United States v. Martoma

894 F.3d 64

Opinion

Judge Pooler dissents in a separate opinion.

Katzmann, Chief Judge:

Defendant-appellant Mathew Martoma was convicted, following a four-week jury trial, of one count of conspiracy to commit securities fraud in violation of 18 U.S.C. § 371 and two counts of securities fraud in violation of 15 U.S.C. §§ 78j(b) & 78ff in connection with an insider trading scheme. On appeal, Martoma argues that the jury was improperly instructed and that there was insufficient evidence to sustain his conviction.

Martoma's contentions focus on the "personal benefit" element of insider trading law. In Dirks v. S.E.C. , the Supreme Court held that a "tippee"-someone who receives confidential information from a corporate insider, or "tipper," and then trades on the information-can be held liable under the insider trading laws "only when the insider has breached his fiduciary duty to the shareholders by disclosing the information to the tippee and the tippee knows or should know that there has been a breach." 463 U.S. 646, 660, 103 S.Ct. 3255, 77 L.Ed.2d 911 (1983). "[T]he test" for whether there has been a breach of the tipper's duty "is whether the [tipper *68] personally will benefit, directly or indirectly, from his disclosure" to the tippee. Id. at 662, 103 S.Ct. 3255. Dirks set forth several personal benefits that could prove the tipper's breach, including, for example, "a relationship" between the tipper and tippee "that suggests a quid pro quo from the latter," the tipper's "intention to benefit" the tippee, and "a gift of confidential information to a trading relative or friend" where "[t]he tip and trade resemble trading by the insider himself followed by a gift of the profits to the recipient." Id. at 664, 103 S.Ct. 3255.

Martoma first argues that the jury in his case was not properly instructed in light of the Second Circuit's decision in United States v. Newman , 773 F.3d 438 (2d Cir. 2014). Martoma asserts that, under Newman , evidence that the tipper made a gift of inside information to a trading relative or friend establishes a "personal benefit" only if tipper and tippee share a "meaningfully close personal relationship." See Newman , 773 F.3d at 452. Martoma contends that the jury instructions were flawed because they did not qualify that evidence of a gift to a trading relative or friend establishes a personal benefit only where there is a "meaningfully close personal relationship." Second, Martoma argues that the evidence at trial was insufficient to sustain a conviction under any theory of personal benefit.

We agree that the jury instructions are inconsistent with Newman , though not for the reasons Martoma advances. Newman held that a personal benefit in the form of "a gift of confidential information to a trading relative or friend," see Dirks , 463 U.S. at 664, 103 S.Ct. 3255, requires proof that the tipper and tippee shared what the decision called a "meaningfully close personal relationship," see Newman , 773 F.3d at 452. The Court explained that this standard "requires evidence of 'a relationship between the insider and the recipient that suggests a quid pro quo from the latter, or an intention to benefit the [latter].' " Id. (quoting United States v. Jiau , 734 F.3d 147, 153 (2d Cir. 2013) (quoting Dirks , 463 U.S. at 664, 103 S.Ct. 3255 ) ). Thus, Martoma's jury instructions were erroneous, not because they omitted the term "meaningfully close personal relationship," but because they allowed the jury to find a personal benefit in the form of a "gift of confidential information to a trading relative or friend" without requiring the jury to find either that tipper and tippee shared a relationship suggesting a quid pro quo or that the tipper gifted confidential information with the intention to benefit the tippee.

We nonetheless conclude that this instructional error did not affect Martoma's substantial rights. At trial, the government presented compelling evidence that at least one tipper received a different type of personal benefit from disclosing inside information: $70,000 in "consulting fees." This evidence establishes the existence of a relationship suggesting a quid pro quo between the tipper and tippee. For this reason, Martoma's challenge to the sufficiency of the personal-benefit evidence fails. Moreover, the government presented sufficient evidence for a rational trier of fact to conclude that at least one tipper received a personal benefit by disclosing inside information with the intention to benefit Martoma. Accordingly, we AFFIRM the judgment of the district court.

BACKGROUND

Martoma's convictions stem from an insider trading scheme involving securities of two pharmaceutical companies, Elan Corporation, plc ("Elan") and Wyeth, that were jointly developing an experimental drug called bapineuzumab to treat Alzheimer's disease. Martoma worked as a *69portfolio manager at S.A.C. Capital Advisors ("SAC"), a hedge fund owned and managed by Steven A. Cohen. In that capacity, Martoma managed an investment portfolio with buying power of between $400 and $500 million that was focused on pharmaceutical and healthcare companies. He also recommended investments to Cohen, who managed SAC's largest portfolio. While at SAC, Martoma began to acquire shares in Elan and Wyeth in his portfolio and recommended that Cohen acquire shares in the companies as well.

In order to obtain information about bapineuzumab, Martoma contacted expert networking firms and arranged paid consultations with doctors knowledgeable about Alzheimer's disease, including two who were working on the bapineuzumab clinical trial. Dr. Sidney Gilman, chair of the safety monitoring committee for the bapineuzumab clinical trial, participated in approximately 43 consultations with Martoma at the rate of around $1,000 per hour.¹ As a member of the safety monitoring committee, Dr. Gilman had an obligation to keep the results of the clinical trial confidential. His consulting contract reiterated that he was not to disclose any confidential information in a consultation. He nevertheless provided Martoma, whom he knew to be an investment manager seeking information to help make securities trading decisions, with confidential updates on the drug's safety that he received during meetings of the safety monitoring committee. Dr. Gilman also shared with Martoma the dates of upcoming safety monitoring committee meetings, which allowed Martoma to schedule consultations with Dr. Gilman shortly after each one. Another consultant, Dr. Joel Ross, one of the principal investigators on the clinical trial, met with Martoma on many occasions between 2006 and July 2008 and charged approximately $1,500 per hour. Like Dr. Gilman, Dr. Ross had an obligation to maintain the confidentiality of information about the bapineuzumab clinical trial. Nevertheless, during their consultations, Dr. Ross provided Martoma with information about the clinical trial, including information about his patients' responses to the drug and the total number of participants in the study, that Dr. Ross recognized was not public.

On June 17, 2008, Elan and Wyeth issued a press release regarding the results of "Phase II" of the bapineuzumab clinical trial. The press release described the preliminary results as "encouraging," with "clinically meaningful benefits in important subgroups" of Alzheimer's patients with certain genetic characteristics, but indicated that the drug had not proven effective in the general population of Alzheimer's patients. J.A. 547. The press release further stated that the results of the trials would be presented in greater detail at the International Conference on Alzehimer's Disease to be held on July 29, 2008. Elan's share price increased following the press release.

In mid-July of 2008, the sponsors of the bapineuzumab trial selected Dr. Gilman to present the results at the July 29 conference. It was only at this point that Dr. Gilman was unblinded as to the final efficacy results of the trial. Dr. Gilman was "initially euphoric" about the results, but identified "two major weaknesses in the data" that called into question the efficacy of the drug as compared to the placebo. Tr. 1419-20. On July 17, 2008, the day after being unblinded to the results, Dr. Gilman spoke with Martoma for about 90 *70minutes by telephone about what he had learned. That same day, Martoma purchased a plane ticket to see Dr. Gilman in person at his office in Ann Arbor, Michigan. That meeting occurred two days later, on July 19, 2008. At that meeting, Dr. Gilman showed Martoma a PowerPoint presentation containing the efficacy results and discussed the data with him in detail.

The next morning, Sunday, July 20, Martoma sent Cohen, the owner of SAC, an email with "It's important" in the subject line and asked to speak with him by telephone. The two had a telephone conversation lasting about twenty minutes, after which Martoma emailed Cohen a summary of SAC's Elan and Wyeth holdings. The day after Martoma spoke to Cohen, on July 21, 2008, SAC began to reduce its position in Elan and Wyeth securities and entered into short-sale and options trades that would be profitable if Elan's and Wyeth's stock fell.

Dr. Gilman publicly presented the final results from the bapineuzumab trial at the International Conference on Alzehimer's Disease in the afternoon of July 29, 2008. Elan's share price began to decline during Dr. Gilman's presentation and at the close of trading the next day, the share prices of Elan's and Wyeth had declined by about 42% and 12%, respectively. The trades that Martoma and Cohen made in advance of the announcement resulted in approximately $80.3 million in gains and $194.6 million in averted losses for SAC. Martoma personally received a $9 million bonus based in large part on his trading activity in Elan and Wyeth.

