J. SKELLY WRIGHT, Circuit Judge:

This case requires us to review the Securities and Exchange Commission's interpretation of its Rule 10b-5 in light of the Supreme Court's most recent discussion of Rule 10b-5 and the statutory provisions upon which it rests. The Commission (SEC) has censured petitioner Raymond Dirks, vice president in a broker-dealer firm, for aiding and abetting violations of Rule 10b-5 by repeating information he had learned from former employees of a corporation to investors likely to sell their shares of the corporation's stock to members of the public without access to the same information. Dirks asks this court to reject the SEC's interpretation of Rule 10b-5 on the authority of Chiarella v. United States, 445 U.S. 222, 100 S.Ct. 1108, 63 L.Ed.2d 348 (1980), claiming that the Supreme Court's analysis in that case forbids the SEC to punish him under Rule 10b-5 for what he did.

We have before us both sensational facts and difficult issues of law and policy. The information Dirks passed on to his clients was no mere "tip": in two weeks of concerted effort, at times resembling something from detective fiction, Dirks investigated and confirmed rumors of massive fraud by the Equity Funding Corporation of America (Equity Funding), an insurance holding company whose stock traded on the New York Stock Exchange. Largely thanks to Dirks one of the most infamous frauds in recent memory was uncovered and exposed, while the record shows that the SEC repeatedly missed opportunities to investigate Equity Funding.

The issues of law and policy concern the role of private securities analysts who investigate corporate frauds. Securities analysts can be an important source of information for the public, supplementing an often overtaxed SEC enforcement staff and the press. There is some danger that the threat of liability under Rule 10b-5 may dampen the zeal of analysts in ferreting out the truth about corporate practices.

We recognize that concern, but we also recognize its limits. Private analysts may not keep information they have discovered from the SEC, while their clients dump fraudulent securities on an uninformed public. In this case, we defer to the SEC's judgment about what will best serve the long-range interests of the public and accomplish the statutory purposes behind Rule 10b-5. Therefore, we reject Dirks' argument based on Chiarella as well as several other challenges he presents.

I

Because the SEC has chosen merely to censure Dirks, and has not taken any other action against him in connection with this case, we need not consider whether all of Dirks' actions discussed by the SEC and the administrative law judge who first heard the case violated Rule 10b-5. A single willful violation of Rule 10b-5 or the statutes upon which it is based provides the SEC with discretion to impose the lightest administrative penalty available to it. See 15 U.S.C. § 78 o(b)(4)(D)(1976). Therefore, we confine our recitation of the facts to Dirks' least ambiguous actions.

The critical events in this case occurred in March 1973. During that period Raymond Dirks was an officer of Delafield Childs, Inc., a registered broker-dealer that served a clientele composed primarily of institutional investors. Dirks himself specialized in providing investment advice about the insurance industry, and he was apparently highly respected within the investment community for his knowledge of insurance companies and his willingness to go beyond mere financial data in evaluating investments.

Dirks regularly provided investment advice to a number of institutions that he knew had invested or might be interested in investing in insurance company stocks. As far as the record shows, he had no formal relationship with these institutions; he simply tried to be aware of their investment objectives and strategies, and to inform them whenever he developed information that he thought would be of interest to them. Under the custom of the industry, Dirks received no direct compensation from these clients. Rather, if they thought the information Dirks gave them was valuable, they would direct some of their brokerage business through Delafield Childs, thus giving Dirks' firm a chance to earn brokerage commissions.

On March 6, 1973 Dirks received a telephone call from Ronald Secrist, who had recently been fired from his job with Bankers National, a New Jersey life insurance company that had been acquired by Equity Funding four years earlier. Secrist told Dirks that he had information about fraud and illegality at Equity Funding, and the two men arranged a meeting for the next day.

Dirks and Secrist met for several hours on March 7. Secrist made a series of detailed but nearly incredible allegations about Equity Funding: mainly that one of its subsidiaries had created false insurance policies and records to inflate its sales figures, but also that it was selling partnerships in nonexistent real estate, that its top officers had Mafia connections which they used to threaten the lives of employees who objected to the fabrications, and that the accounting firm of Haskins Sells had dropped the Equity Funding account out of disagreement with the company's business practices. Although Secrist had neither personal knowledge nor documentation to support any of his charges, Dirks developed a hunch that Secrist was not merely a vindictive former employee trying to get back at the people who fired him. Therefore, Dirks decided to investigate Secrist's allegations.

Dirks' investigation had three major phases, the first two of which are not very important to the issues in this appeal. He began by examining publicly available data on Equity Funding's insurance sales. He compared Equity Funding's ratio of sales to sales force with that of its competitors, but he decided that the figures neither confirmed nor refuted the possibility that Equity Funding was fabricating insurance policies. Second, he contacted other people in the investment community who followed Equity Funding stock — many of them because they had invested in it — to see if they knew anything that could prove or disprove the rumors. This inquiry, too, was generally fruitless; most of the people with whom Dirks spoke did not believe there was any truth in Secrist's story. Through one informant Dirks did learn that Secrist's allegations about Haskins Sells were untrue. The accounting firm had lost Equity Funding's business to a competitor and would be glad to get it back.

Nevertheless, Dirks continued to talk to Secrist regularly, and he continued to suspect that some of Secrist's charges might be justified. Dirks also telephoned Equity Funding's chairman, Stanley Goldblum, who denied that there was any fraud at Equity Funding and invited Dirks to visit the company's headquarters in Los Angeles.

Dirks flew to Los Angeles on March 19 to begin the third phase of his investigation. He spent most of the day on March 20 with Patrick Hopper, a former vice president of Equity Funding who had retired to live the life of a wealthy beach bum. Hopper had been Secrist's superior for a time at Banker's National, and it was Hopper who had suggested to Secrist that he tell his story to Dirks after Secrist lost his job. Hopper told Dirks that Secrist tended to exaggerate things, but that he — Hopper — tended to believe the gist of Secrist's allegations about phony insurance. In 1971 Hopper had been following Equity Funding's life insurance sales on a week-by-week basis until late October, when the company stopped circulating weekly reports. When the final figures for that year appeared in early January, Hopper had noted that the final figure for life insurance sales was almost double what his running total had been in October. He also saw the sales figures for what had previously been the best of the company's five sales districts, and they were far less than a fifth of the total insurance sales figures for 1971.

