JOHN M. WALKER, Jr., Circuit Judge:

In 1996, plaintiffs-appellants filed this securities fraud class action, alleging violations of sections 10(b) and 20(a) of the Securities Exchange Act (“the 1934 Act”) and Rule 10b-5 promulgated thereunder. In two opinions issued in 1998, the district court dismissed both the original complaint and the plaintiffs’ amended complaint pursuant to Fed.R.Civ.P. 12(b)(6) and 15 U.S.C. § 78u-4(b)(3)(A) for failure to plead with sufficient particularity facts supporting a strong inference that the defendants had acted fraudulently. See Novak v. Kasaks, 997 F.Supp. 425 (S.D.N.Y.1998) (“Novak I ”) (dismissing original complaint); Novak v. Kasaks, 26 F.Supp.2d 658 (S.D.N.Y.1998) (“Novak II”) (dismissing amended complaint). On appeal, the appellants contend that the district judge erred in granting the defendants’ motions to dismiss.

In light of Second Circuit precedent and the provisions of the Private Securities Litigation Reform Act (“PSLRA”), we hold that the district court erred in: (1) concluding that the plaintiffs had failed to plead sufficient facts to support a strong inference of fraudulent intent; and (2) imposing an exceedingly onerous burden on the plaintiffs with respect to their obligation to plead facts with particularity. We see no persuasive alternative grounds for upholding the district court’s dismissal of the complaint. Accordingly, we vacate the judgment of the district court and remand for further proceedings consistent with these determinations. In addition, we instruct the district court to allow the plaintiffs to replead to the extent they wish to do so in light of this opinion.

BACKGROUND

On April 25,1996, plaintiffs Carol Novak and Robert Nieman brought this action on behalf of all purchasers of the common stock of the AnnTaylor Stores Corporation between February 3, 1994, and May 4, 1995 (the “Class Period”). In their complaint, the plaintiffs named two groups of defendants: (1) the AnnTaylor defendants, both the corporation itself — which, through its wholly-owned subsidiary, defendant AnnTaylor, Inc., is a specialty retailer of women’s clothing, shoes, and accessories— and several officers at the highest level of management; and (2) the Merrill Lynch defendants, a group of entities and individuals that collectively held a dominant share of AnnTaylor stock and sold a significant fraction of their holdings during the Class Period.

The complaint — in both its original and amended forms — essentially alleges that, during the Class Period, the defendants *304made, or controlled others who made, materially false and misleading statements and omissions concerning the financial performance of AnnTaylor (“the Company”), primarily by failing properly to account for millions of dollars of inventory. According to the plaintiffs, the defendants knowingly and intentionally issued financial statements that overstated AnnTaylor’s financial condition by accounting for inventory that they knew to be obsolete and nearly worthless at inflated values and by deliberately failing to adhere to the Company’s publicly stated markdown policy. The following facts are taken largely from the plaintiffs’ complaint.

The plaintiffs’ specific allegations focus on AnnTaylor’s so-called “Box and Hold” practice, whereby a substantial and growing quantity of out-of-date inventory was stored in several warehouses during the Class Period without being marked down. Internal Company documents (“Weekly Reports”)' — distributed at regular Monday morning merchandise meetings in which the AnnTaylor defendants participated— distinguished between regular inventory and “Box and Hold” inventory. According to the complaint, these reports demonstrated that: (1) much of the “Box and Hold” inventory was several years old and thus unlikely to be sold at full price, if at all; and (2) the levels of such inventory grew significantly during the Class Period, from about 10% to about 34% of total inventory. However, AnnTaylor’s public financial statements did not distinguish between types of inventory, nor did Ann-Taylor write off any of the “Box and Hold” inventory during the Class Period, allegedly in violation of Generally Accepted Accounting Principles (“GAAP”) that required markdowns under these circumstances. Instead, the defendants made or caused to be made a series of positive statements to the public about the status of AnnTaylor’s inventories, describing them at various points during the Class Period as “under control,” “in good shape,” and at “reasonable” or “expected” levels; stating that “no major or unusual markdowns were anticipated”; and attributing rising levels of inventory to growth, expansion, and planned future sales.

The plaintiffs contend that this course of conduct amounts to securities fraud. Had AnnTaylor taken appropriate write-downs, they argue, the Company’s earnings would have been substantially lower than reported. Thus, the AnnTaylor defendants’ alleged deception painted too rosy a picture of the Company’s current performance and future prospects and kept the company’s stock price at an artificially high level during the Class Period. According to the amended complaint, during this time,

many AnnTaylor executives demanded that [the individual AnnTaylor] defendants ... end the Box & Hold practice as it made no business sense and was growing out of control. Defendants’ response ... was that AnnTaylor could not “afford” to eliminate or write-down the Box & Hold inventory because doing so would “kill” the Company’s reported financial results and/or profit margins and damage the Company on “Wall Street.”

Ultimately, the defendants were forced to publicly acknowledge serious inventory problems — i.e., that inventories were too high and liquidation would result in much lower fiscal 1995 earnings than expected— at which point AnnTaylor stock prices fell precipitously, to the plaintiffs’ detriment.

On July 1, 1996, in response to these allegations, the defendants moved to dismiss the action, and on August 16, 1996, the district judge granted a motion by the defendants to stay all discovery pending a ruling on the motions to dismiss pursuant to 15 U.S.C. § 78u-4(b)(3)(B).

On March 10, 1998, the district court issued an opinion and order granting the defendants’ motions to dismiss the complaint. See Novak I, 997 F.Supp. at 426. The court concluded that “the fatal defect in the complaint lies in its allegations of scienter.” Id. at 430. Specifically, the plaintiffs had “fail[ed] to plead facts giving *305rise to a strong inference of fraudulent intent” in that they did not “allege with sufficient specificity that ... defendants ... were aware that much of their inventory was worthless or seriously overvalued, or were reckless as to whether that was the case.” Id. at 430-31. According to the district court, in order to meet the pleading requirement, the plaintiffs needed to identify the confidential sources of their information, see id. at 431-32, include written documentation of the “Box and Hold” practice in their complaint, see id. at 432, and allege facts showing that the Merrill Lynch defendants actually knew about “Box and Hold,” see id. at 434.

On April 9, 1998, the plaintiffs filed an amended complaint. The defendants thereafter served motions to dismiss. On November 9, 1998, the district court dismissed the plaintiffs’ amended complaint with prejudice. See Novak II, 26 F.Supp.2d at 660. In the district court’s view, the amended complaint failed to remedy the defects of the original one, including lack of particularity in pleading, unnamed sources, and lack of specific evidence of the Merrill Lynch defendants’ knowledge of the “Box and Hold” practice. See id. at 660-62. In addition, the district court found “that ■ it would be futile to permit further amendment” of the complaint and thus dismissed it with prejudice. Id. at 663. This appeal followed.

The plaintiffs subsequently reached a settlement with the Merrill Lynch defendants and withdrew the appeal as against them. • Accordingly, our discussion pertains solely to the claims against the Ann-Taylor defendants. In particular, we do not reach the question of control-person liability under § 20(a) of the 1934- Act, since this claim pertains primarily to the Merrill Lynch defendants. We are not concerned with § 20(a) liability as to the AnnTaylor defendants because they are also alleged to be primary violators under the 1934 Act, who may be held directly liable under § 10(b).

DISCUSSION

We review de novo a district court’s order dismissing a complaint on the pleadings and accept as true all facts alleged in the complaint. See Stevelman v. Alias Research Inc., 174 F.3d 79, 83 (2d Cir.1999) (citing Chill v. General Elec. Co., 101 F.3d 263, 267 (2d Cir.1996)). In this case, we are called upon to decide principally whether the district court, in assessing the sufficiency of the pleadings, applied appropriate standards in light of our precedents and the provisions of the PSLRA and whether it erred in concluding that the plaintiffs had failed to state a claim. We must also decide whether, even if the district court erred, there are alternative grounds for affirming the dismissal of the plaintiffs’ § 10(b) claims.

