Richard J. Sullivan, Circuit Judge

SARAH K. EDDY, Assistant United States Attorney (Ian McGinley, Joshua A. Naftalis, Won S. Shin, Assistant United States Attorneys, on the brief), for Geoffrey S. Berman, United States Attorney for the Southern District of New York, New York, NY, for Appellee United States of America.

DONALD B. VERILLI, JR. (Elaine J. Goldenberg, Jonathan S. Meltzer, on the brief), Munger, Tolles & Olson LLP, Washington, D.C., David Esseks, Eugene Ingoglia, Rachel Agress, Alexander Bussey, on the brief, Allen & Overy LLP, New York, NY, for Appellant Robert Olan.

ALEXANDRA A.E. SHAPIRO (Eric S. Onley, on the brief), Shapiro Arato Bach LLP, New York, NY, Dani R. James, on the brief, Kramer Levin Naftalis & Frankel, LLP, New York, NY, for Appellant Theodore Huber.

COLLEEN P. CASSIDY, Federal Defenders of New York, Inc., New York, NY, for Appellant David Blaszczak.

DANIEL M. SULLIVAN (James M. McGuire, on the brief), Holwell Shuster & Goldberg LLP, New York, NY, Stephen Fishbein, John A. Nathanson, on the brief, Shearman & Sterling LLP, New York, NY, for Appellant Christopher Worrall.

Peter Neiman, Wilmer Cutler Pickering Hale and Dore LLP, New York, NY, Lindsay A. Lewis, Joshua L. Dratel, P.C., New York, NY, for Amicus Curiae National Association of Criminal Defense Lawyers.

Kendall Turner, O’Melveny & Myers LLP, Washington, D.C., Anton Metlitsky, O’Melveny & Myers, LLP, New York, NY, for Amici Curiae Law Professors Adam C. Pritchard, Matthew C. Turk, Andrew N. Vollmer, Karen Woody.

Before: Kearse, Droney, and Sullivan, Circuit Judges.

Richard J. Sullivan, Circuit Judge: These consolidated appeals require us to consider whether the federal wire fraud, securities fraud, and conversion statutes, codified at 18 U.S.C. §§ 1343, 1348, and 641, respectively, reach misappropriation of a government agency’s confidential nonpublic information relating to its contemplated rules. Defendants David Blaszczak, Theodore Huber, Robert Olan, and Christopher Worrall were charged with violating these statutes – and with engaging in securities fraud in violation of Section 10(b) of the Securities and Exchange Act, 15 U.S.C. § 78j(b), and SEC Rule 10b-5 ("Title 15 securities fraud") – by misappropriating confidential nonpublic information from the Centers for Medicare & Medicaid Services ("CMS"). The indictment principally alleged that CMS employees, including Worrall, disclosed the agency’s confidential information to Blaszczak, a "political intelligence" consultant for hedge funds, who in turn tipped the information to Huber and Olan, employees of the healthcare-focused hedge fund Deerfield Management Company, L.P. ("Deerfield"), which traded on it. After a one-month trial before the United States District Court for the Southern District of New York (Kaplan, J. ), a jury found Defendants guilty of wire fraud, conversion, and, with the exception of Worrall, Title 18 securities fraud and conspiracy. The jury acquitted Defendants on all counts alleging Title 15 securities fraud.

Defendants now challenge their convictions on various grounds. For the reasons set forth below, we reject these challenges. In doing so, we hold, inter alia , that (1) confidential government information such as the CMS information at issue here may constitute "property" in the hands of the government for purposes of the wire fraud and Title 18 securities fraud statutes, and (2) the "personal-benefit" test established in Dirks v. SEC , 463 U.S. 646, 103 S.Ct. 3255, 77 L.Ed.2d 911 (1983), does not apply to these Title 18 fraud statutes. Because we also discern no prejudicial error with respect to the remaining issues raised on appeal, we affirm the judgments of the district court.

I. BACKGROUND

A. Facts

The jury returned guilty verdicts on counts charging two insider-trading schemes: (1) a scheme relating to Deerfield that involved all defendants to varying degrees, and (2) a scheme relating to another hedge fund investment manager, Visium Asset Management, L.P. ("Visium"), that involved Blaszczak only. We recite the facts pertaining to each of these schemes in turn, construing the evidence at trial underlying the counts of conviction in the light most favorable to the prosecution. See United States v. Kirk Tang Yuk , 885 F.3d 57, 65 (2d Cir. 2018).

1. The Deerfield Scheme

At various times between 2009 and 2014, Olan, Huber, and fellow Deerfield partner Jordan Fogel – a cooperating witness who pleaded guilty and testified at trial – approached Blaszczak for the purpose of obtaining so-called "predecisional" information concerning CMS’s contemplated rules and regulations. The three Deerfield partners knew that Blaszczak, who had worked at CMS before becoming a consultant for hedge funds, enjoyed unique access to the agency’s predecisional information through his inside sources at the agency. Because other consultants did not have access to Blaszczak’s sources, the Deerfield partners counted him as a particularly lucrative fount of illegal market "edge." App’x at 567, 606.

This illegal market edge first paid off for the three Deerfield partners in July 2009, after Blaszczak passed them nonpublic CMS information concerning both the timing and substance of an upcoming proposed CMS rule change that would reduce the reimbursement rate for certain radiation oncology treatments. The Deerfield partners sought to maximize this market edge by trading while "the information wasn’t known to others, and ... wasn’t public." Id. at 593. In late June 2009, Olan, Huber, and Fogel directed Deerfield to enter orders shorting approximately $33 million worth of stock in radiation-device manufacturer Varian Medical Systems ("Varian"), a company that would be hurt by CMS’s proposed rule. Blaszczak’s information was consistent with the proposed rule that CMS ultimately announced on July 1, 2009, and as a result of the Varian trade, Deerfield made $2.76 million in profits.

Deerfield again traded on confidential CMS information obtained from Blaszczak in 2012. This time, Blaszczak obtained the predecisional information at issue from Worrall, a CMS employee who had previously worked with Blaszczak at the agency and remained friends with him after Blaszczak left CMS to become a hedge fund consultant. Blaszczak met Worrall at CMS’s headquarters in Maryland on May 8, 2012; the following day, Blaszczak emailed Fogel to set up a phone call so that he could update him on one of Fogel’s "favorite topics." Id. at 2439. On the call, Blaszczak provided Fogel with predecisional CMS information about additional radiation oncology reimbursement rate changes. Fogel, in turn, shared this information with Huber and Olan, and together the three of them relied on it – in combination with other confidential CMS information that Blaszczak passed them over the next few weeks – in recommending that Deerfield short millions of dollars in the shares of companies that would be hurt by the reimbursement changes. Deerfield earned profits of $2.73 million from trades relating to this radiation oncology rule, which was publicly announced on July 6, 2012.

In February 2013, shortly after Fogel moved to a different group within Deerfield, he reached out to Blaszczak in the hopes of "re-ignit[ing] the Blaszczak-Fogel money printing machine." Supp. App’x at 6. As Fogel testified at trial, the "Blaszczak-Fogel money printing machine" meant that "Blaszczak had a long history of providing [Fogel] and [his] teammates nonpublic information that [they] could trade on, and it was a great asset to get edge for investments." App’x at 581.

Fogel did not have to wait long for the machine to reignite. In June 2013, Blaszczak told Fogel that he expected CMS to propose cutting the reimbursement rate for end-stage renal disease ("ESRD") treatments by 12 percent. Although Blaszczak did not reveal the source of his information to Fogel, the prediction was so specific – and so different from the market consensus – that Fogel believed it came "from a credible source inside of CMS." Id. at 582. Still, Fogel remained anxious about the outlier status of Blaszczak’s prediction and continued to check in with him about his level of certainty. On June 25, 2013, less than a week before CMS announced the ESRD rule, Blaszczak told Fogel that there was "[n]o change in [his] numbers" and that he was "pretty confident" in his information. Id. at 2024. Fogel again took this to mean that Blaszczak obtained the information from a reliable inside source, and further inferred that the public announcement of the proposed rate cut (the timing of which was also nonpublic) was around the corner and thus less likely to change. On the basis of this confidential nonpublic information, Fogel directed Deerfield to enter orders shorting stock in Fresenius Medical Care, a public company that would be hurt by the reimbursement rate cuts. CMS publicly announced the 12 percent rate cut on July 1, 2013, and Deerfield earned approximately $860,000 in profits from the trade.

Blaszczak continued to provide Fogel with predecisional CMS information in advance of CMS’s announcement of the final ESRD rule on November 22, 2013. In particular, Blaszczak informed Fogel that the final ESRD rule would keep the 12 percent rate cut but would be phased in over three to four years. Based on that information, Fogel recommended that Deerfield enter orders to short stock in Fresenius and DaVita Healthcare Partners Inc. Deerfield did so, earning profits of approximately $791,000. Immediately after CMS announced the final ESRD rule, Fogel emailed his colleagues at Deerfield to praise Blaszczak for his ESRD reimbursement predictions: "I told u guys blazcack [sic] is the man. ... [H]e has crushed it on these two rules both times round." Supp. App’x at 10.

