Tippees and Tippers:­­ The Impact of Martoma II

This is a column for insider trading junkies—a special breed who love all the nuances in this very nuanced subject. Late last month, a Second Circuit panel did something fairly unusual: It withdrew a 2017 decision and substituted a new opinion with a new rationale (but still with the same 2-1 division on the panel). The new decision in United States v. Martoma[1] has a less sweeping and more defensible rationale but still deviates from the law in other circuits. In addition, it has some nuances that future cases are certain to explore. Chief among these is the status of gossip: Can it be viewed as a “gift” with the tipper constructively trading and distributing the proceeds to the tippees?

Both the initial and the revised Martoma decisions deal with what had long seemed a footnote to the mainstream of insider trading law. Since Dirks v. SEC in 1983,[2] the mainstream doctrine has required two conditions before a tippee could be held liable: (1) the tipper had to breach a duty to either the company or the source of its information; and (2) the tippee had to pay or promise some “personal benefit” to the tipper—in essence, a bribe for the information. But at the very end of its Dirks decision, almost as an afterthought, the Supreme Court recognized an alternative theory: If the tipper made a gift of the information to the tippee, this could be treated as if the tipper had, itself, traded and given the proceeds to the tippee. This “gift” theory seemed intended to close a loophole, such as cases in which a CEO made gifts of inside information to his or her children.

Increasingly, it has proven difficult to show a “personal benefit” (possibly because both the tipper and tippee anticipate that the tippee will reciprocate in the future, but the prosecution cannot prove this). Thus, prosecutors came to rely increasingly on the “gift” theory. But, for reasons that are unclear, a Second Circuit panel in U.S. v. Newman[3] seemed to shut down this theory by requiring that, in such a “gift” case, the prosecutor had to prove beyond a reasonable doubt “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.”[4] Where this language came from is a mystery (as it had no precedent in Second Circuit decisions), and it seemed to require proof of an expectation of reciprocity—i.e., that the gift-giving tipper expected to get something back. Not only was this requirement unprecedented, but all parents know that gifts to children do not work that way. You do not expect reciprocity but only hope (sometimes futilely) for some gratitude from the next generation. That sort of gift was what Dirks was most clearly addressing.

Not surprisingly, Newman’s reciprocity expectation was quickly rejected by the Supreme Court in 2016 in Salman v. United States.[5] In Salman, the solicitor general actually asked the Supreme Court to overrule Dirks and eliminate its “personal benefit” requirement by holding that persons who knowingly receive material non-public information from an insider should be held to know this was an unlawful gift. But the court refused to go this far and made clear that it wanted to leave Dirks just as it was.[6] My own guess is that at least some on the court were concerned about making insider trading appear even more to be a common law crime that courts could modify and amend as they saw fit.[7]

Still, although the court rejected Newman’s language about the gift-giver expecting a “potential gain” in return, it did not address Newman’s insistence on a “meaningfully close personal relationship.” It had no need to do so on the facts of Salman, which clearly involved a close family relationship.[8] Nor did that requirement then seem controversial, as other circuits had also stressed that the gift-giving tipper and his tippees had a close personal relationship.[9]

This issue of how close the relationship had to be to support an inference that the transfer of the information resembled trading by the tipper and a gift of the proceeds to the tippee finally came to the fore in Martoma. Martoma involved an academic doctor, Sidney Gilman (“Gilman”), who chaired the scientific committee supervising the testing of a new drug for two major drug companies (Elan Corporation and Wyeth) that potentially could revolutionize the treatment of Alzheimer’s disease. Mathew Martoma, who managed a $500 to $600 million portfolio for S.A.C. Capital Advisors (“SAC”), which was run by Steven A. Cohen, a controversial hedge fund entrepreneur, retained an expert networking firm that employed Gilman. Martoma met with Gilman on approximately 43 consultations at the rate of around $1,000 dollars per hour.[10] Although defense counsel argued that these sessions were paid for by SAC, not Martoma, SAC was Martoma’s employer, and few saw any difficulty in the jury finding that Gilman had received a personal benefit arranged by Martoma. Gilman was also contractually bound not to disclose the results of the clinical tests that he was supervising. Thus, all the standard Dirks elements seemed satisfied.