I. Sufficiency of the Pleadings

The landscape of securities fraud litigation has been transformed in recent years by the passage of the PSLRA. This case requires us to determine the impact of two provisions in this legislation on the pleading standard for scienter and the required degree of particularity in pleading in this circuit.

A. The PSLRA and Anti-Fraud Provisions in Federal Securities Laws

Section 10(b) of the 1934 Act, 15 U.S.C. § 78j(b), and Rule 10b-5 promulgated thereunder, 17 C.F.R. § 240.10b-5, prohibit fraudulent activities in connection with securities . transactions. Section 10(b) makes it unlawful

[t]o use or employ, in connection with the purchase or sale of any security ..., 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.

15 U.S.C. § 78j(b). Rule 10b-5 specifies the following actions among the types of behavior proscribed by the statute:

*306To 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, hot misleading....

17 C.F.R. § 240.10b-5.

In order to state a claim under these provisions, a complaint must allege that the defendants acted with scienter. See, e.g., Chill, 101 F.3d at 266. This scienter requirement for a private action under Rule 10b-5 has been firmly established for at least a generation. See Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976) (holding that no “private cause of action for damages will lie under § 10(b) and Rule 10b-5 in the absence of any allegation of ‘scienter’ — intent to deceive, manipulate, or defraud”); Lanza v. Drexel & Co., 479 F.2d 1277, 1301 (2d Cir.1973) (in banc) (“Other cases .in this circuit clearly indicate that ‘facts amounting to scienter, intent to defraud, reckless disregard for the truth, or knowing use of a device, scheme or artifice to defraud’ are essential to the imposition of liability.”) (quoting Shemtob v. Shearson, Hammill & Co., 448 F.2d 442, 445 (2d Cir.1971)). This case pertains not to the scienter requirement itself, but rather to the pleading requirement for scienter in the securities fraud context. Prior to the passage of the PSLRA, we had decided that, in order to state a claim for securities fraud, plaintiffs had to allege facts giving rise to “a strong inference of fraudulent intent.” Acito v. IMCERA Group, Inc., 47 F.3d 47, 52 (2d Cir.1995).

In addition to pleading scienter, it is well-established that a securities fraud complaint must also plead certain facts with particularity in order to state a claim. Fed.R.Civ.P. 9(b) requires that, whenever a complaint contains allegations of fraud, “the circumstances constituting fraud ... shall be stated with particularity.” See also Chill, 101 F.3d at 267 (noting that “the actual fraudulent statements or conduct and the fraud alleged must be stated with particularity”) (internal citations omitted). “[A] complaint making'such allegations must ‘(1) specify the statements that the plaintiff contends were fraudulent, (2) identify the speaker, (3) state where and when the statements were made, and (4)explain why the statements were fraudulent.’ ” Shields v. Citytrust Bancorp, Inc., 25 F.3d 1124, 1128 (2d Cir.1994) (quoting Mills v. Polar Molecular Corp., 12 F.3d 1170, 1175 (2d Cir.1993)).

In 1995, Congress amended the 1934 Act through passage of the PSLRA. See Private Securities Litigation Reform Act of 1995, Pub.L. No. 104-67, 109 Stat. 737 (codified at 15 U.S.C. §§ 77k, 771, 77z-1, 77z-2, 78a, 78j-1, 78t, 78u, 78u-4, 78u-5). Legislators were apparently motivated in large part by a perceived need to deter strike suits wherein opportunistic private plaintiffs file securities fraud claims of dubious merit in order to exact large settlement recoveries. See H.R. Conf. Rep. No. 104-369, at 31 (1995) (noting “significant evidence of-abuse in private securities lawsuits,” including “the routine filing of lawsuits against issuers of securities and others whenever there is a significant change in an issuer’s stock price, without regard to any underlying culpability of the issuer,” and “the abuse of. the discovery process to impose costs so burdensome that- it is often economical for the victimized party to settle”), reprinted in 1995 U.S.C.C.A.N. 730, 730.

In order “to curtail the filing of merit-less lawsuits,” the PSLRA imposed stringent procedural requirements on plaintiffs pursuing private securities fraud actions. See id. at 41. This case concerns two of these provisions in particular. First, the statute requires that,

[i]n any private action arising under this chapter in which the plaintiff may recover money damages only on proof that the defendant acted with a particular state of mind, the complaint shall, with respect to each act or omission alleged to violate this chapter, state with partic *307 ularity facts giving rise to a strong inference that the defendant acted with the required state of mind.

15 U.S.C. § 78u-4(b)(2) (emphasis added) [hereinafter “paragraph (b)(2)”]. Second, the statute requires that,

[i]n any private action arising under this chapter in which the plaintiff alleges that the defendant—

(A) made an untrue statement of a material fact; or

(B) omitted to state a material fact necessary in order to make the statements made, in the light of the circumstances in which they were made, not misleading;

the complaint shall specify each statement alleged to have been misleading, the reason or reasons why the statement is misleading, and, if an allegation regarding the statement or omission is made on information and belief the complaint shall state tvith particularity all facts on which that belief is formed.

15 U.S.C. § 78u-4(b)(l) (emphasis added) [hereinafter “paragraph (b)(1)”]. In addition, § 21D(b)(3)(A) of the PSLRA requires courts to dismiss complaints that fail to meet the pleading requirements of paragraphs (b)(1) and (b)(2). See 15 U.S.C. § 78u-4(b)(3)(A). We must determine the impact of these new requirements in order to decide whether the plaintiffs in this case have pleaded sufficient facts with enough particularity to state a claim under the 1934 Act.

B. The Pleading Standard for Scienter

1. The Second Circuit’s Pre-PSLRA Pleading Standard

We can easily summarize the pleading standard for scienter that prevailed in this circuit prior to the PSLRA:

[Plaintiffs must allege facts that give rise to a strong inference of fraudulent intent. “The requisite ‘strong inference’ of fraud may be established either (a) by alleging facts to show that defendants had both motive and opportunity to commit fraud, or (b) by alleging facts that constitute strong circumstantial evidence of conscious misbehavior or recklessness.”

Acito, 47 F.3d at 52 (quoting Shields, 25 F.3d at 1128) (internal citations omitted). However, this statement of the standard conceals the complexity and uncertainty that often surround its application. This difficulty in application stems, at least in part, from the “inevitable tension” between the interests in deterring securities fraud and deterring strike suits. See In re Time Warner Inc. Sec. Litig., 9 F.3d 259, 263 (2d Cir.1993). As a result, different courts applying the pleading standard to differing factual circumstances may reach seemingly disparate results. See id. at 264. Nevertheless, we discern some basic patterns in our case law under § 10(b) and Rule 10b-5 that help to provide substance to the general language of the standard itself.

We described the type of motive and opportunity required to plead scienter under our pre-reform standard as follows:

Motive would entail concrete benefits that could be realized by one or more of the false statements and wrongful non-disclosures alleged. Opportunity would entail the means and likely prospect of achieving concrete benefits by the means alleged.

Shields, 25 F.3d at 1130. Plaintiffs could not proceed based on motives possessed by virtually all corporate insiders, including: (1) the desire to maintain a high corporate credit rating, see San Leandro Emergency Med. Group Profit Sharing Plan v. Philip Morris Cos., Inc., 75 F.3d 801, 814 (2d Cir.1996), or otherwise sustain “the appearance of corporate profitability, or of the success of an investment,” Chill, 101 F.3d at 268; and (2) the desire to maintain a high stock price in order to increase executive compensation, see Acito, 47 F.3d at 54, or prolong the benefits of holding corporate office, see Shields, 25 F.3d at 1130. Rather, plaintiffs had to allege that defendants benefitted in some concrete and personal way from the purported *308fraud. This requirement was generally met when corporate insiders were alleged to have misrepresented to the public material facts about the corporation’s performance or prospects in order to keep the stock price artificially high while they sold their own shares at a profit. See, e.g., Stevelman, 174 F.3d at 85; Goldman v. Belden, 754 F.2d 1059, 1070 (2d Cir.1985). Accordingly, in the ordinary - case, adequate motive arose from the desire to profit from extensive insider sales.

