UNITED STATES of America, Plaintiff – Appellant, v. Bijan RAFIEKIAN, a/k/a Bijan Kian, Defendant – Appellee.
This case comes to us on appeal for a second time. In 2019, a jury convicted Bijan Rafiekian of one count of acting as an unregistered agent of a foreign government and one count of criminal conspiracy. But the district court granted a judgment of acquittal as to both charges and conditionally granted a new trial in the event the judgment of acquittal was reversed on appeal.
On appeal, in Rafiekian I, we reversed the judgments of acquittal, vacated and remanded the court's new-trial order, and noted that the district court “may have additional justifications for its decision” that it failed to explain. United States v. Rafiekian, 991 F.3d 529, 550 (4th Cir. 2021) (“Rafiekian I”). On remand, the district court provided additional justifications, again ordering a new trial. The government appeals. For the reasons that follow, we affirm.
In July 2016, Rafiekian was an executive at Flynn Intel Group (“FIG”), a consulting and lobbying firm founded by Rafiekian and retired Lt. Gen. Michael T. Flynn.
On July 19, 2016, the Turkish government made a request to the U.S. Department of Justice (“DOJ”) to extradite Turkish preacher and scholar Fethullah Gulen, who lived in Pennsylvania. Turkey painted Gulen as a terrorist, blaming him for a failed coup attempt against Turkish President Recep Tayyip Erdogan earlier that year. DOJ declined to immediately fulfill the request but placed it under review.
Later that month, Rafiekian and Flynn began discussions with Ekim Alptekin, a Turkish businessman who held himself out as being in contact with Turkish officials, about the possibility of hiring FIG for a project related to “Turkey's security and stability.” J.A. 2065.1 Rafiekian indicated that FIG was “ready to engage on what needs to be done” to further these ends, expressing his belief that President Erdogan could “lead the campaign against Radical Islam,” in furtherance of global security. Id. Shortly thereafter, Rafiekian sent Alptekin an initial list of action items for the potential project, which he referred to as “the ‘truth’ campaign” (“Project Truth”). J.A. 2069.
Alptekin responded favorably to the action items and indicated that he had shared the “proposed approach” with Turkey's Foreign Affairs Minister, who was “receptive.” J.A. 2087. Alptekin also highlighted “the depth of the crisis” he felt they were facing concerning global impressions of Turkey, sharing with Flynn and Rafiekian a piece by the New York Times Editorial Board discussing the coup attempt in Turkey and Turkey's request for Gulen's extradition. Id. The op-ed, which was apparently cause for concern for Alptekin, asserted that Turkey needed to be “reminded that Mr. Gulen has a legal right to be in the United States, and that the Justice Department would have to go through a rigorous process before deciding whether he could be handed over, especially to a country where due process is increasingly unlikely and torture is reportedly used against detainees.” Editorial, Turkey's New Anti-Americanism, N.Y. Times (Aug. 4, 2016), https://www.nytimes.com/2016/08/04/opinion/turkeys-new-anti-americanism.html (saved as ECF opinion attachment).
Rafiekian responded to Alptekin by highlighting the threat he believed Gulen posed, employing an analogy to Iran's Ayatollah Khomeini: “A soft spoken cleric sitting under an apple tree in Neauphle-le-Chateau in France looked so harmless.” J.A. 2083. Because he viewed Gulen as a threat, Rafiekian supported FIG's working on behalf of the Turkish government to discredit Gulen in the eyes of the American public.
Alptekin later relayed that he had had more contacts with the Turkish government about the potential project. He indicated that he had met with the Turkish Minister of Economy, who agreed to discuss the engagement with other government officials. Into early August 2016, Alptekin, Rafiekian, and Flynn continued discussing the project, and on August 10, Alptekin informed Rafiekian and Flynn that he had “just finished in Ankara after several meetings today with Min[ister] of Economy Zeybekci and [Minister of Foreign Affairs] Cavusoglu” and that he had “a green light to discuss confidentiality, budget and the scope of the contract.” J.A. 2092.
After August 10, references to Project Truth—and the Turkish government's interest in it—ceased. But the next day, on August 11, Rafiekian emailed Alptekin to discuss an ostensible new project: an “[e]ngagement” with the purpose of “restor[ing] ‘confidence through clarity’ in the trade and investment climate” in Turkey. J.A. 2098. Rafiekian indicated that he had been tinkering with the budget for this project but that he “did not touch the advisory support [they] discussed at 20%.” J.A. 2098. He stated that he and Flynn had “activated the FIG LAB” and were “ready to push the start button immediately.” J.A. 2098.
The same day, Rafiekian emailed others within FIG, stating that they were “about to be engaged by a Dutch client” for the “CONFIDENCE THROUGH CLARITY CAMPAIGN – Operation Confidence.” J.A. 2103. The Dutch client backing the project was a private company called Inovo BV, of which Alptekin was the sole shareholder and employee.
Rafiekian circulated a list of “Phase Zero” action items for Operation Confidence among members of the FIG team. This list substantially mirrored the action-items list previously circulated for the defunct Project Truth, although the items appeared to be largely generic, rather than specific to either project. See, e.g., J.A. 2103, 2069 (including, as an action item, “Define dependencies, uncertainties, expected and unexpected consequences”). He also attached a proposed budget, which included, as one item, $120,000 to Alptekin. J.A. 2104.
The goal of Operation Confidence was to restore confidence in Turkey's investment climate, and the mechanism for accomplishing this goal was discrediting Gulen. Those on the FIG team who worked on the project apparently understood discrediting Gulen to be central to restoring confidence in the Turkish investment climate because it would contextualize the actions President Erdogan had taken in response to the coup attempt—actions that many in the United States viewed as unjustifiably harsh. See Editorial, Turkey's New Anti-Americanism, supra, (New York Times Editorial Board discussing President Erdogan's “use of the crisis” precipitated by the coup attempt “to purge some 66,000 people from the military, government ministries, schools and universities”).