At Martoma's trial, the district court instructed the jury on the personal benefit element of insider trading law as follows:

If you find that Dr. Gilman or Dr. Ross disclosed material, non-public information to Mr. Martoma, you must then determine whether the government proved beyond a reasonable doubt that Dr. Gilman and Dr. Ross received or anticipated receiving some personal benefit, direct or indirect, from disclosing the material, non-public information at issue.

The benefit may, but need not be, financial or tangible in nature; it could include obtaining some future advantage, developing or maintaining a business contact or a friendship, or enhancing the tipper's reputation.

A finding as to benefit should be based on all the objective facts and inferences presented in the case. You may find that Dr. Gilman or Dr. Ross received a direct or indirect personal benefit from providing inside information to Mr. Martoma if you find that Dr. Gilman or Dr. Ross gave the information to Mr. Martoma with the intention of benefiting themselves in some manner, or with the intention of conferring a benefit on Mr. Martoma, or as a gift with the goal of maintaining or developing a personal friendship or a useful networking contact.

Tr. 3191.

After Martoma was convicted and while his appeal was pending, this Court decided United States v. Newman , 773 F.3d 438 (2d Cir. 2014), an insider trading case that considered one of the personal benefits described in Dirks and mentioned in Martoma's jury instructions-making a "gift" of inside information to "a trading relative or friend."² This Court stated:

To the extent Dirks suggests that a personal benefit may be inferred from a personal relationship between the tipper and tippee, where the tippee's trades 'resemble trading by the insider himself *71followed by a gift of the profits to the recipient,' see 463 U.S. at 664 [103 S.Ct. 3255], we hold that such an inference is impermissible in the absence of proof of a meaningfully close personal relationship that generates an exchange that is objective, consequential, and represents at least a potential gain of a pecuniary or similarly valuable nature.

773 F.3d at 452. An initial round of briefing focused in large part on whether Martoma's conviction could stand in light of this passage from Newman .

Shortly thereafter, the Supreme Court decided Salman v. United States , --- U.S. ----, 137 S.Ct. 420, 196 L.Ed.2d 351 (2016), another case involving the gift theory. The defendant, relying on Newman , urged the Supreme Court to hold that a " 'gift of confidential information to a trading relative or friend' " is insufficient to establish insider trading liability "unless the tipper's goal in disclosing inside information [wa]s to obtain money, property, or something of tangible value." Id. at 426 (quoting Dirks , 463 U.S. at 664, 103 S.Ct. 3255 ). The Supreme Court rejected the defendant's argument and "adhere[d] to Dirks ," id. at 427, observing that "[t]o the extent the Second Circuit held that the tipper must also receive something of a 'pecuniary or similarly valuable nature' in exchange for a gift to family or friends, ... this requirement is inconsistent with Dirks ," id. at 428 (quoting Newman , 773 F.3d at 452 ); see also id. ("Here, by disclosing confidential information as a gift to his brother with the expectation that he would trade on it, Maher breached his duty of trust and confidence to Citigroup and its clients ....").

The government now takes the position that Salman fully abrogated Newman 's interpretation of the personal benefit element, whereas Martoma argues that Newman 's"meaningfully close personal relationship" standard survived Salman . However, because there are many ways to establish a personal benefit, we conclude that we need not decide whether Newman 's gloss on the gift theory is inconsistent with Salman . At trial, the government presented compelling evidence that Dr. Gilman received a different type of personal benefit: $70,000 in consulting fees, which can be seen either as evidence of a quid pro quo -like relationship, or simply advance payments for the tips of inside information that Dr. Gilman went on to supply.³ The government also introduced sufficient evidence to prove Dr. Gilman received a personal benefit by disclosing inside information with the intention to benefit Martoma. We accordingly conclude that Martoma has provided no basis for his judgment of conviction to be vacated or reversed.

DISCUSSION

As noted above, Martoma challenges both the adequacy of the district court's jury instructions and the sufficiency of the evidence presented at trial. "We review a jury charge in its entirety and not on the basis of excerpts taken out of context." United States v. Mitchell , 328 F.3d 77, 82 (2d Cir. 2003) (quoting United States v. Zvi , 168 F.3d 49, 58 (2d Cir. 1999) ). "A conviction based on a general verdict is subject to challenge if the jury was instructed on alternative theories of guilt and may have relied on an invalid one." Hedgpeth v. Pulido , 555 U.S. 57, 58, 129 S.Ct. 530, 172 L.Ed.2d 388 (2008). Such a challenge, however, is subject to harmless error review. See ibr.US_Case_Law.Schema.Case_Body:v1" id="p72" href="#p72" data-label="72" data-citation-index="1" class="page-label">*72id. at 58, 61-62, 129 S.Ct. 530. And because Martoma raises his challenge to the jury instructions for the first time on appeal, we review only for plain error. United States v. Vilar , 729 F.3d 62, 70 (2d Cir. 2013). Under the plain error standard, an appellant must demonstrate that "(1) there is an error; (2) the error is clear or obvious, rather than subject to reasonable dispute; (3) the error affected the appellant's substantial rights ...; and (4) the error seriously affects the fairness, integrity or public reputation of judicial proceedings."4 United States v. Marcus , 560 U.S. 258, 262, 130 S.Ct. 2159, 176 L.Ed.2d 1012 (2010) (internal quotation marks and alteration omitted). "[W]e look not to the law at the time of the trial court's decision to assess whether the error was plain, but rather, to the law as it exists at the time of review." Vilar , 729 F.3d at 71. Even with respect to an instructional error that "incorrectly omitted an element of the offense," we will not overturn a conviction "if we find that the jury would have returned the same verdict beyond a reasonable doubt," and thus that "the error did not affect [the defendant's] substantial rights." Nouri , 711 F.3d at 139-40 (internal quotation marks omitted).

With respect to Martoma's second argument, a defendant challenging the sufficiency of the evidence "bears a heavy burden," and "the standard of review is exceedingly deferential." United States v. Coplan , 703 F.3d 46, 62 (2d Cir. 2012) (internal quotation marks omitted). "In evaluating a sufficiency challenge, we 'must view the evidence in the light most favorable to the government, crediting every inference that could have been drawn in the government's favor, and deferring to the jury's assessment of witness credibility and its assessment of the weight of the evidence.' " Id. (quoting United States v. Chavez , 549 F.3d 119, 124 (2d Cir. 2008) ). "Although sufficiency review is de novo , we will uphold the judgment[ ] of conviction if 'any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.' " Id. (citation omitted) (quoting Jackson v. Virginia , 443 U.S. 307, 319, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979) ). "A judgment of acquittal is warranted only if the evidence that the defendant committed the crime alleged is nonexistent or so meager that no reasonable jury could find guilt beyond a reasonable doubt." Jiau , 734 F.3d at 152 (alteration and internal quotation marks omitted).

I.

We first turn to Martoma's challenge to the district court's jury instructions, which focuses on Dirks ' statement that the personal benefit necessary to establish insider trading liability in a tipping case can be inferred from a "gift of confidential information to a trading relative or friend." Dirks , 463 U.S. at 663-64, 103 S.Ct. 3255 ; see also Salman , 137 S.Ct. at 428. Martoma argues that the district court's jury instructions ran afoul of this Court's decision in Newman by permitting the jury to conclude that a gift of confidential information given with the goal of "developing or *73maintaining ... a friendship" qualifies as a personal benefit. According to Martoma, the jury should have been instructed that the tipper and tippee must share a "meaningfully close personal relationship" in order to find a personal benefit based on a gift of inside information to a friend.

