 U.S.C. § 3267(1)(A)(iii)(II) and the “while employed” language of 18 U.S.C. § 3261 as more than a temporal limitation. They argue that MEJA applied “only in the limited capacities or at those limited times” when Blackwater guards actively and directly supported the Defense Department mission. Joint Appellants’ Br. 59. That is, they claim that MEJA required the jury to consider not their employment but instead their challenged actions to determine whether those actions—that is, securing Nisur Square—supported the Defense Department mission. Id. at 41, 58-60. But, as noted, MEJA’s scope is not so narrow. Instead, the most natural conjunctive reading of “while employed by,” as used in 18 U.S.C. § 3261, and “to the extent,” as used in 18 U.S.C. § 3267, is one that interprets these provisions as establishing that the point in time when the defendants’ actions occurred is the benchmark by which their employment’s relation to a DOD mission is measured.5 See supra 10. The defendants’ misreading of the statute to require that their challenged actions must relate to a Defense Department mission violates both MEJA’s text and its purpose. MEJA’s goal, after all, was to close “a dangerous loophole in our criminal law that would have allowed civilian contractors who do the crime to escape doing the time.” 150 CONG. REC. S6863.
Alternatively, the defendants maintain that we should look not to their on-the-ground actions but only to their Blackwater contract to determine whether they were “employed by the Armed Forces outside the United States.” Joint Appellants’ Br. 50-52. Because their contract required them to provide security for State Department personnel, rather than to further a Defense Department mission, they argue that MEJA does not authorize their prosecution. Id. at 53. We decline to take such a cramped view of MEJA’s text given the “deliberately expansive” language used by the Congress. See Greater Wash. Bd. of Trade, 506 U.S. at 129.
Finally, the defendants insist that the rule of lenity requires construing MEJA in their favor. The rule of lenity, however, applies only if, “after considering text, structure, history, and purpose, there remains a grievous ambiguity or uncertainty in the statute such that the Court must simply guess as to what Congress intended.” Maracich v. Spears, 133 S. Ct. 2191, 2209 (2013) (internal quotation marks omitted) (quoting Barber v. Thomas, 560 U.S. 474, 488 (2010)); accord Reno v. Koray, 515 U.S. 50, 65 (1995) (rule of lenity applies “only if . . . [the Court] can make no more than a guess as to what Congress intended” (internal quotation marks omitted)). “The rule [of lenity] comes into operation at the end of the process of construing what Congress has expressed, not at the beginning as an overriding consideration of being lenient to wrongdoers.” Maracich, 133 S. Ct. at 2209 (alteration in original) (quoting Callanan v. United States, 364 U.S. 587, 596 (1961)). Although the phrase “relates to” gives MEJA a broad scope, breadth does not equal ambiguity. See Penn. Dep’t of Corr. v. Yeskey, 524 U.S. 206, 212 (1998) (“[T]he fact that a statute can be applied in situations not expressly anticipated by Congress does not demonstrate ambiguity. It demonstrates breadth.” (internal quotation marks omitted)). Moreover, to the extent—if any—that MEJA’s text is ambiguous, MEJA’s “context, structure, history, and purpose resolve it.” Abramski v. United States, 134 S. Ct. 2259, 2272 n.10 (2014); see supra at 7-9. We conclude that the rule of lenity is inapplicable here.
The defendants also challenge the district court’s jury instructions regarding MEJA. “Whether the district court properly instructed the jury is ‘a question of law that we review de novo.’” United States v. Ring, 706 F.3d 460, 465 (D.C. Cir. 2013) (quoting United States v. Orenuga, 430 F.3d 1158, 1166 (D.C. Cir. 2005)). Our responsibility is to “determine whether, taken as a whole, [the instructions] accurately state the governing law and provide the jury with sufficient understanding of the issues and applicable standards.” United States v. DeFries, 129 F.3d 1293, 1304 (D.C. Cir. 1997) (alteration in original) (emphasis added); accord Ring, 706 F.3d at 465. An “improper instruction on an element of the offense violates the Sixth Amendment’s jury trial guarantee.” Neder v. United States, 527 U.S. 1, 13 (1999).
The district court instructed the jury on the meaning of “employed by the Armed Forces outside the United States” as follows:
[T]he definition of ‘employed by the Armed Forces outside the United States’ includes not only a direct employee or contractor of the Armed Forces of the United States, but also a contractor (including a subcontractor at any tier) or an employee of a contractor (or subcontractor at any tier) of any Federal agency of the United States Government to the extent: (1) such employment relates to supporting the mission of the Department of Defense overseas . . . .
...
[T]he Government may prove that the defendant was ‘employed by the Armed Forces’ by establishing that:
(a) the defendant was employed as a contractor, or an employee of a contractor (including a subcontractor at any tier) of any federal agency, and
(b) that the defendant’s employment related to supporting the mission of the Department of Defense overseas.
JA 497-98.
The challenged jury instruction was not erroneous. First, it quoted MEJA’s “to the extent” clause verbatim: “‘[E]mployed by the Armed Forces outside the United States’ includes . . . an employee of a contractor . . . of any Federal agency of the United States Government to the extent . . . such employment relates to supporting the mission of the Department of Defense overseas.” Id. (emphasis added); see 18 U.S.C. §§ 3261, 3267. Granted, the instruction also stated that the government could establish jurisdiction if the jury found “the defendant’s employment related to supporting the [DOD] mission,” JA 498; taken out of context, a juror could conceivably understand the latter statement to mean jurisdiction would exist if “the defendant’s employment [at any time] related to supporting the mission” of DOD, see id. But we “do not read the language thus criticized in isolation.” Jones v. United States, 404 F.2d 212, 215-16 (D.C. Cir. 1968); see also Cupp v. Naughten, 414 U.S. 141, 147-48 (1973). To the contrary, we have “long recognized that one ambiguous part of an instruction may be made clear by another unambiguous part of the same instruction,” United States v. Gaviria, 116 F.3d 1498, 1510 (D.C. Cir. 1997), and the “to the extent” language unambiguously precludes an erroneous, all-or-nothing understanding of the statute, see John Hancock Mut. Life Ins. Co. v. Harris Trust & Sav. Bank, 510 U.S. 86, 104-05, 109 (1993).
The defendants’ challenge to the instruction largely repeats their argument against the applicability of MEJA itself. For example, they argue the district court erred in failing to instruct the jury that it could consider only their contract employment to determine whether they were “employed by the Armed Forces . . . .” Joint Appellants’ Br. 66-68. The defendants also revive their claim that, even if the jury could consider evidence aliunde their employment contract, it should have been instructed that MEJA applied only when the defendants were acting within the scope of their employment and only if their specific acts supported the DOD mission. Id. at 64-66. To that end, they proposed the following instruction:
[I]f you find that part of a defendant’s contract employment for the Department of State related to supporting the mission of the Department of Defense, and part of his contract employment did not relate to supporting the mission of the Department of Defense, you must consider whether the work the defendant was performing at the time of the conduct charged in the indictment related to supporting the mission of the Department of Defense in Iraq. For purposes of this case, a Defendant is ‘employed by the Armed Forces of the United States’ only if the contract employment he was performing at the time of the charged conduct related to supporting the mission of the Department of Defense in Iraq.
JA 473. In construing MEJA’s text, the Court earlier rejected the premise underlying the defendants’ instruction, see supra 15-16, and continues to do so in this context.
The defendants’ remaining argument is that the district court “grievously erred” by failing to instruct the jury expressly that diplomatic security is a State Department responsibility. Joint Appellants’ Br. 68. They note that 22 U.S.C. §§ 4801-02 assigns to the Secretary of State responsibility for “the security of diplomatic operations . . . abroad,” id. § 4801(b)(1), and requires the Secretary to implement measures “to provide for the security of United States Government operations of a diplomatic nature,” id. § 4802(a)(1). For the defendants, there is a “fundamental conflict between that statutory assignment of responsibility [to the State Department] and MEJA’s requirement that the defendants’ contract employment relate to supporting the Defense Department’s mission.” Joint Appellants’ Br. 74-75. The defendants offered the following instruction:
The Defendants in this case were independent subcontractors employed by the Department of State to provide personal security to State Department personnel in Baghdad, Iraq. By law, the provision of personal security to State Department personnel overseas is the responsibility of the Department of State.
JA 475.
The defendants fail to recognize, however, that State Department contractors—and their employees—could help meet the State Department’s duty to provide security for diplomatic operations abroad and, at the same time, support the Defense Department’s overseas mission. Blackwater without question employed the defendants to protect State Department personnel, see, e.g., JA 1169-74, 1853-54, 3861; the critical question for the jury, however, was whether, in carrying out that responsibility, the defendants’ employment also “relate[d] to supporting the mission of the Department of Defense overseas,” see 18 U.S.C. § 3267(1)(A)(iii)(II). We agree with the district court that the defendants’ proposed instruction “would just be confusing to the jury.” JA 3279-80. The district court’s charge, “taken as a whole . . . accurately state[d] the governing law and provide[d] the jury with sufficient understanding of the issues and applicable standards.” DeFries, 129 F.3d at 1304.
The defendants next complain the District of Columbia was an improper venue for their trials. On November 18, 2008, the United States District Court for the District of Columbia issued an arrest warrant for Ridgeway, and Ridgeway voluntarily flew to Washington, D.C. from California. Once he arrived in Washington, he was met by an FBI agent, formally booked and taken to district court to plead guilty to one count of voluntary manslaughter and one count of attempted voluntary manslaughter. While Ridgeway was not put in handcuffs when apprehended by the FBI, he testified he believed he was under arrest. After pleading guilty, Ridgeway was permitted to return to his home.
If an offense is committed outside the United States and involves charges against multiple people, Congress has declared venue to be proper in the district where any of the joint offenders are first arrested. 18 U.S.C. § 3238. The defendants argue the government improperly used the arrest of Jeremy Ridgeway, one of the other turret gunners who fired in Nisur Square, to satisfy the venue statute because (1) Ridgeway was not arrested in connection with their charged offenses, (2) he was not a “joint offender” with the defendants and (3) the government impermissibly manufactured venue in the District of Columbia.
Since the parties dispute the meaning of the phrases “joint offender” and “is arrested” in the venue statute, we focus on the statute’s text. Section 3238 states, “[t]he trial of all offenses begun or committed . . . out of the jurisdiction of any particular State or district[] shall be in the district in which the offender, or any one of two or more joint offenders, is arrested.” Id. “The Government bears the burden of establishing by a preponderance of the evidence that venue is proper with respect to each count charged against the defendant[s].” United States v. Morgan, 393 F.3d 192, 195 (D.C. Cir. 2004). When reviewing whether venue was properly established, this Court views the evidence “in the light most favorable to the Government.” Id. In order to assure the case would be heard in the District of Columbia, the government entered into a plea agreement with Ridgeway and arranged for him to travel to the District of Columbia from his home in California to be arrested.
While this Court has not specifically defined “arrested” in the context of Section 3238, our sister circuits have consistently interpreted it to mean situations “‘where the defendant is first restrained of his liberty in connection with the offense charged.’” United States v. Wharton, 320 F.3d 526, 537 (5th Cir. 2003) (quoting United States v. Erdos, 474 F.2d 157, 160 (4th Cir. 1973)). We believe this definition is correct and that the test is easily satisfied here. The record shows the district court issued the arrest warrant for Ridgeway. On the same day, he was arrested by the FBI in the District of Columbia and formally booked. The defendants argue Ridgeway’s freedom was never restrained because he voluntarily flew across the country from California and was never put in handcuffs or confined in a cell, but this misconstrues the meaning of arrest. Supreme Court precedent makes clear an arrest can either be carried out with “physical force [against a suspect] . . . or, where that is absent, submission to the assertion of authority.” California v. Hodari D., 499 U.S. 621, 626 (1991). What really matters is whether a “reasonable person would have believed that he was not free to leave.” United States v. Mendenhall, 446 U.S. 544, 554 (1980). Here, Ridgeway testified he understood himself to be under arrest when he was seized by the FBI upon arrival in the District of Columbia. Any reasonable person in Ridgeway’s position would have understood he was not free to leave.6 Ridgeway was first arrested in the District of Columbia; and that arrest established venue here.
The defendants interpret the phrase “joint offender” to mean each offender must possess “a mutual intent” with others to commit a crime. Joint Appellants’ Br. 97–98. Because Ridgeway did not form this mutual intent, they claim he was not a joint offender. They rely primarily on the fact that many of the cases examining Section 3238 have involved collaborative criminal schemes. See, e.g., United States v. Levy Auto Parts of Can., 787 F.2d 946, 948–49 (4th Cir. 1986) (involving a conspiracy to sell munitions); United States v. Hong Vo, 978 F. Supp. 2d 49, 64 (D.D.C. 2013) (involving a conspiracy to commit visa fraud).
However, this interpretation impermissibly narrows Section 3238 to one category of offenses. As noted by the district court, Black’s Law Dictionary defines a joint offense as a crime “committed by the participation of two or more persons.” BLACK’S LAW DICTIONARY 838 (6th ed. 1990). While the defendants are certainly correct that a joint crime can be committed by several defendants with a mutual intent to achieve a criminal goal, this is not the only type of crime in which a group may participate. In fact, Federal Rule of Criminal Procedure 8(b) allows multiple defendants to be charged with the same offense “if they are alleged to have participated in the same act or transaction, or in the same series of acts or transactions, constituting an offense or offenses.”7
Accordingly, instead of limiting “joint offender” to one category of offenses that requires participation by multiple people, a more natural reading of the statutory text encompasses not only people with a mutual intent to commit a crime, but also anyone who has joined others in participating in the same act or transaction constituting a crime or crimes. This interpretation is further supported by this Court’s preference for joint trials in cases involving multiple defendants. See United States v. Manner, 887 F.2d 317, 324 (D.C. Cir. 1989). We have explained joint trials “promote efficiency” and noted that “this preference is especially strong when the respective charges require presentation of much the same evidence, testimony of the same witnesses, and involve [multiple] defendants who are charged . . . with participating in the same illegal acts.” United States v. Wilson, 605 F.3d 985, 1015, 1016 (D.C. Cir. 2010). These rationales are especially compelling in a case like this. Ridgeway was working in the relevant convoy on the day of the Nisur Square attack, and, with other defendants, he opened fire on the civilians in Nisur Square. Thus, in order to convict Ridgeway, the government would be required to present the same evidence and to rely upon testimony from the same witnesses as they would for the other defendants. Also, concerns for efficiency are especially compelling here because many of the witnesses reside in Iraq. Multiple trials would mean arranging multiple international trips for the witnesses, which would likely be both difficult to schedule and costly. Thus, our interpretation of Section 3238 is consistent with both the text of the statute and the general preference for joint trials.8 We conclude “joint offenders” encompasses all defendants who participated in the same act or transaction constituting the charged crimes.
Thus, it is clear Ridgeway was a joint offender. Testimony at trial established Ridgeway was present in Nisur Square as a member of the Raven 23 convoy and that he fired at civilians to the south, to the west and finally to the north. Ridgeway participated in the “same series of acts or transactions” that gave rise to the prosecution, FED. R. CRIM. P. 8(b), which makes him a joint offender. The defendants’ emphasis on personal participation in every count returned by the grand jury focuses on the wrong thing. Although it is true that the government must show that “venue is proper with respect to each count charged,” United States v. Lam Kwong-Wah, 924 F.2d 298, 301 (D.C. Cir. 1991), it does not follow that Ridgeway must have personally participated in each act giving rise to each count. Section 3238 requires that Ridgeway be a “joint offender,” which is satisfied by his participation in the same series of acts or transactions giving rise to those counts, i.e., Ridgeway’s persistent, multi-directional shooting throughout the entire Nisur Square attack. 18 U.S.C. § 3238; see also FED. R. CRIM. P. 8(b) (“All defendants need not be charged in each count.”). Because Ridgeway clearly did participate in the Nisur Square shootings, he was a joint offender within the meaning of Section 3238.
Likewise, the defendants’ claim that the government manufactured venue, while appealing on an intuitive level, fails in light of the congressional design of Section 3238. The text of the statute gives the government a choice regarding prosecution of an extraterritorial crime: either arresting a cooperative defendant in a jurisdiction of the government’s choosing or seeking an indictment in the district where a defendant resides. See 18 U.S.C. § 3238 (stating venue “shall be in the district in which the offender, or any one of two or more joint offenders, is arrested”); see also United States v. Gurr, 471 F.3d 144, 155 (D.C. Cir. 2006) (reading Section 3238’s clauses disjunctively). Thus, by choosing to arrest Ridgeway in the District of Columbia, the government simply exercised the choice given to it under the statute. Something more is required to sustain a claim that venue has been manufactured. See United States v. Spriggs, 102 F.3d 1245, 1250–51 (D.C. Cir. 1996). For example, “where the key events occur in one district, but the prosecution, preferring trial elsewhere, lures a defendant to a distant district for some minor event simply to establish venue,” a claim of manufactured venue might have traction. Id. at 1251. However, Section 3238 forecloses that scenario here by explicitly allowing the government to choose where to arrest a cooperative joint offender. Thus, venue was proper in the District of Columbia.1
The defendants say the district court abused its discretion in denying a new trial based on the victim impact statement (“VIS”) from Officer Monem that appeared to contradict his testimony at trial.
During the trial, the government called Sarhan Dheyab Abdul Monem, an Iraqi police officer, to testify about his observations in Nisur Square during the attack. Before the shooting began, Monem was stationed at a traffic kiosk located close to where the Raven 23 caravan had stopped. Monem testified that, after he heard shots being fired from the Raven 23 caravan, he heard a scream coming from the Kia, so he approached the vehicle. As he neared the Kia, he saw its driver had been shot in the head. After examining the driver’s injury, Monem testified he moved in front of the convoy and attempted to tell them to stop shooting by speaking to them in Arabic and waving his hands. When this had no effect, Monem stated he returned to the Kia and attempted to help the Kia’s passenger, who was weeping and holding the body of the driver. According to Monem, the car began to slowly move forward, which caused the Raven 23 squad to begin firing at the Kia again. When the second burst of gunfire erupted, Monem fled back to his kiosk and hid behind it to shield himself from the bullets.
After the defendants were convicted, the government solicited victim impact evidence from Iraqis who were present in Nisur Square on the day of the attack, including Monem. The purpose of this evidence was to allow victims and witnesses to describe how the Nisur Square shootings had affected them, including “feelings of anger, rage, blaming self, . . . helplessness, [and] vulnerability.” JA 4032. In his VIS, Monem wrote about his guilt for not being able to help the Kia’s occupants; but, he also painted a different picture of what happened that day. Contrary to his testimony at trial, Monem’s VIS stated he “remained in [his] traffic cabin unable to move nor think.” JA 637. The VIS also stated Monem heard the driver of the Kia pleading with his mother to get out of the car before they were both killed. When the government produced Monem’s VIS to the court and defense counsel four days later, the defendants raised concerns about the inconsistency of the VIS with Monem’s trial testimony.
This prompted the government to conduct an ex parte telephone conversation with Monem regarding his VIS. The government did not record this conversation and instead submitted notes to the district court summarizing Monem’s responses. According to these notes, Monem allegedly stated he did not understand his VIS to be a factual statement but rather an “expression” of what he imagined it was like to be the Kia driver. The notes also indicated Monem reaffirmed key portions of his trial testimony, including that he approached the Kia and saw the driver was dead.