Plaintiffs could also meet, the pre-PSLRA pleading standard by alleging facts that constituted strong circumstantial evidence of conscious misbehavior or recklessness on the part of defendants. Intentional misconduct is easily identified since it encompasses deliberate illegal behavior, such as securities trading by insiders privy to undisclosed and material information, see Simon DeBartolo Group, L.P. v. Richard E. Jacobs Group, Inc., 186 F.3d 157, 168-69 (2d Cir.1999), or knowing sale of a company’s stock, at an unwarranted discount, see Schoenbaum v. Firstbrook, 405 F.2d 215, 219 (2d Cir.1968) (in banc).

Recklessness is harder to identify with such precision- and consistency. In 1978, when we first held that recklessness suffices to plead scienter under § 10(b) and Rule 10b-5, we defined reckless conduct as:

at the least, conduct which is “highly unreasonable” and- which represents “an extreme departure from the standards of ordinary care ... to the extent that the danger was either known to the defendant or so obvious that the defendant must have been aware of it.”

Rolf v. Blyth, Eastman Dillon & Co., Inc., 570 F.2d 38, 47 (2d Cir.1978) (quoting Sanders v. John Nuveen & Co., 554 F.2d 790, 793 (7th Cir.1977)) (ellipsis in original). Similarly, we later noted that “ ‘[a]n egregious refusal to see the obvious, or to investigate the doubtful, may in some cases give rise to an inference of ... recklessness.’ ” Chill, 101 F.3d at 269 (quoting Goldman v. McMahan, Brafman, Morgan & Co., 706 F.Supp. 256, 259 (S.D.N.Y.1989)) (ellipsis in original).

However, these general standards offer little insight into precisely what actions and behaviors constitute recklessness sufficient for § 10(b) liability. It is the actual facts of our securities fraud cases that provide the most concrete guidance as to the types of allegations required to meet the pre-PSLRA pleading standard in this circuit.

According, to these eases, securities fraud claims typically have sufficed to state a claim based on recklessness when they have specifically alleged defendants’ knowledge of facts or access to information contradicting their public statements. Under such circumstances, defendants knew or, more importantly, should have known that they were misrepresenting material facts related to the corporation. Thus, for example, the pleading standard was met where the plaintiffs alleged that the defendants made or authorized statements that sales to China would be “an important new source of revenue” when they knew or should have known that Chinese import restrictions in place at the time would severely limit such’ sales. See Cosmas v. Hassett, 886 F.2d 8, 12 (2d Cir.1989). Similarly, the pleading standard was met where the plaintiffs alleged that the defendants released to the investing public several highly positive predictions about the marketing prospects of a computer system to record hotel guests’ long-distance telephone calls when they knew or should have known several facts about the system and its consumers that revealed “grave uncertainties and problems concerning future sales of’ the system. Goldman, 754 F.2d at 1063, 1070.

Under certain circumstances, we have found allegations of recklessness to be sufficient where plaintiffs alleged facts demonstrating that defendants failed to review or check information- that they had a duty to monitor, or ignored obvious signs of fraud. Thus, the pleading standard was met where the plaintiff alleged that the *309defendant, his broker, consistently reassured the plaintiff that the investment ad-visor responsible for the plaintiffs portfolio “knew what he was doing” but never actually investigated the advisor’s decisions to determine “whether there was a basis for the [defendant’s] assertions.” Rolf, 570 F.2d at 47-48. Similarly, the pleading standard was met where the defendant allegedly included false statements in SEC filings despite “the obviously evasive and suspicious statements made to him” by the corporate officials upon whom he was relying for this information and despite outside counsel’s recommendation that these statements not be included. SEC v. McNulty, 137 F.3d 732, 741 (2d Cir.1998).

At the same time, however, we have identified several important limitations on the scope of liability for securities fraud based on reckless conduct. First, we have refused to allow plaintiffs to proceed with allegations of “fraud by hindsight.” See Stevelman, 174 F.3d at 85. Corporate officials need not be clairvoyant; they are only responsible for revealing those material facts reasonably available to them. See Denny v. Barber, 576 F.2d 465, 470 (2d Cir.1978). Thus, allegations that defendants should have anticipated future events and made certain disclosures earlier than they actually did do not suffice to make out a claim of securities fraud. See Acito, 47 F.3d at 53.

Second, as long as the public statements are consistent with reasonably available data, corporate officials need not present an overly gloomy or cautious picture of current performance and future prospects. See Stevelman, 174 F.3d at 85; Shields, 25 F.3d at 1129-30. Where plaintiffs contend defendants had access to contrary facts, they must specifically identify the reports or statements containing this information. See San Leandro, 75 F.3d at 812 (“Plaintiffs’ unsupported general claim of the existence of confidential company sales reports that revealed the larger decline in sales is insufficient to survive a motion to dismiss.”).

Third, there are limits to the scope of liability for failure adequately to monitor the allegedly fraudulent behavior of others. Thus, the failure of a non-fiduciary accounting firm to identify problems with the defendant-company’s internal controls and accounting practices does not constitute reckless conduct sufficient for § 10(b) liability. See Decker v. Massey-Ferguson, Ltd., 681 F.2d 111, 120 (2d Cir.1982). Similarly, the failure of a parent company to interpret extraordinarily positive performance by its subsidiary — specifically, the ‘“unprecedented and dramatically increasing profitability” of a particular form of trading — as a sign of problems and thus to investigate further does not amount to recklessness under the securities laws. See Chill, 101 F.3d at 269-70.

Finally, allegations of GAAP violations or accounting irregularities, standing alone, are insufficient to state a securities fraud claim. See Stevelman, 174 F.3d at 84; Chill, 101 F.3d at 270. Only where such allegations are coupled with evidence of “corresponding fraudulent intent,” Chill, 101 F.3d at 270, might they be sufficient.

We now examine to what extent these lessons from our prior case law have survived the recent reform of the securities laws.

2. Implications of the PSLRA for the Pleading Standard for Scienter in this Circuit

Courts have disagreed on the proper interpretation of the new pleading requirement imposed by paragraph (b)(2) in light of the text of the PSLRA and its legislative history. They have generally come to one of two conclusions:

(1) The statute effectively adopts the Second Circuit’s pleading standard for scienter wholesale, and thus plaintiffs may continue to state a claim by pleading either motive and opportunity or strong circumstantial evidence of reck*310lessness or conscious misbehavior. See In re Advanta Corp. Sec. Litig., 180 F.3d 525 (3d Cir.1999); Press v. Chemical Invest. Servs. Corp., 166 F.3d 529, 538 (2d Cir.1999) (dicta); Rubinstein v. Skyteller, Inc., 48 F.Supp.2d 315, 320 (S.D.N.Y.1999) (following Press).

(2) The statute strengthens the Second Circuit’s standard by rejecting the simple pleading of motive and opportunity. See Bryant v. Avado Brands, Inc., 187 F.3d 1271, 1283 (11th Cir.1999); In re Silicon Graphics Inc. Sec. Litig., 183 F.3d 970, 979 (9th Cir.1999); In re Comshare, Inc. Sec. Litig., 183 F.3d 542, 550-51 (6th Cir.1999); Novak I, 997. F.Supp. at 430; In re Glenayre Tech., Inc. Sec. Litig., 982 F.Supp. 294, 298 (S.D.N.Y.1997); In re Baesa Sec. Litig., 969 F.Supp. 238, 241-42 (S.D.N.Y.1997).