Although Project Truth and Operation Confidence had similarities, there was a key distinction between the projects: while the Turkish government would have been the client for Project Truth, Inovo BV was the client for Operation Confidence. Despite the Turkish government's decision not to go through with engaging FIG on Truth, FIG proceeded with Operation Confidence, in order to accomplish the same goal that would have been pursued in Project Truth—discrediting Gulen—for a different client.
Inovo and FIG signed an agreement solidifying the partnership for the project, and the FIG team worked on Operation Confidence through the rest of August and into September.
In the early stages of Operation Confidence, Flynn and Rafiekian discussed whether and how to register their activities with the federal government. They sought to keep the activities confidential, and Rafiekian was concerned that certain ways they might register could expose their efforts to members of Congress who were favorable to Gulen. Still, in August 2016, Rafiekian contacted attorneys at the law firm Covington & Burling (“Covington”) to inquire as to whether FIG had an obligation to register its activities under the Foreign Agents Registration Act (“FARA”). Due to differences in political alignment, Rafiekian ultimately looked elsewhere for advice on this question.
Rafiekian next consulted his friend Robert Kelley, also an attorney, to ascertain whether FIG needed to register its activities. Rafiekian informed Kelley that he believed FIG had an obligation to register under FARA. But Kelley advised otherwise. Kelley first asked whether the client for the project was a foreign government or foreign political party. Rafiekian responded that it was neither; instead, he informed Kelley, the client was a foreign private business—Inovo BV. Kelley advised Rafiekian that, because the client was a private business rather than a foreign government, FIG could file under the Lobbying Disclosure Act instead of under FARA. So on September 30, 2016, FIG—through Kelley—registered under the Lobbying Disclosure Act, indicating that it would be advising Inovo on U.S. domestic and foreign policy.
While the work on Operation Confidence was ongoing, on September 19, 2016, Rafiekian went to a meeting at a New York hotel attended by Flynn, Alptekin, and Turkey's Energy and Foreign Affairs Ministers, along with other FIG employees. Contemporaneous records—such as checks issued by FIG to compensate members of the FIG team for attending—indicated that the meeting was related to Operation Confidence. Later, when interviewed about the meeting in a subsequent investigation, Rafiekian stated that this meeting had been unrelated to Operation Confidence. Regardless, a member of the FIG team who attended the meeting testified at trial that the meeting focused entirely on Turkey's displeasure with Gulen and its desire for his extradition.
After the meeting, continuing its work on Operation Confidence, FIG hired a public relations firm called Sphere Consulting to assist with investigating Gulen and ultimately producing a video critical of him. Also in furtherance of Operation Confidence, Rafiekian lobbied at least two members of Congress to push for public hearings on Gulen.
Alptekin continued to be heavily involved in the progress of the project. FIG conducted weekly conference calls with Alptekin to provide updates. And throughout the duration of the project, FIG made payments to Alptekin—even though Alptekin's company was the client and therefore was paying FIG for its services. Upon receiving an invoice, Alptekin would pay FIG's fee from a Turkish account, but FIG would then make large payments back to Inovo.
In November 2016, Alptekin met in person with the FIG team to receive an update. Apparently displeased with the lack of results up to that point, he reportedly exclaimed, “What am I going to tell Ankara?” J.A. 1034. That night, Rafiekian emailed Alptekin a draft op-ed he had composed, which was critical of Gulen, and stated: “A promise made is a promise kept.” J.A. 2197. The op-ed was published in The Hill on November 8 under Flynn's name. Lt. Gen. Michael T. Flynn, Our Ally Turkey Is in Crisis and Needs Our Support, The Hill (Nov. 8, 2016), https://thehill.com/blogs/pundits-blog/foreign-policy/305021-our-ally-turkey-is-in-crisis-and-needs-our-support/ (saved as ECF opinion attachment).
The publication of the op-ed sparked attention from DOJ, which sent an inquiry letter to Flynn to ask whether he, FIG, or anyone else related to the entity might have an obligation to register under FARA.
After receiving the DOJ letter in November 2016, FIG again sought the services of Covington, this time hiring the firm to assist in responding to the letter. Covington conducted an investigation, during which it spoke to Rafiekian, among others, about FIG's activities. During the investigation, Rafiekian stated that Operation Confidence was entirely separate from Project Truth, and that Truth—for which the Turkish government would have been the client—had never come to fruition.
After concluding its investigation, in March 2017, Covington submitted a FARA filing on FIG's behalf. The filing stated that FIG did not know the extent to which the government of Turkey was involved in Inovo's retention of FIG for Operation Confidence but acknowledged that the project “could be construed to have principally benefited the Republic of Turkey.” J.A. 2515. The filing also stated that Flynn's op-ed was not written at the request of Inovo, the government of Turkey, or any other party.
Rafiekian was indicted, alongside Alptekin, on two counts. Count 1 charged Rafiekian with criminal conspiracy in violation of 18 U.S.C. § 371, the objects of which were 1) acting as an undisclosed foreign agent in violation of 18 U.S.C. § 951, and 2) making a materially false FARA registration in violation of 22 U.S.C. § 618(a)(2). Count 2 charged Rafiekian with acting as a foreign agent without first notifying the Attorney General, in violation of 18 U.S.C. § 951.