2. The Visium Scheme

Around the same time that Blaszczak was tipping confidential CMS information to his contacts at Deerfield, he also provided similar information to Christopher Plaford, a portfolio manager at the hedge fund Visium. After subsequently pleading guilty pursuant to a cooperation agreement, Plaford testified that he used Blaszczak as a political-intelligence consultant from around 2010 to 2013, during which time Blaszczak would provide him with both public and nonpublic information concerning the healthcare industry. Plaford, like the Deerfield partners, especially valued Blaszczak’s nonpublic CMS information due to the market edge it gave him. Indeed, Plaford considered Blaszczak’s CMS information to be "much more accurate" than the information provided by other consultants, since it came "directly from the horse’s mouth," meaning Blaszczak’s friends and former colleagues at CMS. App’x at 750–51.

In May 2013, for example, Blaszczak tipped Plaford that he expected CMS to propose cutting the reimbursement rate for home healthcare coverage by between three and three-and-a-half percent per year between 2014 and 2017. In the ensuing weeks, Plaford arranged phone calls with Blaszczak to discuss the sources of his information and thus his level of certainty, an issue that Plaford did not want to discuss over email "because it was potentially incriminating." Id. at 752. On the phone call, Blaszczak told Plaford that he had a "high conviction" that his information was accurate because he was "interacting directly with his counterparties in CMS [who] were working on the rule, and they were telling him ... [what] the cut would be." Id. Based on Blaszczak’s information, Plaford directed Visium to maintain its short positions for Amedisys Inc. and Gentiva Health Services Inc., and to buy put-options in those companies. Following CMS’s June 27, 2013 announcement of the proposed home healthcare rule, which included a three-and-a-half percent annual rate cut consistent with Blaszczak’s information, Visium earned approximately $330,000 in trading profits.

B. Procedural History

On March 5, 2018, the government filed an eighteen-count superseding indictment in the United States District Court for the Southern District of New York setting forth allegations relating to the Deerfield scheme (Counts One through Sixteen) and Visium scheme (Counts Seventeen and Eighteen). Counts One and Two charged Defendants with participating in conspiracies centering on the misappropriation of confidential CMS information between 2009 and 2014. In Counts Three through Ten, the indictment charged Defendants with conversion of U.S. property (Count Three), Title 15 securities fraud (Counts Four through Eight), wire fraud (Count Nine), and Title 18 securities fraud (Count Ten), relating to the misappropriation of confidential CMS information that pertained to the July 2012 proposed radiation oncology rule. Counts Eleven and Twelve charged Blaszczak and Worrall with conversion of U.S. property (Count Eleven) and wire fraud (Count Twelve) for allegedly misappropriating confidential CMS information relating to a company called NxStage Medical Inc. The remaining four Deerfield-related counts charged Blaszczak and Worrall with conversion of U.S. property (Count Thirteen), Title 15 securities fraud (Count Fourteen), wire fraud (Count Fifteen), and Title 18 securities fraud (Count Sixteen), based on the misappropriation of confidential CMS information concerning the 2013 proposed and final ESRD rules. Counts Seventeen and Eighteen charged Blaszczak alone with conspiracy and conversion of U.S. property, respectively, for providing confidential CMS information to Plaford as part of the Visium scheme.

On April 2, 2018, the case proceeded to a jury trial before Judge Kaplan. The parties rested their cases three weeks later, on April 23, 2018, and after summations, the district court charged the jury.

In particular, the district court instructed the jury pursuant to Dirks that, (1) in order to convict Worrall of Title 15 securities fraud, it needed to find that he tipped confidential CMS information in exchange for a "personal benefit;" (2) in order to convict Blaszczak of Title 15 securities fraud, it additionally needed to find that he knew that Worrall disclosed the information in exchange for a personal benefit; and (3) in order to convict Huber or Olan of Title 15 securities fraud, it needed to find that Huber or Olan knew that a CMS insider tipped the information in exchange for a personal benefit. App’x at 1042–43. The district court, however, refused to give Dirks -style instructions on the wire fraud and Title 18 securities fraud counts. The district court instead instructed the jury that wire fraud "includes the act of embezzlement, which is ... the fraudulent appropriation to one’s own use of the money or property entrusted to one’s care by someone else." Id. at 1044–45; see Carpenter v. United States , 484 U.S. 19, 27, 108 S.Ct. 316, 98 L.Ed.2d 275 (1987). The district court similarly instructed the jury, for the Title 18 securities fraud counts, that it could find the existence of a scheme to defraud if a defendant "participated in a scheme to embezzle or convert confidential information from CMS by wrongfully taking that information and transferring it to his own use or the use of someone else." App’x at 1045. For both Title 18 fraud offenses, the district court further instructed the jury that it could only convict if it found that the defendant it was considering knowingly and willfully participated in the fraudulent scheme.

On May 3, 2018, after four days of deliberations, the jury returned a split verdict. The jury acquitted all defendants on the Title 15 securities fraud counts; Blaszczak and Worrall on the offenses charged in Counts Eleven and Twelve relating to the NxStage information; and Worrall on the conspiracies charged in Counts One and Two and the substantive offenses charged in Counts Thirteen through Sixteen. The jury nevertheless found all defendants guilty of the conversion and wire fraud offenses charged in Counts Three and Nine, respectively; all defendants but Worrall guilty of the conspiracy offenses charged in Counts One and Two as well as Title 18 securities fraud as charged in Count Ten; and Blaszczak alone guilty of the offenses charged in Counts Thirteen and Fifteen through Eighteen.

On September 13, 2018, the district court denied from the bench Defendants’ post-trial motions for a new trial and/or judgment of acquittal and proceeded to sentencing. The district court sentenced Blaszczak to twelve months and one day of imprisonment, Worrall to twenty months’ imprisonment, and Huber and Olan each to thirty-six months’ imprisonment and fines of $1,250,000. The district court also ordered Blaszczak to forfeit $727,500, Huber to forfeit $87,078, and Olan to forfeit $98,244, and ordered joint and several restitution in the amount of $1,644.26 against all defendants to cover the costs that CMS expended on witnesses’ travel in connection with the criminal investigation and trial. Finally, the district court granted all defendants bail pending appeal on the ground that the forthcoming appeal would present novel and substantial questions. See United States v. Randell , 761 F.2d 122, 125 (2d Cir. 1985). Defendants timely appealed.

II. JURISDICTION AND STANDARD OF REVIEW

We have jurisdiction pursuant to 28 U.S.C. § 1291. We review de novo questions of statutory interpretation, challenges to the district court’s jury instructions, and the propriety of joinder. See United States v. Gayle , 342 F.3d 89, 91 (2d Cir. 2003) ; United States v. Sabhnani , 599 F.3d 215, 237 (2d Cir. 2010) ; United States v. Shellef , 507 F.3d 82, 96 (2d Cir. 2007). We also review de novo the sufficiency of the evidence, Sabhnani , 599 F.3d at 241, recognizing, of course, that a defendant raising such a challenge "bears a heavy burden because a reviewing court must consider the evidence ‘in the light most favorable to the prosecution’ and uphold the conviction if ‘any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt,’ " United States v. Aguilar , 585 F.3d 652, 656 (2d Cir. 2009) (quoting Jackson v. Virginia , 443 U.S. 307, 319, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979) ); accord United States v. Harvey , 746 F.3d 87, 89 (2d Cir. 2014). The district court’s evidentiary rulings are reviewed for abuse of discretion. See United States v. Nektalov , 461 F.3d 309, 318 (2d Cir. 2006).

III. DISCUSSION

Defendants challenge their convictions on several grounds. They argue that (1) the confidential CMS information at issue is not "property" in the hands of CMS for purposes of the wire fraud and Title 18 securities fraud statutes; (2) the district court erred by refusing to instruct the jury on the Dirks personal-benefit test as to the Title 18 fraud counts; (3) Defendants’ convictions for converting U.S. property were infected by a series of legal and factual errors; (4) the evidence at trial was insufficient on all counts; (5) Counts Seventeen and Eighteen, charging Blaszczak alone in the Visium scheme, were misjoined with the other counts; and (6) the district court made a variety of evidentiary errors. We address each of these arguments in turn.

A. "Property" under 18 U.S.C. §§ 1343, 1348

Defendants argue that their convictions for fraud under Title 18 must be reversed because there was insufficient evidence to prove that they engaged in a scheme to defraud CMS of "property." 18 U.S.C. §§ 1343, 1348. The gravamen of their argument is that a government agency’s confidential information is not "property" in the hands of the agency under the Supreme Court’s decision in Cleveland v. United States , 531 U.S. 12, 121 S.Ct. 365, 148 L.Ed.2d 221 (2000), because the agency has a "purely regulatory" interest in such information, id. at 22, 121 S.Ct. 365.

As a preliminary matter, the government contends that Defendants failed to preserve the argument that confidential government information is not "property," since Defendants did not object to the district court’s instruction that "confidential government information may be considered to be property" for purposes of Title 18 securities fraud. App’x at 1045; see also id. (instructing the jury, for purposes of the wire fraud counts, that the government was required to prove that a defendant intended to deprive CMS of "something of value – for example, confidential material, non-public information"). But while Defendants did not challenge the pertinent jury instructions in the district court (and have not done so on appeal), Defendants filed a Rule 29(a) motion for a judgment of acquittal on the ground that the evidence at trial was insufficient to establish that CMS’s information was "property" in the hands of the agency. Contrary to the government’s argument, we do not construe Defendants’ Rule 29(a) motion in the district court as raising a claim distinct from their sufficiency claim on appeal; at both stages, Defendants expressly tied their sufficiency claim to the Supreme Court’s decision in Cleveland , thus raising the broader threshold question of whether a government agency’s confidential regulatory information may constitute "property" in the hands of the agency as a general matter. In answering this question, we are not bound by the district court’s jury instruction that "confidential government information may be considered to be property," id. , since "[a] reviewing court’s limited determination on sufficiency review ... does not rest on how the jury was instructed," Musacchio v. United States , ––– U.S. ––––, 136 S. Ct. 709, 715, 193 L.Ed.2d 639 (2016).