A. The Personal Benefit Requirement

The Supreme Court long ago held that there is no "general duty between all participants in market transactions to forgo actions based on material, nonpublic information." Chiarella v. United States , 445 U.S. 222, 233, 100 S.Ct. 1108, 63 L.Ed.2d 348 (1980). However, the "traditional" or "classical theory" of insider trading provides that a corporate insider violates § 10(b) of the Securities Exchange Act of 1934, 15 U.S.C. § 78j(b), and Rule 10b-5, 17 C.F.R. § 240.10b-5 (2017), when he "trades in the securities of his corporation on the basis of material, non-public information" because "a relationship of trust and confidence [exists] between the shareholders of a corporation and those insiders who have obtained confidential information by reason of their position with that corporation." United States v. O'Hagan , 521 U.S. 642, 651-52, 117 S.Ct. 2199, 138 L.Ed.2d 724 (1997) (alteration in original) (quoting Chiarella , 445 U.S. at 228, 100 S.Ct. 1108 ). Similarly, the "misappropriation theory" of insider trading provides "that a person ... violates § 10(b) and Rule 10b-5[ ] when he misappropriates confidential information for securities trading purposes, in breach of a duty owed to the source of the information." Id. at 652, 117 S.Ct. 2199. It is thus the breach of a fiduciary duty or other "duty of loyalty and confidentiality" that is a necessary predicate to insider trading liability. See id.⁵

The personal benefit element has its origin in Dirks , where the Supreme Court examined how a recipient of inside information who was not himself a corporate insider-i.e. , a tippee-can acquire a duty to disclose or abstain from trading. The Supreme Court held that a tippee acquires the duty to disclose or abstain only if the insider disclosed the confidential information in breach of a fiduciary duty to the firm. Dirks , 463 U.S. at 660-61, 103 S.Ct. 3255. "Whether disclosure is a breach of duty," the Supreme Court explained, "depends in large part on the purpose of the disclosure." Id. at 662, 103 S.Ct. 3255. The personal benefit requirement is designed to test the propriety of the tipper's purpose. See id. at 661-63, 103 S.Ct. 3255. This logic is sound. A firm's confidential information belongs to the firm itself, and an insider entrusted with it has a fiduciary duty to use it only for firm purposes. The insider who personally benefits-i.e. , whose purpose is to help himself-from disclosing confidential information therefore breaches that duty; the insider who discloses for a legitimate corporate purpose does not. Identifying personal benefits is not, however, the central focus of insider trading law, but simply how courts and juries analyze breaches of fiduciary duty.

The Supreme Court defined personal benefit broadly. As noted above, the test for a personal benefit is whether objective evidence shows that "the insider personally will benefit, directly or indirectly, from his disclosure" of confidential information to the tippee.

*74Id. at 662, 103 S.Ct. 3255. Dirks set forth numerous examples of personal benefits that prove the tipper's breach: a "pecuniary gain," a "reputational benefit that will translate into future earning," a "relationship between the insider and the recipient that suggests a quid pro quo from the latter," the tipper's "intention to benefit the particular recipient," and a "gift of confidential information to a trading relative or friend" where "[t]he tip and trade resemble trading by the insider himself followed by a gift of the profits to the recipient." Id. at 663-64, 103 S.Ct. 3255. The tipper's personal benefit need not be pecuniary in nature. See Salman , 137 S.Ct. at 428.

We have applied Dirks to uphold a wide variety of personal benefits. We held that a jury could infer a personal benefit from the fact that a tipper "hoped to curry favor with his boss," Obus , 693 F.3d at 292, and from the fact that another tipper and the tippee "were friends from college," id. at 291. We found evidence of a personal benefit sufficient where the tippee gave one tipper "an iPhone, live lobsters, a gift card, and a jar of honey," and where the tippee had another tipper admitted into an investment club where the tipper "had the opportunity to access information that could yield future pecuniary gain" (even though he never realized that opportunity). Jiau , 734 F.3d at 153. In another case, we held that the government "need not show that the tipper expected or received a specific or tangible benefit in exchange for the tip," and that the personal benefit element is satisfied where there is evidence that the tipper "intend[ed] to benefit the ... recipient." S.E.C. v. Warde , 151 F.3d 42, 48 (2d Cir. 1998) (internal quotation marks omitted).

As we understand the dissent, our core disagreement is over whether intent to benefit is a standalone personal benefit under Dirks . The dissent argues that it is not, claiming instead that the correct formulation is a "relationship ... that suggests ... an intention to benefit" the tippee. See Dissent, op. at 83-84. The key sentence of Dirks is admittedly ambiguous, and we acknowledge that the dissent has offered a plausible reading. See 463 U.S. at 664, 103 S.Ct. 3255 ("For example, there may be a relationship between the insider and the recipient that suggests a quid pro quo from the latter, or an intention to benefit the particular recipient."). But that is not the only reading. The comma separating the "intention to benefit" and "relationship ... suggesting a quid pro quo " phrases can be read to sever any connection between them. The sentence, so understood, effectively reads, "there may be a relationship between the insider and the recipient that suggests a quid pro quo from the latter, or there may be an intention to benefit the particular recipient." And that is the reading this Court adopted in Warde , where we read the "intention to benefit" language independently of the language of relationships: "The 'benefit' element of § 10(b) is satisfied when the tipper 'intend[s] to benefit the ... recipient' or 'makes a gift of confidential information to a trading relative or friend.' "⁶ Warde , 151 F.3d at 48 (quoting Dirks , 463 U.S. at 664, 103 S.Ct. 3255 ). We adhere to Warde .

*75Our understanding is also more consonant with Dirks as a whole. Because the existence of a breach "depends in large part on the purpose of the disclosure," Dirks , 463 U.S. at 662, 103 S.Ct. 3255, it makes perfect sense to permit the government to prove a personal benefit with objective evidence of the tipper's intent, without requiring in every case some additional evidence of the tipper-tippee relationship. Cf. United States v. Falcone , 257 F.3d 226, 230 (2d Cir. 2001) (Sotomayor, J .) (explaining that "the key factor" in proving a personal benefit is "the tipper's intent in providing the information"). For example, suppose a tipper discloses inside information to a perfect stranger and says, in effect, you can make a lot of money by trading on this. Under the dissent's approach, this plain evidence that the tipper intended to benefit the tippee would be insufficient to show a breach of the tipper's fiduciary duty to the firm due to the lack of a personal relationship. Dirks and Warde do not demand such a result. Rather, the statement "you can make a lot of money by trading on this," following the disclosure of material non-public information, suggests an intention to benefit the tippee in breach of the insider's fiduciary duty.

We are not persuaded by our dissenting colleague's arguments to the contrary. The dissent contends that proof that the tipper had an intent to benefit the tippee does not prove that the tipper truly "received" a personal benefit. See Dissent, op. at 84-85. The dissent would evidently have there be proof of something more concrete. However, as we have explained, it is settled law that personal benefits may be indirect and intangible and need not be pecuniary at all. The tipper's intention to benefit the tippee proves a breach of fiduciary duty because it demonstrates that the tipper improperly used inside information for personal ends and thus lacked a legitimate corporate purpose. That is precisely what, under Dirks , the personal benefit element is designed to test. See 463 U.S. at 662, 103 S.Ct. 3255. Is evidence that an insider intended to benefit an outsider with valuable confidential information any less probative of the absence of a legitimate corporate purpose than evidence that the tippee gave the tipper trivialities like shellfish and a gift card? See Jiau , 734 F.3d at 153.⁷

The dissent argues that its formulation is more faithful to the personal benefit standard because evidence of a relationship suggesting an intent to benefit the tippee "provides reason to believe that the tipper benefits by benefitting, since the tipper is understood as contributing to a relationship from which both tipper and tippee benefit," a rationale that does not apply where there has been no proof of a relationship. Dissent, op. at 86. We disagree. That rationale would justify a personal benefit in the form of a relationship suggesting an intention to benefit both tipper and tippee, from which it is straightforward to infer that the tipper personally benefited from the tip. But what Dirks in fact refers to is an intention to benefit the tippee alone. See 463 U.S. at 664, 103 S.Ct. 3255. Whichever way Dirks is read, it recognizes that purposely benefitting the tippee with inside information proves that the tipper has received a personal *76benefit in breach of a fiduciary duty. The question is whether Dirks requires that to be proved with evidence of a relationship or not. We think it clear that the answer is no. And although few reported decisions have relied on the intent to benefit theory, its legitimacy has until today been uncontroversial. To take an example close to home, it featured in the jury instructions in this very case, see Tr. 3191, and no objection was raised, nor was any challenge to this language pressed on appeal.

Finally, we are warned that this approach creates a "subjective" test and allows for convictions based on sheer speculation into the tipper's motives. See Dissent, op. at 83-84, 88. These fears are unwarranted. Intent elements are everywhere in our law and are generally proved with circumstantial evidence. See, e.g. , United States v. Heras , 609 F.3d 101, 106 (2d Cir. 2010) ("The law has long recognized that criminal intent may be proved by circumstantial evidence alone."); United States v. Salameh , 152 F.3d 88, 143 (2d Cir. 1998) ("[A]s a general rule most evidence of intent is circumstantial."). Insider trading is no different. A factfinder may infer the tipper intended to benefit the tippee from the sort of objective evidence that is commonly offered in insider trading cases. To return to the example above, the statement "you can make a lot of money by trading on this" is strong circumstantial evidence of the tipper's intention to benefit the tippee. And the requirement of proof beyond a reasonable doubt remains a formidable barrier to convictions resting on speculation. See United States v. Torres , 604 F.3d 58, 66 (2d Cir. 2010).