We pause here to provide additional detail as to what these statutes require. “To fall within § 951's ambit,” as we made clear in Rafiekian I, “a person must do more than act in parallel with a foreign government's interests or pursue a mutual goal.” 991 F.3d at 538. Instead, a person must “ ‘agree[ ] to operate ․ subject to the direction or control’ of [the foreign] government.” Id. (quoting 18 U.S.C. § 951(d)) (emphasis in Rafiekian I). To exercise “direction or control” under § 951's definition, the foreign government need not dictate every particular act taken by the agent, but the agreement between the government and the agent “cannot be one-sided”—that is, “a person does not become an ‘agent’ for purposes of § 951 simply by acting in accordance with foreign interests or by privately pledging allegiance.” Id. at 541 (emphasis added). There must be “a foreign government or official on one end of the line.” Id. at 540.
Thus, for Rafiekian to be convicted of Count Two—acting as a foreign agent without notifying the Attorney General in violation of § 951—the government had to prove that Rafiekian acted subject to the “direction or control” of the Turkish government, 18 U.S.C. § 951(d), not simply that his actions aligned with Turkey's interests. Rafiekian I, 991 F.3d at 538–41.
To be convicted of Count One—criminal conspiracy—the government had to prove that Rafiekian conspired with another person to commit at least one of the two objects identified in the indictment: first, to violate § 951 by acting as an unregistered foreign agent, or second, to make a materially false FARA registration in violation of 22 U.S.C. § 618(a)(2).
To prove the first object, the government had to show that Rafiekian conspired to act subject to the “direction or control” of the Turkish government which, as explained above, required more than simply showing that Rafiekian agreed with another person to act parallel to Turkey's interests. See Rafiekian I, 991 F.3d at 538–41. To prove the second object, the government had to prove that, in connection with the FARA filing, Rafiekian agreed with another person to willfully make a false statement of material fact, willfully omit a material fact required to be included, or willfully omit a material fact or document necessary to make the filing not misleading. 22 U.S.C. § 618(a)(2).
FARA itself requires individuals to register when they act as an “agent of a foreign principal,” id. § 612(a) (emphasis added), a term that sweeps much more broadly than § 951's reference to an “agent of a foreign government,” 18 U.S.C. § 951(d) (emphasis added); see Rafiekian I, 991 F.3d at 539 (noting that “Congress utilized a more sweeping definition of ‘agent’ ” in FARA than in § 951); 22 U.S.C. § 611(b) (defining “foreign principal” for purposes of FARA to include foreign governments, foreign political parties, certain individuals outside the United States, and foreign private businesses). Registration under FARA is one way to satisfy § 951's registration requirement. See 28 C.F.R. § 73.3(e); Rafiekian I, 991 F.3d at 535 n.6.
Rafiekian's trial began on July 15, 2019, and after the government rested its case, Rafiekian moved for a judgment of acquittal under Federal Rule of Criminal Procedure 29. The court reserved deciding the motion. Rafiekian moved for acquittal again at the close of evidence, and the district court again reserved the decision.
The jury convicted Rafiekian on both counts, including on both objects of the conspiracy charge.
After the verdict, on September 24, 2019, the court granted Rafiekian's motion for a judgment of acquittal as to both counts, finding that there was “insufficient evidence to convict Rafiekian on either count related to § 951, because no rational juror could conclude that Rafiekian had acted ‘on behalf of the Turkish government,’ much less ‘subject to Turkey's direction or control.’ ” Rafiekian I, 991 F.3d at 537 (quoting United States v. Rafiekian, No. 1:18-cr-457-AJT-1, 2019 WL 4647254, at *14 (E.D. Va. Sept. 24, 2019)). Further, it found that there was no evidence of an agreement to avoid filing under FARA or to file a false FARA registration. Id.
The court also conditionally granted Rafiekian a new trial under Federal Rules of Criminal Procedure 29(d)(1) and 33. In relevant part,2 the district court stated merely that, “given the great weight of the evidence, as detailed [in the judgment-of-acquittal analysis], the Court also finds that it would be in the interests of justice to vacate the judgment of conviction, and therefore conditionally orders a new trial in the event the Court's order of acquittal is vacated or set aside as to some or all of the counts of conviction.” Rafiekian, 2019 WL 4647254, at *16; see Fed. R. Crim. P. 29(d)(1) (“If the court enters a judgment of acquittal after a guilty verdict, the court must also conditionally determine whether any motion for a new trial should be granted if the judgment of acquittal is later vacated or reversed.”). The government appealed.
On appeal, this Court reversed the judgment of acquittal, finding that, viewed in the light most favorable to the government, substantial evidence supported the convictions such that a rational jury could convict Rafiekian on both counts. Rafiekian I, 991 F.3d at 544–49. And we vacated and remanded the district court's conditional grant of a new trial. We concluded that the court's explanation was insufficient because it merely pointed to the earlier sufficiency-of-the-evidence reasoning supporting the judgment-of-acquittal analysis. Id. at 549–50. We concluded that the reasoning supporting the judgment of acquittal, “standing alone,” was insufficient to justify granting a new trial. Id. at 550. We also noted, however, that, “having observed the trial in the flesh, the district court may have additional justifications for its decision.” Id. We therefore vacated the new-trial order and remanded to the district court for further proceedings. Id. at 550, 552.
On remand, the district court received additional briefing and heard argument. It then granted a new trial for a second time, this time explaining its reasoning in a thorough, 51-page opinion. The government timely appealed.
We start by setting out some legal background. A district court may order a new trial on the defendant's motion “if the interest of justice so requires.” Fed. R. Crim. P. 33(a). Although only a new-trial order is at issue here, to fully explain the new-trial standard, we must distinguish it from the standard for granting a judgment of acquittal (which the district court initially granted, and which we reversed in Rafiekian I). Though either form of relief could be warranted in a given case on evidentiary grounds, the standards for granting these two different kinds of relief are distinct.