Hopper also reported that he had attended a dinner in New York with several other Equity Funding officers, and that some of them had joked openly at the table about "the Y business" — according to Secrist, the euphemism by which the insurance fabrication program was known within the company. Beyond that, Hopper had no direct knowledge of fraud at Equity Funding. But on the afternoon of March 20, he and Dirks sought out another former Equity Funding employee, Frank Majerus, who Hopper suspected might know something concrete. After a great deal of coaxing, Majerus admitted that he had been involved in altering the company's insurance-in-force figures for 1970.

Dirks spent much of the next day, March 21, with the top management of Equity Funding. The company's officers continued to deny, and even to ridicule, the notion that there was anything amiss at Equity Funding. But over the next two days Dirks contacted four other men whose names had been given him by Hopper and Secrist. Two had worked as computer technicians for Equity Funding, one had worked for the company that programmed Equity Funding's computers, and one was still employed at Equity Funding. All four had independently come to the conclusion that the company's computer files contained large blocks of phony policies. The current employee, Donald Goff, had heard talk about phony insurance and seen figures for insurance-in-force for the years 1970-1972 that did not correspond to the figures ultimately announced by the company.

Throughout his investigation Dirks had been in contact with a number of investors and analysts — some because he hoped they might give him information or support, others because they were clients or potential clients and his records showed they were interested in information about Equity Funding, and still others simply because, having heard rumors about his investigation, they called him. On all occasions Dirks candidly discussed the status of his investigation with anyone who asked.

From Friday, March 23, through Monday, March 26, Dirks spoke repeatedly with members of the investment community. As events developed, these were the last two full days on which it was possible to buy or sell Equity Funding stocks on the New York Stock Exchange. During this period Dirks followed his prior practice of calling companies that he knew were interested in news about Equity Funding and returning calls from anyone else who tried to contact him. At one point, on March 26, Dirks telephoned a non-client, Loew's Inc., because he had heard that it had just purchased a large block of Equity Funding stock from one of his clients, and he wanted to make sure Loew's was aware of his doubts concerning the company. Unsurprisingly, the record shows that many of those with whom Dirks spoke on or after March 20 — by which time he had substantiated the key portions of Secrist's allegations — sold their Equity Funding securities as quickly as possible. They succeeded in unburdening themselves of approximately $15.5 million in Equity Funding stock and $1 million in convertible debentures by the time the New York Stock Exchange finally halted trading in Equity Funding securities.

It is not clear how many of those with whom Dirks spoke promised to direct some brokerage business through Delafield Childs to compensate Dirks, or how many actually did so. But it is clear that the custom of the industry provided that Dirks' regular clients would compensate him in that manner. The record shows that some of those with whom Dirks spoke during this period discussed the idea of directing business through Delafield Childs and did in fact use Delafield Childs as a broker for securities transactions during the ensuing weeks.

Meanwhile, during the entire week that Dirks was in Los Angeles investigating Equity Funding, he was also in touch regularly with William Blundell, the Wall Street Journal's Los Angeles bureau chief. Dirks kept Blundell up to date on the progress of the investigation and badgered him to write a story for the Wall Street Journal on the allegations of fraud at Equity Funding. Blundell, however, was afraid that publishing such damaging rumors supported only by hearsay from former employees might be libelous, so he declined to write the story. Nevertheless, on March 26 Blundell contacted the SEC's Los Angeles office and repeated what Dirks had told him over the course of the previous week. The SEC took no immediate action. Blundell then arranged a meeting for the next day, to include Dirks, Hopper, and the SEC staff in Los Angeles.

Thus, Dirks presented what he knew about Equity Funding to the SEC staff beginning on Tuesday, March 27, and continuing through the next day. The SEC took no action against Equity Funding on Tuesday, but late that day the New York Stock Exchange halted trading in Equity Funding stock after the Salomon Brothers firm lodged a complaint of "disorderliness" in the market for the stock. In less than two weeks — while Dirks was pursuing his investigation and spreading word of Secrist's charges — the trading price of Equity Funding shares had fallen from $26 to less than $15; the price had fallen more than $3 since the market opened on March 26.

On Wednesday, March 28, the Wall Street Journal printed an Equity Funding press release denying "rumors circulating in the financial community about the accuracy of statements by its life insurance subsidiary." That afternoon, while the Los Angeles staff was still closeted with Dirks, the SEC suspended trading in Equity Funding securities for ten days. Later that week, the Illinois and California insurance departments staged surprise inspections of Equity Funding. California impounded Equity Funding's corporate records, and Illinois investigators discovered that $20 million in corporate bonds supposedly owned by Equity Funding — approximately 70 percent of the assets of its major subsidiary — simply did not exist.

On April 2, the SEC filed a complaint against Equity Funding, and the Wall Street Journal published a front-page story written by Blundell but based largely on information assembled by Dirks. Equity Funding immediately went into receivership. Blundell was nominated for a Pulitzer Prize for his coverage of the Equity Funding scandal; Dirks became the object of the disciplinary proceeding on review in this case.

The SEC's disciplinary proceeding originally included not only Dirks but also five of his institutional clients who sold Equity Funding stock between March 12 and March 27. After lengthy hearings, an administrative law judge found that four of the clients and Dirks had violated Rule 10b-5 by trading on the basis of — or in Dirks' case, communicating — inside information. The administrative law judge imposed a censure on each of the four institutions and suspended Dirks from association with any registered broker or dealer for 60 days. The institutions did not appeal their censures to the Commission, but both Dirks and the SEC's Enforcement Division appealed — Dirks claiming that there was insufficient factual and legal basis for finding that he had violated Rule 10b-5, and the Enforcement Division seeking imposition of a more lengthy suspension.