Our own review of the text and legislative history leads us to a middle ground. We conclude that the PSLRA effectively raised the nationwide pleading standard to that previously existing in this circuit and no higher (with the exception of the “with particularity” requirement). At the same time, however, we believe that Congress’s failure to include language about motive and opportunity suggests that we need not be wedded to these concepts in articulating the prevailing standard. We are led to these conclusions by the considerations that follow.

In order to gauge the implications of paragraph (b)(2), we apply familiar canons of statutory construction. We look first to the text of the statute. If that language is plain and its meaning sufficiently clear, we need look no further. See Connecticut Nat’l Bank v. Germain, 503 U.S. 249, 254, 112 S.Ct. 1146, 117 L.Ed.2d 391 (1992). Only if the text of the statute is not unambiguous do we turn for guidance to legislative history and the purposes of the statute. See Dowling v. United States, 473 U.S. 207, 218, 105 S.Ct. 3127, 87 L.Ed.2d 152 (1985). Applying these principles, we conclude that the enactment of paragraph (b)(2) did not change the basic pleading standard for scienter in this circuit.

In this case, our interpretive task begins and ends with the text of the statute. In drafting paragraph (b)(2), Congress specifically incorporated this circuit’s “strong inference” language to define the pleading standard for securities fraud cases. Compare 15 U.S.C. § 78u-4(b)(2) (requiring plaintiffs to “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind”), with Acito, 47 F.3d at 52 (“[Plaintiffs must allege facts that give rise to a strong inference of fraudulent intent.”). We agree with the Third Circuit that this “use of the Second Circuit’s language compels the conclusion that the Reform Act establishes a pleading standard approximately equal in stringency to that of the Second Circuit.” In re Advanta Corp., 180 F.3d at 534. Cf. United States v. Johnson, 14 F.3d 766, 770 (2d Cir.1994) (finding that Congress’s use of “substantially identical language” to that of an earlier statute “bespeaks an intention to import” judicial interpretations of that language into the new statute).

Given the absence of ambiguity in the statutory text, no resort to legislative history or the purposes of the PSLRA is required. In any event, there is nothing in these sources that would alter our conclusion. As far as the general purposes of the PSLRA are concerned, Congress plainly sought to impose a stricter nationwide pleading standard and did so. But this purpose does not require raising the standard above that of this circuit, particularly in light of the explicit Congressional recognition that our pre-PSLRA standard was the most stringent in the nation. See H.R. Conf. Rep. No. 104-369, at 41. “In many jurisdictions, adoption of a ‘strong inference’ standard will substantially heighten the barriers to pleading scienter, a result Congress expressly intended. Moreover, even in jurisdictions already employing the Second Circuit standard, the additional requirement that plaintiffs *311state facts ‘with particularity’ represents a heightening of the standard.” In re Ad-vanta Corp., 180 F.3d at 534.

Meanwhile, in our view, as is so often the case with legislative history generally, the legislative history of the PSLRA contains “conflicting expressions of legislative intent” with respect to the pleading requirement. Id. at 533. For example, while the Conference Committee rejected language from the Senate bill that would have adopted the Second Circuit rule wholesale, including language about motive and opportunity and recklessness, see H.R. Conf. Rep. No. 104-369, at 41 & 48 n. 23, the Senate Committee reporting the bill stated that it was proposing not “a new and untested pleading standard that would generate additional litigation,” but rather “a uniform standard modeled upon the pleading standard of the Second Circuit.” S.Rep. No. 104-98, at 15 (1995), reprinted in 1995 U.S.C.C.A.N. 679, 694 (noting that courts interpreting the proposed “strong inference” pleading standard might find Second Circuit case law “instructive”).

When all is said and done, we believe that the enactment of paragraph (b)(2) did not change the basic pleading standard for scienter in this circuit (except by the addition of the words “with particularity”). Accordingly, we hold that the PSLRA adopted our “strong inference” standard: In order to plead scienter, plaintiffs must “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind,” as required by the language of the Act itself. Although litigants and lower courts need and should not employ or rely on magic words such as “motive and opportunity,” we believe that our prior case law may be helpful in providing guidance as to how the “strong inference” standard may be met. Therefore, in applying this standard, district courts should look to the cases and factors discussed in Section I.B.l above to determine whether plaintiffs have pleaded facts giving rise to the requisite “strong inference.” These cases suggest, in brief, that the inference may arise where the complaint sufficiently alleges that the defendants: (1) benefitted in a concrete and personal way from the purported fraud, see supra at page 307; (2) engaged in deliberately illegal behavior, see supra at page 308; (3) knew facts or had access to information suggesting that their public statements were not accurate, see supra at page 308-09; or (4) failed to check information they had a duty to monitor, see supra at pages 308-09. We now turn to the complaint in this case to determine whether the plaintiffs have met their burden to plead scienter.

3. Strong Inference of Fraudulent Intent on the Part of the AnnTaylor Defendants

The district court concluded that the plaintiffs had failed to plead facts giving rise to a strong inference of the defendants’ fraudulent intent, as required to state a claim under § 10(b). We disagree.

According to the complaint, the AnnTaylor defendants knew at all relevant times that the Company had serious inventory problems that they sought to disguise by adopting the “Box and Hold” scheme. By refusing to mark down inventory they knew to be “worthless,” “obsolete,” and “unsalable,” the defendants acted “intentionally and deliberately” to artificially inflate AnnTaylor’s reported financial results. They discussed the need to mark down inventory but refused to do so because that would damage the Company’s financial prospects. Further, in approving the inventory management practices of “Box and Hold,” the defendants knowingly sanctioned procedures that violated the Company’s own markdown policy, as stated in the Company’s public filings. In doing so, they caused those filings to be materially misleading in that the disclosed policy no longer reflected actual practice. Lastly, despite knowledge of the true reasons for rising inventory levels, the defendants made repeated statements to the investment community either offering false *312reassurances that inventory was under control or giving false explanations for its growth. In short, the Complaint alleges that the defendants engaged in conscious misstatements with the intent to deceive. There is no doubt that this pleading satisfies the standard for scienter under Hoch-felder, and the requirement of the PSLRA that plaintiffs state facts with particularity that give rise to a strong inference of the required state of mind.

In the end, we believe that the district court applied the correct standard but erroneously found that this standard was not met on these pleadings. According to the district court, the scienter requirement can be satisfied by pleading either “conscious recklessness” — i.e., a state of mind “approximating actual intent, and not merely a heightened form of negligence” — or “actual intent.” Novak I, 997 F.Supp. at 430. This was an accurate statement of the law. However, the district court believed that the facts pleaded by the plaintiffs supported nothing more than an inference that the managers of AnnTaylor disagreed over matters of business judgment, such as the valuation of inventory and the timing of markdowns. See Novak II, 26 F.Supp.2d at 660. This was incorrect as a matter of law. When managers deliberately make materially false statements concerning inventory with the intent to deceive the investment community, they have engaged in conduct actionable under the securities laws.

C. Particularity of the Facts Pleaded

The district court also found the facts pleaded by the plaintiffs insufficiently particularized, in large part because they did not reveal the identity of the personal sources of their critical factual allegations. We disagree with the district court’s reasoning and accordingly vacate and remand for further proceedings consistent with the discussion that follows.

As discussed above, Rule 9(b) has long required plaintiffs in securities fraud cases to state “the circumstances constituting fraud ... with particularity.” The PSLRA imposed an additional requirement: whenever plaintiffs allege, on information and belief, that defendants made material misstatements or omissions, the complaint must “state with particularity all facts on which that belief is formed.” 16 U.S.C. § 78u-4(b)(l). This requirement plainly applies in this case. In numerous places in their complaint, the plaintiffs allege, based on information and belief, that the AnnTaylor defendants made materially misleading statements or omissions. Most importantly, they allege that the defendants made false statements concerning the value of inventory because “Box and Hold” merchandise was “unsalable,” “obsolete,” and “nearly worthless,” and its “actual value was nearly zero.” In order to survive at this stage, the complaint must state with particularity sufficient facts to support the belief that the “Box and Hold” inventory was of limited value, and accordingly that the defendants’ positive public statements concerning inventory growth were false and misleading.