On the one hand, a judgment of acquittal is appropriate when the evidence is so deficient that acquittal is “the only proper verdict.” Tibbs v. Florida, 457 U.S. 31, 42, 102 S.Ct. 2211, 72 L.Ed.2d 652 (1982) (emphasis added). That is, if the evidence is so insufficient that no rational trier of fact could convict, the court should enter a judgment of acquittal. Accordingly, in determining whether to grant a judgment of acquittal, the court views “the evidence and inferences therefrom” in the light most favorable to the government. Id. at 41, 102 S.Ct. 2211 (quoting Burks v. United States, 437 U.S. 1, 16, 98 S.Ct. 2141, 57 L.Ed.2d 1 (1978)).
A new trial, on the other hand, may be granted where the government has presented sufficient evidence for a reasonable jury to convict, but the court nevertheless “disagree[s] with the jurors' weighing of the evidence” in finding the defendant guilty. Id. at 42, 102 S.Ct. 2211.3 So in determining whether a new trial is warranted, the district court—“sit[ting] as a ‘thirteenth juror,’ ” id.—conducts its own assessment of the evidence, unconstrained by any requirement to construe the evidence in the government's favor.4 United States v. Arrington, 757 F.2d 1484, 1485 (4th Cir. 1985).
District courts should grant new trials based on the weight of the evidence only in “rare” instances. United States v. Singh, 518 F.3d 236, 249 (4th Cir. 2008) (quoting United States v. Smith, 451 F.3d 209, 217 (4th Cir. 2006)). Merely believing that the case could have come out the other way is not enough to warrant a new trial. Rather, this Court has instructed that a new trial based on the weight of the evidence is warranted “[w]hen the evidence weighs so heavily against the verdict that it would be unjust to enter judgment.” Arrington, 757 F.2d at 1485. Since “determining witness credibility and weighing conflicting evidence are the responsibility of the factfinder,” United States v. Chavez, 894 F.3d 593, 608 (4th Cir. 2018), “the ‘standard for jettisoning a jury verdict in favor of a new trial’ ” is “demanding,” Rafiekian I, 991 F.3d at 549 (quoting United States v. Millender, 970 F.3d 523, 532 (4th Cir. 2020)).
Still, the district court is permitted to conduct its own assessment of witness credibility, Arrington, 757 F.2d at 1485, and to re-weigh the evidence, see United States v. Campos, 306 F.3d 577, 579 (8th Cir. 2002) (citing White v. Pence, 961 F.2d 776, 780 (8th Cir. 1992)), to determine whether a new trial is warranted. And the court may properly conclude that a new trial is warranted based on the “cumulative” weight of the evidence rather than by separately rejecting each individual offer of proof by the government. United States v. Campbell, 977 F.2d 854, 860 n.6 (4th Cir. 1992).
Having set out the legal background of the new-trial order, we now set out the nature of our review of the district court's application of this standard.
We review a district court's grant of a new trial for abuse of discretion. Rafiekian I, 991 F.3d at 549. Under this standard, we do not substitute our judgment for the district court's; we simply ask whether that court exercised its discretion in an arbitrary or capricious manner. United States v. Fulcher, 250 F.3d 244, 249 (4th Cir. 2001).
There are good reasons for our appellate deference to the trial court. We as appellate judges “do not experience the tenor of the testimony at trial. The balance of proof is often close and may hinge on personal evaluations of witness demeanor.” United States v. A. Lanoy Alston, D.M.D., P.C., 974 F.2d 1206, 1212 (9th Cir. 1992). Thus, particularly where the district court's decision rests on “discretionary assessments of the balance of the evidence,” this Court “owe[s] the great deference commanded by the traditional power of trial judges sitting as ‘thirteenth jurors,’ to avoid possible miscarriages of justice by ordering new trials in criminal cases.” United States v. Wolff, No. 89-5551, 1989 WL 152513, at *8 (4th Cir. Dec. 12, 1989) (orally argued, but unpublished, table decision). And when we do affirm a new-trial grant, doing so “does not usurp the jury's function” because holding a new trial still “leaves the final decision in the hands of the jury.” Alston, 974 F.2d at 1212.
In applying the abuse-of-discretion standard here, we are mindful that a district court's power to grant a new trial was “effectively absolute” until a 1984 amendment to 18 U.S.C. § 3731 gave the government the right to appeal new-trial orders. Wolff, 1989 WL 152513, at *8 n.7; Act of Oct. 12, 1984, Pub. L. No. 98-473, Title II, § 1206, 98 Stat. 1837, 2153. And nothing about that amendment suggests that it was intended to significantly curb the district court's historically unreviewable discretion in ordering new trials. Wolff, 1989 WL 152513, at *8 n.7.
Still, although our review must be “highly deferential in view of the wide discretion accorded district courts by [Rule 33],” id. at *8, we will nevertheless find that the district court has abused that discretion if it acted “arbitrarily,” if it failed to adequately “ ‘take into account judicially recognized factors constraining its exercise’ of discretion,” or if it rested its decision on “erroneous factual or legal premises,” Rafiekian I, 991 F.3d at 549 (quoting United States v. Alvarado, 840 F.3d 184, 189 (4th Cir. 2016)). So we must assess whether the district court adhered to the demanding standard for granting a new trial.
With that background established, we turn to the key issue in the present case: whether the district court properly weighed the evidence, recognizing the constraints on its discretion, in determining that a new trial was warranted. The government says that it did not. Instead, the government asserts that the district court erred by reaching its new-trial determination based solely on its disagreement with the jury's inferences of guilt. The government notes that this Court has not previously addressed whether a district court may grant a new trial based solely on the court's disagreement with inferences made by the jury. And the government argues that we should hold that a district court may not do so. Instead, the government suggests, the district court must find “specific defects or indicia of unreliability in the trial evidence” before granting a new trial. Opening Br. at 32 n. 10.