The SEC rendered its decision on January 23, 1981. It concluded

Dirks came into the possession of material, nonpublic corporate information, from persons he knew were insiders, at a time when he knew that the information was confidential and not then publicly available. He communicated that information to those likely to trade before the information became generally available. In committing these acts, Dirks acted with scienter. Accordingly, we conclude that Dirks willfully aided and abetted violations by [his institutional clients] of Section 17(a) of the Securities Act and Section 10(b) of the Securities Exchange Act and Rule 10b-5 thereunder.

JA 1329. However, the SEC reduced the sanction imposed on Dirks to a mere censure, in order not to "hamper legitimate, investigative securities analysis" and in order to recognize Dirks' role in bringing the Equity Funding scandal to light and his theretofore unblemished record in the investment community. See JA 1330.

Dirks submitted a petition for review to this court pursuant to 15 U.S.C. § 78y(a) (1976).

II

Section 15(b) of the Securities Exchange Act authorizes the SEC to impose censures or other penalties on persons associated with registered broker-dealers if the person involved

has willfully violated any provision of the Securities Act of 1933, the Investment Advisers Act of 1940, the Investment Company Act of 1940, this chapter [the Securities Exchange Act of 1934], [or] the rules or regulations under any of such statutes * * *.

15 U.S.C. § 78 o(b)(4)(D) (1976). The three provisions upon which the SEC rested its decision, overlap to a large extent; of these, Rule 10b-5 is the most inclusive. It states:

It shall be unlawful for any person, directly or indirectly, * * *

(a) To employ any device, scheme, or artifice to defraud,

(b) To make any untrue statement of a material fact or to omit to state a material fact necessary in order to make the statements made, in the light of the circumstances under which they were made, not misleading, or

(c) To engage in any act, practice, or course of business which operates or would operate as a fraud or deceit upon any person, in connection with the purchase or sale of any security.

17 C.F.R. § 240.10b-5 (1981). The essence of the SEC's position, and the legal basis for its censure, is that Dirks and his clients had a duty under subsection (c) to disclose the material information they possessed or refrain from trading Equity Funding securities as long as the information had not been disseminated to the public.

The concept of a duty to "disclose-or-refrain" under Rule 10b-5 originated in the SEC's benchmark decision in Cady, Roberts Co., 40 S.E.C. 907 (1961). There the SEC held that a registered broker-dealer violated Rule 10b-5 by selling stock in Curtiss-Wright Corporation based on information that the broker-dealer firm received from one of its partners who was also a director of Curtiss-Wright. Corporate "insiders," such as the director, were bound not to take advantage of information "intended to be available only for a corporate purpose and not for the benefit of anyone" when they knew it was unavailable to those with whom they dealt. 40 S.E.C. at 912. Subsequent cases settled that, even if the insiders themselves did not trade, they were forbidden even to communicate nonpublic "inside information" to others who were likely to trade before the information became public, see SEC v. Texas Gulf Sulphur Co., 401 F.2d 833 (2d Cir. 1968), cert. denied, 394 U.S. 976, 89 S.Ct. 1454, 22 L.Ed.2d 756 (1969), and that Rule 10b-5 required those who received inside information from insiders to disclose it or refrain from trading, see Shapiro v. Merrill Lynch, Pierce, Fenner Smith, 495 F.2d 228 (2d Cir. 1974); Investors Management, Inc., 44 S.E.C. 633 (1971).

In other circumstances, the Supreme Court extended the duty to disclose or refrain to people who were not traditional insiders, in that they had no special access to information about the assets or plans of the corporations that issued the securities in which they traded. See Affiliated Ute Citizens v. United States, 406 U.S. 128, 92 S.Ct. 1456, 31 L.Ed.2d 741 (1972), which involved bank employees who purchased shares in a tribal trust fund from mixed-blood Ute Indians without disclosing that there was a secondary market for the shares, at higher prices, among non-Indians, and SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 84 S.Ct. 275, 11 L.Ed.2d 237 (1963), which involved an investment adviser who purchased stock for his own account just before publishing a recommendation that his clients buy the stock. These cases became known as "market information" (as opposed to inside information) cases, because the information at issue related solely to the market for the securities rather than their intrinsic value. See generally Fleischer, Mundheim Murphy, An Initial Inquiry into the Responsibility to Disclose Market Information, 121 U.PA.L.REV. 798 (1973).

The foregoing cases do not always make clear, however, what circumstances trigger a duty to disclose or refrain from trading. Variants of two general theories appear, often within the same opinion. On the one hand, some of the language in the cases finds the element of unfairness and fraud in conflicts of interest on the part of traders or their informants who profit at the expense or to the exclusion of those who have placed trust in them. The conflict of interest may arise from a traditional fiduciary relationship, as between a corporate director and the corporation's shareholders, or a similar relationship of trust, as between employers and employees or investment bankers and their clients. On the other hand, some of the cases imply that the securities laws impose a duty to disclose or refrain from trading based on the nature of the undisclosed information. The theory is that all investors should have equal access to information that a reasonable investor would consider material to investment decisions, and that any trade in which only one party had an opportunity to learn and did learn such information is inherently unfair.

The Supreme Court has recently sought to resolve the tension between the "fiduciary" theory and the "information" theory of Rule 10b-5's disclose-or-refrain prescription. Chiarella v. United States involved an employee of a financial printer. Because he worked setting type for documents to be used in connection with tender offers, Chiarella was able to learn which companies were about to be the targets of tender offers at above-market prices. He would buy stock in the target companies before the offers were announced and sell afterwards at a considerable profit. When his activities were discovered, the government prosecuted him under 15 U.S.C. § 78ff(a) (1976), which subjects those who willfully violate Rule 10b-5 and Section 10(b) of the Securities Exchange Act to criminal liability. Chiarella was convicted and the Second Circuit affirmed the conviction, holding that anyone who regularly receives material, nonpublic information has a duty to disclose the information or refrain from trading. United States v. Chiarella, 588 F.2d 1358, 1365 (2d Cir. 1978). The Supreme Court reversed.