Nevertheless, the trial court instructed the jury that it could convict Martoma either under the “personal benefit” theory or the “gift” theory, but it did not give the jury any instruction as to whether the relationship between Gilman and Martoma had to be a “meaningfully close personal” one. This gave the defense the opportunity to object that the jury could convict on the “gift” theory, even if jurors believed the two were only casual friends. Ultimately, the panel dealt with this issue (on a 2-1 basis) by finding that this possible error in the jury instruction was harmless, because the jury had overwhelming evidence of Martoma’s guilt under the “personal benefit” standard.[11]

Having so concluded, the panel could have ducked the issue of what the “gift” standard required in terms of the relationship between the tipper and the tippee. But it did not and instead seized the opportunity to reverse Newman on this ground as well. In its initial decision, the Martoma panel ruled that the “gift” theory reached any case in which material non-public information was communicated by an insider to persons who thereafter traded on it, at least when the information was disclosed by the tipper with the expectation that the recipient would trade on it.[12] That seems overboard, as later discussed, and the Martoma panel retreated from this position in its revised opinion. In that revised June 25, 2018, opinion, the “expectation of trading” requirement was dropped in favor of a new standard under which the tippee is liable under the “gift” standard if the tipper intended to benefit the tippee. This seems more closely related to Dirks’ concerns, but, even on this modified basis, the “gift” theory seems to swallow up and overwhelm the “personal benefit” standard. Also, under this revised test, there is no need that the relationship between tipper and tippee be either “close” or “meaningful” so long as there was an intention to benefit the tippee.[13]

Does this comply with Dirks?  Here, it is useful to take a step back and recognize, as a starting point, that Dirks said only two things about the necessary relationship between tipper and tippee. First, it recognized that the prosecution may prove its case by showing “a relationship between the insider and the recipient that suggests…an intention to benefit the particular recipient.”[14] But the stress here is on the “particular recipient,” suggesting that the nature of this relationship matters and that it has to be a gift to a specific person (not a broad class of unknown persons). Then, in the same paragraph describing its “gift” theory, the court added a functional resemblance test: “The tip and trade resemble trading by the insider himself followed by a gift of the profits to the recipient.”[15] This language suggests a possible test: Would the tipper have given profits in this amount to the tippee? This resemblance can only be stretched so far. In Martoma, Gilman, an academic researcher, was presumably a man of fairly modest wealth. Yet, according to the panel majority, the trades that Martoma and Cohen made based on his tips “resulted in approximately $80.3 million in gains and $194.6 million in averted losses for SAC”[16]—or a total of $274.9 million. If one asks, does the trading in Martoma resemble trading by Gilman and a gift by him of $274 million to Martoma and Cohen, it is hard to answer in the affirmative. Even if Gilman could somehow have traded at this level, it still seems implausible that he would have given such amounts to Martoma and Steven Cohen, where he was only a casual acquaintance of Martoma and did not know Cohen. Viewed at least in this literalistic fashion, Gilman’s tipping thus does not seemingly “resemble trading” by Gilman and a gift of possibly 100 times his net worth to Martoma. Of course, Gilman should be found guilty under the “personal benefit” test (because he received lucrative fees for his tips), but it is less clear that the “gift” test should reach him if it is defined as an independent, non-overlapping standard.

Of course, on policy grounds, no one wants to acquit the tippee simply because he made obscene profits that the tipper could not possibly have given him. Nonetheless, in Salman, the court, itself, rejected the solicitor general’s proposed broader theory under which virtually any gift of information by an insider to those who knowingly trade on it creates liability. Martoma essentially reaches this same end position by the back door and thereby trivializes both Dirks’ “personal benefit” test and Salman’s insistence that lower courts must stick with the Dirks standard (unless Congress acts).

Is that so bad? This author has argued in a New York Times op-ed piece that the “personal benefit” test was overly demanding and should be legislatively limited or abolished.[17] But what can be done legislatively cannot necessarily be done by a court of appeals, which seemingly is defying the Supreme Court. Possibly, the revised decision is the product of the panel being sensitized by their fellow judges to the problems in their initial standard. Or conceivably, there might have even been a possibility of a rare rehearing en banc if the panel had not retreated (slightly) to its revised position.