The district court concluded that the plaintiffs had failed to meet these particularity requirements, in substantial part because they failed to reveal their confidential sources for some of the facts on which their belief in the essential worthlessness of the “Box and Hold” inventory was based. See Novak I, 997 F.Supp. at 431-32; Novak II, 26 F.Supp.2d at 660-61. The lower court found the plaintiffs’ allegations in this respect at worst “concluso-ry, unsupported and inflammatory,” and at best “based upon reports by ... anonymous ‘former employees’ ” who should have been identified by name. Novak II, 26 F.Supp.2d at 661. The district court’s reasoning and conclusions were flawed in several respects.

For one thing, the complaint provides specific facts concerning the Company’s significant write-off of inventory directly following the Class Period, which tends to support the plaintiffs’ contention that in*313ventory was seriously overvalued at the time the purportedly misleading statements were made. Specifically, the plaintiffs allege that: (1) in AnnTaylor’s May 1995 reporting of its first quarter fiscal 1995 results, the Company “admitted to analysts that its inventories were too high” and that “inventory liquidation” would follow; (2) in AnnTaylor’s July 29, 1995 10-Q filed with the SEC, it “admitted that the decrease in the Company’s gross profit percentage was attributable to ‘increased cost of goods sold as a percentage of net sales, primarily resulting from markdowns ’ and (3) a January 22, 1996 Weekly Report showed that even six months after the Class Period, substantial amounts of “Box and Hold” inventory still dated from 1993 and 1994, which supports the inference that inventory during the Class Period was similarly dated. Thus, the complaint identifies with particularity several documentary sources that support the plaintiffs’ belief that serious inventory problems existed during the Class Period itself.

We recognize that the complaint does not state with particularity every fact upon which this belief was based, since it is apparent that there were also personal sources who were not specifically identified. However, plaintiffs who rely on confidential sources are not always required to name those sources, even when they make allegations on information and belief concerning false or misleading statements, as here.

First, there is nothing in the caselaw of this circuit that requires plaintiffs to reveal confidential sources at the pleading stage. The defendants rely heavily on Segan v. Dreyfus Corp., 513 F.2d 695 (2d Cir.1975), in which we held that a plaintiffs complaint was insufficiently specific and rejected the argument that further disclosure would, among other things, identify a confidential informant. See id. at 696. But in Dreyfus, we held only that the plaintiff had to plead additional facts, not that the plaintiff was required to reveal the name of the informant. See id. (“A suit charging fraud may not be based on facts so secret that the defendants cannot be told what they are.”) (emphasis added). Some district courts in this circuit have on occasion stated that Rule 9(b) requires plaintiffs in securities fraud ■ case’s to allege the “sources that support the alleged specific facts,” e.g., Blanchard v. Katz, 705 F.Supp. 1011, 1012 (S.D.N.Y.1989); Crystal v. Foy, 562 F.Supp. 422, 425 (S.D.N.Y.1983), but in no case have they dismissed a complaint for failure to identify confidential sources.

Second, while paragraph (b)(1) may compel revelation of confidential sources under certain circumstances, such circumstances, are not necessarily present in this case. The defendants point to district court decisions outside this circuit that hold or iipply that the PSLRA generr ally requires plaintiffs to include the names of their confidential sources. See In re Silicon Graphics Inc. Sec. Litig., 970 F.Supp. 746, 763 (N.D.Cal.1997); In re Aetna Inc. Sec. Litig., No. CIV. A. MDL 1219, 1999 WL 354527, at *4 (E.D.Pa. May 26, 1999). However, this rule is based on a misreading of the legislative history of the PSLRA. Specifically, the court in Silicon Graphics relied primarily on the hyperbolic statements of legislators attempting (unsuccessfully) to amend the proposed Act to lighten plaintiffs’ pleading burden. See Silicon Graphics, 970 F.Supp. at 763-64. In fact, the applicable provision of the law as ultimately enacted requires plaintiffs to plead only facts and makes no mention of the sources of these facts. See 15 U.S.C. § 78u-4(b)(1).

More fundamentally, our reading of the PSLRA rejects any notion that confidential sources must be named as a general matter. In our view, notwithstanding the use of the word “all,” paragraph (b)(1) does not require that plaintiffs plead with particularity every single fact upon which their beliefs concerning false or misleading statements are. based. Rather, plaintiffs need only plead with particularity sufficient facts to support *314those beliefs.1 Accordingly, where plaintiffs rely on confidential personal sources but also on other facts, they need not name their sources as long as the latter facts provide an adequate basis for believing that the defendants’ statements were false. Moreover, even if personal sources must be identified, there is no requirement that they be named, provided they are described in the complaint with sufficient particularity to support the probability that a person, in the position occupied by the source would possess the information alleged. In both of these situations, the plaintiffs will have pleaded enough facts to support their belief, even though some arguably relevant facts have been left out. Accordingly,' a complaint can meet the new pleading requirement imposed by paragraph (b)(1) by providing documentary evidence and/or a sufficient general description of the personal sources of the plaintiffs’ beliefs.

Thus, we find no requirement in existing law that, in the ordinary course, complaints in securities fraud cases must name confidential sources, and we see no reason to impose such a requirement under the circumstances of this case. “The primary purpose of Rule 9(b) is to afford defendant fair notice of the plaintiffs claim and the factual ground upon which it is based.” Ross v. Bolton, 904 F.2d 819, 823 (2d Cir.1990). This purpose, which also underlies paragraph (b)(1), can be served without requiring plaintiffs to name their confidential sources as long as they supply sufficient specific facts to support their allegations. Imposing a general requirement of disclosure of confidential sources serves no legitimate pleading purpose while it could deter informants from providing critical information to investigators in meritorious cases or invite retaliation against them.

We express no view as to whether the plaintiffs’ allegations in this case were sufficiently particularized. Instead, we remand to the district court with instructions to: (1) allow the plaintiffs to replead in light of our discussion above; and (2) reconsider the particularity of the plaintiffs’ pleadings in light of the proper standards.

II. Alternative Grounds for Dismissal

The appellees argue that we still may affirm the district court’s dismissal because: (1) the appellants have not adequately alleged that statements made by securities analysts can be attributed to the AnnTaylor defendants; and (2) the alleged false and misleading statements made during the Class Period are not actionable in any event. We reject both arguments.

Under the law of this circuit, plaintiffs may state a claim against corporate officials for false and misleading information disseminated through analysts’ reports by alleging that the officials either: (1) “intentionally foster[ed] a mistaken belief concerning a material fact” that was incorporated into reports; or (2) adopted or placed their “imprimatur” on the reports. See Elkind v. Liggett & Myers, Inc., 635 F.2d 156, 163-64 (2d Cir.1980). This was plainly the case here. The plaintiffs alleged that the AnnTaylor defendants both made misstatements that were later incorporated into analysts’ reports and subsequently (at least implicitly) adopted the contents of those reports. These allegations were sufficiently detailed to meet the pleading threshold because generally the circumstances of the statements — including dates and participants— were particularized. See Acito, 47 F.3d at 51.

*315-337Moreover, the types of statements the defendants are alleged to have made or adopted are actionable. While statements containing simple economic projections, expressions of optimism, and other puffery are insufficient, see, e.g., Friedman v. Mohasco Corp., 929 F.2d 77, 79 (2d Cir.1991), defendants may be liable for misrepresentations of existing facts, see In re Prudential Sec. Inc. Ltd. Partnerships Litig., 930 F.Supp. 68, 74-75 (S.D.N.Y.1996). Here, the complaint alleges that the defendants did more than just offer rosy predictions; the defendants stated that the inventory situation was “in good shape” or “under control” while they allegedly knew that the contrary was true. Assuming, as we must at this stage, the accuracy of the plaintiffs’ allegations about AnnTaylor’s “Box and Hold” practices, these statements were plainly false and misleading.