Proceeding to the merits, we afford the same meaning to the word "property" in both the wire fraud and Title 18 securities fraud statutes. See S. Rep. No. 107-146, at 20 (2002) (Title 18 securities fraud statute created to be comparable to Title 18 bank and healthcare fraud statutes); Neder v. United States , 527 U.S. 1, 20, 119 S.Ct. 1827, 144 L.Ed.2d 35 (1999) (Title 18 mail, wire, and bank fraud statutes should be analyzed similarly). We may also look to cases interpreting the same word in the mail fraud statute. See, e.g. , Pasquantino v. United States , 544 U.S. 349, 355 n.2, 125 S.Ct. 1766, 161 L.Ed.2d 619 (2005). Under each of these fraud statutes, the word "property" is construed in accordance with its ordinary meaning: "something of value" in the possession of the property holder (in this context, the fraud victim). Pasquantino , 544 U.S. at 355, 125 S.Ct. 1766 (quoting McNally v. United States , 483 U.S. 350, 358, 107 S.Ct. 2875, 97 L.Ed.2d 292 (1987), superseded by statute on other grounds as stated in Skilling v. United States , 561 U.S. 358, 130 S.Ct. 2896, 177 L.Ed.2d 619 (2010) ); see also id. at 356, 125 S.Ct. 1766 (citing Black’s Law Dictionary 1382 (4th ed. 1951) (defining "property" as "extend[ing] to every species of valuable right and interest")). In applying this general notion of property to the facts of this case, in which the fraud victim is a government agency and the claimed property is confidential information regarding contemplated regulatory action, we are guided by two precedents in particular: Carpenter and Cleveland .

In Carpenter , the Supreme Court held that the publication schedule and contents of forthcoming articles in a Wall Street Journal column were the Journal’s "property" because "[t]he Journal had a property right in keeping confidential and making exclusive use" of the information before publication. 484 U.S. at 26, 108 S.Ct. 316. In fact, the Court noted that "[c]onfidential business information ha[d] long been recognized as property." Id. The Court further noted that pre-publication information was "stock in trade, to be gathered at the cost of enterprise, organization, skill, labor, and money, and to be distributed and sold to those who [would] pay money for it." Id. (quoting Int’l News Serv. v. Associated Press , 248 U.S. 215, 236, 39 S.Ct. 68, 63 L.Ed. 211 (1918) ). The Court therefore concluded that a Journal employee fraudulently misappropriated his employer’s "property" in violation of the mail and wire fraud statutes when he knowingly disclosed the Journal’s confidential pre-publication information to a stockbroker who traded on it. Id. at 28, 108 S.Ct. 316.

By contrast, thirteen years later, the Court in Cleveland held that the mail fraud statute did "not reach fraud in obtaining a state or municipal license" to operate video poker machines, holding that "such a license [was] not ‘property’ in the government regulator’s hands." 531 U.S. at 20, 121 S.Ct. 365. The Court reasoned that (1) the licenses themselves had no economic value until they were issued to a private actor, and (2) the state’s right to control the issuance of its licenses "implicated [its] role as sovereign, not as property holder." Id. at 22–24, 121 S.Ct. 365. Thus, the Court concluded that the government’s "theories of property rights ... [both] stray[ed] from traditional concepts of property" and invited a "sweeping expansion of federal criminal jurisdiction in the absence of a clear statement by Congress." Id. at 24, 121 S.Ct. 365.

While Cleveland remains good law, courts have consistently rejected attempts – similar to those advanced by Defendants here – to apply its holding expansively. See, e.g. , Pasquantino , 544 U.S. at 357, 125 S.Ct. 1766 (" Cleveland is different from this case."); Fountain v. United States , 357 F.3d 250, 256 (2d Cir. 2004) (explaining that Cleveland had only a "modest" effect on the existing legal landscape); United States v. Middendorf , No. 18-cr-36 (JPO), 2018 WL 3443117, at *8–9 (S.D.N.Y. July 17, 2018) (rejecting a Cleveland -based argument similar to the one raised here). As the Supreme Court has clarified, Cleveland simply "held that a [s]tate’s interest in an unissued video poker license was not ‘property,’ because the interest in choosing particular licensees was ‘purely regulatory’ and ‘could not be economic.’ " Pasquantino , 544 U.S. at 357, 125 S.Ct. 1766 (emphasis added) (brackets omitted) (quoting Cleveland , 531 U.S. at 22–23, 121 S.Ct. 365 ). Consistent with this formulation, we have observed that Cleveland ’s "particular selection of factors" did not establish "rigid criteria for defining property but instead ... provid[ed] permissible considerations." Fountain , 357 F.3d at 256. The considerations relied upon by the Court in Cleveland are thus in addition to considerations recognized in other cases, such as the "right to exclude" that was "deemed crucial in defining property" in Carpenter . Id.

Here, we find it most significant that CMS possesses a "right to exclude" that is comparable to the proprietary right recognized in Carpenter . Like the private news company in Carpenter , CMS has a "property right in keeping confidential and making exclusive use" of its nonpublic predecisional information. Carpenter , 484 U.S. at 26, 108 S.Ct. 316. In stark contrast to a state’s right to issue or deny a poker license – a "paradigmatic exercise[ ] of the [state’s] traditional police powers" – CMS’s right to exclude the public from accessing its confidential predecisional information squarely implicates the government’s role as property holder, not as sovereign. Cleveland , 531 U.S. at 23, 121 S.Ct. 365. This view is consistent with pre- Cleveland decisions from this and other Circuits. See United States v. Girard , 601 F.2d 69, 71 (2d Cir. 1979) (concluding that "the [g]overnment has a property interest in certain of its private records," including the confidential information contained in those records); United States v. Czubinski , 106 F.3d 1069, 1074 (1st Cir. 1997) (holding that the IRS’s confidential taxpayer information "may constitute intangible ‘property’ " under the wire fraud statute (citing Carpenter , 484 U.S. at 26, 108 S.Ct. 316 )).

Furthermore, although we do not read Cleveland as strictly requiring the government’s property interest to be "economic" in nature, the government presented evidence that CMS does have an economic interest in its confidential predecisional information. For example, the evidence at trial established that CMS invests time and resources into generating and maintaining the confidentiality of its nonpublic predecisional information – resources that are devalued when the information is leaked to members of the public. See Carpenter , 484 U.S. at 26, 108 S.Ct. 316 ; see also, e.g. , Middendorf , 2018 WL 3443117, at *9 (concluding that a statutory non-profit’s confidential inspection lists were "certainly something of value to the [non-profit], which invested time and resources into their creation" (internal quotation marks omitted)). Relatedly, the selective leaking of confidential CMS information risks hampering the agency’s decision-making process. Although this risk obviously implicates CMS’s regulatory interests, it also implicates CMS’s economic interest in making efficient use of its limited time and resources. As former CMS Director Dr. Jonathan Blum testified, leaks of confidential information could result in unbalanced lobbying efforts, which would in turn impede the agency’s efficient functioning by making it "more difficult to manage the process flow and to convince [Blum’s] superiors of the right course for the Medicare program." App’x at 467. Leaks may also require the agency to "tighten up" its internal information-sharing processes, again with the result that the agency would become less efficient. Id. at 766; see also EPA v. Mink , 410 U.S. 73, 87, 93 S.Ct. 827, 35 L.Ed.2d 119 (1973) (explaining that Congress enacted the "deliberative process" exemption to the Freedom of Information Act’s disclosure requirements, 5 U.S.C. § 552(b)(5), because the "efficiency of [g]overnment would be greatly hampered if, with respect to legal and policy matters, all [g]overnment agencies were prematurely forced to ‘operate in a fishbowl.’ " (quoting S. Rep. No. 89-813, at 9 (1965))), superseded by statute on other grounds as stated in CIA v. Sims , 471 U.S. 159, 190 n.5, 105 S.Ct. 1881, 85 L.Ed.2d 173 (1985). Despite CMS’s proprietary right to exclude and well-recognized economic interests, Defendants argue that the confidential CMS information at issue in this case was not "property" because there was no evidence at trial to establish that CMS suffered an actual monetary loss. In support of this argument, Defendants mainly rely on a single sentence in this Court’s decision in Fountain : "[ Cleveland ] indicates that, in the context of government regulation, monetary loss presents a critical, perhaps threshold consideration." 357 F.3d at 257. For two reasons, this sentence cannot bear the weight Defendants place on it.