We are thus satisfied that the personal benefit element can be met by evidence that the tipper's disclosure of inside information was intended to benefit the tippee. And as is clear from the purpose of the personal benefit element, the "broad definition of personal benefit set forth in Dirks ," and the variety of benefits we have upheld, the evidentiary "bar is not a high one." Obus , 693 F.3d at 292.

B. This Court's Decision in Newman

It is against that background that we must assess how Newman affected this Court's insider trading law. The central question in Newman was an issue of scienter on which our district courts had been split: whether a tippee must be aware, not only that the tipper breached a fiduciary duty in disclosing inside information, but also that the tipper received a personal benefit. Newman , 773 F.3d at 447-51. The Court persuasively explained that both were required. Id. at 449 ("[A] tippee's knowledge of the insider's breach necessarily requires knowledge that the insider disclosed confidential information in exchange for personal benefit."). This important teaching of Newman is not before us. We observe that, unlike the defendants in Newman , Martoma received confidential information directly from the tipper, and he does not claim that he was unaware of any personal benefit Dr. Gilman received. Cf. id. at 448 ("In Jiau , the defendant knew about the benefit because she provided it.").

Newman 's second holding is the focus of this appeal. After resolving the scienter question, Newman considered the sufficiency of the personal benefit evidence for two tippers, where the government relied chiefly on evidence that they were friendly with their tippees. The first tipper and tippee were not "close" friends but "had known each other for years, having both attended business school and worked at Dell together," and the tippee had provided modest "career advice and assistance"

*77to the tipper. Id. at 452. The second tipper and tippee were "family friends" that "had met through church and occasionally socialized together." Id. The government argued that these relationships were "sufficient to prove that the tippers derived some benefit from the tip." Id.

The Newman panel rejected the government's argument, holding that the personal benefit "standard, although permissive, does not suggest that the Government may prove the receipt of a personal benefit by the mere fact of a friendship, particularly of a casual or social nature." Id. As the Newman Court reasoned, if that were enough, then "practically anything would qualify," and "the personal benefit requirement would be a nullity." Id. And in the sentence that forms the basis of Martoma's argument on appeal, Newman stated as follows:

To the extent Dirks suggests that a personal benefit may be inferred from a personal relationship between the tipper and tippee, where the tippee's trades 'resemble trading by the insider himself followed by a gift of the profits to the recipient,' we hold that such an inference is impermissible in the absence of proof of a meaningfully close personal relationship ....

Id. at 452 (citation omitted) (quoting Dirks , 463 U.S. at 664, 103 S.Ct. 3255 ). On the facts before it, the Newman Court found that standard had not been satisfied. Id. at 452-53.

Martoma focuses on this single sentence of Newman to argue that a jury may not infer that a tipper received a personal benefit from gifting confidential information in the absence of a "meaningfully close personal relationship." The term "meaningfully close personal relationship" is new to our insider trading jurisprudence, and, viewed in isolation, it might admit multiple interpretations. But Newman provided substantial guidance. Immediately after introducing the "meaningfully close personal relationship" concept, Newman held that it "requires evidence of 'a relationship between the insider and the recipient that suggests a quid pro quo from the latter, or an intention to benefit the [latter].' " Newman , 773 F.3d at 452 (quoting Jiau , 734 F.3d at 153 (quoting Dirks , 463 U.S. at 664, 103 S.Ct. 3255 ) ). As explained above, each of these is an independently sufficient basis to infer a personal benefit under Dirks and its progeny. See, e.g. , Jiau , 734 F.3d at 153 (quid pro quo -like relationship). In other words, Newman cabined the gift theory using two other freestanding personal benefits that have long been recognized by our case law.⁸ And although the dissent urges in strong terms that this reading is mistaken or even improper, its dispute is in truth with the plain language of Dirks , as construed by Warde . We do no more than read literally Newman 's own explanation of its novel standard in light of these decisions, thereby fulfilling our legitimate function to construe and give effect to prior panel decisions.

With that understanding of Newman , we conclude that the personal benefit *78jury instructions in Martoma's trial, issued prior to that decision, were erroneous. The instructions allowed the jury to find a personal benefit based solely on the conclusion that Dr. Gilman tipped Martoma in order to "develop[ ] or maintain[ ] ... a friendship." Under Newman , this articulation of the gift theory is incomplete. A properly instructed jury would have been informed that it could find a personal benefit based on a "gift of confidential information to a trading relative or friend" only if it also found that Dr. Gilman and Martoma shared a relationship suggesting a quid pro quo or that Dr. Gilman intended to benefit Martoma with the inside information. But, of course, there was no error in the district court's instructions that the jury could also find a personal benefit based on either of those two factors alone, i.e. , if it concluded that Dr. Gilman disclosed confidential information "with the intention of conferring a benefit on Mr. Martoma," or "with the intention of benefiting [himself] in some manner." See Tr. 3191. Each of these personal benefits is unaffected by Newman 's interpretation of the gift theory, and neither requires proof that Dr. Gilman and Martoma share any type of "personal relationship."

Although the jury instructions were inaccurate, we conclude that the error did not affect Martoma's substantial rights. See Nouri , 711 F.3d at 139-40. The government produced compelling evidence that Dr. Gilman, the tipper, "entered into a relationship of quid pro quo " with Martoma. See Jiau , 734 F.3d at 153. Dr. Gilman, over the course of approximately 18 months and 43 paid consultation sessions for which he billed $1,000 an hour, regularly and intentionally provided Martoma with confidential information from the bapineuzumab clinical trial. Martoma kept coming back, specifically scheduling consultation sessions so that they would occur shortly after the safety monitoring committee meetings, when Dr. Gilman would have new information to pass along. Starting at least in August 2007, Dr. Gilman would reschedule his conversations with Martoma if he had no new information to reveal at the time they were scheduled to meet. By that point, the consulting relationship between Dr. Gilman and Martoma involved no legitimate service, see Dissent, op. at 88; as Dr. Gilman testified at trial, "the purpose of those consultations was for [him] to disclose to [Martoma] confidential information about the results ... of the last Safety Monitoring Committee [meeting]." Tr. 1274. And because Martoma continued to see Dr. Gilman to receive confidential information, Dr. Gilman continued to receive consulting fees. The fact that Dr. Gilman did not specifically bill for his July 17 and 19, 2008 conversations with Martoma in which Dr. Gilman divulged the final drug efficacy data is also of no moment because, as he admitted at trial, doing so "would [have been] tantamount to confessing that [he] was ... giving [Martoma] inside information." Tr. 1918. In the context of their ongoing "relationship of quid pro quo ," Dr. Gilman's disclosures of confidential information were designed to "make good on the substantial pecuniary benefit he had already earned," Dirks , 463 U.S. at 663, 103 S.Ct. 3255, and as a result, "it is clear beyond a reasonable doubt that a rational jury would have found [Martoma] guilty absent [any] error." Mahaffy , 693 F.3d at 136 (internal quotation marks omitted).

The dissent argues that under our analysis, a fact-finder must always find that tipper and tippee had a quid pro quo -like relationship whenever a tip is exchanged within a paid consulting relationship. Dissent, op. at 88. Not so. We merely hold that on the compelling facts of this case, it is clear beyond a reasonable doubt that a properly instructed jury would have found *79Martoma guilty. Nor does our decision mean that a tipper who accidentally or unknowingly reveals inside information can be found guilty. See id. at 72. Such a tipper would be protected by the requirement that the tipper know (or is reckless in not knowing) that the information is material and non-public, see Obus , 693 F.3d at 286, or by the requirement that the tipper expect the tippee to trade, see United States v. Gansman , 657 F.3d 85, 92 (2d Cir. 2011).

II.

We next turn to Martoma's challenge to the sufficiency of the personal benefit evidence and whether, "evaluating ... the evidence in the light most favorable to the government," a rational jury could have found Martoma guilty of insider trading. See Coplan , 703 F.3d at 62. As an initial matter, it follows from our conclusion that the faulty jury instructions were harmless because of the compelling evidence that Dr. Gilman and Martoma shared a relationship suggesting a quid pro quo that this evidence was also sufficient to support his conviction. In particular, the jury was free to place no weight on the fact that Dr. Gilman did not bill Martoma for the July 17 and 19, 2008 sessions. We reiterate, however, that while the government presented compelling evidence on this point, the evidentiary bar is "modest." Jiau , 734 F.3d at 153.