CONCLUSION

For the foregoing reasons, we hold, first, that: (a) in order to plead scienter in securities fraud cases, plaintiffs must “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind”; and (b) the plaintiffs here pleaded sufficient facts to establish a strong inference of fraudulent intent on the part of the AnnTaylor defendants. Second, we hold that the district court erred, under the circumstances of this case, by requiring the plaintiffs to reveal the names of their confidential sources in order to meet the particularity requirements of the PSLRA. On remand, we instruct the district court to: (a) permit the plaintiffs to replead; and (b) evaluate their pleadings anew in light of our interpretation of these requirements. Finally, we see no alternative grounds for affirming the district court’s dismissal of the plaintiffs’ complaint. Accordingly, the judgment of the district court is vacated and the case remanded for further proceedings consistent with these rulings.

Mr. Justice Powell

delivered the opinion of the Court.

The issue in this case is whether an action for civil damages may lie under § 10 (b) of the Securities Exchange Act of 1934 (1934 Act), 48 Stat. 891, 15 U. S. C. *188§ 78j (b), and Securities and Exchange Commission Rule 10b-5, 17 CFR § 240.1Ob-5 (1975), in the absence of an allegation of intent to deceive, manipulate, or defraud on the part of the defendant.

I

Petitioner, Ernst & Ernst, is an accounting firm. From 1946 through 1967 it was retained by First Securities Company of Chicago (First Securities), a small brokerage firm and member of the Midwest Stock Exchange and of the National Association of Securities Dealers, to perform periodic audits of the firm's books and records. In connection with these audits Ernst & Ernst prepared for filing with the Securities and Exchange Commission (Commission) the annual reports required of First Securities under § 17 (a) of the 1934 Act, 15 U. S. C. § 78q (a).1 It also prepared for First Securities responses to the financial questionnaires of the Midwest Stock Exchange (Exchange).,

*189Respondents were customers of First Securities who invested in a fraudulent securities scheme perpetrated by Leston B. Nay, president of the firm and owner of 92% of its stock. Nay induced the respondents to invest funds in “escrow” accounts that he represented would yield a high rate of return. Respondents did so from 1942 through 1966, with the majority of the transactions occurring in the 1950’s. In fact, there were no escrow accounts as Nay converted respondents’ funds to his own use immediately upon receipt. These transactions were not in the customary form of dealings between First Securities and its customers. The respondents drew their personal checks payable to Nay or a designated bank for his account. No such escrow accounts were reflected on the books and records of First Securities, and none was shown on its periodic accounting to respondents in connection with their other investments. Nor were they included in First Securities’ filings with the Commission or the Exchange.

This fraud came to light in 1968 when Nay committed suicide, leaving a note that described First Securities as bankrupt and the escrow accounts as “spurious.” Respondents subsequently filed this action2 for damages against Ernst & Ernst3 in the United States District Court for the Northern District of Illinois under *190§ 10 (b) of the 1934 Act. The complaint charged that Nay's escrow scheme violated § 10 (b) and Commission Rule 10b-5,4 and that Ernst & Ernst had “aided and abetted” Nay’s violations by its “failure” to conduct proper audits of First Securities. As revealed through discovery, respondents’ cause of action rested on a theory of negligent nonfeasance. The premise was that Ernst & Ernst had failed to utilize “appropriate auditing procedures” in its audits of First Securities, thereby failing to discover internal practices of the firm said to prevent an effective audit. The practice principally relied on was Nay’s rule that only he could open mail addressed to him at First Securities or addressed to First Securities to his attention, even if it arrived in his absence. Respondents contended that if Ernst & Ernst had conducted a proper audit, it would have discovered this “mail rule.” The existence of the rule then would have been disclosed in reports to the Exchange and to the Commission by Ernst & Ernst as an irregular procedure that prevented an effective audit. This would have led to an investigation of Nay that would have revealed the fraudulent scheme. Respondents specifically disclaimed the existence of fraud or intentional misconduct on the part of Ernst & Ernst.5

*191After extensive discovery the District Court granted Ernst & Ernst's motion for summary judgment and dismissed the action. The court rejected Ernst & Ernst's contention that a cause of action for aiding and abetting a securities fraud could not be maintained under § 10 (b) and Rule 10b-5 merely on allegations of negligence. It concluded, however, that there was no genuine issue of material fact with respect to whether Ernst & Ernst had conducted its audits in accordance with generally accepted auditing standards.6

The Court of Appeals for the Seventh Circuit reversed and remanded, holding that one who breaches a duty of inquiry and disclosure owed another is liable in damages for aiding and abetting a third party's violation of Rule 10b-5 if the fraud would have been discovered or prevented but for the breach. 503 F. 2d 1100 (1974).7 *192The court reasoned that Ernst & Ernst had a common-law and statutory duty of inquiry into the adequacy of First Securities’ internal control system because it had contracted to audit First Securities and to prepare for filing with the Commission the annual report of First Securities’ financial condition required under § 17 of the 1934 Act and Rule 17a-5.8 The court further reasoned that respondents were beneficiaries of the statutory duty to inquire9 and the related duty to disclose any material *193irregularities that were discovered. 503 F. 2d, at 1105-1111. The court concluded that there were genuine issues of fact as to whether Ernst & Ernst’s failure to discover and comment upon Nay’s mail rule10 constituted a breach of its duties of inquiry and disclosure, id., at 1111, and whether inquiry and disclosure would have led to the discovery or prevention of Nay’s fraud. Id., at 1115.11

We granted certiorari to resolve the question whether a private cause of action for damages will lie under § 10 (b) and Rule 10b-5 in the absence of any allegation of “scienter” — intent to deceive, manipulate, or defraud.12 421 U. S. 909 (1975). We ednclude that it will not and therefore we reverse.13

*194II

Federal regulation of transactions in securities emerged as part of the aftermath of the market crash in 1929. *195The Securities Act of 1933 (1933 Act), 48 Stat. 74, as amended, 15 U. S. C. § 77a et seq., was designed to provide investors with full disclosure of material information concerning public offerings of securities in commerce, to protect investors against fraud and, through the imposition of specified civil liabilities, to promote ethical standards of honesty and fair dealing. See H. R. Rep. No. 85, 73d Cong., 1st Sess., 1-5 (1933). The 1934 Act was intended principally to protect investors against manipulation of stock prices through regulation of transactions upon securities exchanges and in over-the-counter markets, and to impose regular reporting requirements on companies whose stock is listed on national securities exchanges. See S. Rep. No. 792, 73d Cong., 2d Sess., 1-5 (1934). Although the Acts contain numerous carefully drawn express civil remedies and criminal penalties, Congress recognized that efficient regulation of securities trading could not be accomplished under a rigid statutory program. As part of the 1934 Act Congress created the Commission, which is provided with an arsenal of flexible enforcement powers. See, e. g., 1933 Act §§ 8, 19, 20, 15 U. S. C. §§ 77h, 77s, 77t; 1934 Act §§ 9, 19, 21, 15 U. S. C. §§ 78i, 78s, 78u.

Section 10 of the 1934 Act makes it “unlawful for any person . . . (b) [t]o use or employ, in connection with the purchase or sale of any security . . . 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.” 15 U. S. C. § 78j. In 1942, acting pursuant to the power conferred by § 10 (b), the Commission promulgated Rule 10b-5, which now provides:

“Employment of manipulative and deceptive devices.

*196“It shall be unlawful for any person, directly or indirectly, by the use of any means or instrumentality of interstate commerce, or of the mails or of any facility of any national securities exchange,

“(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.”