Although the government asserts that the law among other circuits is “uniform and unequivocal” in declaring that a court cannot grant a new trial based solely on disagreement with the jury's inferences, United States v. Rafiekian, No. 22-4252, at 1:20 (4th Cir. Jan. 25, 2023), https://www.ca4.uscourts.gov/OAarchive/mp3/22-4252-20230125.mp3, we do not view the other circuits as being so unanimous, or so categorical.
It is true that some of our sister circuits have suggested—consistent with the government's view here—that a serious vulnerability in the evidence must exist to warrant a new trial based on the weight of the evidence. See, e.g., United States v. Burks, 974 F.3d 622, 628 (6th Cir. 2020) (“When it comes to testimony, that means trials featuring accounts that ‘def[y] physical realities,’ or collapse in on themselves due to ‘internal inconsistencies ․’ ” (first quoting United States v. Cote, 544 F.3d 88, 102 (2d Cir. 2008); and then quoting United States v. Lewis, 521 F. App'x 530, 541 (6th Cir. 2013))); United States v. Archer, 977 F.3d 181, 188 (2d Cir. 2020) (finding a new trial to be warranted when testimony is “patently incredible” (quoting United States v. Ferguson, 246 F.3d 129, 134 (2d Cir. 2001))); United States v. Witt, 43 F.4th 1188, 1194 (11th Cir. 2022) (allowing new trials based on the weight of the evidence “only in the rare case in which the evidence of guilt” is “marked by uncertainties and discrepancies” (cleaned up)). Some courts have also concluded that the mere fact that a district court would draw a different inference from a particular piece of evidence than the jury did is not enough to merit granting a new trial. See United States v. Crittenden, 46 F.4th 292, 298–99 (5th Cir. 2022) (en banc); Burks, 974 F.3d at 625; United States v. Landesman, 17 F.4th 298, 333 (2d Cir. 2021). But even these courts have acknowledged that, in the appropriate circumstance, a trial judge can order a new trial when the evidence “weighs heavily against the verdict” and a “miscarriage of justice” would otherwise result, Crittenden, 46 F.4th at 297, and in doing so “may weigh the evidence and credibility of witnesses,” Landesman, 17 F.4th at 330 (quoting Cote, 544 F.3d at 101).
More importantly, this Court has already determined that disagreement with the jury's inferences regarding the evidence can support the district court's decision to grant a new trial. In United States v. Campbell, the district court had “discounted” a certain witness's testimony that conflicted with her own prior statements and had “drawn [other] inferences from the evidence” that undermined the government's case. 977 F.2d at 859. We found that to be inappropriate in the judgment-of-acquittal context and therefore reversed the judgment of acquittal. Id. But we upheld the court's grant of a new trial, making clear that inferences counter to the government's view of the evidence were perfectly permissible in that context. We explained that the district court's “primary error” in granting the judgment of acquittal “was to draw inferences, unfavorable to the Government, from the evidence”—but that “[i]n determining the necessity of a new trial, such inferences are allowed.” Id. at 860 (emphasis added). Thus, even though the jury itself had apparently drawn inferences favorable to the government, the court was free to draw different inferences in making its new-trial determination.
Here, the government implicitly conceded at oral argument that a new-trial order may be based at least in part on the court's disagreement with the jury's inferences regarding a witness's credibility. The problem, according to the government, arises when the new-trial order relies solely on the court's disagreement with the jury's inferences.
We disagree. The key question in determining whether a new trial is warranted under Rule 33 is not what kinds of evidence support the verdict, but the weight of that evidence.
To start, a motion for a new trial based on the “weight of the evidence” invites the court to do precisely what the district court did here: evaluate the persuasiveness of the inculpatory evidence “in comparison with other evidence,” Weight of the Evidence, Black's Law Dictionary (11th ed. 2019) (emphasis added), rather than simply asking whether each individual piece of evidence has an identifiable defect. This exercise is also “far different” from simply determining whether there is sufficient evidence to support the verdict, as is required in the judgment-of-acquittal context. Tibbs, 457 U.S. at 38 n.11, 102 S.Ct. 2211 (quoting United States v. Lincoln, 630 F.2d 1313, 1319 (8th Cir. 1980)). Instead, the court examines the evidence as a whole to determine whether it “preponderates sufficiently heavily against the verdict that a serious miscarriage of justice may have occurred.” Id. (quoting Lincoln, 630 F.2d at 1319). Such a global assessment of the evidence requires the court to weigh whatever evidence it has before it.
And although in some cases, the government will have supplied direct evidence that leads inevitably to the conclusion that the defendant is guilty—the proverbial “smoking gun”—in other cases, the evidence will only suggest guilt by inviting the jury to draw inferences based on the complete picture. In those cases, prohibiting the court from granting a new trial based solely on disagreement with the jury's inferences would make little sense because the entire case rests on such inferences.
As the district court noted here, certain charges—like conspiracy—are often proved almost entirely through circumstantial evidence: “[b]y its very nature, a conspiracy is clandestine and covert, thereby frequently resulting in little direct evidence” of the illicit agreement. J.A. 2016 (quoting United States v. Burgos, 94 F.3d 849, 857 (4th Cir. 1996) (en banc)). In those cases, the jury's verdict is not likely to rest on competing eyewitness accounts of incriminating incidents or dueling recollections as to where, when, and how a crime allegedly occurred. Instead, the government must weave a series of facially innocuous threads into a tapestry of unlawful conduct by inviting the jury to infer guilt.