The defendants moved for a new trial based upon this newly-discovered evidence, but the district court denied their motions without conducting a hearing. United States v. Slough, 144 F. Supp. 3d 4, 5 (D.D.C. 2015). The defendants now appeal, claiming the district court committed reversible error by denying their motions for a new trial. Slatten argues the VIS provides direct evidence of his innocence by establishing that the person he was convicted of murdering was alive after the shooting in Nisur Square began, thus disproving the government’s theory of the case. Additionally, the other defendants argue the VIS shows Monem committed perjury at trial and that this new account refutes many facts vital to the government’s case. Finally, all defendants argue the district court reversibly erred by failing to hold a hearing to examine Monem regarding the conflict his VIS created with his testimony at trial.
Trial courts have broad discretion when deciding whether to grant a new trial based on newly-discovered evidence. Thompson v. United States, 188 F.2d 652, 653 (D.C. Cir. 1951). A district court’s denial of a new trial is reviewed for abuse of discretion. United States v. Oruche, 484 F.3d 590, 595 (D.C. Cir. 2007). In order to obtain a new trial because of newly-discovered evidence, the party seeking a new trial must prove: (1) the evidence was discovered after the trial; (2) the party acted diligently in its attempts to procure the newly-discovered evidence; (3) the evidence relied on is not “merely cumulative or impeaching,” (4) the evidence is “material to the issues involved” in the case and (5) the evidence is “of such nature that in a new trial it would probably produce an acquittal.” Thompson, 188 F.2d at 653. “[W]hen perjury by a prosecution witness is discovered after trial and when the prosecution did not know of the perjury until then,” a defendant is entitled to a new trial only if he can prove he “would probably be acquitted on retrial.” United States v. Williams, 233 F.3d 592, 594 (D.C. Cir. 2000).
We begin by noting the unusual nature of the allegedly exculpatory evidence upon which the defendants rely. In homicide cases, victim impact statements are typically used during the sentencing phase of a trial. They allow the government to either offer a “quick glimpse” into a life taken by the defendant or to “demonstrat[e] the loss to the victim’s family and to society which has resulted from the defendant’s homicide.” Payne v. Tennessee, 501 U.S. 808, 822 (1991). Nothing in the record suggests the government intended to use the VIS in this case as substantive evidence of guilt. See JA 637 (asking Monem to describe how the crime affected him); cf. Payne, 501 U.S. at 856 (Stevens, J., dissenting) (stating victim impact statements “shed[] no light on the defendant’s guilt or moral culpability”). However, this is exactly the purpose for which the defendants now seek to use Monem’s VIS.
Monem’s statements viewed in isolation could be seen as puzzling if not contrary to his testimony at trial, as the defendants suggest. Considered in context, however, as responses to the specific questions posed by the government in preparing for sentencing after the jury had returned its verdicts finding the defendants guilty, his statements take on another cast. Still, the Court is troubled by the government’s conduct upon discovery of what might appear to contradict his trial testimony. Instead of inviting defense counsel to participate in the phone call with Monem or—at a minimum—recording the phone conversation, the government conducted an ex parte phone call and offered nothing but its own notes as evidence of what was said during the call. Because the Court has no way of verifying what was said, we do not believe the notes constitute a repudiation of Monem’s contradictory statements. However, even if we view the statements in the light most favorable to the defendants and consider them to be an admission of perjury and a recantation of Monem’s trial testimony, we do not believe the district court abused its discretion in declining to grant a new trial. In order to succeed on their claims, the defendants must prove Monem’s VIS would probably result in an acquittal at a new trial. Thompson, 188 F.2d at 653. “This is a high bar to cross.” United States v. Celis, 608 F.3d 818, 848 (D.C. Cir. 2010).
Here, even if Monem’s statements did constitute a recantation of his trial testimony, we do not believe they meet this high bar. This holds especially true for Liberty, Slough and Heard, whose convictions regarding victims to the south, east, west and north of Nisur Square did not depend on Monem’s testimony regarding the first moments of the shooting attack. Regarding Dr. Al-Khazali, the Kia passenger, other evidence corroborated Monem’s testimony that the Kia was stopped when the first shots were fired, and Officer Al-Hamidi testified about his own efforts to stop the shooting independent of Monem’s.
The only defendant with even a slight chance of a different outcome based on Monem’s contradictory VIS statements was Slatten. However, even if we were to assume that Monem would reaffirm his VIS testimony, acquittal would still not be likely due to the other record evidence that al-Rubia’y was killed instantly. As discussed in more detail below, testimony from Officer Al-Hamidi established that al-Rubia’y was shot in the head, killing him instantly. Only then did the car begin rolling forward unguided. Comparing this consistent testimony from Officer Al-Hamidi with this new testimony from Monem, which only came to light after he was prompted to describe “feelings of anger, rage, blaming self, . . . helplessness, [and] vulnerability” resulting from the Nisur Square shootings, JA 4032, there is little reason to believe the outcome of the case would have been any different. Thus, it was hardly an abuse of discretion for the district court to refuse to grant a new trial based on evidence unlikely to produce a different outcome.
Furthermore, the district court did not abuse its discretion in declining to hold an evidentiary hearing regarding Monem’s VIS. This Court gives a trial judge “broad discretion in ruling on a motion for a new trial, both in his actual decision and in what he considers before making that decision.” Lam Kwong-Wah, 924 F.2d at 308. “A motion for a new trial can ordinarily be decided . . . without an evidentiary hearing, and a district court’s decision not to hold such a hearing may be reversed only for abuse of discretion.” United States v. Kelly, 790 F.2d 130, 134 (D.C. Cir. 1986); see also United States v. Kearney, 682 F.2d 214, 219 (D.C. Cir. 1982) (noting the need for a hearing is diminished “where the trial judge has had an opportunity to observe the demeanor and weigh the credibility of the witness at trial”). Here, the district court judge presided over the entirety of this multiple-week trial and observed Monem’s testimony when it was given. Also, Monem’s testimony was subject to thorough cross-examination by several defense attorneys and—unlike the VIS—was largely corroborated by other evidence presented at trial. All of these factors combined made the district court “well qualified to rule on the motion for a new trial” based solely on the written motions and the evidence submitted. Kearney, 682 F.2d at 220. While we agree with the defendants that a hearing would have been helpful to clarify what Monem meant when he wrote his VIS, we cannot say it was an abuse of discretion for the district court to decide the motion without a hearing.
Liberty and Slatten challenge the sufficiency of the evidence supporting their convictions. The Court must affirm if, “after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Jackson, 443 U.S. at 319. The jury is “entitled to draw a vast range of reasonable inferences from evidence, but may not base a verdict on mere speculation.” United States v. Harrison, 103 F.3d 986, 991 (D.C. Cir. 1997) (quoting United States v. Long, 905 F.2d 1572, 1576 (D.C. Cir. 1990)). Applying this “highly deferential” standard, United States v. Williams, 836 F.3d 1, 6 (D.C. Cir. 2016), the Court concludes that the evidence supporting the convictions was sufficient, with the exception of one of Liberty’s attempted manslaughter convictions.
Liberty, the driver of the third vehicle in the four-vehicle convoy of Blackwater guards, was convicted of eight counts of voluntary manslaughter, twelve counts of attempted manslaughter and a Section 924(c) weapons count. The jury was also instructed, in view of the charges under 18 U.S.C. § 2, that it could convict on each of these counts if it determined that Liberty aided and abetted their commission. Liberty contends that there was insufficient evidence that he unjustifiably fired his weapon at, or caused the death of, any victim, or that he took some action to aid another defendant’s unjustifiable shooting at any specific victim.
First, Raven 23 member Jeremy Krueger’s testimony provided evidence from which the jury could find that Liberty fired at the white Kia in which the passenger, Dr. Al-Khazali, was killed. Krueger, who was in the vehicle in front of Liberty’s, testified that each member of the Raven 23 team had been assigned roles and that he was responsible for securing one sector of Nisur Square. Krueger testified that he saw shots fired at the Kia from the vehicle behind by “someone sitting in the driver’s position, and [he] assumed it to be Mr. Liberty, just based on [his] knowledge of [Liberty’s] position that day, [of the team members’] assignments.” 8/5/14 (PM) Tr. 34:3-9. Although Krueger was not in a position to see the shooter’s face, he inferred that the shooter was the driver because the shooter was “sitting with his back against the driver’s seat” like a driver would ordinarily sit. 8/5/14 (PM) Tr. 91:10-12.
Liberty maintains that this evidence could just as plausibly describe Jimmy Watson, the Raven 23 leader, who testified that he leaned across Liberty’s body and shot into the Kia from the passenger seat. This, however, ignores that Krueger testified the shooter was sitting with his back “up against the [driver’s] seat,” 8/5/15 (PM) Tr. 91:17-20, and that the shooter’s upper body was above the steering wheel, not “tilting down or out” of the vehicle like someone who was leaning across the driver’s body. Id. at 35:25-36:11. Watson also described Liberty as having his back “up against the seat.” 7/28/14 (PM) Tr. 79:14-15. Although Watson testified that Liberty did not shoot into the Kia “at that time,” 7/28/14 (PM) Tr. 50:4-6, the jury could have reasonably disbelieved him because Watson’s testimony was inconsistent on other key points, such as whether Slatten shot first, and what Liberty did when he exited the vehicle during the tow hook-up. Compare, e.g., 7/28/14 (PM) Tr. 30:18-22, with id. at 30:23-31:20; id. at 95:12-16, with id. at 95:25-96:13. Further, even if the jury credited Watson’s testimony on that point, it could reasonably have understood his other testimony that, after the initial burst of shooting, he told Liberty “to open the door again and fire again,” id. at 50:13-14, to show that Liberty had taken part in the second burst of shooting at the Kia. Given the close proximity of the convoy to the Kia, 7/1/14 (PM) Tr. 138:4, the jury could reasonably find that Liberty’s shots hit Dr. Al-Khazali, contributing to her death.
With regard to the victims shot to the south, Watson testified that Liberty “was engaging in the direction of the south” as the two of them fired simultaneously out of the driver’s side door, which was oriented in that direction. 7/28/14 (PM) Tr. 61:18-62:7. Eddie Randall, another Raven 23 member, testified that he saw shots fired southward from the same door, which, given Liberty’s driving assignment, he too assumed were fired by Liberty. 8/11/14 (AM) Tr. 80:5-82:3. There was also testimony from Jeremy Ridgeway that in the immediate aftermath of the shootings in Nisur Square, Liberty admitted that he had done “another Grey 55,” which Ridgeway explained meant firing blindly out of his porthole with his rifle across his lap. 7/31/14 (AM) Tr. 44:3-9. Liberty maintains nonetheless that the Grey 55 testimony did not establish shooting “in a particular direction at a particular time,” Joint Appellants’ Reply Br. 54, but because Liberty’s door faced south until the convoy pulled away to leave the Square, the jury could reasonably have found that the Grey 55 shots went south.
That said, evidence showing only that Liberty fired south is not especially probative that he hit any particular victim because there were multiple shooters, multiple victims in that area and “millions of square feet to the south.” 7/29/14 (AM) Tr. 31:21-22 (Watson). Even so, and even were the Court to assume that the evidence already discussed was insufficient to show Liberty was directly responsible for the victims to the south, there was sufficient evidence to support Liberty’s convictions under an aiding-and-abetting theory. See United States v. Branch, 91 F.3d 699, 731-32 (5th Cir. 1996). To establish aiding and abetting, the government had to prove, beyond a reasonable doubt, that Liberty intentionally “facilitated any part . . . of [the] criminal venture,” with enough “knowledge [of the crime to] enable[] him to make the relevant legal (and indeed, moral) choice” to opt out instead. Rosemond v. United States, 134 S. Ct. 1240, 1246, 1249 (2014). Given the evidence before the jury, we “find no difficulty in holding that actively participating in a gunbattle in which a gunman kills [multiple victims] can aid and abet that killing” even if the government cannot prove which gunman killed which victim. Branch, 91 F.3d at 732. This is especially true where, as here, the gunfire of each shooter hindered potential escape, leaving victims exposed to the others’ bullets. Cf. Rosemond, 134 S. Ct. at 1247 n.6.
The evidence showed that with Slough, Ridgeway and Heard firing to the south from their location and Watson and Liberty firing south from inside their vehicle, victims in that area had nowhere to turn in order to escape. Krueger, for instance, described people running and “one gentleman particularly hiding behind a car and kind of frantically wondering what to do and how to get away,” as rounds impacted the car and the ground around him. 8/5/14 (AM) Tr. 47:12-16. This unarmed man appeared to be “deciding which way to run, and he just didn’t know what the safe direction was,” before eventually falling as he tried to make a run for safety. Id. at 48:18-49:18. Similarly, Raven 23 member Matthew Murphy described a man near the white Kia that he perceived to be shot while the man was “looking around, . . . trying to think about what he was going to do, you know, [how to] get out of the way” of the gunfire. 7/1/14 (AM) Tr. 11:2-12:10. From this evidence, the jury could reasonably find that Liberty’s southern shooting aided the gunmen who actually inflicted the harm.
Liberty suggests that there is no evidence that he knew what anyone other than Watson was doing, and therefore his shooting could not have knowingly aided in the commission of any crime with the requisite intent. To the extent he relies on the fact that Watson was never charged as a co-defendant, aiding-and-abetting liability can arise even when the principal offense goes uncharged. United States v. Catalan-Roman, 585 F.3d 453, 473 (1st Cir. 2009). The jury could readily find that Watson’s southern shooting was unjustified and thus criminal—for instance, when Watson repeatedly shot at and eventually hit a man running away from the convoy—and further, that Liberty knew of the lack of justification and yet continued to fire his weapon. Liberty’s failure to opt out satisfies the mens rea element, which can arise during the crime’s commission. Rosemond, 134 S. Ct. at 1249. Even assuming that Liberty may not have been able to see Slough, Ridgeway, or Heard, who were firing their weapons from above Liberty, Watson testified he was aware that they were firing their weapons, and the jury could have reasonably imputed that same awareness to Liberty, who was sitting beside Watson. A number of southern-facing Raven 23 members, including Mark Mealy, who was the turret gunner in the lead vehicle, testified to the lack of apparent justification for any southern shooting from the convoy. E.g., 8/4/14 (PM) Tr. 91:18-21 (Ridgeway was unable “to personally identify a legitimate target” as he fired south); 7/15/14 (PM) Tr. 113:16-114:16 (Mealy “didn’t see any reason” for the shots fired at people attempting to flee). Despite Liberty’s claim that Mealy had a different vantage than Liberty, the jury could reasonably find that Liberty, who was looking in the same direction, continued to fire his gun despite the unjustified shooting that was happening around him.
With regard to the two victims shot to the east of the Nisur Square traffic circle, Mealy testified that an unidentified Raven 23 member fired east while the disabled convoy vehicle was being hooked up for evacuation. Mealy saw an Iraqi man with his hands in the air, saw the Raven 23 guard kneeling outside his vehicle holding an M-4 rifle with an ACOG scope, and after he heard two or three shots, Mealy saw the Iraqi man double over with a stomach wound. Watson’s testimony placed Liberty outside their vehicle during the tow hook-up, and although his testimony about what Liberty was doing was inconsistent with his statement to the grand jury that he did not know what Liberty did, at trial he testified Liberty helped with the hook-up. Two rifle magazines later found in Nisur Square bore Liberty’s name and inasmuch the three other guards who were outside during the hook-up testified that they did not fire their weapons, the jury could reasonably infer that Liberty killed Ali Hussein.
Liberty disputes the import of this evidence. First, he maintains that the magazines prove only that he fired his weapon that day, something he does not deny. The jury, however, could have reasonably viewed this evidence to show that Liberty fired his weapon from outside the vehicle, consistent with Mealy’s testimony. Watson did not recall Liberty dropping a spent magazine while shooting inside the vehicle, and it is unclear how else the magazines might have ended up outside the vehicle. Second, Mealy testified that whoever shot Hussein used an ACOG scope. That Liberty had been issued an EOTech scope undercuts the inference that Liberty killed Hussein, but it does nothing to preclude it; the jury heard testimony that swapping scopes “would [not] be that hard,” 7/28/14 (PM) Tr. 97:20-21, and that over time one guard went from using an EOTech to an ACOG and then back again. Third, Liberty further points out that Mealy described the victim as wearing blue, traditional garb, and no victim matched that description. There was, however, testimony that Hussein was shot in the stomach, which is consistent with Mealy’s testimony. Fourth, Liberty maintains that six Raven 23 members testified that no shots were fired during the tow hook-up. This overstates the testimony to a degree, because Frost, Krueger and Rhodes testified that they did not recall or perceive any shots being fired during the hook-up, while Murphy and Ridgeway testified only that no incoming shots (i.e., shooting at the convoy) were fired. Even so, the jury was entitled to credit Mealy’s specific recollection over that of the others. Jackson, 443 U.S. at 319.
In sum, although Liberty may have poked holes in some of the evidence against him, this Court does not review the jury’s verdict de novo. See id. Given Mealy’s testimony and the spent magazines found outside the vehicle, Liberty has not shown that no reasonable factfinder could find him guilty of Hussein’s death. The jury could reasonably have credited Mealy’s testimony and evidence that only Liberty fired his weapon during the hook-up efforts. On the other hand, the government has pointed to no evidence linking Liberty to the attempted manslaughter of Mahdi Al-Faraji, who was also shot to the east of Nisur Square. Mealy testified only to seeing the Blackwater guard taking “two or three shots” to the east, hitting a single victim. 7/15/14 (PM) Tr. 120:1-121:11. An inference that one of those shots also hit a second victim would be based on mere speculation, Harrison, 103 F.3d at 991, and consequently that count of attempted manslaughter must be vacated for insufficient evidence.
Slatten was convicted of first-degree murder in the death of Ahmed Al-Rubia’y, the driver of the white Kia. At the time of the shooting, Slatten was laying across a bench in the back of the third vehicle, aiming his weapon south out of a driver’s side porthole. The government’s theory was that while traffic was at a standstill waiting for the Blackwater convoy to exit the Square, Slatten fired two shots from a sniper rifle into the Kia windshield, killing Al-Rubia’y instantly and setting into motion the day’s horrific events. See, e.g., 6/17/14 (PM) Tr. 7:16-9:19. Slatten maintains there is insufficient evidence to support that theory and that testimony from two government witnesses disproves it.
The jury heard testimony that at the outset, while all traffic was stopped in Nisur Square, there were two distinct pops, after which the Kia started to roll slowly and a woman began to scream. Officer Al-Hamidi testified that he approached the car to see that Al-Rubia’y’s “whole face was full of blood,” that the woman in the passenger seat was holding him and screaming “My son, my son,” and then the car “started moving slowly because the young man was killed, and he did not have control of the car.” 7/2/14 (AM) Tr. 92:11-93:10. Officer Monem similarly testified that, on his approach, he saw that Al-Rubia’y had been shot in the middle of his forehead, while a nearby witness saw a hole in the blood-splattered driver’s side windshield. From this, the jury could reasonably conclude that the first shots were fatal, and Slatten does not dispute this point.