Moreover, even if a jury were inclined to accept Martoma's argument that there was no quid pro quo -like relationship because Dr. Gilman did not bill Martoma for two key sessions, a rational jury could nonetheless find that Dr. Gilman personally benefited by disclosing inside information with the "intention to benefit" Martoma. See Dirks , 463 U.S. at 664, 103 S.Ct. 3255. We think a jury can often infer that a corporate insider receives a personal benefit (i.e. , breaches his fiduciary duty) from deliberately disclosing valuable, confidential information without a corporate purpose and with the expectation that the tippee will trade on it. See id. at 659, 103 S.Ct. 3255 (explaining that "insiders [are] forbidden by their fiduciary relationship" from giving inside information "to an outsider for the ... improper purpose of exploiting the information for their personal gain"); cf. Salman , 137 S.Ct. at 428 (recognizing that where a tipper discloses information with the "expectation that [the recipient will] trade on it," the information is "the equivalent of ... cash"). Here, as previously noted, Dr. Gilman knew that Martoma was an investment manager who was seeking information on which to base securities trading decisions. And Dr. Gilman plainly understood the valuable nature of the information about the bapineuzumab clinical trial, as Martoma had previously paid him $1,000 per hour over the course of 43 consultations to convey his knowledge on the subject, and had visited Dr. Gilman in his Ann Arbor office to receive the key drug efficacy results firsthand. From these facts, a reasonable jury could infer that Dr. Gilman personally benefited by conveying inside information about the trial with the purpose of benefiting Martoma, even if it was not persuaded that the two had a relationship suggesting a quid pro quo (or a personal relationship, for that matter).⁹ See Dirks , 463 U.S. at 667, 103 S.Ct. 3255.

*80For the foregoing reasons, we hold that "a rational trier of fact could have found the essential elements of the crime [of insider trading] beyond a reasonable doubt." Coplan , 703 F.3d at 62 (quoting Jackson , 443 U.S. at 319, 99 S.Ct. 2781 ).¹⁰

CONCLUSION

We have considered Martoma's remaining arguments and find them without merit. Accordingly, we AFFIRM the judgment of the district court.

Pooler, Circuit Judge:

I respectfully dissent. Last year, my colleagues filed an opinion in this matter in which they abrogated our prior decision in United States v. Newman , 773 F.3d 438, 452 (2d Cir. 2014). They declared that a non-insider could be convicted of insider trading on a gift theory even if she did not have a meaningfully close personal relationship with the insider from whom she received the confidential information. See United States v. Martoma , 869 F.3d 58, 69 (2d Cir. 2017). Applying this reasoning to the case at hand, they held that the jury instructions permissibly allowed for conviction based on speculation about Dr. Gilman's desire to be friends with Martoma. I dissented from that opinion because it improperly abrogated a prior panel decision without en banc review or an intervening Supreme Court precedent, undermined the personal benefit rule central to holding corporate outsiders liable for insider trading, and approved of a conviction based on erroneous jury instructions that affected Martoma's substantial rights.

My colleagues now issue a modified opinion. In it, they purport to agree that our precedent prevents a jury from being charged with inferring that a tip was given as a gift unless it finds that there was a meaningfully close personal relationship between the tipper and the tippee. They no longer disclaim Newman . They even agree that the jury instructions were in error.

But these apparent concessions are semantic rather than substantial. My colleagues also attempt to redefine "meaningfully close personal relationship" in subjective rather than objective terms, rendering Newman a relic. To provide support for this move, they improperly construe binding authority. They then hold that the erroneous jury instructions were harmless since the jury could have convicted based on a different theory.

The majority's attempt to undercut the meaningfully close personal relationship requirement is in derogation of circuit precedent and unnecessary to arrive at their disposition. Only by abrogating Newman could my colleagues announce a new rule that a jury can infer a personal benefit based on a freestanding "intention to benefit" and that this "intention to benefit" is at the core of the meaningfully close *81personal relationship standard. Op. at 74-76, 77. Today's opinion must be interpreted consistently with the rule that, as a three-judge panel, we are unable to abrogate prior circuit decisions. See In re Zarnel , 619 F.3d 156, 168 (2d Cir. 2010) ("This panel is bound by the decisions of prior panels until such time as they are overruled either by an en banc panel of our Court or by the Supreme Court.") (internal quotation marks omitted). Newman and a consistent line of cases preceding it make clear that a meaningfully close personal relationship cannot be proven without objective evidence about the nature of the tipper-tippee relationship. Bare speculation into insiders' motives has always been insufficient; it remains so today in spite of the majority's dicta.

Therefore, I continue to respectfully dissent.

I. Gifts and the Law of Insider Trading

Dirks , the foundational case on holding a non-insider liable for insider trading, established that a jury's "initial inquiry" must be whether a corporate insider passed on information to the non-insider "for personal advantage" rather than for the advantage of shareholders. Dirks v. S.E.C. , 463 U.S. 646, 662-63, 103 S.Ct. 3255, 77 L.Ed.2d 911 (1983). Making the inquiry into "whether the insider receives a direct or indirect personal benefit" by disclosing confidential information "requires courts to focus on objective criteria." Id. at 663, 103 S.Ct. 3255. The question for a finder of fact is not whether the insider wished ill on shareholders or wished good on the tippee, but whether she received something in return for her tip.

As the Supreme Court explained,¹ making objective evidence of a personal benefit a prerequisite to holding a non-insider tippee liable serves several purposes. It creates "a guiding principle for those whose daily activities must be limited and instructed by the SEC's inside-trading rules" so that participants in securities markets are not left to the whims of prosecutorial enforcement priorities. Dirks , 463 U.S. at 664, 103 S.Ct. 3255. It protects "persons outside the company such as an analyst or reporter who learns of inside information" from the threat of prosecution for uncovering information about securities issuers just because they also traded on it. Id. at 664 n.24, 103 S.Ct. 3255 (italics omitted). It limits the government's ability to hold non-insiders liable when insiders "mistakenly think...information already has been disclosed or that it is not material enough to affect the market." Id. at 662, 103 S.Ct. 3255.

Restricting proof of a personal benefit to objective evidence avoids turning the rule into a mere formality. Absent objective evidence, a slip of the tongue might be presented to a jury as a purposeful tip with a good cover story, an off-the-record comment to a trusted reporter might be portrayed as a means of bribing a journalist for favorable coverage. The difference between guilty and innocent conduct would be a matter of speculation into what a tippee knew or should have known about the tipper's intent. A trader, journalist, or analyst attempting to avoid running afoul of criminal law would have little to guide her behavior. The conservative thing to do would be to avoid seeking inside information too aggressively, even if the whole *82market could benefit from such investigation. Those who decided to cultivate insider sources would risk prosecution in any case, so they might have fewer scruples about compensating their sources and trading on the information they purchased.

What does objective evidence of a personal benefit consist of? In the easiest case, a tippee has paid the insider for the coveted tidbit. If the government can adduce evidence indicating that money changed hands, it has established all of the objective facts needed to infer that an insider personally benefitted by tipping. In the presence of an obvious quid pro quo, no further facts about the nature of the tipper-tippee relationship will be needed. The insider has effectively made the "secret profits" that securities law has prohibited since its inception, but by selling information to a trader rather than trading on it herself. In re Cady, Roberts & Co. , 40 S.E.C. 907, 916 n.31 (1961) ; see also United States v.O'Hagan , 521 U.S. 642, 653, 117 S.Ct. 2199, 138 L.Ed.2d 724 (1997) (characterizing misappropriation as an insider "secretly converting the [corporation's] information for personal gain").

The majority rightly points out that "[t]he tipper's personal benefit need not be pecuniary in nature." Op. at 74; see also id. at 75. But that does not obviate the requirement that it be provable via "objective evidence." In-kind compensation in goods or services given to the tipper may also constitute a personal benefit, and can be established in court in much the same way monetary compensation can, i.e. with objective evidence pointing to the goods or services received. See , e.g. , United States v. Jiau , 734 F.3d 147, 153 (2d Cir. 2013) (discussing "an iPhone, live lobsters, a gift card, and a jar of honey"). The government can also prove a benefit on the theory that the tipper received potentially profitable social connections, such as admission into an investment club, so long as the prosecution's case relies on evidence of these connections and their potential value to the tipper. Id.