Although § 10 (b) does not by its terms create an express civil remedy for its violation, and there is no indication that Congress,14 or the Commission when adopting Rule 10b-5,15 contemplated such a remedy, the existence of a private cause of action for violations of the statute and the Rule is now well established. Blue Chip Stamps v. Manor Drug Stores, 421 U. S. 723, 730 (1975); Affiliated Ute Citizens v. United States, 406 U. S. 128, 150-154 (1972); Superintendent of Insurance v. Bankers Life & Cas. Co., 404 U. S. 6, 13 n. 9 (1971). During the 30-year period since a private cause of action was first implied under § 10 (b) and Rule 10b-5,16 *197a substantial body of case law and commentary has developed as to its elements. Courts and commentators long.-have differed with regard to whether scienter is a necessary element of such a cause of action, or whether negligent conduct alone is sufficient.17 In addressing this question, we turn first to the language of § 10(b), for “[t]he starting point in every case involving construction of a statute is the language itself.” Blue Chip Stamps, supra, at 756 (Powell, J., concurring); see FTC v. Bunte Bros., Inc., 312 U. S. 349, 350 (1941).

A

Section 10 (b) makes unlawful the use or employment of “any manipulative or deceptive device or contrivance” in contravention of Commission rules. The words “manipulative or deceptive” used in conjunction with “device or contrivance” strongly suggest that § 10 (b) was intended to proscribe knowing or intentional misconduct. See SEC v. Texas Gulf Sulphur Co., 401 F. 2d 833, 868 (CA2 1968) (Friendly, J., concurring), cert. denied sub nom. Coates v. SEC, 394 U. S. 976 (1969); Loss, Summary Remarks, 30 Bus. Law. 163, 165 (Special Issue 1975). See also Kohn v. American Metal Climax, Inc., 458 F. 2d 255, 280 (CA3 1972) (Adams, J., concurring and dissenting).

In its amicus curiae brief, however, the Commission contends that nothing in the language “manipulative or deceptive device or contrivance” limits its operation to *198knowing or intentional practices.18 In support of its view, the Commission cites the overall congressional purpose in the 1933 and 1934 Acts to protect investors against false and deceptive practices that might injure them. See Affiliated Ute Citizens v. United States, supra, at 151; Superintendent of Insurance v. Bankers Life & Cas. Co., supra, at 11-12; J. I. Case Co. v. Borak, 377 U. S. 426, 432-433 (1964). See also SEC v. Capital Gains Res. Bur., 375 U. S. 180, 195 (1963). The Commission then reasons that since the “effect” upon investors of given conduct is the same regardless of whether the conduct is negligent or intentional, Congress must have intended to bar all such practices and not just those done knowingly or intentionally. The logic of this effect-oriented approach would impose liability for wholly faultless conduct where such conduct results in harm to investors, a result the Commission would be unlikely to support. But apart from where its logic might *199lead, the Commission would add a gloss to the operative language of the statute quite different from its commonly accepted meaning. See, e. g., Addison v. Holly Hill Fruit Products, Inc., 322 U. S. 607, 617-618 (1944).19 The argument simply ignores the use of the words “manipulative,” “device,” and “contrivance” — terms that make unmistakable a congressional intent to proscribe a type of conduct quite different from negligence.20 Use of the word “manipulative” is especially significant. It is and was virtually a term of art when used in connection with securities markets. It connotes intentional or willful conduct designed to deceive or defraud investors by controlling or artificially affecting the price of securities.21

In addition to relying upon the Commission’s argument with respect to the operative language of the stat*200ute, respondents contend that since we are dealing with "remedial legislation/’ Tcherepnin v. Knight, 389 U. S. 332, 336 (1967), it must be construed “ 'not technically and restrictively, but flexibly to effectuate its remedial purposes.’ ” Affiliated Ute Citizens v. United States, 406 U. S., at 151, quoting SEC v. Capital Gains Research Bureau, supra, at 195. They argue that the “remedial purposes” of the Acts demand a construction of § 10 (b) that embraces negligence as a standard of liability. But in seeking to accomplish its broad remedial goals, Congress did not adopt uniformly a negligence standard even as to express civil remedies. In some circumstances and with respect to certain classes of defendants, Congress did create express liability predicated upon a failure to exercise reasonable care. E. g., 1933 Act § 11 (b)(3) (B), 48 Stat. 82, as amended, 15 U. S. C. § 77k (b) (3) (B) (liability of “experts,” such as accountants, for misleading statements in portions of registration statements for which they are responsible).22 But in other situations good faith is an absolute defense. 1934 Act § 18, 48 Stat. 897, as amended, 15 U. S. C. § 78r (misleading statements in any document filed pursuant to the 1934 Act). And in still other circumstances Congress created express liability regardless of the defendant’s fault, 1933 Act § 11 (a), 15 U. S. C. § 77k (a) (issuer liability for misleading statements in the registration statement).

It is thus evident that Congress fashioned standards of fault in the express civil remedies in the 1933 and 1934 Acts on a particularized basis. Ascertainment of congressional intent with respect to the standard of liability created by a particular section of the Acts must therefore rest primarily on the language of that section. Where, as here, we deal with a judicially implied liability, the statutory language certainly *201is no less important. In view of the language of § 10 (b), which so clearly connotes intentional misconduct, and mindful that the language of a statute controls when sufficiently clear in its context, United States v. Oregon, 366 U. S. 643, 648 (1961); Packard Motor Car Co. v. NLRB, 330 U. S. 485, 492 (1947), further inquiry may be unnecessary. We turn now, nevertheless, to the legislative history of the 1934 Act to ascertain whether there is support for the meaning attributed to § 10 (b) by the Commission and respondents.

B

Although the extensive legislative history of the 1934 Act is bereft of any explicit explanation of Congress' intent, we think the relevant portions of that history support our conclusion that § 10 (b) was addressed to practices that involve some element of scienter and cannot be read to impose liability for negligent conduct alone.

The original version of what would develop into the 1934 Act was contained in identical bills introduced by Senator Fletcher and Representative Rayburn. S. 2693, 73d Cong., 2d Sess. (1934); H. R. 7852, 73d Cong., 2d Sess. (1934). Section 9 (c) of the bills, from which present § 10 (b) evolved, proscribed as unlawful the use of “any device or contrivance which, or any device or contrivance in a way or manner which the Commission may by its rules and regulations find detrimental to the public interest or to the proper protection of investors.” The other subsections of proposed § 9 listed specific practices that Congress empowered the Commission to regulate through its rulemaking power. See §§ 9 (a) (short sale), (b) (“stop-loss order”). Soon after the hearings on the House bill were held, a substitute bill was introduced in both Houses which abbreviated and modified *202§9 (c)’s operative language to read “any manipulative device or contrivance.” H. R. 8720, 73d Cong., 2d Sess., § 9 (c) (1934); see S. 3420, 73d Cong., 2d Sess., § 10 (b) (1934). Still a third bill, retaining the Commission's power to regulate the specific practices enumerated in the prior bills, and omitting all reference to the Commission’s authority to prescribe rules concerning manipulative or deceptive devices in general, was introduced and passed in the House. H. R. 9323, 73d Cong., 2d Sess., § 9 (1934). The final language of § 10 is a modified version of a Senate amendment to this last House bill. See H. R. Conf. Rep. No. 1838, 73d Cong., 2d Sess., 32-33 (1934).

Neither the intended scope of § 10 (b) nor the reasons for the changes in its operative language are revealed explicitly in the legislative history of the 1934 Act, which deals primarily with other aspects of the legislation. There is no indication, however, that § 10 (b) was intended to proscribe conduct not involving scienter. The extensive hearings that preceded passage of the 1934 Act touched only briefly on § 10, and most of the discussion was devoted to the enumerated devices that the Commission is empowered to proscribe under § 10 (a). The most relevant exposition of the provision that was to become § 10 (b) was by Thomas G. Corcoran, a spokesman for the drafters. Corcoran indicated:

“Subsection (c) [§ 9 (c) of H. R. 7852 — later § 10 (b) ] says, ‘Thou shalt not devise any other cunning devices.’