The jury also heard testimony from Jimmy Watson, who was in the front passenger seat of Slatten’s vehicle. Although unable to recall at trial, Watson had testified before the grand jury to his fairly strong recollection that Slatten fired twice and then the gunners began shooting, and this testimony was admitted into evidence at trial. Watson described Slatten’s first shots as “very rhythmic . . . retort then retort,” 7/28/14 (PM) Tr. 34:14-15, consistent with others’ descriptions of the fatal shots as “two pops,” e.g., 7/14/14 (PM) Tr. 76:2-3. Watson could not see Slatten’s target, but testified that Slatten was aimed generally south, which was “the direction . . . where the [Kia] was,” 7/28/14 (PM) Tr. 38:25-39:2. Similarly, Eddie Randall testified that he heard the first shots come from in front of him, where Slatten’s vehicle was positioned. Slough was in Slatten’s vehicle, and on direct examination Randall testified that nothing he saw in Slough’s appearance indicated to him that Slough had taken the shots.
The jury heard further testimony that Slatten was Raven 23’s best marksman, who carried a sniper rifle that had been modified to be on a hair trigger, and that Slatten was known for his particular disdain for Iraqis, viewing himself as getting payback for 9/11. Indeed, Jeremy Ridgeway testified that Slatten later recounted shooting someone who was taking aim at the convoy, with Slatten saying matter-of-factly that he “popped his grape” and caused him to slump forward. 7/31/14 (AM) Tr. 49:5-16. From this evidence, a reasonable jury could understand this to describe Al-Rubia’y, after being shot in the middle of the forehead by Raven 23’s best marksman.
Slatten’s bias against Iraqis, moreover, provided a basis for finding that Slatten had fired first, in the absence of any insurgent fire or other threat to the heavily armed convoy. Witnesses testified that Slatten had previously engaged in a pattern of preemptively shooting (or encouraging others to preemptively shoot) at targets in order to draw fire from potential adversaries. See United States v. Long, 328 F.3d 655, 661 (D.C. Cir. 2003).
Slatten, like Liberty, pokes some holes in the government’s theory but does not overcome the jury’s reasonable determination of guilt in light of the evidence before it. He makes much of the fact that Ridgeway testified that Slatten confessed to killing an active shooter who slumped forward when shot, while Al-Rubia’y was an unarmed driver who, according to Officer Monem, slumped to the side. The jury could reasonably find that Slatten’s “active shooter” claim to Ridgeway was self-serving and therefore not trustworthy. See Williamson v. United States, 512 U.S. 597, 599-600 (1994). As the district court found, the jury had “ample support in the record to find that Slatten was lying or unreasonably mistaken” about an active shooter. United States v. Slough, 144 F. Supp. 3d 4, 13 (D.D.C. 2015). Aside from two witnesses who thought they heard shots from what sounded like an AK-47, there was no evidence of any active shooters that day, let alone a seated one. And as for Monem’s testimony that Al-Rubia’y was slumped to the side, to the extent it conflicted with Slatten’s recounting the jury was entitled to disregard such a minor discrepancy. Given the lack of evidence that Slatten fired any other shots that day, the jury could reasonably understand his “popped his grape” comment to describe Al-Rubia’y, who had been shot in the middle of his forehead.
With regard to Watson’s testimony, Slatten highlights the equivocation at trial as to who shot first, Slatten or the gunners. He also points out that Watson testified to hearing three AK-47 shots outside the convoy prior to Slatten firing, which Slatten suggests shows that he was returning incoming fire rather than firing at the Kia. Slatten’s attempt to revive the defendants’ discredited self-defense theory lacks merit—the jury necessarily rejected it, and the district court noted that “no witness . . . ever testified that they ever saw [an insurgent’s] weapon at the scene,” 4/13/15 Tr. 152:6-8. In his reply brief, Slatten suggests that the initial shots Watson heard might have come from the gunners rather than insurgents, but Watson testified that he first heard AK-47 rounds in the distance, at which point either Slatten or the gunners began to fire. The jury could reasonably conclude that, despite his equivocation, Watson’s testimony supported the government’s theory that Slatten fired first, and also, in light of the overwhelming evidence to the contrary, that there was no incoming fire directed at the convoy.
Slatten points out that Jeremy Krueger testified hearing 5.56 caliber rounds as the first shots fired, which Krueger claimed he could distinguish from the sound of 7.62 caliber rounds, the caliber that Slatten’s sniper rifle would have fired. This testimony is probative, but not forcefully so in view of Krueger’s acknowledgment that his hearing was limited by noise-reducing ear protection and being inside of a different vehicle than the shooter. Still, it was for the jury to resolve the credibility of Krueger’s testimony that depending on the situation and circumstances, he “still [thought he] could” distinguish caliber rounds even when inside another vehicle and while wearing ear protection. 8/5/14 (AM) Tr. 21:22-22:2.
Slatten’s strongest counterevidence comes from Officers Monem and Al-Hamidi, who testified that the first shots came from the gunners. Al-Hamidi was “100 percent certain” that the first shots came from a gunner on top of a vehicle, 7/2/14 (PM) Tr. 35:4-15, while Monem “did not see the explosion from the mouth of [a gunner’s] rifle, but it was so close” that he could tell from the sound that it did. 6/23/14 (AM) Tr. 12:12-13. This testimony, however, does not “disprove[]” the government’s theory of Slatten’s guilt. Slatten’s Br. 47. It simply creates a dispute of fact, and it was the jury’s responsibility to weigh the officers’ conflicting testimony against that of Watson to resolve the dispute. Jackson, 443 U.S. at 319. That a different jury might have resolved the conflict differently is not tantamount to showing that no reasonable fact-finder could conclude that Slatten shot first. See id. Without any other plausible target for Slatten’s first shots, and given the proximity of the Kia, it would have been reasonable for the jury to find that Slatten killed Al-Rubia’y.
Slatten further contends that his re-indictment for first-degree murder, after he successfully challenged his previous indictment for manslaughter, attempted manslaughter and weapons charges, constituted vindictive prosecution. Our review of the district court’s contrary finding is for clear error. United States v. Safavian, 649 F.3d 688, 692 (D.C. Cir. 2011).
In December 2008, Slatten was indicted jointly with his co-defendants for identical counts of manslaughter, attempted manslaughter and weapons charges. When the government later concluded that “tainted” testimony against Slatten had been presented to the grand jury, see generally Kastigar v. United States, 406 U.S. 441 (1972), it moved to voluntarily dismiss the indictment as to Slatten. The district court granted defendants’ motion to dismiss the indictment as to all defendants on related Kastigar grounds. United States v. Slough, 677 F. Supp. 2d 112, 166 & n.67 (D.D.C. 2009). On appeal, this Court reversed and remanded the dismissal as to all defendants except Slatten, concluding that the district court had already granted the government’s motion to dismiss and “taken Slatten out of the case for now.” Slough, 641 F.3d at 547.
Two years later, the government secured a superseding indictment charging Slatten with the manslaughter of Al-Rubia’y, and jointly charging all defendants with various other manslaughter, attempted manslaughter and weapons counts. Slatten moved to dismiss the charges as time-barred because this Court’s earlier reversal of dismissal had not applied to him and the limitations period had continued to run. The district court denied his motion, and Slatten filed a petition for a writ of mandamus. This Court granted the writ upon concluding that its earlier reversal “clearly applied” only to Slatten’s co-defendants. In re Slatten, No. 14-3007, at 1 (D.C. Cir. Apr. 7, 2014). It denied the government’s own petition for rehearing, observing that the government’s concern about a miscarriage of justice if its prosecution of Slatten were time-barred was caused by the government’s “inexplicable failure to [timely] reindict Slatten.” In re Slatten, No. 14-3007, at 2 (D.C. Cir. Apr. 18, 2014). The government subsequently obtained an indictment charging Slatten with first-degree murder in the death of Al-Rubia’y, a charge not subject to the statute of limitations. 18 U.S.C. §§ 1111(b), 3281. The prosecutor conveyed to Slatten’s counsel an offer to reduce the charge to manslaughter if Slatten would waive any limitations defense, explaining that the murder charge was the government’s only remaining option for holding Slatten accountable.
Slatten moved to dismiss the first-degree murder charge on due process grounds, arguing that the increased charge constituted vindictive prosecution. The district court denied the motion, finding that the facts did not raise a presumption of vindictive prosecution. It found that Slatten exercised his rights in a pre-trial context, in which courts are far more hesitant to presume vindictiveness. It further found that the prosecutor’s offer to reduce the charge was a permissible pre-trial negotiation, akin to plea bargaining, and that no other facts suggested that the government was improperly motivated. Instead, the government simply sought to hold Slatten accountable for a heinous crime it believed he committed. The district court also rejected Slatten’s argument that the government was required to provide a contemporaneous explanation of its decision to increase the charge. United States v. Slatten, 22 F. Supp. 3d 9, 12-16 (D.D.C. 2014).
The Due Process Clause prohibits prosecutors from “upping the ante” by filing increased charges in order to retaliate against a defendant for exercising a legal right. Blackledge v. Perry, 417 U.S. 21, 27-28 (1974). At the same time, however, prosecutors have broad discretion to enforce the law, and their decisions are presumed to be proper absent clear evidence to the contrary. United States v. Armstrong, 517 U.S. 456, 464 (1996). Thus, to succeed on a claim of vindictive prosecution, a defendant must establish that the increased charge was “brought solely to ‘penalize’ [him] and could not be justified as a proper exercise of prosecutorial discretion.” United States v. Goodwin, 457 U.S. 368, 380 n.12 (1982) (emphasis added). This can be accomplished in two ways: through objective evidence showing actual vindictiveness, or through evidence “indicat[ing] a ‘realistic likelihood of vindictiveness,’” which gives rise to a presumption that the government must then attempt to rebut. United States v. Meyer, 810 F.2d 1242, 1245 (D.C. Cir. 1987) (quoting Blackledge, 417 U.S. at 27). Slatten relies on the latter, presumptive route.
In presumption cases, the Supreme Court has distinguished between pre-trial and post-trial settings. Goodwin, 457 U.S. at 381. In a pre-trial setting, “the prosecutor’s assessment of the proper extent of prosecution may not have crystallized,” so an increase in charges may be the result of additional information or further consideration of known information, rather than a vindictive motive. Id. The routine exercise of many pre-trial rights also weakens any inference of vindictiveness, i.e., that a prosecutor would retaliate simply because a defendant sought a jury trial or pleaded an affirmative defense. Id. On the other hand, a post-trial increase in charges is unlikely to be based on new information, and thus it is “much more likely to be improperly motivated than is a pretrial decision.” Id. For this reason, a presumption of vindictiveness will “automatically” arise whenever charges are increased post-trial, but in the pre-trial context, a defendant must provide additional facts sufficient to show that “all of the circumstances, when taken together, support a realistic likelihood of vindictiveness.” Meyer, 810 F.2d at 1245-46.
The parties dispute whether the first degree murder indictment is properly characterized as occurring in a pre-trial or post-trial setting. The government maintains that as a factual matter the charging decision was unquestionably made prior to Slatten’s trial, while Slatten maintains that it was more akin to a post-trial decision because it followed a hotly contested mandamus proceeding in which this Court chastised the government for failing timely to reindict him. Slatten also points out that this case was closely watched by U.S. and Iraqi leaders, citing former Vice President Biden’s assurance to former Iraqi President Talabani that the earlier Kastigar dismissal would be appealed. Anthony Shadid, Biden Says U.S. Will Appeal Blackwater Case Dismissal, N.Y. TIMES, Jan. 23, 2010. This Court has acknowledged that particularly in an important, highly publicized case, a prosecutor “being but human ‘may have a personal stake in [obtaining a] conviction and a motivation to engage in self-vindication.’” Safavian, 649 F.3d at 692 (quoting United States v. Stanfield, 360 F.3d 1346, 1362 (D.C. Cir. 2004)). Especially when compared to the routine pretrial motions identified in Goodwin, 457 U.S. at 381, there can be little question that the extraordinary mandamus grant here, followed by a rather sharply-worded criticism in denying reconsideration, in a high-profile prosecution with international ramifications no less, had far greater potential to give rise to a vindictive motive. But these unusual facts do not convert the pre-trial setting into a post-trial one in which a presumption would automatically apply; rather they constitute “additional facts” that support the finding of a presumption. Meyer, 810 F.2d at 1245-46.
Slatten’s other contentions, derived from the Court’s analysis in Meyer, 810 F.2d at 1246-47, do not fare as well. He maintains that he received disparate treatment from his co-defendants, but he ignores that his co-defendants had no viable limitations defense and were not similarly situated, as the Meyer defendants were. See 810 F.2d at 1246. Next, although the government had twice considered the facts and twice charged manslaughter, “the initial charges filed by a prosecutor may not reflect the extent to which an individual is legitimately subject to prosecution.” Goodwin, 457 U.S. at 382. Here, the government’s decision in the superseding indictment to charge Slatten alone in the death of Al-Rubia’y indicates that it continued to develop facts after its initial charging decision. But even where the government has full knowledge of the facts, it can initially exercise its discretion to bring lesser charges. E.g., United States v. Saltzman, 537 F.3d 353, 361 (5th Cir. 2008). Moreover, Slatten is incorrect that, as in Meyer, “[t]he only relevant intervening event” before the charge increase was Slatten’s assertion of rights. Slatten’s Br. 21-22. Here, Slatten exercised his right to file a mandamus petition and this Court granted it, nullifying the government’s ability to proceed on the existing charges. Finally, the government’s offer to charge manslaughter in exchange for Slatten waiving his limitations defense was not improper. As the district court found, Slatten was advised by competent counsel and was free to accept or reject the government’s offer, which was a permissible give-and-take. See Bordenkircher v. Hayes, 434 U.S. 357, 363 (1978); see also Paradise v. CCI Warden, 136 F.3d 331, 335 n.6 (2d Cir. 1998).
Still, although it is a close question, the unusual, high-profile and potentially embarrassing context surrounding Slatten’s mandamus petition could be viewed to “support a realistic likelihood of prosecutorial vindictiveness.” Meyer, 810 F.2d at 1246. In that situation, the burden would shift to the government to provide any objective evidence showing a non-retaliatory justification for the increased charge. Safavian, 649 F.3d at 694. The government has met this “admittedly minimal” burden, id., pointing to this Court’s grant of Slatten’s mandamus petition that left the government with no alternative but to charge him with murder or else see “a heinous crime” go unpunished. Appellee’s Br. 88 (quoting Slatten, 22 F. Supp. 3d at 14). In closely analogous circumstances, the Second Circuit found no vindictiveness when a defendant successfully pursued a statute of limitations defense in the state’s highest court, and the prosecution then reindicted him for a capital charge not subject to any limitations period. Paradise, 136 F.3d at 334, 336. As here, the capital charge “was simply the only charge available[] after the other charges had been dismissed . . . as time barred,” and the government’s desire to see the crime punished “does not amount to a constitutional violation.” Id. at 336.
This does not mean, as amicus asserts, that prosecutors can permissibly “up the ante” any time a defendant succeeds on appeal. Amicus Br. 27. In many cases, the same charges will remain available to the prosecution after a defendant’s successful appeal, and any increase in the charges will still give rise to the specter of vindictiveness. See Meyer, 810 F.2d at 1245-46. And even if the same charges are unavailable on retrial, a defendant can still marshal any available evidence of actual vindictiveness to show that the prosecution’s purported desire to see the crime punished is mere pretext. Nor should this result cause doubt about whether Slatten was punished for exercising a legal right. Again, the Court relies little on the government’s stated desire to see the crime punished, and instead places dispositive weight on the intervening grant of mandamus, as this Court has held that an adverse appellate ruling can provide an objective basis for the prosecution’s new charging decision. Safavian, 649 F.3d at 694. It is also immaterial that the new charge was the result of the prosecution’s initial mistake in allowing the limitations period to run. See Paradise, 136 F.3d at 336 n.7. Slatten and amicus urge that the government can only increase charges when, “through no fault of its own,” the government learns of new information after the initial charging decision. United States v. Jamison, 505 F.2d 407, 416-17 (D.C. Cir. 1974). But the Supreme Court has rejected the “presum[ption] that every prosecutor is infallible.” Goodwin, 457 U.S. at 382 n.14; see also Paradise, 136 F.3d at 336 n.7. Finally, as the district court ruled, the government was not required to state its justification when it obtained the first-degree murder indictment because “the prosecutor is not required to sustain any burden of justification” until after the defendant comes forward with evidence of vindictiveness. Goodwin, 457 U.S. at 384 n.19.
With the presumption rebutted, Slatten’s vindictive prosecution challenge fails because he does not offer any evidence to support a finding of actual vindictiveness. Safavian, 649 F.3d at 694. The district court reached the same conclusion, albeit by considering the government’s objective justification to rule out a presumption of vindictiveness at step one, rather than to rebut it at step two. Otherwise, the substance of its analysis is much the same as our own, and as such, we hold that the district did not err, let alone clearly err, in rejecting Slatten’s defense of prosecutorial vindictiveness.
We next turn to Slatten’s challenge to the district court’s denial of his Rule 14 motion to sever his trial from that of a co-defendant. Slatten argued for severance because he sought to introduce exculpatory evidence—the co-defendant’s admissions that he, not Slatten, initiated the Nisur Square attack by firing on the white Kia—evidence inadmissible in a joint trial with the co-defendant. See Kastigar v. United States, 406 U.S. 441, 458-61 (1972) (government cannot prosecute declarant based on immunized statement). The district court denied Slatten’s motion to sever, finding the co-defendant’s admissions constituted inadmissible hearsay. We disagree. Because the co-defendant’s admissions were vital to Slatten’s defense and possessed sufficient circumstantial guarantees of trustworthiness, we believe they were admissible under Federal Rule of Evidence 807. Accordingly, because the district court erroneously denied severance, we reverse Slatten’s first-degree murder conviction—Count One of the superseding indictment—and remand his case for a new trial.
As we outlined earlier, Slatten’s first-degree murder conviction arose from the killing of the driver of the white Kia. As the Raven 23 convoy entered Nisur Square on September 16, 2007, shift leader Jimmy Watson gave the command to “lock[] down” the area to aid the movement of other Blackwater teams operating nearby. JA 1776, 1846-48, 1856-57, 2351-52. With the help of Iraqi policemen, the Blackwater convoy brought traffic in the Square to a halt, as was their usual procedure. After the traffic stopped, shots rang out. The shots, originating from the Raven 23 convoy, targeted and hit a white Kia, shattering its windshield and striking its driver, Ahmed Haithem Ahmed Al-Rubia’y, in the head.10 General gunfire then began as Raven 23 team members fired on Iraqi civilian pedestrians in several directions in Nisur Square and the surrounding area.
As noted, the government maintained that Slatten’s shot was the match that ignited the Nisur Square firestorm—that Slatten intentionally opened fire on the white Kia because of an anti-Iraqi animus. See also Appellee’s Br. 103 (“The evidence also showed that Slatten had both the intent and motive to open the firing in the Square. His hatred toward Iraqis stood out, even among those who held such views.”). The government insisted that the Nisur Square attack was part of Slatten’s plan to “get[] payback for 9/11,” JA 2117, and the white Kia presented him with the target for which he had been waiting.
But in the hours and days following the Nisur Square attack, it was another member of the Raven 23 team—a co-defendant here—who said that he had fired the first shots at the white Kia. SA 1, 4, 6-7. Just hours after the shooting, the co-defendant was interviewed and debriefed by State Department investigators operating in Baghdad. SA 1. Before his interview, the investigators told the co-defendant that if he was “honest and truthful, that nothing would be used against [him], and that they were there to gather information not to be used in a criminal setting.” SA 22. During his first debriefing, the co-defendant told the investigators that he had “engaged and hit the driver” of the white Kia sedan. SA 1. The investigators’ corresponding report states:
[T]he team came into and locked down the circle. Traffic was very heavy, but responded to their commands to stop. A white vehicle approached the team at a high rate of speed and would not stop despite [the co-defendant’s] hand signals and throwing a water bottle. Other civilians tried to waive the vehicle down, but it still would not stop. [The co-defendant] engaged and hit the driver.