When the alleged benefit to the tipper is less concrete, objective evidence about the nature of the relationship between tipper and tippee takes on more importance in identifying the benefit. For instance, unlike with money, goods, services, and connections, one cannot directly trace a "reputational benefit that will translate into future earnings." Dirks , 463 U.S. at 663, 103 S.Ct. 3255. Instead, one must draw upon circumstantial evidence about the power of a tippee to materially benefit someone in the tipper's position and the inclination of a tippee to view the tipper in a better light based on the tipper's provision of inside information. Evidence about the tipper-tippee relationship will make these conclusions easier, as illustrated by the facts of S.E.C. v. Obus , 693 F.3d 276 (2d Cir. 2012). In that case, we concluded that it was sufficient that the tipper, an employee of a hedge fund that "was a large holder" of the stock in question, "hoped to curry favor with his boss," the tippee, who was the principal of that hedge fund. Id. at 280, 292. Bosses, of course, have substantial say over subordinates' future earnings, and the boss of a hedge fund trading in a particular stock is likely to value employees that can get him information about that stock.

More directly on point, if the government fails to put forward evidence of any particular quo that was provided in exchange for the quid of inside information, it can still establish objective facts that point to a "relationship between the insider and the [tippee] that suggests a quid pro quo ." Dirks , 463 U.S. at 664, 103 S.Ct. 3255 (emphasis added). That is, an apparently gratuitous tip can reasonably be understood as recompense for past benefits *83or as a means of keeping a good thing going so long as there is objective evidence of a history of mutually enriching exchanges or favors between tipper and tippee.

The personal benefit rule is also satisfied by other "relationships between the insider and the recipient" that "suggest an intention to benefit the particular recipient" even when the insider receives no immediately discernible compensation. Id. In particular, as relevant here, an apparently uncompensated tip can be said to "resemble trading by the insider himself followed by a gift of the profits to the recipient" when it is given to a "trading relative or friend." Id. ; see also Salman v. United States , --- U.S. ----, 137 S.Ct. 420, 427-28, 196 L.Ed.2d 351 (2016). Friends and relatives tend to internalize each other's interests, see Transcript of Oral Argument at 8, Salman v. United States , --- U.S. ----, 137 S.Ct. 420, 196 L.Ed.2d 351 (2016) ("[T]o help a close family member [or friend] is like helping yourself."), to give each other things of value to demonstrate care, or to commit acts of generosity with the assumption that the other would do the same in a rough sort of quid-pro-quo.² For any of these reasons, a tipper can be said to benefit himself by giving something valuable to somebody with who he shares a "meaningfully close personal relationship."³ Newman , 773 F.3d at 452. Without objective evidence of such a relationship, however, the inference that a gratuitous tip functioned as a gift will not be available. Newman made clear that the "gift theory" is not applicable to casual acquaintances or mere members of the same club, church, or alumni association-or, it should go without saying, to perfect strangers-at least not without additional evidence indicating meaningful closeness. 773 F.3d at 452-55. Other boundaries of the concept of meaningful closeness remain to be developed. See Salman , 137 S.Ct. at 429 ("...there is no need for us to address those difficult cases today...").

II. The Majority's Error

Last year the majority attempted to rewrite this doctrine explicitly. Today they attempt to do so more subtly. In their now withdrawn opinion, they held that a gratuitous tip could be understood as beneficial to the tipper so long as a jury were to conclude that a tipper expected the tippee to trade on it. Martoma , 869 F.3d at 70-71. Now they hold that an uncompensated tip can be found to personally benefit the tipper so long as the jury were to concludes that the tipper intended to benefit the tippee. Op. at 74-76, 77. All that *84"meaningfully close personal relationship" means, they inform us, is a tipper-tippee pairing in which the tipper has such an intention.⁴ Id. at 77.

This interpretation would eliminate the rule that has been with us since Dirks that the government must prove objective facts indicating that the tipper benefitted from her relationship with the tippee. On the majority's proposal, the prosecution could pile up insinuations about the tipper's subjective understanding of the purpose of the tip, and the jury would be charged with resting their inferences about her benefit on those wobbly foundations. The only objective facts the government would have to prove would be the communication of material non-public information. All of the protections of the personal benefit rule-a clear guide for conduct, preventing liability for slip ups and other innocent disclosures-would erode.

It is good news, then, that binding precedent stands for the opposite principle. The only time Dirks refers to an "intention to benefit" is when it discusses the need to prove "a relationship between the insider and the recipient that suggests ...an intention to benefit the particular recipient." 463 U.S. at 664, 103 S.Ct. 3255 (emphasis added). Reading "intention to benefit" out of context, my colleagues assert that, under Dirks , an intention can be inferred without any objective evidence about relationships. Op. at 88. But Dirks does not say that, and it has never been applied to allow such a freestanding inference of intent in this Circuit or elsewhere. Salman , 137 S.Ct. at 427 (applying gift theory to sibling relationship) Jiau , 734 F.3d at 153 (discussing gift theory as relationship-based before finding quid pro quo); Obus , 693 F.3d at 285 (discussing "trading relative or friend" standard); United States v. Bray , 853 F.3d 18, 26-27 (1st Cir. 2017) ("good friends"); United States v. Parigian , 824 F.3d 5, 16 (1st Cir. 2016) (friendship and quid pro quo); S.E.C. v. Rocklage , 470 F.3d 1, 7 n.4 (1st Cir. 2006) (siblings); S.E.C. v. Sargent , 229 F.3d 68, 77 (1st Cir. 2000) ("reconciliation" between friends and reputational benefit); S.E.C. v. Maio , 51 F.3d 623, 632-33 (7th Cir. 1995) (exchange of favors within a friendship); S.E.C. v. Yun , 327 F.3d 1263, 1280 (11th Cir. 2003) ("a friend and frequent partner in real estate deals").

S.E.C. v. Warde , 151 F.3d 42 (2d Cir. 1998), cited in the majority opinion, is not to the contrary. Op. at 74-75. In that case, "[t]he evidence showed that Warde[, the tippee,] was a good friend of Edward Downe," the tipper. Warde , 151 F.3d at 45 (emphasis added). We found that the "close friendship between Downe and Warde suggests that Downe's tip was 'inten[ded] to benefit' Warde..." Id. at 49 (emphasis added, brackets in original). Thus, we did not find that a freestanding "intention to benefit" would have been sufficient to prove Downe's personal benefit. Instead, we followed the principle that an intention to benefit can only be inferred from objective facts about the nature of the relationship between tipper and tippee. Again, the majority extracts the phrase "intention to benefit" from its context, suggesting that the relationship between tipper and tippee did not matter when it was the central focus of our inquiry.

The majority offers an alternative interpretation in which the sentence at issue in Dirks "effectively reads, 'there may be a relationship between the insider and the *85recipient that suggests a quid pro quo from the latter, or there may be an intention to benefit the particular recipient.' " Op. at 74. Perhaps one could read the sentence in that way in isolation, but doing so would certainly not be "more consonant with Dirks as a whole" or with the subsequent case law relying on Dirks . Id. The Dirks court included that sentence to provide examples of "objective facts and circumstances that often justify...an inference" that "the insider receive[d] a direct or indirect personal benefit from the disclosure." Dirks , 463 U.S. at 663-64, 103 S.Ct. 3255. It is difficult to understand why the Court would have mentioned an intention to benefit, which is a subjective fact, as an example of a personal benefit, which is an objective fact. Nearly as difficult to understand is why the Dirks court would have provided an intention to benefit a tippee as an example of a benefit to the tipper . Intending to benefit somebody is not in itself a benefit. That is, not unless one has reason to believe that the person with the intention to benefit benefits from the beneficiary's benefit or one adopts the trivializing view of human psychology wherein everything any individual does is to benefit herself.

Perhaps the majority's theory is that an intention to benefit a tippee is circumstantial evidence that a tipper is receiving some other benefit by providing the information. On this theory, so long as objective evidence would allow a jury to infer that a tipper intended to benefit the tippee, the jury should be allowed to infer from that inference that the tipper somehow benefitted by benefitting the tippee without actually having to determine what that benefit might be. This theory fails to deal with the fact that an intention to benefit is not itself an "objective fact or circumstance," as Dirks requires, but rather an inference drawn from objective facts or circumstances. Additionally, this theory makes it difficult to understand why the Dirks court would have adopted the personal benefit test in the first place. If a jury can conclude that a tipper breached his duty so long as it concludes that she intended to benefit the tippee, why should it have to go through the tortuous process of concluding that the tipper received a personal benefit based on its conclusion that the tipper intended to benefit the tippee? Why should we care about the tipper's benefit at all?