Five Justices joined a majority opinion written by Justice Powell. It begins by noting that neither Section 10(b) nor its legislative history, nor the history of Rule 10b-5 itself, offers clear authority for imposing a disclose-or-refrain obligation on anyone. The opinion then traces the progression of cases that recognize such an obligation, rationalizing them on the principle of fiduciary duty, and rejecting the argument that there is "a general duty between all participants in market transactions to forgo actions based on material, nonpublic information," 445 U.S. at 233, 100 S.Ct. 1117. The majority opinion states:

Thus, administrative and judicial interpretations have established that silence in connection with the purchase or sale of securities may operate as a fraud actionable under § 10(b) despite the absence of statutory language or legislative history specifically addressing the legality of nondisclosure. But such liability is premised upon a duty to disclose arising from a relationship of trust and confidence between parties to a transaction. Application of a duty to disclose prior to trading guarantees that corporate insiders, who have an obligation to place the shareholder's welfare before their own, will not benefit personally through fraudulent use of material, nonpublic information.

Id. at 230, 100 S.Ct. at 1115.

The majority opinion deals with the liability of those who receive information, see Shapiro v. Merrill Lynch, supra, in a footnote appended to the above statement:

"Tippees" [persons who receive information] of corporate insiders have been held liable under § 10(b) because they have a duty not to profit from the use of inside information that they know is confidential and know or should know came from a corporate insider. * * * The tippee's obligation has been viewed as arising from his role as a participant after the fact in the insider's breach of a fiduciary duty.

445 U.S. at 230 n. 12, 100 S.Ct. at 1115 n. 12.

Other Justices, however, advanced slightly different theories of when a duty to disclose-or-refrain should arise. Four Justices stated that they thought the duty could be grounded on something other than a relationship of trust between the parties to a securities transaction. The Chief Justice and Justice Brennan, each writing separately, stated their belief that "a person violates § 10(b) whenever he improperly obtains or converts to his own benefit nonpublic information which he then uses in connection with the purchase or sale of securities." 445 U.S. at 239, 100 S.Ct. at 1120 (Brennan, J., concurring in the judgment); cf. id. at 249, 100 S.Ct. at 1125 (Burger, C. J., dissenting) (duty to disclose-or-refrain arises "when an informational advantage is obtained * * * by some unlawful means"). Justice Blackmun, joined by Justice Marshall, agreed that misappropriation of information could give rise to a duty to disclose-or-refrain, but he stated that he would go further, imposing the duty on all "persons having access to confidential material information not available to others generally," id. at 251, 100 S.Ct. at 1126, or indeed perhaps on anyone who "turn[ed] secret information to account for personal profit," id. at 249, 100 S.Ct. at 1125. And Justice Stevens, whose vote was necessary to give Justice Powell's opinion a majority, indicated that he considered it an open question whether the duty to disclose-or-refrain could be premised on yet another theory: a duty of confidentiality owed to someone other than a purchaser or seller of the security at issue. See id. at 237-238, 100 S.Ct. at 1119.

The Chiarella majority's statement of its holding, nonetheless, is narrower than the reasoning in any of the opinions: "a duty to disclose under § 10(b) does not arise from the mere possession of nonpublic market information." 445 U.S. at 235, 100 S.Ct. at 1118. Although it is clear that the reasoning of the Justices in Chiarella must inform our consideration of the question before us, the narrowness of the holding in Chiarella cautions against applying the majority's rationale rigidly or exclusively to contexts involving information other than "market information. The disputes among the Justices in Chiarella lead us to doubt that it will be the Supreme Court's last word on Rule 10b-5.

Therefore, reading Chiarella in light of the case law that preceded it, and extracting those views that seem to command a clear majority of the Court, we take the following lessons from Chiarella: Rule 10b-5 and its statutory sources, standing alone, do not require "any person" who is a party to a securities transaction to disclose all material, nonpublic information or refrain from trading, and a mere failure to disclose material information, absent other compelling legal circumstances, does not "operate as a fraud." Thus, the "information" theory is rejected. Because the disclose-or-refrain duty is extraordinary, it attaches only when a party has legal obligations other than a mere duty to comply with the general antifraud proscriptions in the federal securities laws.

The duty of loyalty owed by corporate officers to a corporation and its shareholders is one obvious example of a legal duty that, in conjunction with Rule 10b-5, gives rise to an obligation to disclose-or-refrain. But the law of fiduciary duties is not coextensive with Rule 10b-5. See Goldberg v. Meridor, 567 F.2d 209, 220-221 (2d Cir. 1977), cert. denied, 434 U.S. 1069, 98 S.Ct. 1249, 54 L.Ed.2d 771 (1978). Fiduciaries may violate Rule 10b-5's disclose-or-refrain prescription even if their actions do not violate their fiduciary duties as defined by state law. E.g., Chasins v. Smith, Barney Co., 438 F.2d 1167 (2d Cir. 1971). Furthermore, the disclose-or-refrain rule extends to other relationships of trust, whether or not they are subject to fiduciary duties under state law. See, e.g., Affiliated Ute Citizens, supra (employees of a bank acting as stock transfer agent who "facilitated" sales of shares); Capital Gains Research Bureau, supra (financial adviser, under the Investment Advisers Act of 1940); Zweig v. Hearst Corp., 594 F.2d 1261 (9th Cir. 1979) (financial columnist).

As we show in Part III, Dirks had a duty to disclose what he knew about Equity Funding to the public or to refrain from trading (or fostering trades) in Equity Funding securities. He violated that duty, and with it Rule 10b-5, when he passed his information to investors who were likely to sell their Equity Funding securities before the public learned about the Equity Funding fraud.