SA 1. Two days later, on September 18, 2007, the co-defendant signed a sworn written statement regarding the Nisur Square attack. SA 3-5. As with his earlier statement, the co-defendant’s September 18 statement was made with the understanding that “neither [the co-defendant’s] statements nor any information or evidence gained by reason of [his] statements [could] be used against [him] in a criminal proceeding, except that if [he] knowingly and willfully provide[d] false statements or information, [he could] be criminally prosecuted for that action under 18 United States Code, Section 1001.” SA 3. In his second statement, the co-defendant repeated his earlier statement:
As our motorcade pulled into the intersection I noticed a white four door sedan driving directly at our motorcade from the west bound lane. I and others were yelling, and using hand signals for the car to stop and the driver looked directly at me and kept moving toward our motorcade. Fearing for my life and the lives of my teammates, I engaged the driver and stopped the threat.
SA 4. On September 20, 2007, the co-defendant again spoke to State Department investigators and with the same limited use condition as obtained in his first two interviews. SA 6-7; 22-23. The investigators’ report recounted the co-defendant’s statement made at that time:
On the day of the incident . . . [the co-defendant] was positioned just west of the police booth that is located near the north end of the median south of the Circle. A white car was moving north on Jinub Street toward the motorcade, and [the co-defendant] gave commands for the driver to stop. The car did not stop, and [he] engaged it with his M4. [The co-defendant] is not sure whether he was the first one to fire during this incident. He is not aware of any shots being fired before his. The car kept moving straight toward the motorcade without braking. [The co-defendant] used one magazine of M4 ammunition to engage the white car.
SA 6-7.
Taken together, then, the co-defendant’s statements relate a different version of the Nisur Square events from that presented by the government at trial. The government’s case against Slatten hinged on his having fired the first shots, his animosity toward the Iraqis having led him to target the white Kia unprovoked. See supra 40-44. The co-defendant’s statements, however, strike at the heart of that theory and instead point to the co-defendant, not Slatten, as the Blackwater convoy member who first “engaged and hit the driver” of the white Kia. SA 1.
At Slatten’s arraignment, the district court granted the government’s motion to join Slatten’s trial with that of Liberty, Heard and Slough. JA 388-91. Slatten asked the district court to reconsider joinder on two grounds, insisting, first, that, because of his need for a co-defendant’s testimony, severance was essential so that the co-defendant could be called as a witness for Slatten at the latter’s separate trial. See SA 42-43. Further, if, in a joint trial, the co-defendant statements were deemed admissible as exculpatory evidence as to Slatten, then severance was appropriate to protect the co-defendant’s Fifth Amendment right. SA 43.
The district court rejected both rationales and denied the motion to sever. Regarding Slatten’s first argument, the district court concluded that Slatten had failed to show a “reasonable probability” that the co-defendant would be willing to testify at a separate trial, as required by United States v. Ford, 870 F.2d 729, 731 (D.C. Cir. 1989) (when weighing appropriateness of severance based on alleged need for co-defendant’s testimony, court should consider, inter alia, “the likelihood that the co-defendant will testify if the cases are severed”). SA 42-43. The district court further found no constitutional problem in joining Slatten’s and his co-defendant’s trials because the latter’s “statements [were] . . . inadmissible hearsay.” SA 43. Slatten challenges only the second ruling on appeal. See Slatten’s Br. 36-46.
Hearsay is an out-of-court statement that is inadmissible at trial to establish the truth thereof. See FED. R. EVID. 801(c) (defining hearsay); FED. R. EVID. 802 (hearsay generally inadmissible). The hearsay rule is rooted in the belief that an out-of-court statement lacks necessary assurances of veracity. See Williamson v. United States, 512 U.S. 594, 598 (1994) (“The hearsay rule . . . is premised on the theory that out-of-court statements are subject to particular hazards.”). With any statement, a “declarant might be lying; he might have misperceived the events which he relates; he might have faulty memory; [or] his words might be misunderstood or taken out of context by the listener.” Id. To avoid these shortcomings, our judicial system chooses in-court statements that can be tested by “the oath, the witness’ awareness of the gravity of the proceedings, the jury’s ability to observe the witness’ demeanor, and, most importantly, the right of the opponent to cross-examine.” Id. Admitting hearsay would prevent opposing parties, and our judicial system as a whole, from using these checks. United States v. Evans, 216 F.3d 80, 85 (D.C. Cir. 2000) (“The problem with hearsay is that it deprives the defendant of the opportunity to cross-examine the person who uttered the statement at issue.”).
“Nonetheless, the Federal Rules of Evidence also recognize that some kinds of out-of-court statements are less subject to these hearsay dangers, and therefore except them from the general rule that hearsay is inadmissible.” Williamson, 512 U.S. at 598. The enumerated exceptions apply to hearsay that possesses certain guarantees of trustworthiness. See FED. R. EVID. 803-04 (enumerating exceptions and exclusions to hearsay rule).
On appeal, Slatten does not argue his co-defendant’s statements fall outside the definition of hearsay. See FED. R. EVID. 801(c). Indeed, he could not succeed if he did so argue—Slatten acknowledges that he seeks to use his co-defendant’s out-of-court statements to establish the truth thereof, that is, that his co-defendant fired the first shots at the white Kia. See Slatten’s Br. 36. Slatten does, however, challenge the district court’s conclusions that his co-defendant’s statements do not fit within any of three exceptions to the hearsay rule: 1) Rule 804(b)(3)’s statement against interest exception; 2) Rule 803(6)’s business record exception; and 3) Rule 807’s residual hearsay exception.2 SA 43-45.
Ordinarily, the Court reviews the exclusion of a hearsay statement under the abuse of discretion standard. United States v. Moore, 651 F.3d 30, 83 (D.C. Cir. 2011) (per curiam). Nevertheless, for Rule 807, we have enunciated a slightly different standard; namely, we should be “particularly hesitant to overturn a trial court’s admissibility ruling under the residual hearsay exception absent a definite and firm conviction that the court made a clear error of judgment in the conclusion it reached based upon a weighing of the relevant factors.” United States v. North, 910 F.2d 843, 909 (D.C. Cir. 1990) (internal quotation marks omitted) (quoting Balogh’s of Coral Gables, Inc. v. Getz, 798 F.2d 1356, 1358 (11th Cir. 1986) (en banc)).
Because “the legislative history of [Rule 807] indicates that it should be applied sparingly,” we believe it appropriate to engage in a Rule 807 analysis only if it is apparent that no other exception renders a hearsay statement admissible. See SEC v. First City Fin. Corp., 890 F.2d 1215, 1225 (D.C. Cir. 1989); accord United States v. Kim, 595 F.2d 755, 759-66 (D.C. Cir. 1979) (analyzing admissibility of statement under Rule 803(6) before residual hearsay exception analysis). Therefore, before discussing the residual hearsay exception, we briefly turn to Slatten’s arguments that his co-defendant’s statements are admissible under Rule 804(b)(3) and Rule 803(6).
Rule 804(b)(3) provides an exception to the hearsay rule if: “(1) the declarant [is] unavailable, (2) the statement [is] against the declarant’s interest, and (3) corroborating circumstances clearly indicate the trustworthiness of the statement.” Moore, 651 F.3d at 82 (internal quotation marks omitted); see FED. R. EVID. 804(b)(3). Although we agree with Slatten that, at their joint trial, his co-defendant qualified as “an unavailable witness,” see United States v. Harris, 846 F. Supp. 121, 124 n.6 (D.D.C. 1994) (witness “on the advice of counsel, invoked his Fifth Amendment privilege against self-incrimination and did not testify at trial . . . [a]s a result, he became an unavailable witness”), and that his co-defendant’s statements do possess indicia of trustworthiness,12 see infra at Part VII.B.3, Slatten could not show that his co-defendant’s statements were so inculpatory that a reasonable person in the latter’s position would have made the statements only if he believed them to be true; his co-defendant’s statements were immunized and, as a general matter, a self-defense claim is not “clearly” against a declarant’s interest, see United States v. Henley, 766 F.3d 893, 915 (8th Cir. 2014) (affidavit of unavailable declarant “was not clearly against his own interest because in it he claims he shot [the victim] in self defense”); United States v. Shryock, 342 F.3d 948, 981 (9th Cir. 2003) (“The district court did not abuse its discretion by excluding [a declarant’s] statement that he shot the victims in self-defense because the statement was exculpatory, and not against his penal interest.”). Accordingly, it was not an abuse of discretion to conclude, as the district court did, that the co-defendant’s statements did not fit within Rule 804(b)(3)’s exception. See FED. R. EVID. 804(b)(3).
Rule 803(6) provides an exception to the hearsay rule for records that are, inter alia, “kept in the course of a regularly conducted activity of a business.” FED. R. EVID. 803(6). Rule 803(6) does not support the admissibility of the co-defendant’s statements because he himself was not acting in the regular course of business when he made his statements to State Department investigators.13 United States v. Warren, 42 F.3d 647, 656 (D.C. Cir. 1994) (Rule 803(6) “allows admission of statements in [police] reports only if they reflect the maker’s personal knowledge, or if they were reported to the maker, directly or through others, by one who is himself acting in the regular course of business, and who has personal knowledge” (emphasis added) (internal quotation marks omitted)). A “witness’[s] description of [an incident], recorded by [a public official] in his report, is not made in the regular course of the witness’[s] business and does not deserve the presumption of regularity accorded a business record.” United States v. Smith, 521 F.2d 957, 964 (D.C. Cir. 1975). Having rejected Rule 804(b)(3)’s and Rule 803(6)’s applicability, we turn to Rule 807.
Using the United States v. North standard of review, we consider Slatten’s argument that his co-defendant’s statements are admissible under Federal Rule of Evidence 807 (“Rule 807”)—the residual hearsay exception. Rule 807 makes admissible a statement otherwise violative of the hearsay rule if the statement meets five criteria. First, the statement must have “equivalent circumstantial guarantees of trustworthiness” comparable to those found in Rule 803’s and Rule 804’s enumerated hearsay exceptions. FED R. EVID. 807(a)(1). Second, it must be “offered as evidence of a material fact.” Id. § 807(a)(2). Third, the statement must be “more probative on the point for which it is offered than any other evidence that the proponent can obtain through reasonable efforts.” Id. § 807(a)(3). Fourth, “admitting it [must] . . . serve the purposes of these rules and the interests of justice.” Id. § 807(a)(4). And finally, the proponent of the statement must have given “an adverse party reasonable notice of the intent to offer the statement and its particulars, including the declarant’s name and address, so that the party has a fair opportunity to meet it.” Id. § 807(b).
The residual hearsay exception “was designed to encourage the progressive growth and development of federal evidentiary law by giving courts the flexibility to deal with new evidentiary situations which may not be pigeon-holed elsewhere.” United States v. Mathis, 559 F.2d 294, 299 (5th Cir. 1977); see also Dallas Cty. v. Commercial Union Assoc., 286 F.2d 388 (5th Cir. 1961). As the Federal Rules of Evidence Advisory Committee noted, the enumerated hearsay exceptions of Rules 803 and 804, “while they reflect the most typical and well recognized exceptions to the hearsay rule, may not encompass every situation in which the reliability and appropriateness of a particular piece of hearsay evidence make clear that it should be heard and considered by the trier of fact.” FED. R. EVID. 803(24) (advisory committee’s note to 1974 enactment).14
That said, we also recognize that the residual hearsay exception is “extremely narrow and require[s] testimony to be ‘very important and very reliable.’” United States v. Washington, 106 F.3d 983, 1001 (D.C. Cir. 1997) (per curiam) (quoting Kim, 595 F.2d at 766); accord First City Fin. Corp., 890 F.2d at 1225 (“[T]he legislative history of the [residual hearsay] exception indicates that it should be applied sparingly.”). Indeed, were Rule 807 to be liberally applied, the exception might read out the rule. See Akrabawi v. Carnes Co., 152 F.3d 688, 697 (7th Cir. 1998) (“We . . . narrowly constru[e] the residual provision to prevent it from becoming the exception that swallows the hearsay rule.”); Mathis, 559 F.2d at 299 (“[T]ight reins must be held to insure that this provision does not emasculate our well developed body of law and the notions underlying our evidentiary rules.”). Thus, only in the most “exceptional circumstances” does Rule 807 make admissible a statement that does not fall within one of Rule 803’s or Rule 804’s enumerated hearsay exceptions. See Kim, 595 F.2d at 765-66; United States v. Phillips, 219 F.3d 404, 419 & n.23 (5th Cir. 2000) (“The [residual hearsay] exception is to be used only rarely, in truly exceptional cases.” (internal quotation marks omitted)).
We believe this case presents one of those exceptional circumstances. Our analysis begins with Rule 807’s first element—the requirement that the co-defendant’s statements contain “equivalent circumstantial guarantees of trustworthiness” to those ensured by the Rule 803 and Rule 804 hearsay exceptions. See FED R. EVID. 807(a)(1). In assessing trustworthiness, we look to the “totality of circumstances . . . that surround the making of the statement and that render the declarant particularly worthy of belief”; and drawing parallels from the enumerated hearsay exceptions, we must gauge whether the declarant was “highly unlikely to lie.” Idaho v. Wright, 497 U.S. 805, 819-20 (1990). As we have recognized before, “in order to find [a] statement trustworthy, a court must find that the declarant of the prior statement ‘was particularly likely to be telling the truth when the statement was made.’” Washington, 106 F.3d at 1002 (quoting United States v. Tome, 61 F.3d 1446, 1453 (10th Cir. 1995)); accord Rivers v. United States, 777 F.3d 1306, 1314-15 (11th Cir. 2015) (“By requiring hearsay admitted under the residual exception to have circumstantial guarantees of trustworthiness that are like the guarantees of the specific exceptions, Rule 807 is clearly concerned, first and foremost, about whether the declarant originally made the statements under circumstances that render the statements more trustworthy.”).
Several of the circumstances surrounding the co-defendant’s declarations indicate their reliability and manifest that he was likely telling the truth at the time he made his statements. See Washington, 106 F.3d at 1002. For one, during his debriefing interviews with the State Department, the co-defendant had “the incentive . . . to speak truthfully . . . .” See United States v. Bailey, 581 F.2d 341, 349 (3d Cir. 1978) (emphasis added). He was almost completely immunized when he made his statements—he faced no criminal liability (absent one exception discussed below) as a result of his providing the investigators his account of the Nisur Square attack. See SA 1, 4, 6-7, 22-23. Immunity can indicate trustworthiness, particularly if the immunized statements do not cast blame or “divert attention” to another. See, e.g., Curro v. United States, 4 F.3d 436, 437 (6th Cir. 1993); see also United States v. Henderson, 406 F. Supp. 417, 428 n.19 (D. Del. 1975) (“The purpose of an immunity statute is to obtain truthful information, most frequently regarding otherwise undiscoverable offenses.”). But cf. United States v. Gomez–Lemos, 939 F.2d 326, 333-34 (6th Cir. 1991) (expressing skepticism that immunity makes trustworthy statement “divert[ing] attention to another”). More importantly, the one exception to the co-defendant’s immunity may have been an even greater incentive encouraging his honesty; that is, he faced criminal liability under 18 U.S.C. § 1001 if he made a materially false statement to the investigators and he expressly acknowledged that he could be so prosecuted. See SA 3 (“I further understand . . . that if I knowingly and willfully provide false statements or information, I may be criminally prosecuted for that action under 18 United States Code, Section 1001.”); SA 22-23 (co-defendant testified that his understanding was that if he was “honest and truthful, that nothing would be used against [him]. . . .” (emphasis added)). We have previously concluded that the threat of 18 U.S.C. § 1001 liability bolsters the trustworthiness of a declaration for the residual hearsay exception. First City Fin. Corp., 890 F.2d at 1225 (affirming district court’s application of residual hearsay exception where, inter alia, statement was “subject to criminal prosecution under 18 U.S.C. § 1001”); see United States v. Int’l Bhd. of Teamsters, Chauffeurs, Warehousemen & Helpers of Am., AFL-CIO, 964 F.2d 1308, 1312-13 (2d Cir. 1992) (hearsay statement was reliable because, inter alia, declarants “faced possible criminal sanctions for making false statements” under 18 U.S.C. § 1001).
Additional factors point to the trustworthiness of the co-defendant’s statements. He “consistently reported the essential details of [his] story . . . over the course of multiple [interviews]” on September 16, see SA 1, September 18, see SA 4, and September 20, see SA 6-7. See Al Alwi v. Obama, 653 F.3d 11, 19 (D.C. Cir. 2011). Consistency supports the reliability of his multiple statements and, consequently, his veracity. See United States v. Bumpass, 60 F.3d 1099, 1102 (4th Cir. 1995) (listing consistency of declarant’s statements as a factor in assessing trustworthiness under Rule 804(b)(3)). Other circuits have reached the same conclusion in applying the Rule 807 exception. See United States v. Harrison, 296 F.3d 994, 1005 (10th Cir. 2002) (noting, in Rule 807 analysis, “the consistency of the declarant’s statements” is “a factor that we find particularly persuasive”).
The record also contains evidence “corroborating the veracity of the statement[s].” See Rivers, 777 F.3d at 1315 (quoting Bailey, 581 F.2d at 349) (naming corroboration of veracity of statements as factor to be considered in assessing guarantee of truthfulness under Rule 807); United States v. Hall, 165 F.3d 1095, 1110-11 (7th Cir. 1999) (in gauging trustworthiness of statement under residual hearsay exception, corroboration of declarant’s statement, inter alia, is considered). Iraqi Police Officer Ali Ghalaf Salman Mansur Al-Hamidi was “within feet” of the Blackwater convoy in Nisur Square on the day of the attack. JA 1248. Al-Hamidi testified that the Raven 23 team “started throwing bottles of water” in order to stop traffic. JA 1247-48; accord SA 1 (co-defendant stated that he had “throw[n] a water bottle” to stop traffic). Additionally, Al-Hamidi agreed that, from his proximity to the convoy, he was “100 percent certain” that a man in the co-defendant’s precise position fired the first shots.15 JA 1270. Al-Hamidi’s partner, Sarhan Dheyab Abdul Monem, also testified that, from his “very close” vantage point “about three to four meters away from [the] armored cars,” he also witnessed the first shots coming from the co-defendant’s precise position and “not from the holes or the windows that are in the [Raven 23] vehicles.” JA 797. Blackwater convoy member Jeremy Krueger also provided corroboration, testifying that the first shots he heard in Nisur Square sounded like “5.56 rounds,” the co-defendant’s ammunition, not 7.62 rounds, Slatten’s ammunition. JA 2302-03. Collectively, then, this evidence corroborates the co-defendant’s statements that he “engaged and hit the driver,” of the white Kia, see SA 1, and was unaware “of any shots being fired before his,” see SA 7.