First, Fountain , like Cleveland , was not a case about confidential government information – it simply held that taxes owed to a government may constitute "property" in its hands – and thus we do not believe that Fountain ’s reference to "the context of government regulation" contemplated the circumstances presented here. Second, and more fundamentally, while monetary loss may generally be a useful tool for distinguishing the government’s property interests from its "purely regulatory" interests, Cleveland did not, we emphasize, establish any "rigid criteria for defining property." Id. at 256. Nor do we see any reason to impose a rigid "monetary loss" criterion here. Such a requirement would be at odds with Carpenter , which squarely rejected the argument "that a scheme to defraud requires a monetary loss," and instead found it "sufficient that the Journal ha[d] been deprived of its right to exclusive use of the information" because "exclusivity is an important aspect of confidential business information and most private property for that matter." 484 U.S. at 26–27, 108 S.Ct. 316. Although CMS is not a private entity, Carpenter ’s reasoning applies with equal force, since exclusivity is no less important in the context of confidential government information. See, e.g. , Girard , 601 F.2d at 71 ; see also Pasquantino , 544 U.S. at 356, 125 S.Ct. 1766 ("The fact that the victim of the fraud happens to be the government, rather than a private party, does not lessen the injury."); Middendorf , 2018 WL 3443117, at *8 (explaining that the "reasoning of Carpenter supports the conclusion that confidential information – whether held by the government [or] a private entity ... – is ‘property’ "). It is abundantly clear that government agencies have strong interests – both regulatory and economic – in controlling whether, when, and how to disclose confidential information relating to their contemplated rules. See Mink , 410 U.S. at 87, 93 S.Ct. 827 (recognizing the important "public policy ... of open, frank discussion between subordinate and chief concerning administrative action" (internal quotation marks omitted)); supra pp. 33–34. Although fraudulent interference with these interests may at times result in monetary loss to the fraud victim, nothing in the Title 18 fraud statutes requires that to be so.

In sum, the government’s theory of property rights over a regulatory agency’s confidential predecisional information does not "stray from traditional concepts of property," Cleveland , 531 U.S. at 24, 121 S.Ct. 365, but rather is entirely consistent with them. We therefore hold that, in general, confidential government information may constitute government "property" for purposes of 18 U.S.C. §§ 1343 and 1348, and that here, there was sufficient evidence to establish that the CMS information at issue was "property" in the hands of CMS.

B. Whether Dirks v. SEC applies to 18 U.S.C. §§ 1343 and 1348

Under Dirks , an insider may not be convicted of Title 15 securities fraud unless the government proves that he breached a duty of trust and confidence by disclosing material, nonpublic information in exchange for a "personal benefit." 463 U.S. at 663, 103 S.Ct. 3255. Similarly, a tippee may not be convicted of such fraud unless he utilized the inside information knowing that it had been obtained in breach of the insider’s duty. See United States v. Newman , 773 F.3d 438, 447–49 (2d Cir. 2014), abrogated on other grounds by Salman v. United States , ––– U.S. ––––, 137 S. Ct. 420, 196 L.Ed.2d 351 (2016). Here, Defendants claim that the district court erred by not instructing the jury that Dirks ’s personal-benefit test also applied to the wire fraud and Title 18 securities fraud counts. In essence, Defendants argue that the term "defraud" should be construed to have the same meaning across the Title 18 fraud provisions and Rule 10b-5, so that the elements of insider-trading fraud are the same under each of these provisions. We disagree.

We begin by noting what the Title 18 fraud statutes and Title 15 fraud provisions have in common: their text does not mention a "personal benefit" test. Rather, these provisions prohibit, with certain variations, schemes to "defraud." 18 U.S.C. §§ 1343, 1348(1) ; 17 C.F.R. § 240.10b-5(a) ; see 18 U.S.C. § 1348(2) (prohibiting schemes to obtain certain property "by means of false or fraudulent pretenses"); 15 U.S.C. § 78j(b) (prohibiting the use of any "manipulative or deceptive device"). For each of these provisions, the term "defraud" encompasses the so-called "embezzlement" or "misappropriation" theory of fraud. See United States v. O'Hagan , 521 U.S. 642, 653–54, 117 S.Ct. 2199, 138 L.Ed.2d 724 (1997) (Title 15 securities fraud); Carpenter , 484 U.S. at 27, 108 S.Ct. 316 (mail and wire fraud); see also, e.g. , United States v. Mahaffy , 693 F.3d 113, 123 (2d Cir. 2012) (Title 18 securities fraud). According to this theory, "[t]he concept of ‘fraud’ includes the act of embezzlement, which is ‘the fraudulent appropriation to one’s own use of the money or goods entrusted to one’s care by another.’ " Carpenter , 484 U.S. at 27, 108 S.Ct. 316 (quoting Grin v. Shine , 187 U.S. 181, 189, 23 S.Ct. 98, 47 L.Ed. 130 (1902) ). The undisclosed misappropriation of confidential information, in breach of a fiduciary or similar duty of trust and confidence, "constitutes fraud akin to embezzlement." O'Hagan , 521 U.S. at 654, 117 S.Ct. 2199 ; see also United States v. Chestman , 947 F.2d 551, 566–67, 571 (2d Cir. 1991) (en banc).

While the Title 18 fraud statutes and Title 15 fraud provisions thus share similar text and proscribe similar theories of fraud, these common features have little to do with the personal-benefit test. Rather, the personal-benefit test is a judge-made doctrine premised on the Exchange Act’s statutory purpose. As Dirks explained, in order to protect the free flow of information into the securities markets, Congress enacted the Title 15 fraud provisions with the limited "purpose of ... eliminat[ing] [the] use of inside information for personal advantage ." 463 U.S. at 662, 103 S.Ct. 3255 (emphasis added) (internal quotation marks omitted). Dirks effectuated this purpose by holding that an insider could not breach his fiduciary duties by tipping confidential information unless he did so in exchange for a personal benefit. Id. at 662–64, 103 S.Ct. 3255 ; see also Chestman , 947 F.2d at 581 (Winter, J. , concurring in part and dissenting in part) (observing that whereas the theory of fraud recognized in Carpenter "is derived from the law of theft or embezzlement," the " Dirks rule is derived from securities law, and ... [is] influenced by the need to allow persons to profit from generating information about firms so that the pricing of securities is efficient"); United States v. Pinto-Thomaz , 352 F. Supp. 3d 287, 298 (S.D.N.Y. 2018) (Rakoff, J. ) ("Although [the Dirks personal-benefit test] was novel law, the Court reasoned that this test was consistent with the ‘purpose of the [Title 15] securities laws ... to eliminate use of inside information for personal advantage.’ " (quoting Dirks , 463 U.S. at 662, 103 S.Ct. 3255 )).

But once untethered from the statutory context in which it arose, the personal-benefit test finds no support in the embezzlement theory of fraud recognized in Carpenter . In the context of embezzlement, there is no additional requirement that an insider breach a duty to the owner of the property, since "it is impossible for a person to embezzle the money of another without committing a fraud upon him." Grin , 187 U.S. at 189, 23 S.Ct. 98. Because a breach of duty is thus inherent in Carpenter ’s formulation of embezzlement, there is likewise no additional requirement that the government prove a breach of duty in a specific manner, let alone through evidence that an insider tipped confidential information in exchange for a personal benefit. See O'Hagan , 521 U.S. at 682 n.1, 117 S.Ct. 2199 (Thomas, J. , concurring in the judgment in part and dissenting in part) ("Of course, the ‘use’ to which one puts misappropriated property need not be one designed to bring profit to the misappropriator: Any ‘fraudulent appropriation to one’s own use’ constitutes embezzlement, regardless of what the embezzler chooses to do with the money."); see also United States v. Bryan , 58 F.3d 933, 953 (4th Cir. 1995) ("Those who trade on purloined information but who do not come within the ... definition of ‘insider’ [set forth in Chiarella v. United States , 445 U.S. 222, 100 S.Ct. 1108, 63 L.Ed.2d 348 (1980), and Dirks ] are still almost certain to be subject to criminal liability for federal mail or wire fraud."), abrogated on other grounds by O'Hagan , 521 U.S. 642, 117 S.Ct. 2199. In short, because the personal-benefit test is not grounded in the embezzlement theory of fraud, but rather depends entirely on the purpose of the Exchange Act, we decline to extend Dirks beyond the context of that statute.

Our conclusion is the same for both the wire fraud and Title 18 securities fraud statutes. While it is true that Section 1348 of Title 18, unlike the wire fraud statute, concerns the general subject matter of securities law, Section 1348 and the Exchange Act do not share the same statutory purpose. See United States v. Mills , 850 F.3d 693, 699 (4th Cir. 2017) ("The doctrine of in pari materia is inapplicable when statutes have different purposes."). Indeed, Section 1348 was added to the criminal code by the Sarbanes-Oxley Act of 2002 in large part to overcome the "technical legal requirements" of the Title 15 fraud provisions. S. Rep. No. 107-146, at 6 ; see United States v. Hoskins , 902 F.3d 69, 81 n.5 (2d Cir. 2018) ("As a general matter, we may consider reliable legislative history where, as here, the statute is susceptible to divergent understandings and, equally important, where there exists authoritative legislative history that assists in discerning what Congress actually meant." (internal quotation marks omitted)). In particular, Congress intended for Section 1348 to "supplement the patchwork of existing technical securities law violations with a more general and less technical provision, with elements and intent requirements comparable to current bank fraud and health care fraud statutes." S. Rep. No. 107-146, at 14. Given that Section 1348 was intended to provide prosecutors with a different – and broader – enforcement mechanism to address securities fraud than what had been previously provided in the Title 15 fraud provisions, we decline to graft the Dirks personal-benefit test onto the elements of Title 18 securities fraud.