III

The SEC's opinion states:

It is well established that corporate insiders who trade on the basis of material, non-public information, without first disclosing it, violate Section 10(b) of the Securities Exchange Act and Rule 10b-5 thereunder. "Tippees" of corporate insiders, who themselves trade, are equally liable. Moreover, both corporate insiders and their "tippees" are liable for trading violations of those whom they, in turn, "tip." The dissemination of inside information to those who utilize it in dealing with uninformed public investors constitutes an act, practice, or course of conduct which operates as a fraud on investors. Even where the tipper does not himself engage in trading, he aids and abets the violation by providing the means by which the wrongful act occurs.

JA 1321 (footnotes omitted). The SEC's theory of this case rests almost entirely on the fourth sentence of this passage. In the SEC's view, Dirks aided and abetted violations of Rule 10b-5 by selectively disseminating inside information to investors likely to sell their Equity Funding stock without making a public disclosure of what they learned from Dirks.

Relying on Chiarella, Dirks attacks this line of reasoning in a number of ways. He argues that his actions did not violate Rule 10b-5 unless the corporate insiders who disclosed the information in the first place also violated Rule 10b-5, and that they did not violate Rule 10b-5 because their fiduciary duties under California law did not require them to keep the information they gave Dirks confidential. He also argues that information is not "inside information" unless there is an enforceable duty to keep it confidential for corporate purposes, or that selective dissemination does not operate as a fraud unless those responsible have a duty to maintain confidentiality. Each of these arguments comes down to the same thing. After Chiarella, Dirks argues, neither he nor his informants can be held to violate Rule 10b-5 unless they breached some legal duty to Equity Funding to keep what they knew of its crimes secret. And Dirks asserts that neither California law nor any other law creates such a duty.

The SEC does not dispute that Dirks' informants — Secrist, Hopper, Goff, et al. — had no duty under California law to maintain confidentiality about the Equity Funding fraud, and we may assume, without deciding the question, that California law relieves corporate employees of their duty to keep corporate information confidential when it relates to crimes or frauds perpetrated by the corporation. Even if none of Dirks' informants breached their fiduciary duties in disclosing what they knew of Equity Funding's false insurance scheme to Dirks, Rule 10b-5 may prohibit Dirks and his clients from profiting on the basis of their knowledge before public exposure of the fraud.

First, there is a clear fallacy in Dirks' argument that he cannot have violated Rule 10b-5 because his informants did not violate their fiduciary duties as defined by California law. While the standards of Rule 10b-5 have always duplicated state fiduciary obligations to a certain extent, courts have never regarded the two as identical. In Chasins v. Smith, Barney Co., supra, the court sustained a verdict that the defendant had violated Rule 10b-5 but that it had not violated New York's law of fiduciary duty. And in Goldberg v. Meridor, supra, the Second Circuit explained that, in a case involving failure to disclose material information, the existence of parallel state remedies neither defined nor precluded application of Rule 10b-5. See 567 F.2d at 221. The Chiarella majority focused on the existence of a set of fiduciary obligations as a prerequisite to the addition of a disclosure-or-refrain duty, but it did not hold that breach of the fiduciary obligations was required to bring Rule 10b-5 to bear on a case, nor did it hold that state fiduciary law was the sole source of the duty to disclose-or-refrain.

To say that Dirks' informants did not breach their fiduciary obligations under California law is not to say that the federal government must permit them to enter the market for securities and, on the basis of their knowledge that Equity Funding was defrauding the public, extract a profit from uninformed investors. No matter what standard of conduct state law sets, Rule 10b-5 may require those who are fiduciaries under state law to disclose material information they have learned in the course of their fiduciary relationships before trading in securities. Other courts interpreting Chiarella have reached similar conclusions. See Staffin v. Greenberg, 672 F.2d 1196, 1202 (3d Cir. 1981); United States v. Newman, 664 F.2d 12, 17-18 (2d Cir. 1981). And under the doctrine of Shapiro v. Merrill Lynch, supra, and Investors Management, supra, the obligations of corporate fiduciaries pass to all those to whom they disclose their information before it has been disseminated to the public at large. Thus Dirks (and through him his clients) became subject to his informants' disclose-or-refrain obligation.

That is not to say that it will always be irrelevant under Rule 10b-5 that no fiduciary duty was breached in communicating information. In this case — assuming still that Dirks is correct about California law — California had good reason to release corporate employees from their duty to keep corporate information confidential when the information concerns corporate crimes or frauds. Both the state and the public at large have an interest in exposing corporate misconduct. Ordinarily, we would expect that official law enforcement agencies would be sufficient for that task, but this case shows that the organs of government are not always able to accomplish swift investigation of possible crimes. The press also has an historic role in discovering and exposing wrongdoing, but here, too, the press failed to move as quickly as Dirks, who had more immediate financial and reputational incentives to discover the truth about Equity Funding. Interpreting Rule 10b-5 so as to categorically prevent private sector investigators from contacting corporate employees and asking them about crimes or frauds that their employer is committing may not, in the end, serve the public interest.

To paraphrase Justice Cardozo in another context: With respect to imposition of the disclose-or-refrain rule where fiduciary obligations are not violated, we do not fix the outermost line. Wherever the line may be, this case is within it. Steward Machine Co. v. Davis, 301 U.S. 548, 591, 57 S.Ct. 883, 892, 81 L.Ed. 1279 (1937). In the first place, we do not write on a clean slate. The SEC, charged with administering the nation's securities laws, is clearly entitled to deference when it interprets its own regulations in an administrative adjudication. See Udall v. Tallman, 380 U.S. 1, 16-17, 85 S.Ct. 792, 801, 13 L.Ed.2d 616 (1965). It has decided to censure Dirks under Rule 10b-5, and we cannot say that the SEC's disposition is "inconsistent with the statutory mandate" behind Rule 10b-5 or that it "frustrate[s] the policy that Congress sought to implement." See FEC v. Democratic Senatorial Campaign Committee, 454 U.S. 27, 32, 102 S.Ct. 38, 42, 70 L.Ed. 2d 23 (1981); SEC v. Sloan, 436 U.S. 103, 118, 98 S.Ct. 1702, 1711, 56 L.Ed.2d 148 (1978).