We find that Rule 807’s remaining requirements are also met; indeed, the government raises no dispute in this respect on appeal. There is no doubt that Slatten seeks to offer his co-defendant’s statements “as evidence of a material fact.” FED. R. EVID. 807(a)(2). That is, Slatten seeks to introduce the statements to bolster his defense that his co-defendant—not he—fired the first shots at the white Kia. See Slatten’s Br. 42. After thorough review of the record, we are not aware of evidence “more probative on the point for” which Slatten seeks to admit his co-defendant’s statements.16 See FED. R. EVID. 807(a)(3). The co-defendant’s statements contradict the core of the homicide count against Slatten, charging him with “willfully, deliberately, maliciously, and with premeditation and malice aforethought, [unlawfully killing] the driver of a white Kia sedan.” JA 383. Indeed, the co-defendant acknowledged that he was “not aware of any shots being fired before his,” SA 7, and that he “engaged the driver” to respond to the active threat posed by the white Kia, SA 4. We also believe that “admitting [the co-defendant’s statements] serve[s] the purposes of [the federal evidentiary] rules and the interests of justice.” FED. R. EVID. 807(a)(4). Allowing the jury to weigh the statements—to determine their weight, if any, as against the evidence incriminating Slatten—advances the Federal Rules of Evidence’s goal of “ascertaining the truth and securing a just determination.” Id. § 102. Finally, the record demonstrates that Slatten gave the government “reasonable notice of [his] intent to offer the statement[s].” Id. § 807(b).
In finding Rule 807’s residual hearsay exception inapplicable to the co-defendant’s statements, the district court relied on two points: 1) its determination that the statements lacked “equivalent circumstantial guarantees of trustworthiness” because the co-defendant “provided his statements under the specter of dismissal from his position, or even criminal penalty,” and 2) its belief that Slatten had no additional guarantees of trustworthiness. SA 44. Regarding the first point, the only criminal penalty that the co-defendant faced was 18 U.S.C. § 1001 false statement liability, a factor that weighs in favor (not against) the trustworthiness of the statements. See First City Fin. Corp., 890 F.2d at 1225 (application of residual hearsay exception appropriate where, inter alia, statement was “subject to criminal prosecution under 18 U.S.C. § 1001”). Regarding the second, Slatten possessed additional guarantees of the trustworthiness of his co-defendant’s statements given their consistent repetition and factual corroboration. See supra at 65-66.
In sum, we are left with a “definite and firm conviction” that the district court clearly erred in excluding the co-defendant’s statements as inadmissible hearsay. See North, 910 F.2d at 909; see also United States v. Sanchez-Lima, 161 F.3d 545, 547-48 (9th Cir. 1998) (reversing district court’s refusal to admit statements under Rule 807 where, inter alia, the statements in question were made “under oath and subject to the penalty of perjury,” were made voluntarily, were based “on facts within [the declarants’] own personal knowledge” and “did not contradict any of their previous statements to government agents and defense investigators”). Moreover, because of the critical nature of the co-defendant’s statements, we believe their exclusion had a “substantial and injurious effect or influence in determining the jury’s verdict” and was therefore not harmless error. See United States v. Mahdi, 598 F.3d 883, 892 (D.C. Cir. 2010) (“[E]rror is harmless unless it has substantial and injurious effect or influence in determining the jury’s verdict . . . .” (internal quotation marks omitted)). Having found the co-defendant’s statements admissible, we leave it to the “jury [to] . . . make the ultimate determination concerning the truth of the statements” in light of all of the evidence. United States v. Price, 134 F.3d 340, 348 (6th Cir. 1998).
In view of our conclusion that the co-defendant’s statements were admissible, we return to Slatten’s motion to sever his trial from that of the co-defendant. We review the district court’s ruling on a motion to sever under the abuse of discretion standard as Federal Rule of Criminal Procedure 14 “leaves the determination of risk of prejudice and any remedy that may be necessary to the sound discretion of the district courts.” Zafiro v. United States, 506 U.S. 534, 541 (1993). That said, the district court recognized that the severance issue here largely hinged on the admissibility of the co-defendant’s statements; in fact, it expressly acknowledged the government’s concession that “[i]f admissible, [the co-defendant’s] Garrity statements would justify severance of Slatten’s case from [the former’s] case in deference to [the co-defendant’s] Fifth Amendment rights as enunciated in Kastigar.” SA 43. There is no record indication that the government has changed its position on this point.
The Supreme Court has instructed that “a district court should grant a severance . . . if there is a serious risk that a joint trial would compromise a specific trial right of one of the defendants, or prevent the jury from making a reliable judgment about guilt or innocence” such as when “essential exculpatory evidence that would be available to a defendant tried alone were unavailable in a joint trial.” Zafiro, 506 U.S. at 539. Because joinder of Slatten’s and his co-defendant’s trial rendered the latter’s otherwise admissible statements—“essential exculpatory evidence,” id.—unavailable to Slatten, it was an abuse of discretion to deny Slatten’s motion to sever. Accordingly, we reverse Slatten’s conviction on Count One (first-degree murder) and remand for a new trial thereon.
Slough, Liberty and Heard also claim the application of 18 U.S.C. § 924(c)’s mandatory 30-year sentence to their convictions violates the Eighth Amendment’s prohibition against cruel and unusual punishment. We review this question de novo. United States v. Said, 798 F.3d 182, 196 (4th Cir. 2015); Pharaon v. Bd. of Governors of Fed. Reserve Sys., 135 F.3d 148, 157 (D.C. Cir. 1998).
Under 18 U.S.C. § 924(c)(1)(B)(ii), anyone who uses a machine gun or a destructive device during and in furtherance of a crime of violence is subject to a mandatory sentence of no less than thirty years. Here, the jury found defendants Slough and Heard violated Section 924(c) by discharging machine guns and destructive devices during the Nisur Square shootings, and it found Liberty violated Section 924(c) by discharging a machine gun during the same attack. In response to these findings, Slough, Heard and Liberty were each sentenced to imprisonment for thirty years for their Section 924(c) conviction plus one day for their remaining voluntary manslaughter and attempted voluntary manslaughter convictions. They now challenge their sentences as being cruel and unusual punishments because the sentences are “unconstitutionally rigid and grossly disproportionate.” Joint Appellants’ Br. 110. We conclude the mandatory 30-year sentence imposed by Section 924(c) based solely on the type of weapons Slough, Heard and Liberty used during the Nisur Square shooting is grossly disproportionate to their culpability for using government-issued weapons in a war zone. We therefore also conclude these sentences violate the Eighth Amendment and remand for resentencing.
The Eighth Amendment prohibits the infliction of “cruel and unusual punishments.” U.S. CONST. amend. VIII. Central to this prohibition is the requirement that the punishment for crime “be graduated and proportioned to the offense.” Graham v. Florida, 560 U.S. 48, 59 (2010). However, this proportionality principle is narrow, and it only forbids “extreme sentences that are grossly disproportionate to the crime.” Harmelin v. Michigan, 501 U.S. 957, 1001 (1991) (Kennedy, J., concurring in part and concurring in judgment).
There are two types of Eighth Amendment challenges to sentences: 1) challenges to sentences as applied to an individual defendant based on “all the circumstances in a particular case” and 2) categorical challenges to sentences imposed based on the nature of the offense or the “characteristics of the offender.” See Graham, 560 U.S. at 59–61. Slough, Liberty and Heard assert their sentences are disproportionate both as applied to their situations individually and categorically to all defendants who have discharged government-issued weapons in a war zone. We begin by addressing the as-applied challenges.
When addressing an as-applied challenge, courts begin “by comparing the gravity of the offense and the severity of the sentence” based on “all of the circumstances of the case.” Id. at 59, 60. When engaging in this comparison, courts are to give “substantial deference to the broad authority that legislatures necessarily possess in determining the types and limits of punishments for crimes.” Solem v. Helm, 463 U.S. 277, 290 (1983). Also, the imposition of a severe mandatory sentence does not in itself make a sentence unconstitutional. See Harmelin, 501 U.S. 994 (“Severe, mandatory penalties may be cruel, but they are not unusual in the constitutional sense.”); see also id. at 1006–07 (Kennedy, J., concurring in part and concurring in judgment) (“We have never invalidated a penalty mandated by a legislature based only on the length of sentence . . . .”). Thus, courts should be “reluctant to review legislatively mandated terms of imprisonment,” and “successful challenges to the proportionality of particular sentences should be exceedingly rare.” Hutto v. Davis, 454 U.S. 370, 374 (1982) (per curiam). However, the unusual circumstances of this case make it one of those “exceedingly rare” instances.
We begin by evaluating the gravity of the defendants’ crime. When evaluating the severity of a crime, we consider “the harm caused or threatened to the victim or society and the culpability [and degree of involvement] of the [defendant].” See Solem, 463 U.S. at 292. When examining a defendant’s culpability, the Court may look to the defendant’s intent and motive in committing the crime. See id. at 293. The Court may also consider the defendant’s criminal history. See Rummel v. Estelle, 445 U.S. 263, 276 (1980).
Here, we believe it is important to distinguish between the predicate crimes of violence for which Slough, Heard and Liberty were convicted and the conviction under Section 924(c) that carries with it a mandatory 30-year sentence. We agree with the district court that the actions of these defendants, which killed fourteen Iraqi civilians and injured seventeen others, constitute very serious offenses. We also agree the use of automatic weapons or explosives during a crime of violence typically does increase the severity of that crime. Moreover, under normal circumstances, we would be “reluctant to review Congress’s legislatively mandated terms of imprisonment.” Hutto, 454 U.S. at 374. However, we do not believe such deference is owed when a statute’s application only tangentially relates to Congress’s purpose for creating the statute in the first place. See Gonzalez v. Duncan, 551 F.3d 875, 884–86 (9th Cir. 2008) (holding the application of a statute to a defendant that was only tangentially related to the legislature’s reason for creating the law undermined the gravity of the offense).
The Supreme Court has described Section 924(c)’s basic purpose as an effort to combat the “dangerous combination” of “drugs and guns.” Smith v. United States, 508 U.S. 223, 240 (1993). For this reason, the text of the statute applies to any person who “uses or carries a firearm” “during and in relation to any crime of violence or drug trafficking crime.” 18 U.S.C. § 924(c)(1)(A). Furthermore, the Supreme Court has recognized Section 924(c) was created “‘to persuade the man who is tempted to commit a Federal felony to leave his gun at home.’” Muscarello v. United States, 524 U.S. 125, 132 (1998) (quoting Representative Poff, the chief legislative sponsor of Section 924(c)); see also Busic v. United States, 446 U.S. 398, 405 (1980) (describing Representative Poff’s comments as “crucial material” in interpreting the purpose of Section 924(c)). Thus, precedent clarifies Section 924(c) applies against those who intentionally bring dangerous guns with them to facilitate the commission of a crime.
None of these concerns are remotely implicated by this case. On the day of the Nisur Square attack, Slough, Heard and Liberty were providing diplomatic security for the Department of State in Iraq. As part of their jobs, they were required to carry the very weapons they have now been sentenced to thirty years of imprisonment for using. While we acknowledge some courts have held the text of 924(c) is broad enough to allow the statute to be applied against individuals using government-issued weapons while on duty, see, e.g., United States v. Ramos, 537 F.3d 439, 457 (5th Cir. 2008) (upholding the application of Section 924(c) against Border Patrol agents who shot a fleeing felon); see also S. Rep. No. 98-225, at 314 n.10 (1983), reprinted in 1984 U.S.C.C.A.N. 3182, 3,492 (stating that “persons who are licensed to carry firearms and abuse that privilege by committing a crime with the weapon, as in the extremely rare case of the armed police officer who commits a crime, are as deserving of punishment as a person whose possession of the gun violates a state or local ordinance”), there is no evidence Congress intended for Section 924(c) to be applied against those required to be armed with dangerous guns who discharge their weapons in a war zone.
When Congress amended Section 924(c) in 1984 so it would also apply against those who were licensed to carry firearms, compare 18 U.S.C. § 924(c) (1982), with Pub. L. No. 98-473, 98 Stat. 1837 (1984), MEJA did not exist. In fact, Congress did not create MEJA until over a decade later in 2000. See Pub. L. 106-523, 114 Stat. 2488 (2000). Because Congress had not yet considered the extra-territorial application of federal criminal law to employees of the Armed Forces at all, Congress could not have possibly contemplated applying Section 924(c) against private contractors providing diplomatic security for a federal agency. Thus, combining the public interests Section 924(c) was intended to advance with the lack of evidence Congress ever intended the law to apply against military employees in a war zone, we conclude this case does not involve the usual legislative judgments on the severity of a crime that would cause us to defer to Congress’s determinations regarding the punishments for crimes.
This conclusion is further supported by the events preceding the Nisur Square shootings. When the Raven 23 convoy arrived in Nisur Square on the day of the incident, it was responding to the explosion of a car bomb near a U.S. diplomat under its protection. Accordingly, this is not a case where the defendants went out with the intention of committing a crime and brought their weapons with them to assist them in the commission of that crime. This is not even a case where these three defendants acted recklessly by inserting themselves into a dangerous situation in a place filled with innocent bystanders. The decision to go to Nisur Square was made by Watson, the Raven 23 shift leader, and once he decided to ignore his orders and proceed to Nisur Square, they had no choice but to follow their commander’s lead. Once they arrived in Nisur Square, they found themselves in a crowded environment, where the ability to differentiate between civilians and enemies was significantly diminished. The tragedy that unfolded shortly after their arrival in Nisur Square owed more to panic and poor judgment than to any coordinated plan to murder Iraqi civilians. While we agree the defendants are responsible for their exaggerated response to perceived threats, the crime’s severity and Defendant’s culpability flow from the harm caused by their hypervigilance, not from the use of weapons which would have been appropriate had they not misperceived the threat.
The government argues Slough, Heard and Liberty could have used less deadly weapons, such as pistols or the semi-automatic setting on their rifles, in response to perceived threats. But this argument mistakenly applies the “20/20 vision of hindsight,” an approach the Supreme Court has explicitly rejected when evaluating a police officer’s use of force. See Graham v. Connor, 490 U.S. 386, 396 (1989). Instead, this Court applies an analysis that “‘allow[s] for the fact that police officers are often forced to make split-second judgments—in circumstances that are tense, uncertain and rapidly evolving—about the amount of force that is necessary in a particular situation.’” Robinson v. Pezzat, 818 F.3d 1, 8 (D.C. Cir. 2016) (quoting Connor, 490 U.S. at 396–97). If courts are to give police officers this type of leeway in making split-second judgments about which of their tools to use based upon tense and uncertain situations, we must give an even greater amount of latitude to decisions made by those supporting our military overseas in a hostile environment. Here, we believe it is imprudent to second-guess the defendants’ choice of firearm in responding to what they believed to be an approaching car bomb or enemy fire. We emphasize they are still culpable for their decision to fire at all, as encompassed by their manslaughter and attempted manslaughter convictions, but the type of weapon used should not be more determinative of their punishments than the death and destruction that resulted from their decisions to fire.
We also find it highly significant that none of the defendants sentenced under Section 924(c) have any prior convictions. Although the government is free to impose harsh, mandatory penalties for first-time offenders, see Harmelin, 501 U.S. at 994–95, a regime of strict liability resulting in draconian punishment is usually reserved for hardened criminals. As the Supreme Court has noted, recidivism is a legitimate consideration to support the imposition of a more severe penalty. See Ewing v. California, 538 U.S. 11, 29 (2003) (“In weighing the gravity of [the defendant’s] offense, we must place on the scales not only his current felony, but also his . . . history”); Rummel, 445 U.S. at 276 (stating legislatures have a legitimate interest in dealing more harshly with recidivists). In fact, in virtually every instance where the Supreme Court has upheld the imposition of a harsh sentence for a relatively minor nonviolent crime for an as-applied challenge, it has done so in the context of a recidivist criminal.17 Here, none of these defendants have a criminal record at all. The district court noted they were “good young men who [had] never been in trouble.” JA 3330. It also stated they had “served their country honorably in the military and nothing in their backgrounds suggest[ed] that they would have ever committed offenses such as these.” Ibid. Based upon these observations and the distinctions made by the Supreme Court, we hold the defendants’ clean criminal records weigh against the imposition of a harsh, mandatory sentence.
Additionally, the imposition of a mandatory 30-year sentence through Section 924(c) fails to truly account for the culpability of Slough, Heard and Liberty individually. Because these men were not convicted of the same counts, it makes little sense for the sentences to be identical. See Koon v. United States, 518 U.S. 81, 113 (1996) (stating a sentencing judge must “consider every convicted person as an individual and every case as a unique study in the human failings that sometimes mitigate, sometimes magnify, the crime and the punishment to ensue”). Thus, a more prudent way to sentence would be to examine each defendant as an individual, taking into account all of the aggravating and mitigating factors typically considered by sentencing judges. See 18 U.S.C. § 3553. While it does appear the sentencing judge might have been inclined to align sentences more closely to the circumstances, his hands were tied by Section 924(c)’s mandatory minimum. Thus, we do not know if he would have imposed the same sentence on each of these three defendants or if he would have allowed for the differing number of victims and the presence of other mitigating factors, like the existence of post-traumatic stress disorder at the time of the shootings, to lead to a reduced sentence for some of them. Because the mandatory sentence hindered the sentencing judge’s ability to individually examine the severity of each defendant’s crime, we find the one-size-fits-all nature of these sentences troubling.
Turning now to the severity of the sentence, we consider the actual severity of the penalty, not the penalty’s name. In the context of life sentences, the Supreme Court has acknowledged there is an important distinction between a life sentence with the possibility of parole and a life sentence without the possibility of parole. See Rummel, 445 U.S. at 280–81; Solem, 463 U.S. at 297. Thus, we evaluate Slough, Heard and Liberty’s sentences based upon the amount of time they will actually spend in prison and the possibility of early release.
Here, there is no doubt that a mandatory, 30-year sentence is a severe sanction. United States v. Spencer, 25 F.3d 1105, 1110 (D.C. Cir. 1994) (“Thirty years’ imprisonment is, by anyone’s lights, a severe sanction.”). With the exception of the death penalty or a life sentence, a 30-year sentence is the harshest mandatory sentence the federal criminal law can impose on a first-time offender. The severity of these sentences is amplified by the fact that there is no possibility of parole in the federal system. See Pub. L. No. 98-473, 98 Stat. 1837 (1984). Even if we were to presume the defendants would receive fifty-four days of good-time credit each year for the duration of their incarceration, see 18 U.S.C. § 3624(b)(1), the most their sentences could possibly be reduced is approximately four years. Thus, even with the maximum amount of good-time credit available, these sentences are among the harshest in existence for first-time offenders.
Combining all of these considerations, we conclude Slough, Heard and Liberty’s mandatory 30-year sentences create the “rare case” that “leads to an inference of gross disproportionality.” Graham, 560 U.S. at 60. We do not believe their culpability in this case—based solely on using weapons they were required to carry when performing diplomatic security missions—is on par with the typical culpability of defendants convicted under Section 924(c), and we are troubled by the imposition of such a harsh mandatory sentence without any individualized examination of each defendant’s underlying crimes.