This was one such case. The government highlighted arguably suspicious—but not outright incriminating—details: the murky line between Project Truth (for which Turkey was unambiguously the would-be client) and Operation Confidence; the odd payment arrangement between FIG and Alptekin; the New York meeting with Turkish officials; and the fact that Operation Confidence's aims aligned with Turkey's stated wishes. It invited the jury to infer that Turkey was funding the whole arrangement, secretly pulling Rafiekian's strings. We do not suggest that the government's reliance on circumstantial evidence was improper; indeed, in the appropriate case, the government may rely entirely on circumstantial evidence to prove the charges. Burgos, 94 F.3d at 858 (“[A] conspiracy may be proved wholly by circumstantial evidence.”); see also Rafiekian I, 991 F.3d at 545 (“[I]f the prosecution is to prove that a defendant acted as an ‘agent of a foreign government,’ it may need to rely on circumstantial evidence and reasonable inferences to make its case—as it is entitled to do.”).
But because the government's case relied on the jury's drawing inferences of guilt, the district court had no choice but to examine those inferences in considering the new-trial motion. Barring the district court from granting a new trial based solely on disagreement with the jury's inferences of guilt would place this class of cases beyond the reach of the new-trial standard, which would mean that when the government has introduced less direct evidence, district courts are more constrained in their ability to grant a new trial. That can't be right. The government is entitled to rely on circumstantial evidence, but it is not entitled to special deference when it does so.
The “ultimate test” of whether a new trial is warranted is “whether letting a guilty verdict stand would be a manifest injustice.” Ferguson, 246 F.3d at 134. And a guilty verdict can effect a manifest injustice regardless of the nature of the evidence supporting it.
In this case, because the district court determined that a new trial was warranted based on the weight of the evidence, our role is only to ask whether the court abused its discretion in doing so. Exercising “great deference” to the district court's “discretionary assessments of the balance of the evidence,” Wolff, 1989 WL 152513, at *8, we hold that it did not.
The court explicitly recognized the “limits on [its] discretion in granting a new trial,” J.A. 2013, noting that new trials should be granted only “sparingly” and that the standard for granting a new trial is “demanding,” J.A. 2016 (quoting Millender, 970 F.3d at 531–32). Still, at every turn, the court concluded that the exculpatory inferences were more persuasive than the inferences of guilt—and powerfully so. It explained that all of the key evidence in the case “points to Rafiekian's innocence,” while the convictions relied on “weak inferences, many built upon one another, drawn from narrowly framed circumstantial evidence, without regard to a broader context that substantially undercuts any inculpatory inferences.” J.A. 2059–60.
This conclusion was well supported by the court's explanation of its own reasonable inferences drawn from the evidence. We briefly summarize this explanation as to each count.
We turn first to Count Two, which charged Rafiekian with acting as an agent of a foreign government without notifying the Attorney General, in violation of 18 U.S.C. § 951. In considering Count Two, while the district court made clear that it was considering the evidence as a whole, it also explained why each inference of guilt that could have supported the jury's verdict was outweighed by a competing, more persuasive inference. The court explained how the communications between Rafiekian and Alptekin—many of which were admitted not for their truth but only to show their effect on Rafiekian—painted Alptekin as a mere “wannabe emissary,” J.A. 2030 (quoting Rafiekian I, 991 F.3d at 540), rather than as an intermediary through which Rafiekian received marching orders from the Turkish government. In support of this conclusion, the court pointed out that the language of the agreement between FIG and Inovo was facially non-incriminating and that the direct evidence in the case showed only that Inovo—not Turkey—was the client for Operation Confidence.
The court also rejected any inference of guilt based on the supposed “rebranding” of Project Truth to Operation Confidence. It concluded that, even if FIG renamed and repackaged Project Truth as Operation Confidence, that move was consistent with the possibility that FIG simply renamed the project when it became clear that Turkey would not be the client. And although the government argued that investigating Gulen was necessarily central to Project Truth and had “nothing to do with Project Confidence's purported purpose of restoring confidence in the Turkish trade and investment climate,” in fact, “the evidence was quite the opposite.” J.A. 2032–33. Indeed, “everyone on the FIG team” understood that discrediting Gulen was key to the objective of restoring confidence in the Turkish investment climate. J.A. 2033. The mere fact that discrediting Gulen was also consistent with the Turkish government's wishes is not enough to show guilt under § 951, as explained above.
Further, the district court explained that there was little evidence supporting the inference that Turkish officials delivered directions to Rafiekian through the September 19, 2016, meeting in New York. The court pointed out that “FIG's engagement was not mentioned or acknowledged” at the meeting, that “no one on the Turkish side made any request of any one on the FIG side to do anything,” and that there was “no evidence that any attending Turkish official ever directly spoke to Rafiekian.” J.A. 2041. And, the court pointed out, Rafiekian had expressed views regarding Gulen consistent with the publicly stated views of Turkey “[l]ong before” the September 19 meeting. J.A. 2042. The court also noted that there was no evidence of contact between Rafiekian and Turkey outside of this meeting.
The district court also highlighted several pieces of evidence indicating that FIG pushed back on requests from Alptekin to undertake certain services, such as getting “dirt” on Gulen and initiating a financial fraud investigation into him, including requests made following up on the New York meeting. J.A. 2039. The court concluded that this evidence contradicted any inference that FIG—and Rafiekian himself—were taking orders from Turkey through Alptekin.
As to contacts between Alptekin and FIG, the district court concluded that FIG's periodic updates to Alptekin were “hardly surprising or suspicious” given that he was the client on a commercial-services contract. J.A. 2038. And the court similarly concluded that the payments to Alptekin did not allow “much of an inference, if any, that Rafiekian had acted or agreed to act subject to the direction or control of the Turkish government.” J.A. 2045.
Summarizing its view of the evidence related to Count Two, the court concluded that the agreement between FIG and Inovo would have been “an ill-suited mechanism through which to install Rafiekian as a Turkish agent subject to its direction or control” because it was “far from an open-ended, ill-defined arrangement that would allow Turkey, acting through Alptekin, to direct and control what work would be performed.” Id. “Overall,” the court explained, “the great evidentiary weight is that Rafiekian did not agree to act, or would not have understood that Alptekin was proposing that he act, or that he would be seen as having agreed to act, subject to the direction or control of Turkey.” J.A. 2048.