Finally, Defendants argue that we should extend Dirks beyond the Title 15 fraud provisions because otherwise the government may avoid the personal-benefit test altogether by prosecuting insider-trading fraud with less difficulty under the Title 18 fraud statutes – particularly the Title 18 securities fraud statute, which (unlike the wire fraud statute) does not require proof that wires were used to carry out the fraud. But whatever the force of this argument as a policy matter, we may not rest our interpretation of the Title 18 fraud provisions "on such enforcement policy considerations." O’Hagan , 521 U.S. at 678 n.25, 117 S.Ct. 2199. "The Federal Criminal Code is replete with provisions that criminalize overlapping conduct," and so "[t]he mere fact that two federal criminal statutes criminalize similar conduct says little about the scope of either." Pasquantino , 544 U.S. at 358 n.4, 125 S.Ct. 1766. Congress was certainly authorized to enact a broader securities fraud provision, and it is not the place of courts to check that decision on policy grounds.

Accordingly, we hold that the personal-benefit test does not apply to the wire fraud and Title 18 securities fraud statutes, and thus the district court did not err by refusing to instruct the jury on the personal-benefit test for those offenses.

C. Conversion of U.S. Property

The federal conversion statute proscribes "knowingly convert[ing] to [one’s] use or the use of another ... any ... thing of value of the United States," or "receiv[ing] ... the same with intent to convert it to [one’s] use or gain, knowing it to have been ... converted." 18 U.S.C. § 641. Defendants challenge their convictions under this statute on five grounds. All defendants argue that (1) the evidence was insufficient to establish that they "seriously interfered" with CMS’s ownership of its confidential information, as required to prove conversion, and (2) information is not a "thing of value" for purposes of Section 641. Olan, Huber, and Blaszczak further argue that (3) the conversion statute is unconstitutionally vague as applied to them, and (4) the evidence was insufficient to establish scienter. Finally, Olan and Huber contend that (5) the district court erred in giving a conscious avoidance jury instruction. We address each of these arguments in turn.

1. "Serious Interference"

Defendants first argue that there was insufficient evidence at trial to prove conversion of U.S. property because the government presented no evidence that Defendants interfered, let alone "seriously interfered," with CMS’s ability to use its confidential information in the rulemaking process. Although the government agrees that "serious interference" is required, it responds that "the interference is complete when the [confidential] information is disclosed, and the interference is serious when the government has demonstrated a strong interest in maintaining confidentiality of that species of information." Appellee’s Br. at 109. We disagree with Defendants’ view of how the "serious interference" standard applies when, as here, the property at issue is confidential information. By focusing on the fact that their misappropriation of confidential CMS information did not ultimately affect the rules that CMS subsequently announced, Defendants disregard the Supreme Court’s teaching in Morissette v. United States that conversion under Section 641 extends broadly to the "misuse or abuse of [government] property." 342 U.S. 246, 272, 72 S.Ct. 240, 96 L.Ed. 288 (1952). Moreover, Defendants’ argument overlooks the fact that the unauthorized disclosure of CMS’s confidential nonpublic information by definition interferes with the agency’s right to exclude the public from accessing such information. See Carpenter , 484 U.S. at 26, 108 S.Ct. 316 (rejecting the defendants’ argument that they "did not interfere with the Journal’s use of the [pre-publication] information" as "miss[ing] the point," because it sufficed that the defendants interfered with the Journal’s "right to decide how to use [the information] prior to disclosing it to the public"). Thus, we agree with the government that the relevant "interference" with CMS’s ownership of confidential information was complete upon the unauthorized disclosure.

As for the "seriousness" of the interference, we also reject Defendants’ contention that their misappropriation of confidential CMS information exceeded the reach of the conversion statute simply because CMS was able to keep using the information. Defendants’ argument is inconsistent with the Restatement, which sets forth a multi-factor test for determining the "seriousness of the interference" that lists "the harm done to the [property]" and "the inconvenience and expense caused to the [property owner]" as only two of six non-exhaustive factors, none of which "is always predominant." Restatement (Second) of Torts § 222A(2) & cmt. d (1965) (hereinafter "Restatement"); see also United States v. Collins , 56 F.3d 1416, 1420 (D.C. Cir. 1995) (citing the Restatement to interpret Section 641 ); United States v. May , 625 F.2d 186, 192 (8th Cir. 1980) (same). Moreover, Defendants’ view is also in stark tension with our holding in Girard , where we upheld the defendants’ convictions under Section 641 for engaging in a scheme to sell confidential DEA information that identified the agency’s informants, even though the scheme was unsuccessful and there was no suggestion that the informants were in fact compromised. 601 F.2d at 70, 73 ; see also Morissette , 342 U.S. at 272, 72 S.Ct. 240 (explaining that "merely ... commingling" money may constitute conversion where the custodian is "under a duty to keep it separate and intact").

Thus, while the jury in this case was free to consider the fact that CMS was able to use the misappropriated information and did not suffer any monetary loss, it was also free to consider other factors, including (1) the strength of the government’s interest in maintaining confidentiality, (2) the risk of harm to the government’s interests posed by the unauthorized disclosure, and (3) the extent of the unauthorized disclosure. See Restatement § 222A(2) ; see also, e.g. , Girard , 601 F.2d at 70, 73.

Applying this standard here, we conclude that there was sufficient evidence to support the jury’s finding of serious interference with CMS’s ownership of its confidential information. Dr. Blum testified that "[i]t’s a very strong precedent and a very strong principle that every stakeholder has the right to receive the materials [concerning a rule] at the same time," because the "rule-making process is based upon the notion that the entire public that can be affected ... ha[s] the right to comment" in a manner that is fair to all stakeholders. App’x at 467. The leaking of predecisional information, Dr. Blum explained, could thus tilt the playing field against interest groups (and the public) who were not yet privy to the information, and also prematurely "trigger powerful [lobbying] forces to try and stop decisions." Id. CMS employee Amy Bassano echoed these views in her testimony, while adding that CMS employees were more "wary of what [stakeholders were] going to be sharing" with the agency after predecisional information had leaked. Id . at 767. This increased wariness, combined with the agency’s tightening up of internal information-sharing protocols, "sometimes result[ed] in suboptimal [policy] outcomes." Id. Furthermore, all of these adverse effects harmed CMS economically by making the agency function less efficiently. See supra pp. 33–34.

As for other relevant factors, the jury could reasonably infer that the disclosure of confidential information to a Washington D.C. consultant like Blaszczak – and ultimately to Blaszczak’s clients – seriously risked harming the government’s interests by threatening wider disclosure of the information to interested stakeholders. Indeed, the government presented evidence that Blaszczak tipped confidential information not only to hedge fund partners, who sought to use the information for trading purposes, but also to employees of healthcare companies such as Amgen, a regulated entity that stood to benefit from the very informational asymmetry that the government’s confidentiality rules for predecisional information were designed to prevent. Taken together, this evidence was sufficient to support a finding that Defendants’ misappropriation of CMS’s confidential nonpublic information "seriously interfered" with CMS’s ownership rights for purposes of the conversion statute.

2. "Thing of Value"

Defendants next argue that confidential information is not a "thing of value" within the meaning of the conversion statute. 18 U.S.C. § 641 (emphasis added). But as this Court explained in Girard , "[t]he word ‘thing’ notwithstanding, the phrase is generally construed to cover intangibles as well as tangibles." 601 F.2d at 71 (collecting cases). Thus, "[a]lthough the content of a writing is an intangible, it is nonetheless a thing a value." Id. Contrary to Defendants’ strained reading of the case, we read Girard to hold that confidential information can itself be a "thing of value" under Section 641. Id. ; see also United States v. Matzkin , 14 F.3d 1014, 1021 (4th Cir. 1994) (holding that confidential information was a "thing of value"); United States v. Barger , 931 F.2d 359, 368 (6th Cir. 1991) (citing, inter alia , Girard for the proposition that "information itself is enough to meet the property or ‘thing of value’ element of the statute."). Thus, whatever the merit of Defendants’ textual argument, we are not at liberty to reconsider Girard here. See, e.g. , Deem v. DiMella-Deem , 941 F.3d 618, 623 (2d Cir. 2019) ("[A] published panel decision is binding on future panels unless and until it is overruled by the Court en banc or by the Supreme Court." (internal quotation marks omitted)).

3. Vagueness

Olan, Huber, and Blaszczak further argue that Section 641 is unconstitutionally vague as applied to them because there was no rule or regulation making clear that Worrall’s disclosure of CMS’s confidential information was "without authority." This argument too lacks merit.

"Where, as here, we are not dealing with defendants’ exercise of a first amendment freedom, we should not search for statutory vagueness that did not exist for the defendants themselves." Girard , 601 F.2d at 71 ; see also United States v. Mazurie , 419 U.S. 544, 550, 95 S.Ct. 710, 42 L.Ed.2d 706 (1975) ("[V]agueness challenges to statutes which do not involve First Amendment freedoms must be examined in the light of the facts of the case at hand."). In Girard , we held that "statutory vagueness ... did not exist for the defendants themselves" because the defendants "must have known" that the disclosure of the identity of DEA informants was unauthorized. 601 F.2d at 71. Although we noted that the "DEA’s own rules and regulations forbidding such disclosure" were relevant to the inquiry, id. , we did not, contrary to Defendants’ suggestion, require the existence of a published rule or regulation on point. See United States v. McAusland , 979 F.2d 970, 975 (4th Cir. 1992) ("We do not read [ Girard ] as requiring the disclosure to be specifically proscribed by published regulations."). Nor will we impose such a sweeping extra-textual requirement here. Rather, we agree with the Fourth Circuit that "the existence of a published regulation proscribing disclosure" is not "the exclusive method of preventing vagueness." Id. ; see also, e.g. , id. at 975–76 (rejecting defendants’ as-applied vagueness challenge in light of "legends restricting disclosure" on the converted documents, "[d]efendants’ behavior," and witnesses’ testimony at trial that defendants "would have known that the information was not to be disclosed"); United States v. Jones , 677 F. Supp. 238, 241 (S.D.N.Y. 1988) ("Given the government’s long[-]standing practice of maintaining the confidentiality of information relevant to on-going criminal investigations, and given the government’s obvious interest in maintaining such confidentiality, the defendant could reasonably know the proscribed nature of his alleged actions.").