All this is speculative, so let us instead focus on what the revised standard will mean. Clearly, the defense seems entitled to an instruction that the jury must find that the defendant intended to benefit its tippees. When would this not be the case? Consider the following cases and hypotheticals:

Case One: Tipping Based on Spite or Hostility

Suppose our tipper is an attorney or investment banker who has come to dislike, even hate, the company’s CEO. He wants to drive its stock price down (but not for his own financial benefit), and he will not trade.[18] He knows very damaging material non-public information from his work for the company. As such a constructive insider, he leaks this information to tippees (in breach of his duty to the corporation), not to benefit these tippees, but rather to injure the CEO once the adverse information becomes public and drives down the stock price. If these facts are credited by the jury, it should not be able to find the requisite intention to benefit. Even if we think the tipper received some vicarious benefit here in seeing the CEO embarrassed, this benefit was not knowingly given by the tippees.

Case Two: The Tipper as an Obsessive Believer

Suppose Dr. Gilman in the Martoma case is dedicated to saving lives and is convinced that the new drug under testing is a wonder drug, but he fears that the company will not produce it because the company sees little profit in it. Thus, he wants to get the favorable test results out (so that the drug will not be abandoned), and so he leaks the results to hedge funds that, he hopes, will publicize the positive findings (after buying the stock). Again, there is little, if any, intention to benefit tippees here. The tipper’s goal is to save lives (but anonymously).

Case Three: The Tipper as an Incurable Gossip

Suppose now our tipper is a gossip, who leaks information, not to benefit the tippee, but out of a compulsion to gossip. Possibly, he thinks he will receive recognition for his knowledge, expertise, and contacts, but this is not networking and implies no future benefit to him.[19] Whatever the motive is for such gossiping, we all know such people (and academia is overpopulated with them). If we assume that our tipper just wants to be the center of attention, it is hard to see his tips as motivated by an intention to benefit.

Case Four: The Tippee as Whistleblower

Suppose now that an insider is tipping others to end a fraud. These are the actual facts of Dirks, which is still good law, and neither the tipper nor tippee is liable in such a case. Therefore the tippee can trade freely (just as Ray Dirks did).

Case Five: Tipping to an Unknown Audience

Comments made by an insider on social media (particularly when the media has a limited circulation) can be regarded as tipping,[20] but the tipper may have no idea in such a case whom he has tipped. In this setting, it is difficult to believe that he intended to benefit unknown strangers.

As a generalization, the more distant and tenuous the relationship between the tipper and the tippee becomes, the more difficult it should be to find an intention to benefit. To be sure, the information recipient may have received a benefit, but it was arguably fortuitous, and the information transfer may reflect no intention to make a gift.

It is important here to distinguish tipper and tippee liability. In some of these cases, it may be that the tipper should be liable (because he breached a duty and caused trading by the tippees based on material non-public information that he has leaked), but this does not imply that his tippees are also liable.[21] The positions of the tipper and the tippees are independent.

Defense tactics will also predictably respond to Martoma. Going forward, defense counsel will likely express concerns about the possibility of a duplicitous verdict. That is, if the court charges the jury that it may convict under either the “personal benefit” theory or the “gift” theory, there is the possibility that half the jury accepted one theory and half the other, but not all the jury agreed on any one theory. Defense counsel may therefore ask for a special verdict (possibly to confuse and split the jury). Otherwise defense counsel will seek to prove no intention to benefit, perhaps even by introducing evidence that the tipper disliked or did not know the tippees.


Both Newman and the first Martoma decision seem overwritten opinions that attempt to legislate a better rule (in each panel’s polarly differing estimation). Although the revised Martoma decision frames an improved and narrower test, other circuits may still hesitate about relying on Martoma and may continue to require some form of a close relationship. There is some justice in such a requirement, because gifts of information to strangers do not carry the same clear indication of culpability as does a tip to a close relative. Such gifts to a casual acquaintance (where no personal benefit is involved) seem more likely to have been inadvertent or less than fully considered (whereas a tip to a close friend or relative better indicates an attempt to transfer wealth through misappropriation from the issuer). Ideally, the “gift” theory should require the prosecution to show the tipper’s motive for the gift, and here gifts to unknown classes should be viewed skeptically.[22]

Finally, all these difficulties again indicate that someday Congress should take up the difficult and dangerous task of turning a common law crime into a real statute.


[1] The original decision was United States v. Martoma, 869 F.3d 58 (2d Cir. 2017), and was announced on August 25, 2017. Ten months later, the revised or “amended” decision was issued on June 25, 2018. See 2017 U.S. App. LEXIS 27880 (2d Cir. June 25, 2018).

[2] 463 U.S. 646 (1983).

[3] 773 F.3d 438 (2d Cir. 2014).