The district court did not abuse its discretion in ordering a new trial on Count Two. We recur to the standard we outlined in Rafiekian I for determining when an individual operates subject to the “direction or control” of a foreign government: there must be “a foreign government or official on one end of the line,” and the alleged agent must do more than “simply ․ act[ ] in accordance with foreign interests” or “privately pledg[e] allegiance.” 991 F.3d at 540–41. Instead, he must agree with that foreign government to act in furtherance of its interests. For all of the reasons just outlined, the district court reasonably concluded that “any inference that Turkey was, in fact, the undisclosed client, giving directions to Rafiekian through Alptekin,” was heavily outweighed by competing inferences. J.A. 2040. Recognizing the “highly deferential” standard of review appropriate here, Wolff, 1989 WL 152513, *8, we conclude that the district court did not abuse its discretion in determining that a new trial was warranted on Count Two.
Count One charged Rafiekian with criminal conspiracy in violation of 18 U.S.C. § 371. In concluding that the verdict as to Count One ran counter to the weight of the evidence, the district court's overarching conclusion was that the evidence did not show an agreement between Rafiekian and anyone else as to either object of the alleged conspiracy.
The court concluded that Rafiekian's desire for confidentiality did not provide much support for an inference of conspiracy; the more persuasive inference, the court concluded, was that FIG was concerned about keeping things quiet because it feared publicity would make its efforts more challenging by alerting individuals who supported Gulen to FIG's efforts to paint him in a negative light. The court also found it notable that there was very little evidence of communications between Rafiekian and Alptekin following DOJ's inquiry letter.5
Regarding the first object of the conspiracy—conspiring to act as an unregistered foreign agent—the court pointed to its conclusion that the evidence did not support Rafiekian's conviction for acting as an unregistered foreign agent under Count Two and explained that substantially the same reasoning indicated there was only a “weak inference of any conspiratorial agreement to operate subject to Turkish direction or control.” J.A. 2050 (emphasis added). The court explained that any inference of conspiracy to operate subject to Turkey's control was “heavily outweighed by the evidence that Rafiekian did not contemplate any relationship other than an arms-length, highly defined and focused engagement ․ with no outside direction or control.” Id.
Additionally, the court noted that there was undisputed evidence that Turkey “had a long history of registering openly under FARA with respect to its use of consultants, such as FIG, including with respect to contracts related to Gulen.” J.A. 2052. It found this fact to undermine any inference that Turkey would want Rafiekian to secretly act as its agent. Instead, the court found the stronger inference to be that Rafiekian merely acted in parallel with Turkey's goals, given that he had previously expressed sentiments that aligned with Turkey's publicly stated views.
Further, the court found it “hardly surprising or probative of an illicit conspiracy that there is no mention of Turkey” in the agreements between FIG and Inovo—this absence is equally consistent with the conclusion that Inovo, not Turkey, was in fact FIG's client. J.A. 2053.
Regarding the second object of the alleged conspiracy—conspiring to file a materially false or misleading FARA registration—the district court listed the ways that the evidence more persuasively supported contrary inferences. The court pointed to the role of a law firm, Covington—including its investigation and its control over the FARA filing—as undercutting any inference that Rafiekian participated in a conspiracy to make a false filing. It also pointed to the fact that Rafiekian sought the advice of Robert Kelley, an attorney; answered Kelley's questions about the engagement; and permitted Kelley to make a Lobbying Disclosure Act filing consistent with Kelley's understanding of the engagement as undercutting any inference that Rafiekian was conspiring to obscure his true efforts. Finally, the court did not find Rafiekian's own statements—some of which the government characterized as misleading—to provide much support for an inference of conspiracy, particularly in light of Rafiekian's other “extensive disclosures” to Covington in connection with its investigation. J.A. 2056.
The district court did not abuse its discretion in ordering a new trial on Count One. Proving the conspiracy charge required the government to prove that Rafiekian and at least one other person agreed to break the law. For the reasons stated above, the district court found the evidence that Rafiekian conspired with anyone in this way to be “weak.” J.A. 2050. Because the court found that the inferences supporting Rafiekian's conviction for acting as an unregistered foreign agent were outweighed by innocent inferences, the court similarly found the government's narrative that Rafiekian conspired to act as an unregistered foreign agent to be outweighed by competing explanations. And the court further found that the evidence of Rafiekian's attempts to properly register FIG's activities, among other evidence, weighed heavily against the government's narrative that Rafiekian conspired with others to hide those activities. Particularly where, as here, a district court's new-trial order rests on “discretionary assessments of the balance of the evidence,” Wolff, 1989 WL 152513, at *8, we defer to the district court's determination.
* * *
Having concluded that it would be “unjust to enter judgment of conviction” for the reasons outlined above, J.A. 2048, the district court did not abuse its broad discretion in granting a new trial.6
To be sure, as we explained in Rafiekian I, there was evidence supporting the jury's verdict. Giving the government “the benefit of all reasonable inferences,” we concluded then that the evidence was sufficient to render a judgment of acquittal inappropriate. Rafiekian I, 991 F.3d at 544 (quoting United States v. Savage, 885 F.3d 212, 219–20 (4th Cir. 2018)); see id. at 549.
But here, our role is simply to ask whether the district court abused its discretion in determining that the evidence weighed heavily against the verdict such that a new trial was warranted. And, unbound by the requirement that we make inferences in favor of the government, we conclude that the court did not abuse that discretion.
Based on the foregoing, we affirm the judgment of the district court.