Even more important, Dirks himself had obligations to the SEC and to the public completely independent of any obligations he acquired under the Shapiro v. Merrill Lynch doctrine. Those obligations, implicit in the scheme of broker-dealer registration under the federal securities laws, provide a basis for imposing a duty to disclose-or-refrain on Dirks even if we would not impose it on his sources at Equity Funding.

The Securities Exchange Act of 1934 subjects registered broker-dealers, like Delafield Childs, and those associated with them, like Dirks, to myriad duties not imposed on corporate officers or other members of the general public. These include a registration requirements, 15 U.S.C. § 78 o(a)-(b) (1976), and special requirements to avoid practices that would operate as a fraud, id. § 78 o(c)(1)-(6); see 17 C.F.R. §§ 240.10b-3, .15c1-2, .15c1-4(1981). Even before the promulgation of Rule 10b-5, the SEC and the courts held that, because of the importance of broker-dealers within the securities markets and the general scheme of securities regulation, broker-dealers were required to meet a high standard of ethical behavior in their activities. See Charles Hughes Co. v. SEC, 139 F.2d 434 (2d Cir. 1943), cert. denied, 321 U.S. 786, 64 S.Ct. 781, 88 L.Ed. 1077 (1944); Duker Duker, 6 S.E.C. 386 (1939). The Second Circuit has even stated that Rule 10b-5, as applied to broker-dealers, might reach the sort of nondeceptive breach of fiduciary duty that does not come within the rule as applied to corporate managers. Compare O'Neill v. Maytag, 339 F.2d 764, 768-769 (2d Cir. 1964), with Birnbaum v. Newport Steel Corp., 193 F.2d 461 (2d Cir.), cert. denied, 343 U.S. 956, 72 S.Ct. 1051, 96 L.Ed. 1356 (1952), and Santa Fe Industries, Inc. v. Green, 430 U.S. 462, 97 S.Ct. 1292, 51 L.Ed.2d 480 (1977).

To be sure, the main focus of the SEC's regulation of broker-dealers has been to ensure that broker-dealers treat their customers honestly and fairly. But fairness to their customers is not the only obligation that the securities laws impose on broker-dealers. The Supreme Court has held repeatedly that the scheme of federal securities regulation, taken as a whole, has a major objective "to achieve a high standard of business ethics * * * in every facet of the securities industry." United States v. Naftalin, 441 U.S. 768, 775, 99 S.Ct. 2077, 2082, 60 L.Ed.2d 624 (1979) (emphasis in original); SEC v. Capital Gains Research Bureau, 375 U.S. 180, 186-187, 84 S.Ct. 275, 279-280, 11 L.Ed.2d 237 (1963); cf. Ernst Ernst v. Hochfelder, 425 U.S. 185, 195, 96 S.Ct. 1375, 1382, 47 L.Ed.2d 668 (1976). And in enacting the first major piece of securities legislation, the Securities Act of 1933, Congress clearly announced the fundamental purpose of its regulation of the securities markets:

The purpose of this bill is to protect the investing public and honest business. * * The aim is to prevent further exploitation of the public by the sale of unsound, fraudulent, and worthless securities through misrepresentation; to place adequate and true information before the investor; to protect honest enterprise, seeking capital by honest presentation, against the competition afforded by dishonest securities offered to the public through crooked promotion; to restore the confidence of the prospective investor in his ability to select sound securities; to bring into productive channels of industry and development capital which has grown timid to the point of hoarding; and to aid in providing employment and restoring buying and consuming power.

S.Rep. No. 47, 73d Cong., 1st Sess. 1 (1933); see United States v. Naftalin, supra, 441 U.S. at 775-776, 99 S.Ct. at 2082-2083.

There is no identifiable segment of the securities industry whose ethical conduct is more crucial to the attainment of Congress' goals than the ethical conduct of broker-dealers. The high standard of business ethics required by the securities laws restrains broker-dealers not only in their dealings with their customers, but also in their dealings with the SEC and the public at large, at least when the alternative is direct frustration of the federal scheme of securities regulation.

This case provides a clear example of a duty to the SEC and the public created by the ethical standard that applies to broker-dealers. Dirks discovered an ongoing fraud that meant, at the very least, that the stock of Equity Funding was being traded at prices far in excess of the value which any reasonable investor, fully informed, would place upon it. As the public learned the full extent of the Equity Funding fraud, its stock became first illiquid, then worthless. By the time Dirks entered the picture, a loss amounting to many millions of dollars to the investing public was already unavoidable. But, instead of going straight to the SEC with what he had learned, Dirks paused to make certain that his clients and those who might some day become his clients — all large institutional investors — were not the ones left "holding the bag." Dirks therefore risked that much of the loss would be borne by relatively smaller, less sophisticated investors.

Dirks' conduct was fundamentally inconsistent with "prevent[ing] further exploitation of the public by the sale of unsound, fraudulent, and worthless securities." There is no question that securities industry professionals may legitimately engage in securities analysis for the benefit of their clients, and they may accumulate and sell information in order to make a profit. In the common run of day-to-day business, professional securities analysis is perhaps the best way to aggregate and evaluate information in the securities markets; the analysts thus serve themselves, their clients, and the public interest in efficient capital markets all at the same time. But it is intrinsic in our notion of "business ethics" that at some point we are not content to let securities analysts' quest for fees and commissions define their obligations to the public at large. In this case, the role of analysis-for-hire ceased when its unavoidable result was to foster the sale of "unsound, fraudulent, and worthless securities" to the uninformed public. At the very least, Dirks owed a duty to report what he had learned to the SEC and not to foster such sales.