Here, as in Girard , there was ample evidence at trial to establish that Defendants "must have known" that the disclosure of the predecisional CMS information at issue was prohibited. Although Worrall does not raise a vagueness challenge himself, it bears noting that CMS employees were subject to 5 C.F.R. § 2635.703(a) (the text of which was introduced into evidence at trial), which forbids the "improper use of nonpublic information to further [the employee’s] own private interest or that of another ... by knowing unauthorized disclosure." The regulation further provides that "nonpublic information is information that the employee gains by reason of Federal employment and that he knows or reasonably should know has not been made available to the general public." Id. § 2635.703(b). In addition, CMS employees received extensive training on the rules prohibiting disclosure of nonpublic predecisional information.

As a former employee, Blaszczak was previously subject to these same rules and presumably had also received training on the confidential nature of predecisional information. At trial, moreover, the government’s witnesses consistently testified to the fact that Blaszczak, Olan, and Huber – and consultants and securities traders in the healthcare space more generally – knew that predecisional CMS information was nonpublic and confidential. Indeed, Fogel testified that the Deerfield defendants valued predecisional CMS information precisely because it was not available to other traders. Plaford testified similarly as to his own motivations.

That testimony was corroborated by evidence of Defendants’ own communications and behavior. In one episode in 2012, for example, Olan, Huber, Fogel, and Blaszczak attempted to extract predecisional CMS information from CMS consultant Dr. Niles Rosen, prompting an email discussion of the fact that Rosen was unlikely to disclose such information. Olan commented that he thought the odds of Blaszczak "getting shut down by [R]osen [were] 103%," but nevertheless Blaszczak and the Deerfield partners pushed ahead in the hopes that Blaszczak might get Rosen to "bite[ ]," since he was "the man with the keys to [the radiation-oncology device] companies’ coffins." App’x at 1982, 2428. Ultimately, Rosen rebuffed Blaszczak’s efforts, writing in an email, "As you clearly understand, I cannot share with you our recommendations to CMS." Id. at 2431.

Thus, construing the evidence in the light most favorable to the prosecution, we conclude that there was sufficient evidence to establish that Olan, Huber, and Blaszczak knew that the CMS information at issue was disclosed "without authority." Accordingly, their as-applied vagueness challenge fails.

4. Scienter

Olan and Huber next argue that there was insufficient evidence at trial to establish that they received confidential CMS information "knowing it to have been ... converted," as required by 18 U.S.C. § 641. Blaszczak similarly argues that the evidence was insufficient to prove his "intent to convert [such information] to his use or gain." Id. Again, we disagree, and find that the evidence at trial was sufficient to establish Olan’s and Huber’s knowledge that they received converted property.

Specifically, we reject, for the reasons just mentioned, Olan’s argument (joined by Huber) that the evidence was insufficient to prove his knowledge of unauthorized disclosure. We also reject Olan’s claim that the evidence was insufficient to prove his knowledge of "serious interference" with CMS’s ownership of its confidential information. Despite Olan’s bald assertion that "[t]here was no way for [him] ... to know that disclosure of the information" could affect CMS’s rulemaking process given that he had "never worked for CMS," Olan Br. at 45, fellow Deerfield partner Fogel – who had also never worked for CMS – testified that he understood that disclosure of CMS’s confidential information "had the potential to disrupt CMS’s process," App’x at 564. Indeed, Fogel specifically acknowledged that if CMS’s confidential "information was out there, it would give industry lobbyists and others a chance to ... stop a proposed cut or increase from happening." Id. Most notably, Fogel testified that he "discuss[ed] th[e] impact on the CMS process" with Huber and Olan. Id. This detailed testimony alone was enough to establish Huber’s and Olan’s knowledge of serious interference. See United States v. Hamilton , 334 F.3d 170, 179 (2d Cir. 2003) ("The testimony of a single accomplice is sufficient to sustain a conviction so long as that testimony is not incredible on its face and is capable of establishing guilt beyond a reasonable doubt." (internal quotation marks omitted)).

As to Blaszczak’s sufficiency challenge, there was ample evidence to support a finding that Blaszczak intended to convert the confidential CMS information that he received from CMS insiders to his use or gain. Although Blaszczak argues that there was insufficient evidence to establish that he specifically "intend[ed] [for] his predictions and analyses ... to interfere ... with CMS’s work," Blaszczak Br. at 57, the requisite intent was established by evidence that Blaszczak, himself a former CMS employee, obviously knew that the disclosure of the predecisional CMS information he received was unauthorized and could spawn interference with CMS’s processes, but he nevertheless intentionally proceeded to appropriate such information to his own use by disclosing it to his hedge fund clients. See Morissette , 342 U.S. at 270–72, 72 S.Ct. 240.

5. Conscious Avoidance Instruction

Last, we reject Olan’s and Huber’s claim that the district court erred in giving a "conscious avoidance" instruction. As relevant here, a conscious avoidance instruction may only be given if "the appropriate factual predicate for the charge exists, i.e. the evidence is such that a rational juror may reach the conclusion beyond a reasonable doubt that the defendant was aware of a high probability of the fact in dispute and consciously avoided confirming that fact." United States v. Goffer , 721 F.3d 113, 126–27 (2d Cir. 2013) (internal quotation marks omitted). This standard is easily satisfied here. To repeat, the evidence at trial established that Olan and Huber sought out Blaszczak’s services precisely so they could trade on information that other analysts and consultants did not possess. And as Fogel testified, when Blaszczak gave the Deerfield partners the nonpublic information they sought, he either told them "explicitly" that it came from CMS insiders, or that fact was "implied or obvious" given the context in which the information was conveyed. App’x at 555. In addition, Fogel testified that he, Olan, and Huber specifically discussed the fact that disclosure of CMS’s confidential predecisional information could harm the agency’s regulatory process. In these circumstances, a rational juror could find that, even if Olan and Huber did not have actual knowledge that Blaszczak’s predictions were based on confidential CMS information that had been converted, Olan and Huber were at least aware of a high probability of that fact and yet consciously avoided confirming it.

D. Other Sufficiency Arguments

Blaszczak, joined by Olan and Huber, next argues that at most the evidence established that he passed along information that was already public, or that was disclosed by CMS insiders who had the authority to disclose it. This argument is meritless. The fact that Blaszczak had access to legitimate sources of information that could have supported his predictions hardly compels the conclusion that he in fact relied on those sources, rather than on CMS insiders who disclosed confidential information without authority, as Fogel and Plaford testified. And while Blaszczak makes much of the fact that his predictions were not always accurate, his lack of perfection does not compel an inference that his sources were legitimate and public. As the evidence reflected, there were various reasons why CMS might adjust its position between the time that confidential predecisional information leaked and the time that a rule was publicly announced. Moreover, despite Blaszczak’s imperfect record, his predictions were still more accurate (and valuable) than those of other market consultants. Put simply, Blaszczak invites us to choose "between competing inferences," but this is a fact-finding function that lies "solely within the province of the jury." United States v. Payne , 591 F.3d 46, 60 (2d Cir. 2010).

For similar reasons, we reject Worrall’s argument that the evidence was insufficient to establish that he was the source of leaked CMS information in 2012. Contrary to Worrall’s suggestion, the government was not required to prove the precise way in which he became aware of predecisional information concerning the proposed radiation oncology rule. Rather, the government was entitled to prove Worrall’s knowledge of the information through circumstantial evidence, including evidence that Worrall had access to the information because he worked closely with Blum and his job responsibilities exposed him to various matters within the agency. As to whether Worrall disclosed this information to Blaszczak, the government introduced into evidence a May 8, 2012 CMS sign-in sheet establishing that Blaszczak met Worrall the day before relaying confidential information concerning the proposed radiation oncology rule to Fogel. This evidence was buttressed by testimony from Marc Samuels, Blaszczak’s consulting partner between 2008 and 2012, who recalled that Blaszczak had specifically named Worrall as a source of confidential CMS information. The government also presented evidence that Blaszczak and Worrall remained close in 2013 and 2014; for example, Blaszczak’s research analyst during that period, Timothy Epple, testified that Blaszczak "would reference his friend Chris most often" as his source of nonpublic CMS information. App’x at 872. Epple further testified that, after Blaszczak learned he was under investigation by the SEC, he pointedly asked Worrall whether investigators had been questioning people at CMS. While Worrall argues that Blaszczak could nevertheless have obtained information about the 2012 radiation oncology rule from other people at CMS, the above-referenced evidence was more than sufficient to support the jury’s contrary finding on this point.

Thus, having carefully reviewed the record, we conclude that the evidence at trial was sufficient to support the jury’s verdict on each count of conviction.