At times the majority seems to suggest that Dirks does not really require proof of a personal benefit. Rather, the personal benefit test is mentioned merely as a guide to prosecutors regarding the sort of evidence that will help them establish the tipper's intention to benefit the tippee. Thus, when Dirks says that "a breach of duty...depends in large part on the purpose of the disclosure," it is announcing the real test for a breach of the duty to shareholders. Id. at 662, 103 S.Ct. 3255. "Identifying personal benefits is...simply how courts and juries analyze breaches of" this duty. Op. at 73. When the government can adduce other evidence that a tipper "lacked a legitimate corporate purpose," Op. at 75, then "it makes perfect sense to permit the government to prove a personal benefit with [only] objective evidence of the tipper's intent." Op. at 75.

But Dirks is entirely unambiguous that "the test [for whether duty has been breached] is whether the insider personally will benefit, directly or indirectly, from his disclosure." 463 U.S. at 662, 103 S.Ct. 3255. Whatever the insider's purpose in disclosing the information, "[a]bsent some personal gain [to the insider], there has been no breach of duty to stockholders." Id. Nowhere does Dirks suggest that the need to prove personal benefit can be ignored simply because a tipper's intent or purpose can be independently demonstrated.

*86⁵ And Dirks expressly disclaims the idea that "personal benefit" is merely a synonym for a tipper "not act[ing] simply out of the goodness of his heart." Op. at 75 n.7; Dirks , 463 U.S. at 663, 103 S.Ct. 3255 (stating that proof of personal benefit should not be focused on mind reading). The personal benefit test may well be a way to get at a tipper's purpose, but it is the former and not the latter that the prosecution must prove.

None of these puzzles is presented if one reads the relevant sentence in Dirks the way I have suggested. It is easy to understand why the Dirks court would have mentioned a relationship suggesting an intention to benefit, an objective circumstance, when it was providing examples of objective facts and circumstances. Unlike a standalone intention to benefit, a relationship suggesting an intention to benefit provides reason to believe that the tipper benefits by benefitting, since the tipper is understood as contributing to a relationship from which both tipper and tippee benefit. See supra at 83. And the focus on relationships rather than bare intentions fits neatly with Dirks 's cabining of the gift theory to disclosures to "trading relative[s] or friend[s]." 463 U.S. at 664, 103 S.Ct. 3255.

This cannot be so, my colleagues protest. They ask us to imagine a situation where a tipper "discloses inside information to a perfect stranger and says, in effect, you can make a lot of money by trading on this." Op. at 75. Wouldn't it be absurd if this perfect stranger could not be held liable for insider trading if he went ahead and traded on this information? No, it would not be. At least, not if one takes the personal benefit rule seriously. Ex hypothesi, the fictional tipper in their scenario receives absolutely nothing in return for his disclosure, except, I suppose, the warmth that comes with knowing that somebody else might have made some money because of his actions (or perhaps the schadenfreude that comes with knowing that shareholders were defrauded). But if those sorts of "benefits" were enough, then every disclosure of inside information without affirmative indication of a pure heart would be presumptively beneficial to the tipper. Dirks rejected that possibility, and every appellate court to have considered the issue, including us, has consistently done the same. That is the law whether we like or not, but, for what it's worth, I see no reason to worry that truly random acts of enrichment can go unpunished.

Even assuming arguendo that there was any ambiguity on the topic in our precedents, Newman removed it by requiring a "meaningfully close personal relationship" in order to prove personal benefit via the gift theory. 773 F.3d at 452. In the majority's withdrawn opinion, they candidly acknowledged that they were abrogating Newman , relying on a justification for doing so that they no longer advance. Martoma , 869 F.3d at 68-70. Today they do not even attempt to argue they can do so. Instead, they call into question settled law *87in non-binding dicta. Newman remains good law.

III. The Jury Instructions

Turning to the case at hand, I agree with my colleagues' updated view that the jury was erroneously instructed. However, in light of the foregoing, I disagree with their formulation of the proper instruction. A properly instructed jury would have instead been asked whether Dr. Gilman and Martoma shared a relationship suggesting a quid pro quo or were close enough friends that it would be reasonable to understand Dr. Gilman's provision of information to Martoma as a gift. The jury could not conclude that their relationship was meaningfully close based on the mere possibility of a future friendship. Nor could it make a relationship-independent inference about Dr. Gilman's intentions, contrary to the majority's dicta. That is because Newman 's interpretation of the gift theory does "require[ ] proof that Dr. Gilman and Martoma share[d] any type of personal relationship." Op. at 78 (internal quotation marks omitted).

Moreover, I disagree that the error in the jury instructions was harmless. The majority rightly states that we can only find harmlessness in this context if "it is clear beyond a reasonable doubt that a rational jury would have found the defendant guilty absent the error." United States v. Mahaffy , 693 F.3d 113, 136 (2d Cir. 2012) (internal quotation marks omitted).⁶ This record provides plenty of reasons to doubt that Dr. Gilman and Martoma shared a meaningfully close personal relationship. The government itself repeatedly denied that Dr. Gilman and Martoma had anything other than a "commercial, pecuniary relationship." Recording of Oral Argument at 34:18-34:27, 26:27-26:58, United States v. Martoma , No. 14-3599 (2d Cir. Oct. 28, 2015). Dr. Gilman testified that he shared almost nothing about his personal life with Martoma and that Martoma acted friendlier than Dr. Gilman thought appropriate for a professional relationship. Tr. at 1238, 1236. As the government itself pointed out to the district court, there is no evidence that Martoma and Dr. Gilman ever interacted outside of their consulting sessions. Recording of Oral Argument at 34:18-34:27, United States v. Martoma , No. 14-3599 (2d Cir. Oct. 28, 2015).

A reasonable jury could also have doubted whether the relationship between Dr. Gilman and Martoma suggested a quid pro quo. Dr. Gilman took no payment for the consulting sessions in which he provided the inside information at issue here, and there is no evidence in the record that his compensation before or after that session was higher than usual. He was in high demand as an expert and a researcher, so *88there is reason to doubt that he would have risked prosecution just to keep up his consulting relationship with SAC and Martoma. See Tr. at 1552-60. A reasonable jury could have found a relationship suggesting a quid pro quo, but it was not required to. Ruling otherwise would lead to the holding that whenever inside information is revealed within a paid consulting relationship where other, legitimate services are rendered, a fact-finder must infer that the insider was paid to breach his duties. That rule would allow convictions for erroneously revealed information or for information revealed based on a misunderstanding about its materiality or its confidentiality.

IV. Sufficiency of the Evidence

Because the jury instructions amount to reversible error, I would not reach the sufficiency of the evidence question. But even were the majority correct that there was sufficient evidence here, it is incorrect and ill-advised to go on to speculate that the jury could have inferred an intention to benefit merely because "a corporate insider...deliberately disclos[ed] valuable, confidential information without a corporate purpose and with the expectation that the tippee will trade on it." Op. at 79.

In addition to undermining Newman in the manner already discussed, this musing flirts with the possibility that the personal benefit test that goes back to Dirks may no longer be good law. The very reason the government must establish a personal benefit is to allow for the possible conclusion that the insider provided information without a "corporate purpose." Dirks , 463 U.S. at 654, 655 n.14, 103 S.Ct. 3255. If the government can put forward evidence that an insider did not have a corporate purpose in order to establish that the insider personally benefitted from providing the information, it can convict based on circular reasoning. "Personal benefit" would then no longer have any independent meaning. The government would need only convince the jury to "read [the tipper's] mind[ ]," Dirks , 463 U.S. at 663, 103 S.Ct. 3255. The tipper would need not have benefitted in any objective sense so long as the prosecution could convince a jury that she was not thinking of the corporation's interest. Dirks stands for the opposite proposition.⁷

Setting our disagreement about the harmfulness of the district court's error to one side, my colleagues could have reached the conclusion they did by following the path our precedent provides. They need only have held that the jury instructions were erroneous because they allowed for conviction absent objective evidence of a meaningfully close personal relationship but harmless because there was objective evidence of a relationship suggesting a quid pro quo. Instead, they have taken a detour to declare that subjective evidence could have worked as well. This detour calls into question well-established principles of insider trading law that we have neither reason nor power to abrogate.

1 Martoma did not pay Dr. Gilman or any other consultant directly. Instead, SAC would pay the expert networking firm, and the expert networking firm would in turn pay Dr. Gilman and the other consultants.