Typically, once we have found an inference of gross disproportionality, we would “compare the defendant’s sentence with the sentences received by other offenders in the same jurisdiction and with the sentences imposed for the same crime in other jurisdictions.” Id. Here, such a comparison is of little value because Section 924(c)’s penalty for using a machine gun or explosive device is the same for all defendants—thirty years’ imprisonment. This case also presents a unique challenge for comparison purposes because of its unusual facts. The parties have not identified a single instance in which a defendant was convicted and sentenced under Section 924(c) in a manner similar to this case. Moreover, the closest this Court has come to locating a similar situation is United States v. Drotleff, where two Department of Defense contractors were convicted of a single count of involuntary manslaughter for killing two civilians and sentenced to 30 and 37 months of imprisonment. 497 F. App’x 357, 358–59 (4th Cir. 2012) (per curiam); see also United States v. Drotleff, No. 10cr00001-002, 2011 WL 2610190 (E.D. Va. June 21, 2011); United States v. Cannon, No. 2:10cr00001-001, 2011 WL 2610188 (E.D. Va. June 30, 2011). The case is similar because—like the Nisur Square attack—the shooting began when a vehicle began driving towards the contractors in what they perceived to be a threatening manner. Drotleff, 497 F. App’x at 358–59. Also like this case, the government charged the contractors with violating Section 924(c). United States v. Cannon, 711 F. Supp. 2d 602, 603 (E.D. Va. 2010). However, the similarities end there because the number of victims was substantially lower and because the jury did not convict on the Section 924(c) counts. See Drotleff, 497 F. App’x at 359. Thus, it appears this case presents a novel application of Section 924(c) to government contractors in a war zone, and direct comparisons to another case are therefore not possible.
Notwithstanding the uniqueness of this case, we find it helpful to examine the other instances in which Section 924(c) has been applied against people who were licensed to carry the weapon that they were later convicted of carrying or using. In doing so, the Court has located numerous instances in which the government has applied Section 924(c) against law enforcement personnel. The overwhelming majority of cases in which the statute has been applied against those carrying government-issued firearms have involved instances in which the defendant made a conscious decision to commit a crime outside the scope of their duties as police. See, e.g., United States v. Washington, 106 F.3d 983, 1010 (D.C. Cir. 1997) (applying Section 924(c) to police officers carrying government-issued firearms while engaging in drug trafficking); United States v. Guidry, 456 F.3d 493, 507–09 (5th Cir. 2006) (applying Section 924(c) against a police officer who carried a government-issued firearm while committing sexual assault). However, there are also instances where Section 924(c) has applied against law enforcement officials who commit a crime of violence while on duty. See Ramos, 537 F.3d at 457 (applying Section 924(c) against a police officer who shot a felon without justification); United States v. Williams, 343 F.3d 423, 429–34 (5th Cir. 2003) (affirming a Section 924(c) conviction against an officer who shot a fleeing suspect in the back after he had surrendered); United States v. Winters, 105 F.3d 200, 202 (5th Cir. 1997) (affirming the conviction of a prison guard convicted under Section 924(c) for hitting a recaptured inmate in the back of the head with his service revolver after the inmate had attempted to escape). While the government urges us to treat this case identically to the cases discussed above, this argument overlooks the different environments in which domestic law enforcement and private international security contractors live and the different functions they serve. Law enforcement officers are a vital part of any community. They live and work among the community’s citizens and are tasked with performing a variety of functions, including “reduc[ing] the opportunities for the commission of some crimes . . . , aid[ing] individuals who are in danger of physical harm, assist[ing] those who cannot care for themselves, resolv[ing] conflict, creat[ing] and maintain[ing] a feeling of security in the community, and provid[ing] other services on an emergency basis.” 3 WAYNE R. LAFAVE, SEARCH AND SEIZURE: A TREATISE ON THE FOURTH AMENDMENT § 6.6 (5th ed. Oct. 2016). While they may sometimes be called upon to use lethal force in the line of duty, it is not a routine part of their job and is instead reserved only for situations in which a suspect poses a substantial risk to law enforcement personnel or the community. See Tennessee v. Garner, 471 U.S. 1, 11 (1985) (stating law enforcement officials must have “probable cause to believe that the suspect poses a threat of serious physical harm, either to the officer or to others” before using deadly force).
Conversely, private security contractors work in places that are “extremely dangerous” because of “conflicts, wars, political unrest, and . . . terrorist activity.” JA 3861. Accordingly, they live and work in a hostile environment in a war zone in which the enemy could strike at any moment. Because of this ever-present danger, they are often required to use lethal force. In fact, using lethal force to eliminate hostile forces is a central component of assuring the safety of any American personnel they are tasked with protecting. They are issued powerful weapons to assist them in performing this task. Thus, because these three defendants were living in a much more dangerous environment and performing a substantially different function than law enforcement officials, we find the government’s attempts to analogize this case to other applications of Section 924(c) to be unpersuasive.
Because comparisons to other applications of Section 924(c) are of little value, we now broaden our comparison to encompass other types of crimes that bear similar types of penalties. We are mindful of the fact that each crime is unique and that it is difficult to quantify the harm done by a crime, but the Supreme Court has recognized courts are competent to make these kinds of determinations “on a relative scale.” Solem, 463 U.S. at 292. In doing so, we consider factors traditionally applied by courts, such as whether the crime involves violence, the gravity of the harm caused by the crime and the intent of the offender. See id. at 292–94. “If more serious crimes are subject to the same penalty, or to less serious penalties, that is some indication that the punishment at issue may be excessive.” Id. at 291; see also Weems v. United States, 217 U.S. 349, 380–81 (1910) (listing more severe crimes subject to less serious penalties than the offense at issue).
Here, Slough, Heard and Liberty each received a 30-year sentence based on their use of government-issued weapons during the Nisur Square attack. While their crimes obviously did involve violence, we note the gravity of the harm done would be essentially the same regardless of whether they used an automatic rifle, a semi-automatic rifle, or a pistol. Moreover, neither their conviction under Section 924(c) nor their underlying crimes of violence were intentional. The defendants used weapons their profession required them to carry, and their convictions for voluntary manslaughter involved extreme recklessness and gross misjudgments, not an intention to kill innocent people.
Comparing their sentences to other federal crimes with similar sentences for first-time offenders, we find it significant that other crimes with comparable sentences involve the intentional commission of serious crimes. For example, the federal criminal code contains numerous 30-year sentences for first-time offenses involving the intentional infliction of harm to children. See, e.g., 18 U.S.C. § 2251A(a) (30-year sentence for a parent or legal guardian who sells his child for the purpose of sexual exploitation); id. § 2251A(b) (30-year sentence for purchasing a child for the purpose of sexual exploitation); id. § 2241(c) (30-year sentence for engaging in a sexual act with a child under the age of twelve); id. § 3559(f)(1) (30-year sentence for murdering a child under eighteen). Likewise, a person who causes or conspires to cause damage to or destruction of a motor vehicle carrying high-level radioactive waste or spent nuclear fuel with intent to endanger the safety of others will receive an identical 30-year sentence. 18 U.S.C. § 33(b). Perhaps most extreme of all, a person who attempts or threatens to use an atomic weapon while in possession of one also receives a minimum sentence of 30 years. 42 U.S.C. § 2272(b). Thus, it appears that outside of Section 924(c), a 30-year mandatory sentence is typically reserved for instances where the defendant has intentionally committed a heinous crime that either harms the most vulnerable of our society or has the potential to result in wide-spread devastation. The use of government-issued rifles and explosives in a war zone is simply not comparable. While the weapons these three defendants fired do have the potential to—and in this case did—unleash wide-spread destruction, they are the tools our government gave to them to adequately perform their job. If circumstances had been as they believed them to be, it would have been negligent to rely on less effective weapons.
In reaching this conclusion, we by no means intend to minimize the carnage attributable to Slough, Heard and Liberty’s actions. Their poor judgments resulted in the deaths of many innocent people. What happened in Nisur Square defies civilized description. However, none of the penological justifications our society relies upon when sentencing criminals—incapacitation, rehabilitation, retribution, or deterrence—are properly served here by a sentence whose length is determined solely based on the type of weapon used during the crime. See Ewing, 538 U.S. at 25 (discussing the penological goals of criminal punishments). While we acknowledge our Constitution “does not mandate adoption of any one penological theory” and that sentencing rationales should generally be made by legislatures and not federal courts, id. at 24–25, the Supreme Court’s examination of penological goals in previous cases suggests those goals should be a relevant part of our analysis. See id. at 25–28; Harmelin, 501 U.S. at 999 (Kennedy, J., concurring in part and concurring in judgment); see also Graham, 560 U.S. at 71 (stating “[a] sentence lacking any legitimate penological justification is by its nature disproportionate to the offense”).
Regarding incapacitation, nothing in any of these defendants’ records suggests they pose a danger to society such that they must remain in prison to prevent them from committing more crimes. Before the Nisur Square shootings, none of them had any prior convictions, and nothing in the record or their backgrounds suggests they are likely to commit more crimes in the future. For similar reasons, rehabilitation is not an issue. No doubt Nisur Square and its haunting aftermath will provide reason enough for these defendants to avoid any analogous circumstances. As to retribution, we recognize the 30-year sentence does punish the defendants for their crimes and allows society “to express its condemnation of [their] crime[s] and to seek restoration of the moral imbalance caused by [their] offense[s].” Graham, 560 U.S. at 71. However, “[t]he heart of the retribution rationale is that a criminal sentence must be directly related to the personal culpability of the criminal offender.” Id. at 71; see also Ewing, 538 U.S. at 31 (Scalia, J., concurring in judgment) (“Proportionality—the notion that the punishment should fit the crime—is inherently a concept tied to the penological goal of retribution.”). Here, we have concluded the mandatory 30-year sentence imposed by Section 924(c) is grossly disproportionate as applied to Slough, Heard and Liberty and that such a sentence actually prevents the sentencing judge from directly examining the personal culpability of each defendant in this case. Furthermore, society’s interest in retribution can be equally served by a sentence imposed based solely on the voluntary manslaughter and attempted voluntary manslaughter convictions. Therefore, this sentence cannot be justified based on retribution.
Regarding deterrence, the district court observed there was no need to deter the defendants individually. JA 3332. We agree with this observation based on the defendants’ lack of criminal background. Thus, we are left with examining whether this sentence serves the penological goal of general deterrence. Under the theory of general deterrence, the government essentially seeks to make an example of an offender through punishing him so that other potential offenders are intimidated into refraining from committing the contemplated crime. 1 WHARTON’S CRIMINAL LAW § 3 (15th ed. Sept. 2016); see also Pell v. Procunier, 417 U.S. 817, 822 (1974) (stating the premise of general deterrence is that “by confining criminal offenders in a facility where they are isolated from the rest of society, a condition that most people presumably find undesirable, they and others will be deterred from committing additional criminal offenses”). The harsh sentences imposed under Section 924(c) generally do operate as strong deterrents against using firearms when committing a crime of violence or a drug trafficking offense. In fact, this is precisely what Congress envisioned when it first passed the law. See Muscarello, 524 U.S. at 132 (stating Section 924(c) was created “to persuade the man who is tempted to commit a Federal felony to leave his gun at home”). However, as discussed above, the application of Section 924(c)’s mandatory sentence does little to advance this purpose. Instead, it will only deter future private security contractors from quickly making the split-second decisions their jobs require them to make. In theory, if they are wrong even once about a potential threat and use their machine gun in response, they are potentially subject to this penalty. In the dangerous environments in which these contractors live and work, even a single moment’s hesitation because of fear of such a harsh criminal sanction could be the difference between life and death for themselves, their fellow contractors and the diplomats they were hired to protect. Thus, deterrence is both an irrational and unjust reason to justify these sentences under Section 924(c). This is especially true given that contractors will already be deterred from recklessly firing their firearms based on the possibility of receiving other criminal sanctions, such as manslaughter charges, for any severe lapses in judgment. Thus, these sentences cannot be justified under any of our society’s penological goals.
For the foregoing reasons, we conclude the application of Section 924(c) to Slough, Heard and Liberty is cruel and unusual punishment.3 The sentences are cruel in that they impose a 30-year sentence based on the fact that private security contractors in a war zone were armed with government-issued automatic rifles and explosives. They are unusual because they apply Section 924(c) in a manner it has never been applied before to a situation which Congress never contemplated. We again emphasize these defendants can and should be held accountable for the death and destruction they unleashed on the innocent Iraqi civilians who were harmed by their actions. But instead of using the sledgehammer of a mandatory 30-year sentence, the sentencing court should instead use more nuanced tools to impose sentences proportionally tailored to the culpability of each defendant.
For the foregoing reasons, we vacate defendant Nicholas Slatten’s first degree murder conviction and remand for a new trial. Further, we vacate defendant Evan Liberty’s conviction for the attempted manslaughter of Mahdi Al-Faraji. The Court remands the sentences of Liberty, defendant Paul Slough and defendant Dustin Heard for resentencing consistent with this opinion. In all other respects, the Court affirms the judgment of the district court.
So ordered.
I write separately to express my view that the aftermath of Slatten’s mandamus petition did not, and could not, give rise to a realistic likelihood of prosecutorial vindictiveness. United States v. Meyer, 810 F.2d 1242, 1246 (D.C. Cir. 1987). Although the majority notes that “the extraordinary mandamus grant here, followed by a rather sharply-worded criticism in denying reconsideration, in a high-profile prosecution with international ramifications no less, had [great] potential to give rise to a vindictive motive,” Maj Op. 48, that description fails to account for our Court’s own mistake leading to the mandamus petition (and its aftermath) in the first place.
The Nisur Square attack took place on September 16, 2007. Under MEJA, then, the government had until September 16, 2012 to indict (or reindict) Slatten on a non-capital offense. 18 U.S.C. § 3282(a).
On December 4, 2008, a grand jury indicted Slatten and his co-defendants on, inter alia, multiple manslaughter charges. One year later, on December 31, 2009, the district court dismissed the indictment based on the government’s violation of Kastigar v. United States, 406 U.S. 441 (1972), and Garrity v. New Jersey, 385 U.S. 493 (1967). See United States v. Slough, 677 F. Supp. 2d 112 (D.D.C. 2009). Although the government had earlier moved to voluntarily dismiss Slatten’s indictment, the district court expressly denied that request as moot given its simultaneous dismissal of all charges against all defendants.19 Id. at 166 n.67 (“Because the court dismisses the indictment against all of the defendants, including defendant Slatten, it denies as moot the government’s motion for leave to dismiss the indictment against defendant Slatten without prejudice.”).
The government appealed the dismissal. Our Court reversed the district court, concluding that Kastigar required it to determine with greater specificity the taint, if any, each defendant’s compelled statements had on the grand jury evidence. United States v. Slough, 641 F.3d 544, 550-55 (D.C. Cir. 2011). Importantly, however, it did so assuming—mistakenly—that “the government itself moved to dismiss the indictment against Nicholas Slatten, without prejudice to possible later re-indictment, and the district court’s grant of the motion has taken Slatten out of the case for now.” Id. at 547 (emphasis added). Stated differently, it did not overturn the district court’s dismissal of the manslaughter charges against Slatten because it incorrectly believed that the district court had granted the government’s motion to dismiss and therefore Slatten was already out of the case. See id. (reversing and remanding “as to four of the defendants”). Slatten’s co-defendants unsuccessfully petitioned for certiorari and, because of that delay, the Slough mandate did not issue until June 5, 2012.
Over one year later, on October 17, 2013, a grand jury reindicted Slatten and his co-defendants on, inter alia, multiple manslaughter charges. JA 314. Although the superseding indictment issued after September 16, 2012—the date MEJA’s five-year statute of limitations for non-capital offenses was set to expire—it related back to the filing of the original indictment for statute of limitations purposes because it did not broaden the original indictment. See, e.g., United States v. Grady, 544 F.2d 598, 602-03 (2d Cir. 1976) (“Since the statute stops running with the bringing of the first indictment, a superseding indictment brought at any time while the first indictment is still validly pending, if and only if it does not broaden the charges made in the first indictment, cannot be barred by the statute of limitations.” (footnote omitted)); United States v. Yielding, 657 F.3d 688, 703 (8th Cir. 2011) (“For limitations purposes, ‘a superseding indictment filed while the original indictment is validly pending relates back to the time of filing of the original indictment if it does not substantially broaden or amend the original charges.’”); JA 323. The critical question regarding Slatten, however, was whether Slatten had remained in the case and was therefore covered by the relation back. See Opposition to Slatten’s Motion to Dismiss Superseding Indictment on Statute of Limitations Grounds, United States v. Slatten, et al., Docket No. 1:08-cr-00360-RCL, Doc. 352 (November 29, 2013). On the one hand, the district court’s December 2009 order manifested that Slatten’s dismissal was based on the same Kastigar/Garrity rationale applicable to his co-defendants, the rationale we rejected in reversing the dismissal. On the other hand, our Court declared (incorrectly) that Slatten was no longer in the case because it mistakenly believed the district court had granted the government’s motion to dismiss Slatten’s indictment in its December 2009 order. Slough, 641 F.3d at 547.
Once the superseding indictment against him issued, Slatten moved to dismiss it on the ground that he was no longer in the case based on our Court’s Slough opinion. The district court—a successor judge—denied Slatten’s motion to dismiss. In doing so, it expressly discussed the mistake caused by our Slough language. See Memorandum Opinion, United States v. Slatten, No. 1:08-cr-00360-RCL, Doc. 388 (February 18, 2014). It noted that the “only ambiguity in [Slough] was the judgment appealed from,” pointing out that “[e]ven though the Government conceded that Slatten’s indictment was deficient, [the original judge] found that it was deficient on broader grounds than the Government sought in its own motion” and, thus, “[i]t was those broader grounds that the Government appealed from as to all five defendants including Slatten.” Id. at 2-3 (emphasis added). Attempting to adjust to our Court’s error, the district court reasoned that the “letter” of the district court’s December 2009 order and the “spirit” of Slough’s mandate established that Slatten remained a defendant and therefore the non-capital MEJA charges against Slatten remained timely. Id. at 2-4.
Following the district court’s decision, Slatten petitioned this Court for mandamus relief “to prevent the district court from expanding the scope of Slough.” See Emergency Petition for Writ of Mandamus, In re Nicholas Abram Slatten, Case No. 14-3007 (D.C. Cir. March 6, 2014). In reviewing that petition, our Court again failed to focus on the relevant portion of the December 2009 dismissal order (which expressly denied the government’s motion to voluntarily dismiss Slatten) and instead granted Slatten’s petition, stating that the “mandate reversing and remanding [to] the district court clearly applied only to Slatten’s four co-defendants” because the “government conceded to us, both in its briefs and at oral argument, that Slatten’s indictment was infirm.” See Per Curiam Order, In re Nicholas Abram Slatten, No. 14-3007 (D.C. Cir. April 7, 2014). Given the mandamus’s limitation on what charges could (and could not) be brought against Slatten, the government pressed the issue by petitioning for rehearing. See Petition, In re Nicholas Abram Slatten, No. 14-3007 (D.C. Cir. April 17, 2014). In its petition, the government expressly drew the Court’s attention to the fact that “[t]he [Slough] opinion incorrectly stated that the district court granted the government’s motion to dismiss the indictment as to Slatten” and that “the record compels the conclusion . . . that this Court in Slough reversed the dismissal order with respect to all five defendants.” Id. at 1-2 (emphasis added).