That fundamental obligation would inhere in the securities laws enacted by Congress even if they included no generally applicable antifraud provisions. As it is, Rule 10b-5 operates in conjunction with the ethical duty implied by the other provisions of the securities laws to forbid broker-dealers, and those who obtain information from them, to trade in unsound, worthless, or fraudulent securities without first disclosing the material information they possess. The public and the SEC both have a right to expect good-faith cooperation from securities industry professionals in accomplishing the core purposes of the securities laws. Dirks violated his duties to the SEC and the public by failing to report promptly what he knew. When his clients sold their Equity Funding stock without first disclosing the information Dirks had given them, they became "participant[s] after the fact" in Dirks' breach of duty, see Chiarella, supra, 445 U.S. at 230 n. 12, 100 S.Ct. at 1115 n. 12, and they violated Rule 10b-5.

IV

Dirks raises two further challenges to the SEC's decision in this case. First, he disputes that the information he passed on to his clients constituted "material facts." Second, he claims that the SEC applied the wrong scienter standard in finding that he had aided or abetted his clients' violation of Rule 10b-5. Neither of these contentions has merit.

A. Material Facts

The disclose-or-refrain obligation does not extend to everything a trader knows or believes. It applies only to "material facts" that have not yet been disseminated to the public. See SEC v. Texas Gulf Sulphur Co., supra, 401 F.2d at 848-850; List v. Fashion Park, Inc., 340 F.2d 457, 462 (2d Cir.), cert. denied, 382 U.S. 811, 86 S.Ct. 23, 15 L.Ed.2d 60 (1965). Dirks argues both that the information he gave his clients was not "factual" enough to come within the disclose-or-refrain rule and that it could not be deemed "material" to his clients' decisions to sell their Equity Funding stock.

To the extent that Dirks attacks the SEC's findings of fact, our review is strictly limited. "The findings of the Commission as to the facts, if supported by substantial evidence, are conclusive." 15 U.S.C. § 78y(a)(4) (1976). We review the SEC's findings only to judge whether there is substantial evidence in the record from which a reasonable person might find a violation of Rule 10b-5 by a preponderance of the evidence. See Steadman v. SEC, 450 U.S. 91, 101 S.Ct. 999, 67 L.Ed.2d 69 (1981).

Dirks challenges both the legal and factual adequacy of the SEC's finding that the information he communicated was material. TSC Industries, Inc. v. Northway, Inc., 426 U.S. 438, 449, 96 S.Ct. 2126, 2132, 48 L.Ed.2d 757 (1976), states the test for materiality. An omission is material if "the disclosure of the omitted fact would have been viewed by the reasonable investor as having significantly altered the `total mix' of information made available." Although the SEC's opinion cites the correct language in TSC Industries, see JA 1324, Dirks claims that the SEC was really applying a test derived from the Texas Gulf Sulphur case, weighing the probability that the omitted fact is true against the magnitude of its effect on the market if it is true, see 401 F.2d at 849. According to Dirks, the "probability/magnitude" test makes even the least probable rumors "material" if they are sensational enough, and Dirks maintains that TSC Industries rejected the probability/magnitude test in Texas Gulf Sulphur.

Although the TSC Industries test is more sensitive than the probability/magnitude test, since it takes into account not only how a reasonable investor would evaluate a single piece of information in isolation but also how one fact relates to all the others available to a reasonable investor, it is simply a case of the greater including the lesser. TSC Industries did not reject any balancing of probability and magnitude of effect; indeed, it is hard to conceive of a better description of what a "reasonable investor" would do. It merely injects another factor into the overall consideration — how the probability/magnitude calculation with respect to one piece of information changes the calculation as applied to all available information. The SEC's opinion (and the hearing examiner upon whom the SEC relied) takes both factors into account, as have many courts since TSC Industries.

Dirks asserts that one of the footnotes in the SEC's opinion demonstrates further departure from the TSC Industries standard and also shows that the SEC based its decision on factual predicates that were not established by substantial evidence on the record. The SEC's opinion states, "The fact that some tipped advisers `acted immediately or very shortly after receipt of the information to effect sales * * * is in itself evidence of its materiality.'" JA 1324 n. 33 (citing Investors Management, supra, 44 S.E.C. at 642). The SEC regarded this as only one indication that investors, who might have been reasonable, considered the information Dirks gave them material. Many courts have looked to similar evidence, both before and after TSC Industries. Where, as here, the SEC does not treat evidence of actual behavior as dispositive, TSC Industries surely does not forbid adverting to it.

Not all of Dirks' clients sold their Equity Funding stock immediately or shortly after receiving his information. Some continued to hold Equity Funding securities even after learning of Dirks' suspicions, and at least one sophisticated investor, Loew's, Inc., continued to buy Equity Funding stock after being warned by Dirks of the possibility of fraud. Yet those who ignored Dirks' information were generally those with whom Dirks spoke before the crucial period of March 21-26, when Dirks obtained corroboration for Secrist's charges; investors with whom Dirks communicated during the final week of his investigation, especially on March 23 and March 26, understandably tended to sell whatever Equity Funding stock they could.

Since the SEC imposed the mildest penalty available to it under 15 U.S.C. § 78 o(b)(4) (1976), the evidence in the record is substantial enough to support the disposition if a reasonable person could infer from it the facts necessary to find a single violation of Rule 10b-5 by a preponderance of the evidence. The record before the SEC easily passes that test. The extensive sales by Dirks' clients on March 23 and March 26-27 render irrelevant any ambiguities in the record relating to sales before those dates.