[4] Id. at 452.

[5] 137 S. Ct. 420 (2016).

[6] Id. at 426-27. The government had argued that “a tipper personally benefits whenever the tipper discloses confidential information for a noncorporate purpose.” Id. at 426. Declining to adapt this broader theory (or the defendant’s very narrow theory), the court wrote: “We adhere to Dirks, which easily resolves the narrow issue presented here.” Id. at 427.

[7] Common law crimes were abolished in 1812 in United States v. Hudson & Goodwin, 11 U.S. (7 Cranch) 32 (1812). Although the idea that a federal conviction requires a legislative act is constitutionally enshrined, Congress has never defined insider trading.

[8] In Salman, one brother tipped his sibling, who in turn passed the information on to their brother-in-law. No one raised any question about the closeness of these relationships.

[9] See, for example, United States v. Bray, 853 F.3d 18, 26-27 (1st Cir. 2017); United States v. Panigian, 824 F.3d 5,16 (1st Cir. 2016); 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. Rockledge, 470 F.3d 1,7 n.4 (1st Cir. 2006) (siblings).

[10] See 869 F.3d at 61.

[11] Id. at 65 and 72 (any error in jury instruction did not impair “Martoma’s substantial rights in light of the compelling evidence…”).

[12] Id. at 69. (“Thus, we hold that an insider or tipper personally benefits whenever the information was disclosed with the expectation that the recipient would trade on it.”)

[13] Id. at 70-71.

[14] 464 U.S. at 664.

[15] Id.

[16] 869 F.3d at 62.

[17] John C. Coffee, Jr., “How to Get Away with Insider Trading.” New York Times, May 23, 2016 (op-ed column).

[18] Some will answer that the tipper receives a “personal benefit” from the humiliation of his enemy. Perhaps, but this benefit was not provided by the tippees, who know nothing about his motivation. This column is focused on tippee liability.

[19] The cases have permitted some fairly soft factors to amount to a personal benefit. See S.E.C. v. Obus 693 F. 3d 276, 280, 292 (2d Cir. 2012) (defendant “hoped to curry favor with his boss”); United States v. Jiau, 734 F.3d 147, 153 (2nd Cir. 2013) (admission to a social and investment club that offered valuable business contacts). But the pure gossip, talking at the equivalent of the water cooler, is receiving none of these intangible benefits.

[20] The SEC has recognized that information becomes “public” if it is disseminated “in a manner calculated to reach the securities marketplace in general through recognized channels of distribution and public investors [are] afforded a reasonable waiting period to react to the information.” See “Commission Guidance on the Use of Company Websites,” 73 Fed. Reg. 45, 862 (August 7, 2008). Even if a company website may amount (sometimes) to a “recognized channel of distribution,” smaller blogs do not, and they will predictably tip those who anticipate new information surfacing on them. See also “Report of Investigation Pursuant to Section  21(a) of the Securities Exchange Act of 1934: Netflix, Inc. and Reed Hastings,” Securities Exchange Act Release No. 34-69279 (April 2, 2013). This column will not discuss Regulation FD, which will apply to many issues in these facts.

[21] See United States v. Gupta, 111 F.Supp.3d 557 (S.D.N.Y. 2015) (tipper intention to benefit the tippee is sufficient to satisfy the benefit standard for the tipper, even if it would not create liability for the tippee).

[22] For example, a casual comment (but containing material non-public information) on the alumni blog of a prominent university (say, Notre Dame) should not be viewed as an intent to make a gift to all the alumni of Notre Dame. This is overcriminalization, as the remark was probably more negligent than fraudulent. In any event, intention to benefit cannot be inferred that easily if Dirks is our guide.

This post comes to us from John C. Coffee, Jr., the Adolf A. Berle Professor of Law at Columbia University Law School and Director of its Center on Corporate Governance.

1 Comment

  1. Professor Michael Guttentag

    Thanks for posting on this (from another insider trading/personal benefit nerd). Generally agree on all counts with one caveat. For reasons discussed in detail elsewhere, Selective Disclosure and Insider Trading, 69 Florida Law Review 519 (2017), I think it is a mistake to describe the “gift” theory as a doctrine distinct from the Dirks Court’s personal benefit test. The Dirks Court’s personal benefit test was always just a mishmash based on four different rationales for imposing tipper/tippee liability, and one of those rationales was always the “gift standard.”

Comments are closed.