E. Misjoinder

Olan and Huber next argue that the district court erred in denying their motion under Federal Rule of Criminal Procedure 8(b) to sever Counts Seventeen and Eighteen, which charged Blaszczak alone in the Visium scheme, from the remaining counts.

Rule 8(b) provides that an indictment "may charge [two] or more defendants 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." Fed. R. Crim. P. 8(b). Under this rule, "joinder of defendants is proper when the alleged acts are ‘unified by some substantial identity of facts or participants, or arise out of a common plan or scheme.’ " United States v. Feyrer , 333 F.3d 110, 114 (2d Cir. 2003) (quoting United States v. Attanasio , 870 F.2d 809, 815 (2d Cir. 1989) ). In administering this standard, we "apply a ‘commonsense rule’ to decide whether, in light of the factual overlap among charges, joint proceedings would produce sufficient efficiencies such that joinder is proper notwithstanding the possibility of prejudice to either or both of the defendants resulting from the joinder." United States v. Rittweger , 524 F.3d 171, 177 (2d Cir. 2008) (Sotomayor, J. ) (quoting Shellef , 507 F.3d at 96 ). Even where joinder is erroneous, we will not reverse unless the "misjoinder results in actual prejudice because it had [a] substantial and injurious effect or influence in determining the jury’s verdict." Shellef , 507 F.3d at 100 (internal quotation marks omitted).

Here, the district court did not err in concluding that the Visium-related charges against Blaszczak were properly joined with the Deerfield-related charges against Blaszczak, Olan, Huber, and Worrall. Although these two sets of charges involved distinct schemes, there was substantial temporal overlap between the Visium scheme (2011 to 2013) and Deerfield scheme (mainly 2012 to 2014); the schemes involved nearly identical conduct, i.e., misappropriation and insider trading of confidential government information concerning healthcare rules; and in both schemes, Blaszczak was the key player and CMS was the victim. These similarities alone were sufficient to render Rule 8(b) joinder both efficient and proper. See Rittweger , 524 F.3d at 177 ; Feyrer , 333 F.3d at 114 ; Attanasio , 870 F.2d at 815.

In any event, even if joinder were improper, any error would be harmless because much of the evidence relating to the Visium scheme would have been admissible against Olan and Huber on Counts One through Sixteen. See Shellef , 507 F.3d at 101–02. The district court correctly determined that Plaford’s testimony, which both corroborated Fogel’s testimony and provided useful background on Blaszczak’s methods and sources during the same time period as the Deerfield conspiracy, was relevant evidence on the charges against Olan and Huber. See Fed. R. Evid. 401 ; see also id. 404(b). While the court also recognized that the probative value of Plaford’s testimony "may [have been] somewhat attenuated" in relation to the Deerfield scheme, the court permissibly concluded that such testimony would not result in any undue prejudice for purposes of Rule 403(b). App’x at 996; see United States v. Awadallah , 436 F.3d 125, 134 (2d Cir. 2006) ("Only rarely – and in extraordinarily compelling circumstances – will we, from the vista of a cold appellate record, reverse a district court’s on-the-spot judgment concerning the relative weighing of probative value and unfair effect." (internal quotation marks omitted)).

F. Evidentiary Issues

Blaszczak, again joined by Olan and Huber, also argues that the district court committed several evidentiary errors warranting a new trial. Specifically, Blaszczak contends that the district court erred by (1) limiting as cumulative the defense’s cross-examination of CMS employee Mark Hartstein concerning the fact that CMS’s 2012 proposed radiation oncology rule was based on published recommendations of the American College for Radiology; (2) precluding cross-examination of Plaford as to a prior inconsistent statement; (3) admitting into evidence statements made by Amgen employee Ruth Hoffman under the coconspirator exclusion set forth in Rule 801(d)(2)(E); and (4) admitting into evidence minutes of a 2007 Deerfield meeting as a business record for the purpose of proving Olan’s and Huber’s states of mind.

Having considered these arguments in the context of the record as a whole, we discern no error warranting a new trial. The district court acted within its discretion in limiting Hartstein’s testimony as to the basis for CMS’s proposed radiation oncology rule, since other evidence had indeed been introduced on this subject and Hartstein’s testimony would have been cumulative. Regarding Plaford’s prior inconsistent statement that the market’s prediction for the home healthcare cuts was 2.5% rather than 3.5% as he recalled at trial, the district court did not err in concluding that Plaford’s recollection as to the actual market consensus was a collateral issue. As for Hoffman’s email statements, the evidence at trial was sufficient to establish Hoffman’s status as an unindicted coconspirator for purposes of Rule 801(d)(2)(E) based on her implied agreement with Blaszczak to misappropriate confidential CMS information. See, e.g. , United States v. Downing , 297 F.3d 52, 57–58 (2d Cir. 2002). Finally, the district court properly admitted the minutes of the 2007 Deerfield meeting – reflecting that someone at the meeting had opined that "Blazacks [sic] comments pre-news suggest he had a read of draft documents," App’x at 2039 – as a business record probative of Olan’s and Huber’s states of mind during the years of the charged conspiracy, see Fed. R. Evid. 803(6), and subject to a clear limiting instruction that such evidence could not be considered against Blaszczak.

We therefore discern no error in the district court’s evidentiary rulings. Moreover, even assuming that one or more of these rulings were erroneous, any errors would fall well short of prejudicial. Over the course of the month-long trial, the government presented various forms of evidence establishing that Blaszczak’s predictions were based on confidential nonpublic CMS information obtained directly from CMS insiders, and that Olan and Huber were aware of that fact when they sought out this information, received it, and directed Deerfield to trade on it.

IV. CONCLUSION

In upholding the jury’s verdict, we pause to reject Defendants’ thematic claim that the government’s positions, if accepted, would herald an unprecedented expansion of federal criminal law. It is Defendants who ask us to break new ground by rejecting well-recognized theories of property rights and by adding, in effect, a "personal benefit" element to the Title 18 fraud statutes. We decline these requests, holding instead that (1) a government agency’s confidential information relating to its contemplated rules may constitute "property" for purposes of the wire fraud and Title 18 securities fraud statutes, and (2) Dirks ’s "personal-benefit" framework does not apply to these Title 18 fraud statutes. Our remaining holdings confirm that Defendants’ misappropriation of CMS’s predecisional information, as proven at trial, fall comfortably within the Title 18 securities fraud, wire fraud, conversion, and conspiracy statutes. To the extent that the government’s decision to prosecute any or all of these crimes in this case raises broader enforcement policy concerns, that is a matter for Congress and the Executive, not the Judiciary. Our inquiry is a more limited one, and having now completed it, we AFFIRM the judgments of the district court.

KEARSE, Circuit Judge, dissenting:

I respectfully dissent from the majority's affirmance of the convictions of these four defendants for substantive crimes of conversion of government property in violation of 18 U.S.C. § 641 and wire fraud in violation of 18 U.S.C. § 1343, as well as the convictions of three of the defendants for substantive crimes of securities fraud in violation of 18 U.S.C. § 1348, for conspiracy in violation of 18 U.S.C. § 1349 to commit Title 18 crimes of wire fraud and securities fraud, and for conspiracy in violation of 18 U.S.C. § 371 to commit offenses under § 641 and other provisions, including Title 15 securities fraud in violation of 15 U.S.C. § 78j(b) and SEC Rule 10b-5 promulgated thereunder.

Section 641, one of the sections under which all four defendants were convicted, provides that it is unlawful to

embezzle[ ], steal[ ], purloin[ ], or knowingly convert [ ] to his use or the use of another, or without authority, sell[ ], convey[ ] or dispose[ ] of any record, voucher, money, or thing of value of the United States or of any department or agency thereof ....

18 U.S.C. § 641 (emphases added). Section 1343, under which all four defendants were also convicted, provides in part that

[w]hoever, having devised or intending to devise any scheme or artifice ... for obtaining money or property by means of false or fraudulent pretenses ... transmits or causes to be transmitted by means of wire ... any writings, signs, signals ... for the purpose of executing such scheme or artifice, shall be fined under this title or imprisoned not more than 20 years, or both.

18 U.S.C. § 1343 (emphases added). Section 1348, under which three defendants were convicted, is similar to § 1343. It provides in part that

[w]hoever knowingly executes, or attempts to execute, a scheme or artifice--

....

(2) to obtain , by means of false or fraudulent pretenses, representations, or promises, any money or property in connection with the purchase or sale of ... any security of an issuer with a class of securities registered under Section 12 of the Securities Exchange Act of 1934 ....

shall be fined under this title or imprisoned not more than 25 years, or both.

18 U.S.C. § 1348(2) (emphases added).

With respect to the issue dividing us, the majority treats the relevant elements of §§ 1343 and 1348 as the same: the property that the defendant is charged with obtaining by false or fraudulent pretenses must be the property of the defrauded victim. While this has been held to be so with respect to the mail fraud statute, 18 U.S.C. § 1341, see , e.g. , Cleveland v. United States , 531 U.S. 12, 15, 121 S.Ct. 365, 148 L.Ed.2d 221 (2000) ("the thing obtained must be property in the hands of the [fraud] victim"), and §§ 1341 and 1343 "share the same language in relevant part" and are subject to the same analysis, Carpenter v. United States , 484 U.S. 19, 25 n.6, 108 S.Ct. 316, 98 L.Ed.2d 275 (1987), it is not entirely clear to me that this is true of § 1348. However, for purposes of this opinion, I accept that both §§ 1343 and 1348 prohibit obtaining property belonging to the victim of the fraud.