On April 18, 2014, our Court—for the first time—recognized that it “erred in stating that the district court’s dismissal as to Slatten had come in response to the Government’s own motion to dismiss, rather than to Slatten’s motion.” See Per Curiam Order, In re Nicholas Abram Slatten, No. 14-3007 (D.C. Cir. April 18, 2014). Nevertheless, it denied the government’s petition, declaring that “the dispositional posture following [Slough] was unaffected” by the error because “Slatten’s indictment had been dismissed, and we had reversed only as to the other four.” Id at 1. It appears the error was seen as harmless because, if Slatten had been within the scope of the mandate, the district court—which, in its December 2009 order, had denied as moot the government’s motion to dismiss Slatten—would have then granted that motion, leaving the same scenario, that is, Slatten would no longer be a defendant. Id.20
I believe we incorrectly—albeit in good faith—contributed to the Hobson’s choice facing the government at that point. First, and critically, MEJA’s statute-of-limitations clock would not have run on any non-capital offense had Slatten remained a defendant throughout the Slough appeal and remand because, to repeat, “a superseding indictment brought at any time while the first indictment is still validly pending, if . . . it does not broaden the charges made in the first indictment, cannot be barred by the statute of limitations.”21 Grady, 544 F.2d at 601-02. “[T]he dispositional posture following [Slough],” then, was anything but “unaffected” by the misreading of the December 2009 dismissal order; it disabled the government from filing any non-capital charge against Slatten in the superseding indictment. Moreover, our Slough language left the government uncertain regarding Slatten’s status. 641 F.3d at 547. Indeed, on July 25, 2012—two months before MEJA’s five-year statute of limitations clock ran—the government announced it intended to seek a superseding indictment covering all five Slough defendants (including Slatten), indicating no recognition of the need to omit Slatten on all non-capital counts. It took almost the next two years for it to recognize definitively that Slatten could not be indicted on a non-capital offense. Although the wiser move would have been for the government to reindict Slatten immediately upon remand, I believe our Court’s mistaken reading of the December 2009 order contributed to the government’s failure to do so.
The government faced a forced choice—indict Slatten on the only charge unaffected by MEJA’s five-year deadline or completely forego prosecution of him. “[These] circumstances, when taken together,” plainly fail to “support a realistic likelihood of vindictiveness.” Meyer, 810 F.2d at 1246.
I join the Court’s opinion with two exceptions. First, in accordance with the Supreme Court’s instruction, portions of a co-defendant’s statements to investigators22 should have been admitted in Slatten’s defense, but not as a result of unduly expanding a narrow residual hearsay exception when the statements are covered by an established exception. Second, defendants’ Eighth Amendment challenge lacks any merit whatsoever, especially in view of the district court judge’s express assessment, which my colleagues ignore, that the sentences were an appropriate response to the human carnage for which these defendants were convicted by a jury.
Concurring in the judgment in Part VII. I agree that the district court’s exclusion of certain statements by a co-defendant as inadmissible hearsay requires reversal of Slatten’s conviction. Op. 67–68. In my view, however, the district court did not abuse its discretion in finding the co-defendant’s compelled statements untrustworthy to the extent they offered an exculpatory narrative of self-defense. Rather, the district court abused its discretion by failing, as a matter of law, to isolate certain inculpatory statements within that broader narrative to consider whether they were sufficiently trustworthy to be admitted under an exception to the hearsay rule. See Williamson v. United States, 512 U.S. 594, 600 (1994); Koon v. United States, 518 U.S. 81, 100 (1996). The co-defendant’s inculpatory statements were admissible in Slatten’s defense under Federal Rule of Evidence 804(b)(3),23 and because they concerned the single most important issue underlying Slatten’s conviction — who fired the first shots that day — their exclusion implicated Slatten’s due process right to present a complete defense and was not harmless beyond a reasonable doubt. See United States v. Whitmore, 359 F.3d 609, 616 (D.C. Cir. 2004). Indeed, even if the statements’ exclusion did not impinge on Slatten’s constitutional right to present a complete defense, see Appellee Br. 128, the exclusion was not harmless because it had a “substantial and injurious effect” on the jury’s consideration of this close question. See United States v. Mahdi, 598 F.3d 883, 892 (D.C. Cir. 2010) (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946)).
In the immediate aftermath of the Nisur Square massacre, a co-defendant of Slatten’s offered statements on four different days to State Department investigators, one on a written departmental form and three oral. The oral statements, which were incorporated into written reports by State Department investigators, constitute hearsay within hearsay. Op. 57 n.7; Fed. R. Evid. 805. As a preliminary matter, the investigators’ contemporaneous recounting in their reports of what the co-defendant said falls within the business records exception under FRE 803(6). See United States v. Smith, 521 F.2d 957, 962–65 (D.C. Cir. 1975); Michael H. Graham & Kenneth W. Graham, 30C Fed. Prac. & Proc. Evid. § 7047 & n.29 (2017 ed.); see also United States v. Warren, 42 F.3d 647, 657 n.7 (D.C. Cir. 1994). At this second level of hearsay, the question is whether the co-defendant said what he is reported to have said, not whether he was being truthful. See Smith, 521 F.2d at 965. The investigators personally witnessed the co-defendant making these statements. See Fed. R. Evid. 803(6)(A). Furthermore, testimony established that State Department investigators regularly took such statements any time a contractor was involved in a shooting incident, see Fed. R. Evid. 803(6)(B)–(C); 11/2/09 Hearing Tr. 48:23–49:14, and it is “presumed that [the investigators] accurately transcribed and reported” the co-defendant’s statements. Smith, 521 F.2d at 965.
The difficulty with relying on the evaluative reports prong of the public records exception, see Op. 57 n.7, is that the records reflect only unverified witness statements about the Nisur Square massacre, rather than the investigators’ own “factual findings” about what occurred. See Fed. R. Evid. 803(8)(A)(iii). A “factual finding[]” in this context means a public official’s “conclusion by way of reasonable inference from the evidence,” not a piece of evidence gathered in aid of a potential conclusion down the road. See Beech Aircraft Corp. v. Rainey, 488 U.S. 153, 164 (1988) (quoting BLACK’S LAW DICTIONARY 569 (5th ed. 1979)). For this reason, the relevant factors identified by Advisory Committee under this exception focus on the trustworthiness of the investigator’s conclusions (e.g., the skill or experience of the investigator, the investigator’s potential bias). See Notes of Advisory Committee on Proposed Rules, Fed. R. Evid. 803. Had State Department investigators concluded that the co-defendant’s version of events was credible and adopted it as their own, then FRE 803(8)(A)(iii) would likely come into play. See Beech Aircraft Corp., 488 U.S. at 169. Nothing in the reports, however, indicates that the investigators found any facts to be as the co-defendant portrayed them, and FRE 803(8)(A)(iii) “bars the admission of statements not based on factual investigation,” such as an eyewitness’s unverified statements to investigators. Beech Aircraft Corp., 488 U.S. at 169.
That leaves only the second level of hearsay, the co-defendant’s statements themselves. Although he was informed that the statements, if truthful, could not be used directly or indirectly against him in a criminal proceeding, he was also informed that they could be used in the course of a disciplinary proceeding and could result in termination of his employment. A statement that jeopardizes the declarant’s employment can be sufficient to trigger FRE 803’s pecuniary interest exception, provided it is so contrary to that interest that a reasonable person would not have made it unless it were true. Gichner v. Antonio Troiano Tile & Marble Co., 410 F.2d 238, 242 (D.C. Cir. 1969).
Taken together, the co-defendant’s statements offered a generally exculpatory version of events, in which the white Kia sped dangerously toward the convoy and ignored repeated warnings to stop, until it became necessary to fire upon and disable the Kia in order to protect the Raven 23 convoy. Within his narrative of self-defense, however, he offered details that had the potential to jeopardize his employment. Namely, he admitted that he “engaged and hit the driver,” Mem. Report of Interview at 1 (Sept. 16, 2007), “fir[ing] two rounds at the driver from his M-4 rifle . . . [that] impacted the driver’s area of the windshield,” Mem. Report of Interview at 1 (Sept. 23, 2007). Most crucially, he acknowledged that he was “not aware of any shots being fired before his,” Mem. Report of Interview at 2 (Sept. 20, 2007), and that he made eye contact with the driver just before firing, which further suggests that he was the first to fire. That is, following an incident in which multiple Raven 23 members were seen firing into the Kia, the co-defendant voluntarily singled himself out as the first shooter — the one likely responsible for the death of Al-Rubia’y and, in the government’s words, “the one who lit the match that ignited the firestorm.” 8/27/14 (AM) Tr. 27:1-4. Thus, if investigators doubted the claim that the Kia represented a threat, then the co-defendant’s statements all but ensured that he would lose his job.
The district court ruled that the co-defendant’s statements constituted inadmissible hearsay because their lack of trustworthiness disqualified them from the statement against interest exception, Fed. R. Evid. 804(b)(3), the business records exception, Fed. R. Evid. 803(6), and the residual hearsay exception, Fed. R. Evid. 807. In particular, it found that the co-defendant, “facing the threat of job loss or worse, had great incentive to provide a story of self-defense rather than a statement against his interest.” United States v. Slatten, Crim. No. 14-107, at 6 (D.D.C. June 16, 2014). This is true as a general matter, but it only answers part of the question. The co-defendant’s incentive to keep his job indicates why he might invent a self-defense scenario, and it illustrates why self-serving, exculpatory statements are inadmissible under FRE 804(b)(3). See Williamson, 512 U.S. at 599–600. On the other hand, the co-defendant’s incentive to keep his job does little to explain why he would falsely claim to have shot first and hit the driver, admissions that had the potential to single him out for greater scrutiny and punishment. To the contrary, the threat of job loss magnifies the likelihood that the co-defendant was telling the truth as to those details. See id.; Gichner, 410 F.2d at 242.
This failure to distinguish between inculpatory and exculpatory statements within this co-defendant’s larger narrative, and instead treating the entire four-part narrative as a single “statement” to be admitted or excluded as a whole, was legal error. Williamson, 512 U.S. at 599–600; United States v. Smalls, 605 F.3d 765, 780–87 (10th Cir. 2010). A statement, within the meaning of FRE 804(b)(3), is a “single declaration or remark” rather than a “report or narrative,” Williamson, 512 U.S. at 599 (quoting WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY 2229 (1961)), and thus it was incumbent upon the district court to isolate and admit any “declarations or remarks within the [narrative] that are individually self-inculpatory.” Id. As the Supreme Court has noted, the fact that the narrative was generally exculpatory and untrustworthy does not mean it was entirely untrue: “One of the most effective ways to lie is to mix falsehood with truth, especially truth that seems particularly persuasive because of its self-inculpatory nature.” Id. at 599–600.
To determine whether the error was harmless requires consideration of the antecedent question whether any statements within the co-defendant’s narrative were sufficiently self-inculpatory to be admissible as statements against interest. See Fed. R. Evid. 804(b)(3). If so, then in light of the numerous other pieces of evidence suggesting that the co-defendant fired first, the error could not possibly have been harmless. Although the jury could have reasonably credited Jimmy Watson’s testimony that Slatten fired first over the traffic officers’ testimony that someone in the co-defendant’s position did so, the officers’ testimony would take on new significance if buttressed by the co-defendant’s own admission to firing first. The co-defendant’s claim to have fired first and hit the Kia driver was admissible because no reasonable person would have falsely so claimed, thereby setting the day’s tragic events in motion, especially given the near-certainty that such statements would cost him his job if the self-defense claim were disbelieved. See Fed. R. Evid. 804(b)(3)(A).
The government maintains that the “I shot first and hit the driver” statement cannot be separated from the self-defense statement, i.e., “I shot first and hit the driver in order to protect my team from an imminent threat.” See Appellee Br. 121. The government is correct that a statement’s context must be carefully considered in determining whether the statement is truly self-inculpatory, Williamson, 512 U.S. at 603, but to the extent the government suggests that the court can only consider for admission the conjoined self-defense statement in his narrative, rather than considering for admission only the “I shot first and hit the driver” statement, Williamson instructs to the contrary. The Supreme Court made clear that courts must narrowly parse statements submitted under FRE 804(b)(3) and independently analyze each “declaration[] or remark” within such a statement for admissibility; parts of statements that are not self-inculpatory may not be admitted solely based on their proximity to other self-inculpatory declarations. Id. at 599–601. Here, the inverse is true — the district court erred in excluding self-inculpatory declarations solely based on their proximity to other self-exculpatory declarations.
The relevant self-inculpatory “declarations or remarks” within the co-defendant’s narrative statement, id. at 599, are:
Certainly, the reliability of those statements must be considered in light of the broader self-defense context, but the context is not actually a part of those inculpatory statements. On the other hand, the co-defendant’s statement “Fearing for my life and the lives of my teammates, I engaged the driver and stopped the threat” is generally self-exculpatory and therefore inadmissible. Sworn Statement at 2 (Sept. 18, 2007).
The government makes the related point that an assertion of self-defense automatically removes the underlying admission from the statement against interest exception. In its view, a self-defense claim can never be sufficiently contrary to self-interest under FRE 804(b)(3) because, if true, it would wholly exonerate the declarant. See also Op. 59–60 (citing United States v. Henley, 766 F.3d 893, 915 (8th Cir. 2014); United States v. Shryock, 342 F.3d 948, 981 (9th Cir. 2003)). Once again, Williamson controls: any such generalization is inappropriate because determining whether a statement is self-inculpatory is a “fact-intensive inquiry, which . . . require[s] careful examination of all the circumstances surrounding the criminal activity involved.” Williamson, 512 U.S. at 604.
In Shryock, 342 F.3d at 966–67, for instance, the police already had multiple pieces of evidence showing that the declarant shot the victims, and thus the declarant’s self-defense statement to police was much more obviously exculpatory — the admission “I shot the victims” provided the police with nothing they did not already know or strongly suspect. Here, on the other hand, the record indicates that in the immediate aftermath of the Nisur Square massacre, the co-defendant provided investigators with the very first evidence that he fired before anyone else in the convoy and that he also hit the driver. One can imagine circumstances in which a self-defense claim would be even more devastating, e.g., a person walking into a police station and claiming self-defense in a long-forgotten murder, leading police to reopen the case and immediately find evidence disproving the implausible self-defense claim. Invoking a blanket rule to the contrary seems to suggest that a statement can only be sufficiently damaging to self-interest if, standing alone, it is enough to support a conviction, a civil judgment, or termination. See Op. [59-60]. Thus, an admission to four elements of a crime (but not the fifth) would not qualify as a statement against interest, nor, as here, would an admission to a killing so long as self-defense is also claimed. This court has rejected that notion: “[T]he mere fact that the statements alone do not create an inference of guilt beyond a reasonable doubt does not remove them from the ambit of Rule 804(b)(3).” United States v. Wilson, 160 F.3d 732, 739 & n.4 (D.C. Cir. 1998).
Turning to the statements at issue, the question is whether a reasonable person in the co-defendant’s position would falsely claim to have shot first and hit the driver, even in the context of a self-defense narrative. Fed. R. Evid. 804(b)(3)(A). As the government points out, State Department investigators likely knew (or would soon learn) that this co-defendant fired at the Kia, so he would have been ill-advised to deny he had. That said, investigators also knew (or would soon learn) that numerous other Blackwater guards fired at the Kia around the same time. Thus, with investigators facing the daunting task of piecing together who did what in the midst of a melee, would a reasonable person in the co-defendant’s position claim (1) ignorance as to who shot when and where, essentially hiding behind the “fog of war;” (2) to have fired defensive shots into the Kia’s engine block only after other Raven 23 members had opened fire engaging the driver; or (3) affirmatively admit that he was the first, and likely fatal, shooter? In other words, if in reality this co-defendant had fired fourth and hit only the engine block, then why would he falsely claim responsibility for shots that likely killed a man?
The government suggests that because State Department protocol required guards to shoot occupants of cars that refused to stop despite warnings, the co-defendant could have falsely claimed to have shot first and hit the driver in order to portray himself as having “done precisely the right thing (the heroic thing, even).” Appellee Br. 122. Such an implausible high-risk high-reward strategy would only make sense if one were absolutely confident the self-defense claim would hold up; if not, the heroic narrative would give way to something far more troubling, with devastating consequences for the co-defendant. On the other hand, if at all concerned about the self-defense narrative being rejected, a reasonable person in the co-defendant’s position would have claimed to have fired fourth and hit the engine block because such shots would ultimately be of little consequence even if fired without justification. Based on the evidence adduced at trial showing that the Kia had come to a stop before any shots were fired, the co-defendant would have had little reason to feel so confident in the self-defense claim. Instead, it is far more likely that the co-defendant offered these inculpatory details because they were true, in order to lend credence to his flimsy self-defense narrative. See Williamson, 512 U.S. at 599–600. Indeed, even the co-defendant’s expression of uncertainty as to shooting first (“[I am] not aware of any shots being fired before [mine]”) indicates that he was being truthful in that he would be unlikely to undercut his own lie by expressing doubt about it. As such, the co-defendant’s inculpatory statements were sufficiently trustworthy and contrary to his pecuniary interest to qualify for admission under FRE 804(b)(3).
The requirement to show “corroborating circumstances that clearly indicate . . . trustworthiness” does not apply to statements against pecuniary interest, Fed. R. Evid. 804(b)(3)(B), but the existence of such corroborating circumstances here further demonstrates the admissibility of the inculpatory portion of the co-defendant’s statements. Not only did multiple witnesses similarly describe the first shots as hitting the driver’s side windshield, and multiple others similarly described two initial shots, but Officer Monem testified that a specifically-located gunner fired the first shots, and, crucially, Officer Al-Hamidi testified he was “100 percent certain” that the first shots came from “the [same location] of [a particular] vehicle,” 7/2/14 (PM) Tr. 35:4–15, which was the co-defendant’s position that day. Moreover, Jeremy Krueger, in the second convoy vehicle, testified that the first shots sounded like the 5.56 ammunition used by the co-defendant, rather than the 7.62 ammunition used by Slatten, while Jeremy Ridgeway testified that a few days after the massacre the co-defendant said to him “I feel like this is my fault.” 8/4/14 (PM) Tr. 13:15–14:16. It is difficult to imagine why the co-defendant would seem to accept such responsibility, especially to someone like Ridgeway who would later confess to multiple killings, unless the co-defendant believed he was the one who started the shooting.
That the co-defendant’s inculpatory statements are sufficiently trustworthy to be admitted under an exception to the hearsay rule does not necessarily mean they are true. Nor does it mean that, even if the co-defendant believed they were true, he was correct in thinking he fired first. Rather, it simply means that the statements are trustworthy enough to be presented to a jury, which on retrial can consider all of the available evidence in determining whether or not Slatten fired the first shots that day. The government’s position that any error was harmless because of the strength of the evidence that Slatten fired first, see Appellee Br. 128–29, highlights the importance of a jury making this determination. On remand, the government can make its argument to the jury that the co-defendant’s “equivocal out-of-court response to an investigator,” given the falsity of other aspects of his statement, “surely, [should not] turn[] the tide” and prevent a second conviction of Slatten. Id. at 129.
In sum, because the co-defendant’s inculpatory statements are admissible as statements contrary to pecuniary interest and thus “specifically covered,” see Fed. R. Evid. 807, FRE 804(b)(3) is properly relied upon rather than FRE 807’s residual exception to the hearsay rule. See United States v. Earles, 113 F.3d 796, 800 (8th Cir. 1997).