2 For convenience, we sometimes refer to this as the "gift theory" of personal benefit.

3 The parties focus primarily on Dr. Gilman because it was Dr. Gilman, not Dr. Ross, who gave Martoma the final efficacy data that led Martoma to reduce SAC's position in Elan and Wyeth.

4 In the past, we have stated that "[w]here ... the source of an alleged jury instruction error is a supervening decision, we employ a 'modified plain-error rule, under which the government, not the defendant, bears the burden to demonstrate that the error ... was harmless.' " United States v. Mahaffy , 693 F.3d 113, 136 (2d Cir. 2012) (second alteration in original). We have "on at least twenty-two occasions," Vilar , 729 F.3d at 71 n.5, observed that the Supreme Court's decision in Johnson v. United States , 520 U.S. 461, 117 S.Ct. 1544, 137 L.Ed.2d 718 (1997), "called into question the modified plain error standard of review," United States v. Botti , 711 F.3d 299, 308 (2d Cir. 2013). Here, as in the past, "[b]ecause we would reach the same conclusion under either standard, we need not resolve that question." United States v. Nouri , 711 F.3d 129, 138 n.2 (2d Cir. 2013).

5 Although many of the cases refer to "insiders" and "fiduciary" duties because those cases involve the "classical theory" of insider trading, the Dirks articulation of tipper and tippee liability also applies under the misappropriation theory, where the misappropriator violates some duty owed to the source of the information. See S.E.C. v. Obus , 693 F.3d 276, 286-88 (2d Cir. 2012) ; see also Newman , 773 F.3d at 445-46.

6 Warde 's use of friendship to find the evidence of intent to benefit sufficient does not prove otherwise. See 151 F.3d at 49. Warde teaches only that such evidence is relevant and may even be sufficient in an appropriate case. It nowhere suggests that it is necessary. Cf. S.E.C. v. Payton , 219 F.Supp.3d 485, 490 (S.D.N.Y. 2016) (finding evidence of tipper's intent to benefit the tippee sufficient where tipper gave tippee a Post-It note with the stock ticker symbol, told the tippee they "could make some money on" the stock, and said the stock was a "good opportunity"), aff'd , 2018 WL 832917 (2d Cir. Feb. 13, 2018) (summary order).

7 In any event, even assuming arguendo that a more concrete benefit is required, the tipper's intention to benefit the tippee would still be an appropriate personal benefit. A tipper's disclosure of valuable confidential information with the intent to benefit the tippee can satisfy the personal benefit requirement because it can allow for the inference that the tipper has not acted simply out of the goodness of his heart, but because he expects to receive some future benefit. Cf. Obus , 693 F.3d at 292 (finding evidence of personal benefit sufficient where tipper "hoped to curry favor with his boss").

8 Our cases applying Dirks demonstrate that the government can prove a personal benefit in several ways that do not require proof of any sort of personal relationship. Consider the underling who disclosed inside information to "curry favor with his boss," see Obus , 693 F.3d at 292, or the tipper's admission into an investment club that yielded the possibility of future benefits that were never realized, see Jiau , 734 F.3d at 153, or the tipper's receipt of a cell phone, gift card, and various foodstuffs from the tippee, see id. In none of these situations was the government required to show any degree of personal closeness between tipper and tippee. Newman , with its focus on the gift theory, does not require a different result in these cases.

9 As this discussion demonstrates, the dissent's concern that intent to benefit can be shown only with subjective evidence or speculation is unfounded. See supra at 75-76. We conclude that the evidence was sufficient to infer that Dr. Gilman intended to benefit Martoma based, not on "subjective" evidence or speculation, but on the circumstantial evidence surrounding the tip.

10 We further find that even if the district court erroneously excluded the testimony of Steven Cohen under Federal Rule of Evidence 804(b)(1), such error was harmless because much of the testimony was inculpatory. See United States v. Dukagjini , 326 F.3d 45, 61 (2d Cir. 2003) ("In order to uphold a verdict in the face of an evidentiary error, it must be 'highly probable' that the error did not affect the verdict." (quoting United States v. Forrester , 60 F.3d 52, 64 (2d Cir. 1995) ) ). Cohen's testimony before the SEC described Martoma as having played a substantial role in his decision both to accumulate large positions in Elan and Wyeth, and then to sell them off. J.A. 91-92, 96-97. This was consistent with other trial evidence that Martoma received credit for SAC's trades in those companies. Tr. 497-99. It is therefore highly improbable that Cohen's vague statements that Martoma told him that he was "getting uncomfortable with the Elan position," J.A. 95, and that he heard that Martoma's reasons for being uncomfortable were "normal" and "typical," J.A. 97, would have affected the jury's verdict.

1 And as I explained in my previous dissent. See Martoma , 869 F.3d at 75-78 (Pooler, J., dissenting).

2 Following Newman 's suggestion, the majority holds that a jury could conclude that tipper and tippee shared a meaningfully close relationship so long as they shared a relationship suggesting quid pro quo. Op. at 77 (citing Newman , 773 F.3d at 452 ). But, as the majority rightly points out, Dirks and subsequent cases established that a relationship suggesting quid pro quo can itself give rise to the inference of a personal benefit to the tipper, without any need to determine whether it gives rise to the intermediate inference of a meaningfully close personal relationship. Op. at 77. It would make for a less confusing bit of doctrine to cleanly separate quid-pro-quo and meaningfully close personal relationships, and I do not think Newman requires conflating them. But the majority's interpretation is consistent with Newman , and, in any case, a jury can infer personal benefit from the former whether or not it gives rise to an inference of the latter.

3 Salman made clear that the tipper need not also "receive something of pecuniary or similarly valuable nature in exchange for a gift to family or friends." 137 S.Ct. at 428 (internal quotation marks omitted). While my colleagues earlier opinion read Salman to have eliminated the meaningfully close personal relationship standard entirely, Martoma , 869 F.3d at 68-71, they now agree that that aspect of Newman remains good law.

4 As discussed supra in note 2, the majority, following Newman 's suggestion, holds that a relationship suggesting a quid pro quo is also a meaningfully close personal relationship. I do not find this part of their analysis objectionable (except in the sense discussed in note 2).

5 I do not deny that "[i]ntent elements are everywhere in our law and are generally proved with circumstantial evidence." Op. at 76. I deny that one can replace proof of personal benefit to the tipper with proof of the tipper's intention to benefit the tippee. As I discussed in my previous dissent, insider trading law separately requires that the insider expect the tippee will trade on the information, and this expectation can be proven with circumstantial evidence. See Martoma , 869 F.3d at 82 (Pooler, J., dissenting) (citing Obus , 693 F.3d at 286-87 ; United States v. Gansman , 657 F.3d 85, 92 (2d Cir. 2011) ). The majority's approach comes close to conflating this element of insider trading liability and the separate personal benefit test.

6 As I discussed in my previous dissent, I would hold that the modified plain error rule applies here. Martoma , 869 F.3d at 87-88 (Pooler, J., dissenting). We have long held that where "the source of an alleged jury instruction is a supervening decision, we employ a modified plain error rule, under which the government, not the defendant, bears the burden to demonstrate that the error was harmless." Mahaffy , 693 F.3d at 136 (internal quotation marks omitted). The majority points out that multiple panels in this Circuit have called into question the continued applicability of the modified plain error rule after Johnson v. United States , 520 U.S. 461, 117 S.Ct. 1544, 137 L.Ed.2d 718 (1997), without deciding the matter either way. Op. at 72 n.4. But Johnson did not provide any reason to abandon the well-established modified plain error rule. Johnson only "cautioned against any unwarranted expansion of Rule 52(b)," which provides for plain error review in criminal matters in which an issue has not been raised below. 520 U.S. at 466, 117 S.Ct. 1544. The modified plain error rule does not expand Rule 52(b). It merely allocates the burden of proof, a matter on which Rule 52 is silent.

7 The majority suggests that the abundance of objective evidence in this case demonstrates that my concern about differentiating guilty from innocent conduct based entirely on inferences about intent is "unfounded." Op. 79 n.8. If so, then Dirks 's and Newman 's similar concerns are also unfounded. see Dirks , 463 U.S. at 663, 103 S.Ct. 3255 ; Newman , 773 F.3d at 452. Anyway, assuming arguendo that the evidence here was more than sufficient, that fact alone does not mean that we should endorse a rule that would allow for convictions based on speculation in cases where the evidence is thinner or more ambiguous. See supra at 81-82.