My disagreement with the majority is focused on the charges of the operative superseding indictment ("Indictment") that defendants violated §§ 1343 and 1348 by obtaining something that was government "property" and violated § 641 by "converting" something that was a "thing of value" to the government.

The alleged conduct underlying virtually all of these charges was that defendants Blaszczak, Huber, and Olan obtained directly or indirectly from defendant Worrall, an employee of the federal agency Centers for Medicare & Medicaid Services ("CMS"), confidential information as to the substance and timing of upcoming changes to CMS rules governing reimbursement rates for certain medical treatments. CMS is not a business; it does not sell, or offer for sale, a service or a product; it is a regulatory agency. It adopts regulations that affect, inter alia , business organizations or health industry entities--whether the affected persons or entities favor the regulations or not. While CMS seeks to maintain confidentiality as to its planned regulations--and the regulations can plainly have either a favorable or an adverse effect on certain business entities' fortunes--I do not view a planned CMS regulation as a "thing of value" to CMS, 18 U.S.C. § 641, that is susceptible to conversion. Unlike the information that was planned for publication by the news publisher victim in Carpenter , information is not CMS's "stock in trade," 484 U.S. at 26, 108 S.Ct. 316 (internal quotation marks omitted). CMS does not seek buyers or subscribers; it is not in a competition; it is an agency of the government that regulates the conduct of others. It does so whether or not any information on which its regulation is premised is confidential. Further, regardless of whether information as to the substance or timing of a planned regulation remains confidential as CMS prefers or is disclosed to unauthorized listeners, CMS adopts its preferred planned regulation and--subject to legal requirements as to timing, e.g. , 42 U.S.C. § 1395w-4(b)(1) (requiring that reimbursement rates for a given year be announced prior to November 1 of the preceding year)--can do so in accordance with its own timetable. I cannot see that predecisional regulatory information is subject to conversion within the contemplation of § 641.

Although the majority views our decision upholding a § 641 conviction in United States v. Girard , 601 F.2d 69, 71 (2d Cir. 1979), as compelling the conclusion that CMS's desire for predecisional confidentiality is a thing of value, I disagree. Girard involved a drug dealer's attempt to purchase confidential records of the United States Drug Enforcement Administration ("DEA") as to what persons were DEA informants. Confidential information as to the identities of informants and cooperators is clearly "[some]thing of value" to a government agency whose mission is law enforcement. That confidential information has inherent value because it enables the agency to, inter alia , collect evidence upon which the Department of Justice may obtain authorizations to conduct electronic surveillance, obtain warrants for arrests, and commence prosecutions. Confidentiality in that context enhances the value of the information because, inter alia , it reduces the chances that suspects will alter their observable behavior, hide their contraband, flee into hiding, or tamper with--or harm--witnesses before the law enforcement agency has an opportunity to fully act upon the information it possesses.

An agency such as CMS whose brief is to issue regulations is entirely different. It may either carry out or deviate from its planned adoption of regulations even if its plans, and/or the information that affects those plans, become public knowledge before CMS prefers that such disclosures occur. There has been no conversion.

For similar reasons, I do not view CMS's interest in issuing a regulation, or in doing so on a particular date, or in keeping the planned regulation a secret until its issuance, as constituting government "property" within the meaning of §§ 1343 and 1348. Given that CMS, notwithstanding any premature disclosure of its predecisional regulatory information, can issue a regulation that adheres to its preliminary inclination or can issue a different regulation, I cannot see that CMS has been deprived of anything that could be considered property.

Nor do I see merit in the government's contention that predecisional regulatory information should be considered government property because CMS is "responsible for allocating $1 trillion in federal funds every year," and that "[b]ecause a large part of" CMS's "mission" to "develop[ ] and maintain[ ] effective health care policy .... is centered on cost-effective allocation of health care spending, interference with CMS's right to exclusive use of its confidential information necessarily creates the potential for significant economic consequences" (Government brief on appeal at 92). Whatever economic consequences actually occur will be based on what CMS actually decides as to the substance and the timing of the regulation it adopts. The Cleveland Court rejected the government's argument that a property right of the State of Louisiana had been interfered with because the defendant "frustrated the State's right to control the issuance" of gaming licenses. 531 U.S. at 23, 121 S.Ct. 365. The Court held that "these intangible rights of allocation, exclusion, and control amount to no more and no less than Louisiana's sovereign power to regulate." Id .

Like the gaming licenses in question in Cleveland , which the State had the right to control or withhold--but which had no property status or effect until they were issued (and even when issued were not the property of the State)--the predecisional CMS information has no economic impact on the government until after CMS has actually decided what regulation to issue and when the regulation will take effect. And at the point when the regulation has economic impact on the government fisc, its impact will be in accordance with whatever regulation CMS ultimately decided to adopt. Thus, I cannot agree that a premature disclosure of predecisional regulatory information has taken any property from CMS or the government.

As the majority notes, all four defendants were acquitted on all of the counts charging them with substantive securities fraud violations of Title 15 and SEC Rule 10b-5 promulgated thereunder. The only substantive counts on which the jury found any defendant guilty were those charging violations of 18 U.S.C. §§ 641, 1343, and 1348. Since, in my view, the predecisional regulatory information at issue here did not constitute CMS property within the meaning of §§ 1343 and 1348, or a thing of value stolen from CMS in violation of § 641, none of defendants' convictions on substantive counts should stand.

The Indictment also contained three conspiracy counts: Counts 1 and 2 against all four defendants (on both of which Worrall was acquitted), and Count 17 against Blaszczak alone. Count 2 charged all defendants with violating 18 U.S.C. § 1349, which prohibits conspiracy "to commit any offense under this chapter," to wit, Chapter 63 of Title 18, i.e. , 18 U.S.C. §§ 1341 - 1351. Count 17 charged Blaszczak with violating 18 U.S.C. § 371 by conspiring with a cooperating coconspirator to violate § 641. Since in my view the Indictment's allegations of substantive violations of §§ 1343, 1348, and 641 charged defendants only with conduct that was not prohibited by those sections, defendants could not properly be convicted of conspiring to violate them. Thus, I would conclude that the convictions on Counts 2 and 17 should also be reversed.

The conspiracy charged in Count 1, however, was not limited to a conspiracy to violate §§ 641, 1343, and 1348. Count 1 (Indictment ¶¶ 1-76) charged defendants with agreeing to commit "conversion of United States property, in violation of Title 18, United States Code, Section 641 ; securities fraud, in violation of Title 15, United States Code, Sections 78j(b) and 78ff, and Title 17, Code of Federal Regulations, Section 240.10b-5 ; and to defraud the United States and an agency thereof , to wit, CMS, in violation of Title 18, United States Code, Section 371 and Title 5, Code of Federal Regulations, Section 2635.703(a) ." (Indictment ¶ 72 (emphases added)). The latter Code of Federal Regulations provision states in part that "[a]n employee shall not ... allow the improper use of nonpublic information to further his own private interest or that of another ... by knowing unauthorized disclosure." 5 C.F.R. § 2635.703(a) (emphases added). Count 1 alleged that defendants agreed to, inter alia , defraud CMS by obtaining from its employee Worrall confidential information about CMS's predecisional regulatory information (see Indictment ¶ 75) and engage in purchases and sales of securities in violation of 15 U.S.C. § 78j(b) and 78ff (see id . ¶ 74), and that pursuant to their conspiracy certain overt acts, including short sales of the shares of specified companies, were committed, all in violation of 18 U.S.C. § 371 (see id . ¶ 76).

The defendants other than Worrall were found guilty on this count. The jury was not given questions to answer that would reveal, with respect to Count 1, whether it found that the three convicted defendants had conspired to violate the securities fraud provisions of Title 15 and SEC Rule 10b-5 promulgated under that Title or to violate a government employee's duty of confidentiality, or instead had only conspired to violate § 641. When, as here, the jury has been presented with several bases for conviction, one or more of which is invalid as a matter of law, and it is impossible to tell which ground the jury selected, the conviction should be vacated. See , e.g. , Yates v. United States , 354 U.S. 298, 312, 77 S.Ct. 1064, 1 L.Ed.2d 1356 (1957) (prosecution for conduct beyond statute-of-limitations period invalid as a matter of law), partially overruled on other grounds by Burks v. United States , 437 U.S. 1, 7-10, 98 S.Ct. 2141, 57 L.Ed.2d 1 (1978) ; see generally Griffin v. United States , 502 U.S. 46, 52-56, 112 S.Ct. 466, 116 L.Ed.2d 371 (1991). While the mere insufficiency of the evidence to support one of the bases submitted to the jury does not fall within this principle, see id . at 56, 112 S.Ct. 466, a basis is invalid as a matter of law when the conduct in question "fails to come within the statutory definition of the crime," id . at 59, 112 S.Ct. 466.

As the jury could have found that the three defendants it convicted under Count 1 agreed to commit crimes prohibited by Title 15 and the regulations promulgated under that Title, but may instead have found only that they agreed to engage in conduct that was alleged to violate 18 U.S.C. § 641, 1343, or 1348 but that did not come within the definitions of those sections, the convictions of Blaszczak, Huber, and Olan on Count 1 should be vacated.

Accordingly, I respectfully dissent.