Finally, Dirks argues that, as a matter of law, the information he communicated was not specific enough to rise to the level of "fact." Unless a "tip" comes from a source whose word would not be doubted, there can be no violation of Rule 10b-5 for failure to disclose information that "lack[s] the basic elements of specificity." See SEC v. Monarch Fund, 608 F.2d 938, 942 (2d Cir. 1979). But by the end of Dirks' investigation there is no doubt that the information he possessed and passed on to his clients had enough specificity to satisfy Rule 10b-5. He knew of specific blocks of false insurance, and he had a very detailed idea of how Equity Funding had gone about fooling its auditors and regulators. Thus, the information involved in this case is totally unlike the general rumors that the Monarch Fund court found not specific enough to support Rule 10b-5 liability. Cf. Elkind v. Liggett Myers, Inc., 635 F.2d 156, 164 (2d Cir. 1980) (comments by corporate officers like "we expect another good year" were not specific enough to mislead the sophisticated investors to whom they were addressed); Pachter v. Merrill Lynch, Pierce, Fenner Smith, 444 F.Supp. 417, 422 (E.D.N.Y.), aff'd mem., 594 F.2d 852 (2d Cir. 1978) (a broker who had heard rumors of fraud at Equity Funding did not have a duty to disclose them to a client who submitted an unsolicited "buy" order on the morning of March 27).

Therefore, we affirm the SEC's determination that Dirks communicated "material facts" to his clients.

B. Aiding or Abetting Scienter

To be liable for aiding or abetting a violation of Rule 10b-5, the purported aider-or-abettor must have "a general awareness that his role was part of an overall activity that was improper" and must "knowingly and substantially assist[ ] the principal violation." Investors Research Corp. v. SEC, 628 F.2d 168, 178 (D.C. Cir.), cert. denied, 449 U.S. 919, 101 S.Ct. 317, 66 L.Ed.2d 146 (1980); Woodward v. Metro Bank of Dallas, 522 F.2d 84, 94 (5th Cir. 1975). The SEC's opinion contains the somewhat conclusory finding that "Dirks did act with the requisite scienter." JA 1328.

Dirks attacks the SEC's scienter finding on two grounds. First, he argues that at worst he was reckless as to the possibility that his clients would trade without disclosing the information he gave them, and that Investors Research does not permit aiding-or-abetting liability for mere recklessness. Second, he argues that the fact that he was constantly badgering the Wall Street Journal's Los Angeles bureau to do a story on Equity Funding rebuts any evidence as to scienter, as it shows that he could not have been aware that his role was part of an overall activity that was improper.

Even if all the record showed were recklessness, the SEC would be justified in censuring Dirks. In both Ernst Ernst v. Hochfelder, supra, 425 U.S. at 193 n. 12, 96 S.Ct. at 1381 n. 12, and Aaron v. SEC, 446 U.S. 680, 686 n. 5, 100 S.Ct. 1945, 1950 n. 5, 64 L.Ed.2d 611 (1980), the Supreme Court expressly reserved the question whether the Rule 10b-5 scienter standard includes recklessness. But the overwhelming rule in the Courts of Appeals — including the Fifth Circuit, whose Metro Bank language this circuit borrowed in Investors Research — is that recklessness satisfies the Rule 10b-5 scienter requirement. Investors Research and Metro Bank do not imply that "knowingly * * * assist" or "general awareness" require a higher standard for aiding or abetting liability than the general scienter standard required by Ernst Ernst. Recklessness would have been enough.

Dirks cannot maintain, however, that he did not know that his clients would sell their Equity Funding stock on the basis of his information without making a disclosure, and his arguments about recklessness or his good faith in trying to get the Wall Street Journal to publish a story miss the point about what he did that was wrong. Dirks was in the securities business. He made his living collecting information, analyzing it, and selling his work product to investors who would consider it valuable. The record shows that Dirks spent almost half his working hours in Los Angeles on the phone to clients or potential clients, telling them about Equity Funding. The record also shows that Dirks expected to be compensated for his efforts, and that indeed he was compensated by some of those to whom he passed information. The only basis upon which he could have expected compensation was that his clients would consider his information worth paying for, and the only information his clients were willing to pay for was information that would help them make money in the securities markets. The Equity Funding information would have been worth little to them unless they could sell their holdings, and it would have lost most of its value if they had disclosed it before selling.

The Second Circuit, in Elkind v. Liggett Myers, supra, has held that the scienter standard in Rule 10b-5 was satisfied when corporate officers knowingly disclosed information they knew was material and not public to one who might "reasonably be expected to use it to his advantage." 635 F.2d at 167. The Liggett Myers officers, who had a duty to keep corporate information confidential or disclose it to all shareholders, acted knowingly with respect to disclosure, even though they may only have been reckless or negligent with respect to the trading of those whom they informed. Dirks also acted knowingly when he passed on his information to clients before going to the SEC, in violation of his duty to the public and the SEC and in violation of his informants' disclose-or-refrain obligations. Therefore it is not precisely relevant whether Dirks subjectively "knew" that his clients would trade. He knowingly took improper actions that put parties who were reasonably likely to trade without disclosure in a position to do so.

The record thus amply supports the SEC's finding that Dirks acted with the requisite scienter for aiding or abetting liability under Rule 10b-5.

V

By selling his information to his clients before informing the SEC, Dirks deepened the Equity Funding tragedy at the same time he was working to end it. The SEC, appropriately, has expressed the hope that its disposition in this case will not discourage legitimate efforts by securities analysts to discover the truth and communicate it to their clients. We concur. But the SEC has also concluded that Dirks' actions in this case aided and abetted a violation of Rule 10b-5, and we defer to its judgment in making the policy choices required for this disposition. Securities analysts associated with registered broker-dealers must respect the disclose-or-refrain obligations of their corporate sources, and they must stop short of transferring the losses from a fraud to those least able to avoid them. Accordingly, the petition for review is dismissed.

* * * * * *

(b) To use or employ, in connection with the purchase or sale of any security registered on a national securities exchange or any security not so registered, any manipulative or deceptive device or contrivance in contravention of such rules and regulations as the Commission may prescribe as necessary or appropriate in the public interest or for the protection of investors.