The government charged Bijan Rafiekian with acting covertly as an agent of Turkey in seeking to discredit the public image of a well-known Turkish dissident in the United States, Fethullah Gulen, with the ultimate goal of having Gulen extradited to Turkey, in violation of 18 U.S.C. § 951 (providing that anyone who acts as a foreign agent without giving the Attorney General prior notification shall be punished with up to 10 years' imprisonment). The government also charged Rafiekian with conspiracy to violate § 951 and to make false statements.
Following a week-long trial, a jury convicted Rafiekian on all counts. Nonetheless, the district court granted Rafiekian's motion for judgment of acquittal, ruling that the evidence was insufficient to convict him. Conditionally, the court also granted Rafiekian's motion for a new trial, ruling that the jury verdict was against the great weight of the evidence, among other things.
On the government's appeal, we reversed the judgment of acquittal, vacated the conditional grant of a new trial, and remanded. See United States v. Rafiekian, 991 F.3d 529, 533, 552 (4th Cir. 2021). After a lengthy recital of the evidence that the government introduced against Rafiekian, we concluded that the evidence was sufficient for a rational jury to have convicted him — indeed, that there was “considerable evidence” to do so. Id. at 544–49. And as to the conditional new trial order, we vacated the order, explaining that the district court, in finding the verdict against the great weight of evidence, “simply pointed back to its acquittal analysis without further elaboration” and thereby failed to recognize that the standards for granting a motion for acquittal and granting a new trial were distinct. Id. at 549–50. We remanded the case to the district court to give it an opportunity to elaborate on its new trial order. Id. at 552.
On remand, the district court again granted a new trial, finding mainly that the verdict was against the great weight of evidence. The majority now affirms, concluding that the district court did not abuse its discretion.
I come to a different conclusion because the district court, in its expanded analysis on the motion for a new trial, essentially conducted the same type of analysis that it did in granting the motion to acquit, albeit couching its discussion in the standard for a new trial. At bottom, the district court simply disagreed with the jury and concluded that the government's evidence was insufficient. In doing so, the district court, in my judgment, failed to give any weight to some especially material evidence presented by the government, thereby avoiding the consequences of the inferences to be drawn therefrom. For instance, the court simply recited the text of an August 25, 2016 Skype message, which appeared to refer to a meeting with Turkey's President Erdogan and instructions from the Turkish Foreign Minister, without noting that this was an especially important document that laid the foundation of culpability for much of what came thereafter. Based on my reading of the district court's opinion, the court appeared simply to disagree with the jury despite the existence of strong circumstantial evidence supporting the jury's verdict.
It is a fundamental policy to yield to juries and their roles in assessing and weighing evidence and finding facts. While district judges oversee the reception of evidence and make rulings of law, they are given only a very narrow role when disagreeing with a jury's factfinding. We have stated that a new trial should be granted only “[w]hen the evidence weighs so heavily against the verdict that it would be unjust to enter judgment.” United States v. Arrington, 757 F.2d 1484, 1485–86 (4th Cir. 1985). For this reason, we have admonished that new trials where juries return verdicts are “disfavored.” United States v. Chavez, 894 F.3d 593, 607 (4th Cir. 2018). And the standard is demanding, requiring that it be clear that the jury reached a seriously erroneous result such that there was a “serious miscarriage of justice.” Tibbs v. Florida, 457 U.S. 31, 38 n.11, 102 S.Ct. 2211, 72 L.Ed.2d 652 (1982) (emphasis added) (quoting United States v. Lincoln, 630 F.2d 1313, 1319 (8th Cir. 1980)). Moreover, this demanding standard must be most carefully applied because it would be inappropriate for a district judge to set aside a jury verdict simply because the judge felt that some other result would be more reasonable.
I do not believe that the new trial standard was satisfied here. Taking into account the district court's order granting acquittal, its first order granting a new trial based on its earlier acquittal order, and now its second order granting a new trial, the district court revealed essentially that from the beginning, it disagreed with the jury, even though the jury had more than ample evidence from which to find the defendant guilty. Granting the new trial now is, I believe, an abuse of discretion. Accordingly, I would reverse, remand, and direct that the district court reenter the jury's verdict convicting Rafiekian.
1. Citations to the “J.A.” refer to the Joint Appendix filed by the parties in this appeal.
2. The court also concluded that the jury instructions were flawed in three separate ways. Rafiekian, 2019 WL 4647254, at *16–17. We rejected that reasoning, Rafiekian I, 991 F.3d at 550–52, and the district court did not rely on it on remand.
3. We note that the focus of the Supreme Court's decision in Tibbs was whether retrial following a court's determination that a guilty verdict was not supported by the weight of the evidence violated the Double Jeopardy Clause—an issue not before us in this case. Still, the Court's description of the inquiry that a court faces when assessing whether a new trial is warranted based on the weight of the evidence is instructive here.
4. Further illustrating the differences between these two standards, in some instances, this Court has reversed a judgment of acquittal but upheld a new-trial order as to the same underlying conviction. See, e.g., United States v. Campbell, 977 F.2d 854, 860 (4th Cir. 1992); United States v. Wolff, No. 89-5551, 1989 WL 152513 (4th Cir. Dec. 12, 1989) (orally argued, but unpublished, table decision).
5. As the court noted, on January 18, 2017, Alptekin emailed Rafiekian a copy of an opinion letter he obtained from his own lawyer.
6. The government also challenges the district court's determination that a new trial was warranted in part because of certain statements the prosecutor made in closing arguments regarding a summary exhibit of classified information submitted as part of the defense's evidence. We need not reach this issue. The district court adequately supported its determination that a new trial was warranted based on the weight of the evidence, even without taking into consideration its reasoning on this exhibit.
WYNN, Circuit Judge:
Affirmed by published opinion. Judge Wynn wrote the opinion, in which Judge Bredar joined. Judge Niemeyer wrote a dissenting opinion.
Was this helpful?