In any event, there appears good reason not to rely on FRE 807 here. Not only is the district court is “vested with considerable discretion” to apply the residual hearsay exception, United States v. Kim, 595 F.2d 755, 766 (D.C. Cir. 1979), this court has repeatedly emphasized that FRE 807 “was intended to be a narrow exception to the hearsay rule, applied only in exceptional cases,” when the district court would otherwise be forced to exclude evidence that is “very important and very reliable.” Id. FRE 807 requires “circumstantial guarantees of trustworthiness” equivalent to those ensured by FRE 803 and FRE 804, and if my colleagues do not trust the veracity of self-serving statements under FRE 804(b)(3), see Op. 59–60, then it seems illogical to admit those same statements under FRE 807 merely because the co-defendant was under oath, immunized from criminal liability except for the remote possibility of prosecution for making false statements under 18 U.S.C. § 1001, and repeated his statements several times. But see Op. 63–65. Underscoring this point, the investigators’ reports reflect a certain skepticism about the co-defendant’s trustworthiness, at one point noting that he had failed to mention firing his M-203 grenade launcher in prior statements, which the co-defendant “claimed that he had not documented firing . . . because he ‘didn’t think it was important.’” Mem. Report of Interview at 3 (Sept. 23, 2007).
Further, the evidence identified as corroborating the self-defense narrative — which the jury necessarily rejected in reaching its verdict — is misconstrued by my colleagues. See Op. 65–66. The co-defendant told the State Department investigators that traffic in Nisur Square came to a stop upon the convoy’s command, but when the white Kia then approached the convoy at a high rate of speed, he threw a water bottle as a warning prior to firing the first shots. By contrast, Officer Al-Hamidi testified that when Raven 23 members threw water bottles, he turned to see that all traffic was stopped, then he turned back to the convoy as the first shots were fired. Al-Hamidi’s testimony thus in no way corroborates the co-defendant’s self-serving claim that he threw a water bottle at a speeding Kia as a warning, and to construe it otherwise only amplifies the error in deeming the self-defense portions of the narrative admissible.
Finally, reliance on FRE 807 for admission of the co-defendant’s statements as a whole generally ignores both Williamson, 512 U.S. at 599–600, and Slatten’s clarification that he is not seeking admission of the portions of the co-defendant’s statements that the government claims are self-serving, such as that the Kia approached at high speed, did not stop despite the co-defendant’s attempts to stop it, or that the co-defendant feared for his life. See Slatten Br. 40. But see Op. 54, 68–69. Although I too conclude that the district court abused its discretion, Slatten’s challenge to the denial of his motion for severance does not require the Court to reach FRE 807 in order to grant him the relief he seeks (or, indeed, relief he expressly does not seek), and therefore I would not do so.
Accordingly, I concur in the judgment that the district court abused its discretion in denying admission of the co-defendant’s statements, to the extent I have identified in view of the analysis Williamson requires.
Dissenting from Part VIII. Paul Slough was convicted by a jury of killing thirteen (13) people and attempting to kill seventeen (17) others. Evan Liberty was convicted by a jury of killing eight (8) people and attempting to kill twelve (12) others. Dustin Heard was convicted by a jury of killing six (6) people and attempting to kill eleven (11) others. Even leaving aside their firearms convictions under 18 U.S.C. § 924(c), the maximum sentences on their remaining convictions were 249 years for Slough, 164 years for Liberty, and 137 years for Heard. Especially in light of that congressionally determined exposure, the thirty-year-and-one-day sentences imposed by the district court were not unconstitutionally “grossly disproportionate to the crime[s].” Graham v. Florida, 560 U.S. 48, 60 (2010) (internal quotation marks omitted).
Today my colleagues hold that the mandatory sentence of thirty years under Section 924(c), as applied to these three private security guards for using government-issued weapons in a war zone, is cruel and unusual punishment in violation of the Eighth Amendment to the U.S. Constitution. Op. 86–87. In so doing, they have failed to account, as they must, for “all of the circumstances of the case.” Graham, 560 U.S. at 59. Most crucially, my colleagues make no mention of the fact that the district court judge, who presided at the months-long trial, imposed sentencing packages that the judge concluded “achieved an overall appropriate sentence [for each of these defendants] rather than calculating individual sentences for each component.” Sent. Tr. 150:18–25 (Apr. 13, 2015) (citing United States v. Townsend, 178 F.3d 558, 567 (D.C. Cir. 1999)). That is, in consideration of the mandatory minimum under Section 924(c), the district court imposed only a one-day sentence for all of these defendants’ many manslaughter and attempted manslaughter convictions. The Supreme Court has affirmed the district court’s discretionary authority to impose such a sentencing package in Dean v. United States, 137 S. Ct. 1170 (2017), holding that nothing in Section 924(c) prevents a district court from, as here, mitigating the harshness of a mandatory thirty-year minimum by imposing a one-day sentence for the predicate convictions. See id. at 1176–77.
My colleagues’ conclusion that there has been a constitutional violation, by contrast, rests on the mistaken premise that the thirty years allocated to the Section 924(c) convictions represent freestanding sentences distinct from the one-day sentences on the remaining manslaughter and attempted manslaughter convictions. See Op. 72; Townsend, 178 F.3d at 567. In disregarding the basic structure of these defendants’ sentences, my colleagues fail to recognize that the district court already mitigated any disproportionality. Indeed, the district court judge stated on the record that he was “very satisfied” with the thirty-year sentences in light of the “many killings and woundings” for which these defendants were responsible. Sent. Tr. 154:9–22. My colleagues ignore this fact too, particularly when they suggest that the district court judge felt constrained to impose an unduly harsh sentence. See Op. 76–77. And whatever their concern with the “one-size-fits-all nature of these sentences,” see id. at 77, that is not an Eighth Amendment concern because none of the sentences are disproportionate to the enormity of the crimes that the jury found the defendants had committed. Graham, 560 U.S. at 60. Again, Congress has determined that the least culpable defendant here — who was convicted of killing six people and wounding eleven others — should be subject to a maximum of 167 years in prison, in addition to the thirty years under Section 924(c).
Although it is possible to imagine circumstances in which a thirty-year minimum sentence for a private security guard working in a war zone would approach the outer bounds of constitutionality under the Eighth Amendment, this is not that case. The jury rejected these defendants’ claim that they fired in self-defense, and far more of their fellow security guards chose not to fire their weapons at all that day. Yet as my colleagues apparently see it, Congress should have included an exception for all such military contractor employees, or, rather, it would have included such an exception if it had only considered the issue. See Op. 72–74. Perhaps so, but that is not the question before us. The district court judge made an individualized assessment of an appropriate sentencing package for each of these defendants, and the result is not disproportionate to the defendants’ crimes, let alone grossly, unconstitutionally disproportionate.
Accordingly, I respectfully dissent from Part VIII.
While the Court’s ultimate conclusions follow inexorably from its broad reading of the Military Extraterritorial Jurisdiction Act (“MEJA”), 18 U.S.C. §§ 3261 et seq., the Court’s initial premise seems faulty. MEJA was amended in 2004 to close a loophole that allowed non-Department of Defense (“DOD”) contractors to escape criminal liability for crimes committed overseas. I agree Congress used “deliberately expansive” language in MEJA so contractors working to support the DOD in its mission would not escape prosecution for crimes committed while performing their duties, regardless of which federal agency was their employer. See Maj. Op. at 11. However, I am not convinced that any federal contractor whose employment relates—even minimally—to the DOD’s mission is automatically subject to MEJA. The Court’s interpretation unnecessarily broadens that which the statutory language seems designed to limit.
When interpreting a statute, the analysis begins—and often ends—with its text. See, e.g., Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 438 (1999). Here, the text of MEJA extends the jurisdiction of federal courts to crimes committed in foreign countries if the crime was committed while the defendant was “employed by . . . the Armed Forces outside the United States.” 18 U.S.C. § 3261(a)(1). The statute further defines the time period of being “employed by the Armed Forces outside the United States” to include acts committed while a person is the employee of a contractor of “(I) the [DOD]. . . ; or (II) any other Federal agency . . . to the extent such employment relates to supporting the mission of the [DOD] overseas . . . .” 18 U.S.C. § 3267(1)(A)(iii). Thus, by MEJA’s plain terms, the employee of a DOD contractor is automatically subject to prosecution under MEJA for any offense committed while working overseas without any qualifications. See id. § 3267(1)(A)(iii)(I). This suggests DOD contractors and their employees are subject to MEJA for crimes committed while on or off duty. The same is not true for non-DOD contractors though. If the perpetrator of a crime is an employee or contractor of any federal agency other than the DOD, he is subject to MEJA only “to the extent [his] employment relates to supporting the mission of the [DOD].” Id. § 3267(1)(A)(iii)(II) (emphasis added). The phrase “to the extent” does no work unless it implies the criminal liability of non-DOD contractors is more limited than DOD contractors. After all, the word “extent” is defined as “the range (as of inclusiveness or application) over which something extends.” WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY UNABRIDGED 805 (1993). Moreover, the Supreme Court has taken a similar view of this phrase when interpreting its statutory meaning. See John Hancock Mut. Life Ins. Co. v. Harris Tr. & Sav. Bank, 510 U.S. 86, 104–05 (1993) (describing “to the extent” as “words of limitation” in a statute).
Accordingly, if MEJA targets contractors (or their employees) performing specific military roles, it makes little sense to turn our inquiry into an essentially all-or-nothing analysis. However, this is precisely the effect of concluding that MEJA applies to all of a federal contractor’s conduct for the duration of the time that some aspect of that contractor’s employment supports the DOD’s mission. In situations such as this case—where the mission of the DOD is characterized to broadly encompass all activities related to nation-building—the limitation the Court purports to create is virtually boundless. Instead, the more logical reading of the statute is that a non-DOD contractor is subject to MEJA only when a specific task being performed by that contractor is integral to the DOD’s mission. Had Congress wished MEJA to apply more broadly to non-DOD contractors, it could have substituted the word “if” for the phrase “to the extent” to give MEJA the expansive wording necessary to achieve such a result. See id. (contrasting the word “if” with the phrase “to the extent”); see also In re Silveira, 141 F.3d 34, 36 (1st Cir. 1998) (“If Congress intended for [the statute at issue] to be an ‘all-or-nothing’ matter, one might wonder why the provisions’ drafters chose to use the connective phrase ‘to the extent that,’ in lieu of the word ‘if,’ which obviously would have been a simpler construction.”). To hold this difference of language is nothing more than a “temporal limitation,” Maj. Op. at 10, ignores the distinction Congress made between those who work directly for the DOD and those who do not.
Moreover, the Court’s interpretation goes beyond the problem Congress was attempting to solve when it amended MEJA in 2004. As noted by the Court, Congress sought to amend MEJA in response to the atrocities committed by Interior Department contractors at the Abu Ghraib prison in Baghdad. Maj. Op. at 8. These contractors working in Abu Gharib were soldiers in all but name, and they were directly assisting the DOD in running a prison for detained enemy combatants. See Saleh v. Titan Corp., 580 F.3d 1, 6–7 (D.C. Cir. 2009) (stating the Abu Ghraib contractors were “integrated [with the military] and performing a common mission with the military under ultimate military command”). Viewed with this context in mind, it is clear that what Congress sought to do when it amended MEJA was to assure that contractors of any federal agency who were performing tasks conventionally done by soldiers could not elude U.S. jurisdiction. Creating criminal liability for all federal employees or contractors whose employment relates—even tangentially so—to the DOD’s mission goes beyond a plain reading of the text. Because we are to “scrupulously confine [our] own jurisdiction to the precise limits which (a federal) statute has defined,” Victory Carriers, Inc. v. Law, 404 U.S. 202, 212 (1971) (quoting Healy v. Ratta, 292 U.S. 263, 270 (1934)), we must use extreme caution when expanding our jurisdiction—particularly when doing so results in our criminal law applying extraterritorially.
Here, I believe Congress said what it meant and meant what it said, see Conn. Nat’l Bank v. Germain, 503 U.S. 249, 253–54 (1992), and I would not dismiss the distinctions made in the text in favor of aspirational goals set forth by the statute’s sponsors. See Maj. Op. at 9 (citing Senator Schumer’s floor statement declaring MEJA was amended to address “a dangerous loophole in our criminal law that would have allowed civilian contractors who do the crime to escape doing the time”). The Court may be correct that Congress intended for MEJA’s 2004 amendment to treat DOD and non-DOD contractors and their employees exactly the same when a non-DOD contractor’s employment relates to the DOD’s mission, but “[i]t is not for us to rewrite the statute so that it covers . . . what we think is necessary to achieve what we think Congress really intended.” Lewis v. City of Chicago, 560 U.S. 205, 215 (2010). If the government truly desires this result, the proper course of action is to petition Congress to amend the statute, not advocate for courts to read problematic language out of its text.
Because MEJA’s text compels the conclusion that Congress meant to treat DOD and non-DOD employees and contractors differently, the next inquiry is to determine which actions of non-DOD contractors are subject to MEJA and which are not. The text once again provides a clear answer: only crimes committed while “employed by . . . the Armed Forces outside the United States” falls within MEJA’s purview. 18 U.S.C. § 3261(a)(1). MEJA specifically provides non-DOD contractors are only “employed by . . . the Armed Forces” for the purposes of the statute when, though acting within the scope of their employment, they are “supporting the mission of the [DOD].” Id. §§ 3261, 3267. The phrase “relating to” is “deliberately expansive” and must be given broad scope. Maj. Op. at 11. However, its broad scope is not so expansive as to swallow up the “words of limitation” immediately preceding them. See John Hancock Mut. Life Ins. Co., 510 U.S. at 104–05. Instead, these competing phrases must be balanced in ways that give both full meaning. Therefore, MEJA logically encompasses those actions taken by non-DOD employees pursuant to their employment that either directly or indirectly support the DOD’s mission. The statutory framework focuses on military employment and thus limits the scope of jurisdiction not just temporally but factually. This interpretation gives full meaning to the broad language of the text without making virtually all potential crimes committed by a non-DOD employee subject to the federal criminal law. The proper question is whether the Defendants were either directly or indirectly supporting the DOD when they entered Nisur Square on the day of the incident.
Under this Court’s precedent, we examine the jury findings for each element of MEJA under the deferential sufficiency-of-the-evidence standard. United States v. Williams, 836 F.3d 1, 7 (D.C. Cir. 2016). In doing so, we determine “whether the evidence, considered in the light most favorable to the government, was sufficient to permit a rational trier of fact to find all of the essential elements of the [statute were met] beyond a reasonable doubt.” United States v. Wilson, 240 F.3d 39, 43 (D.C. Cir. 2001). Here, the government has arguably met its burden. I do not join the Court in holding that any actions deemed to facilitate rebuilding the war-torn nation of Iraq automatically relates to the DOD’s mission based on the text and history discussed above. For this reason, I also find the evidence of the Defendants performing other tasks to support the DOD—such as assisting distressed military units and training Army escorts, Maj. Op. at 14—to be of doubtful relevance in determining whether the Defendants were supporting the DOD on the day of the Nisur Square incident.
However, neither of these pieces of evidence are necessary to uphold the jury’s finding under the deferential sufficiency of the evidence standard, and we need not decide whether they would be sufficient on their own to meet MEJA’s criteria. As posited by the Court, the Defendants’ employment—providing diplomatic security for the Department of State—indirectly supported the DOD’s mission by allowing military personnel previously responsible for providing State Department security to concentrate exclusively on the DOD’s rebuilding mission. Maj. Op. at 14–15. The relatively small size of America’s active, volunteer military and the breadth of its commitments may blur the lines, but it does not erase them. Although statements from Deputy Secretary of Defense Gordon England unequivocally stating that the Defendants were not supporting DOD’s mission contradicted the prosecution’s narrative, JA 2919–20, 2932, 2936, contrary evidence is not enough to overcome this deferential standard. While I would interpret MEJA more narrowly and find the question close, arguably sufficient evidence existed for a rational juror to conclude that MEJA applied to the Defendants.
One question remains. Did the district court properly instruct the jury on MEJA’s application to this case? When examining a challenge to jury instructions, we must determine “whether, taken as a whole, the [district court’s] instructions accurately state the governing law and provide the jury with sufficient understanding of the issues and applicable standards.” United States v. DeFries, 129 F.3d 1293, 1304 (D.C. Cir. 1997). Because an “improper instruction on an element of the offense violates the Sixth Amendment’s jury trial guarantee,” it is a reversible error requiring a new trial unless the error was harmless. Neder v. United States, 527 U.S. 1, 12, 15 (1999). The burden is on the government to prove the error was harmless beyond a reasonable doubt. Wilson, 240 F.3d at 44.
Here, the district court interpreted MEJA as an all-or-nothing proposition. When explaining how MEJA should apply to the case, the court emphasized the broad nature of certain words in the text by defining the word “relates” to mean “a connection with” and defining “supporting” to mean “to promote the interest or cause of something or someone.” JA 3293. After giving these definitions, the culmination of the instruction advised that employment relating to supporting the DOD’s mission included a contractor of “any federal agency whose employment in the Republic of Iraq bears some relationship to supporting the mission of the Department of Defense in that country.” Ibid. While the instruction did contain the phrase “to the extent,” the presence of this phrase does little work because the overall framing of the issue is erroneous. By describing MEJA in this manner, the district court, essentially read the limiting effect of the phrase “to the extent” right out of the statute and instead substituted “if” in its place. Moreover, this expansive view effectively eliminates the connection to military employment. As discussed above, this is a dubious interpretation of the statute.
At the jury instruction stage, the imprecision inherent in a sweeping view of the DOD’s mission becomes apparent. The jury instruction, if erroneous, was prejudicial because it affected a central issue in a close case where persuasive evidence was presented by both the prosecution and the defense. See Williams, 836 F.3d at 16 (reversing a murder conviction under MEJA because a misstatement of the law by the prosecution during its closing statement involved a “central and close issue in the case” that was “insufficiently cured”). Because the question of whether MEJA applied to the Defendants was a threshold issue for each conviction, there is no issue more central to the entire case than this. Furthermore, even if sufficient evidence existed to find jurisdiction under MEJA in this case, the same would be true if the jury had reached the opposite conclusion based upon Deputy Secretary England’s testimony and the representations he made that the Defendants were not supporting the DOD’s mission, see JA 2953, 3843, 3858. Thus, the importance of an accurate statement of the law cannot be gainsaid.
However, given the district court’s instructions, it was entirely possible for the jurors to begin deliberations believing that if any aspect of the Defendants’ employment related to supporting the DOD’s mission, then any supporting action taken during the course of that employment made the Defendants subject to MEJA. Accordingly, the jurors could find MEJA applied solely on the basis of actions taken during the course of the Defendants’ employment—even actions completely unrelated to the events that transpired in Nisur Square, such as providing assistance to distressed military units or training Army security escorts. This is a significantly different calculus than attempting to determine if the Defendants’ action on the day of the Nisur Square incident related to supporting the DOD’s mission. The difference is stark. The jury conceivably could have reached a different conclusion had it been correctly instructed. At a minimum, the government cannot prove beyond a reasonable doubt the erroneous jury instruction was harmless error. See Wilson, 240 F.3d at 44. Therefore, I would have reversed the Defendants’ convictions and remanded the case for a new trial.
The question of how our criminal justice system should treat private contractors who commit crimes overseas in war time is a difficult one. However, Congress has made the determination that such individuals should be held responsible for their actions in federal courts if they either work for the military or commit a crime during the performance of a task related to supporting the military, such as the atrocities committed at Abu Ghraib. Today’s opinion expands MEJA beyond the limits defined by this history and clearly laid out in the text. Because it is not possible to conclude, beyond a reasonable doubt, that the erroneous instruction did not improperly influence the ultimate outcome of the case, I respectfully dissent from this portion of the Court’s decision.