“Of course subsection (c) is a catch-all clause to prevent manipulative devices. I do not think there is any objection to that kind of clause. The Commission should have the authority to deal with new manipulative devices.” Hearings on H. R. 7852 *203and H. R. 8720 before the House Committee on Interstate and Foreign Commerce, 73d Cong., 2d Sess., 115 (1934).

This brief explanation of § 10 (b) by a spokesman for its drafters is significant. The section was described rightly as a “catchall” clause to enable the Commission “to deal with new manipulative [or cunning] devices.” It is difficult to believe that any lawyer, legislative draftsman, or legislator would use these words if the intent was to create liability for merely negligent acts or omissions.23 Neither the legislative history nor the briefs supporting respondents identify any usage or authority for construing “manipulative [or cunning] devices” to include negligence.24

*204The legislative reports do not address the scope of § 10 (b) or its catchall function directly. In considering specific manipulative practices left to Commission regulation, however, the reports indicate that liability would not attach absent scienter, supporting the conclusion that Congress intended no lesser standard under § 10 (b). The Senate Report of S. 3420 discusses generally the various abuses that precipitated the need for the legislation and the inadequacy of self-regulation by the stock exchanges. The Report then analyzes the component provisions of the statute, but does not parse § 10. The only specific reference to § 10 is the following:

“In addition to the discretionary and elastic powers conferred on the administrative authority, effective regulation must include several clear statutory provisions reinforced by penal and civil sanctions, aimed at those manipulative and deceptive practices which have been demonstrated to fulfill no useful *205function. These sanctions are found in sections 9,10 and 16.” S. Rep. No. 792, 73d Cong., 2d Sess., 6 (1934).

In the portion of the general-analysis section of the Report entitled Manipulative Practices, however, there is a discussion of specific practices that were considered so inimical to the public interest as to require express prohibition, such as “wash” sales and “matched” orders,25 and of other practices that might in some cases serve legitimate purposes, such as stabilization of security prices and grants of options. Id., at 7-9. These latter practices were left to regulation by the Commission. 1934 Act §§ 9 (a) (6), (c), 48 Stat. 890, 15 U. S. C. §§ 78i (a)(6), (c). Significantly, we think, in the discussion of the need to regulate even the latter category of practices when they are manipulative, there is no indication that any type of criminal or civil liability is to attach in the absence of scienter. Furthermore, in commenting on the express civil liabilities provided in the 1934 Act, the Report explains:

“[I]f air investor has suffered loss by reason of illicit practices, it is equitable that he should be allowed to recover damages from the guilty party.... [T]he bill provides that any person who unlaw*206fully manipulates the price of a security, or who induces transactions in a security by means of false or misleading statements, or who makes a false or misleading statement in the report of a corporation, shall be liable in damages to those who have bought or sold the security at prices affected by such violation or statement. In such case the burden is on the plaintiff to show the violation or the fact that the statement was false or misleading, and that he relied thereon to his damage. The defendant may escape liability by showing that the statement was made in good faith.” S. Rep. No. 792, supra, at 12-13 (emphasis supplied).

The Report therefore reveals with respect to the specified practices, an overall congressional intent to prevent “manipulative and deceptive practices which . . . fulfill no useful function” and to create private actions for damages stemming from “illicit practices,” where the defendant has not acted in good faith. The views expressed in the House Report are consistent with this interpretation. H. R. Rep. No. 1383, 73d Cong., 2d Sess., 10-11, 20-21 (1934) (H. R. 9323). There is no indication that Congress intended anyone to be made liable for such practices unless he acted other than in good faith. The catchall provision of § 10 (b) should be interpreted no more broadly.

C

The 1933 and 1934 Acts constitute interrelated components of the federal regulatory scheme governing transactions in securities. See Blue Chip Stamps, 421 U. S., at 727-730. As the Court indicated in SEC v. National Securities, Inc., 393 U. S. 453, 466 (1969), “the interdependence of the various sections of the securities laws is certainly a relevant factor in any interpretation of the language Congress has chosen . . . .” Recognizing this, *207respondents and the Commission contrast § 10 (b) with other sections of the Acts to support their contention that civil liability may be imposed upon proof of negligent conduct. We think they misconceive the significance of the other provisions of the Acts.

The Commission argues that Congress has been explicit in requiring willful conduct when that was the standard of fault intended, citing § 9 of the 1934 Act, 48 Stat. 889, 15 U. S. C. § 78i, which generally proscribes manipulation of securities prices. Sections 9(a)(1) and (a)(2), for example, respectively prohibit manipulation of security prices “[f]or the purpose of creating a false or misleading appearance of active trading in any security . . . or . . . with respect to the market for any such security,” and “for the purpose of inducing the purchase or sale of such security by others.” See also § 9 (a)(4). Section 9 (e) then imposes upon “[a]ny person who willfully participates in any act or transaction in violation of” other provisions of § 9 civil liability to anyone who purchased or sold a security at a price affected by the manipulative activities. From this the Commission concludes that since § 10 (b) is not by its terms explicitly restricted to willful, knowing, or purposeful conduct, it should not be construed in all cases to require more than negligent action or inaction as a precondition for civil liability.

The structure of the Acts does not support the Commission’s argument. In each instance that Congress created express civil liability in favor of purchasers or sellers of securities it clearly specified whether recovery was to be premised on knowing or intentional conduct, negligence, or entirely innocent mistake. See 1933 Act, §§11, 12, 15, 48 Stat. 82, 84, as amended, 15 U. S. C. §§ 77k, 771, 77o; 1934 Act §§ 9, 18, 20, 48 Stat. 889, 897, 899, as amended, 15 U. S. C. §§ 78i, 78r, 78t. For example, § 11 of the 1933 Act unambigu*208ously creates a private action for damages when a registration statement includes untrue statements of material facts or fails to state material facts necessary to make the statements therein not misleading. Within the limits specified by § 11 (e), the issuer of the securities is held absolutely liable for any damages resulting from such misstatement or omission. But experts such as accountants who have prepared portions of the registration statement are accorded a “due diligence” defense. In effect, this is a negligence standard. An expert may avoid civil liability with respect to the portions of the registration statement for which he was responsible by showing that “after reasonable investigation” he had “reasonable ground [s] to believe” that the statements'for which he was responsible were true and there was no omission of a material fact.26 § 11 (b) (3) (B) (i). See, e. g., Escott v. Barchris Constr. Corp., 283 F. Supp. 643, 697-703 (SDNY 1968). The express recognition of a cause of action premised on negligent behavior in § 11 stands in sharp contrast to the language of § 10 (b), and significantly undercuts the Commission’s argument.

We also consider it significant that each of the express civil remedies in the 1933 Act allowing recovery for negligent conduct, see §§ 11, 12 (2), 15, 15 U. S. C. §§ 77k, 771 *209(2), 77o,27 is subject to significant procedural restrictions not applicable under § 10 (b).28 Section 11 (e) of the 1933 Act, for example, authorizes the court to require a *210plaintiff bringing a suit under § 11, § 12 (2), or § 15 thereof to post a bond for costs, including attorneys’ fees, and in specified circumstances to assess costs at the conclusion of the litigation. Section 13 specifies a statute of limitations of one year from the time the violation was or should have been discovered, in no event to exceed three years from the time of offer or sale, applicable to actions brought under § 11, § 12 (2), or § 15. These restrictions, significantly, were imposed by amendments to the 1933 Act adopted as part of the 1934 Act. Prior to amendment § 11 (e) contained no provision for payment of costs. Act of May 27, 1933, c. 38, § 11 (e), 48 Stat. 83. See Act of June 6, 1934, c. 404, § 206 (e), 48 Stat. 907. The amendments also substantially shortened the statute of limitations provided by § 13. Compare § 13, 48 Stat. 84, with 15 U. S. C. § 77m. See 1934 Act, § 207, 48 Stat. 908. We think these procedural limitations indicate that the judicially created private damages remedy under § 10 (b) — which has no comparable restrictions29 — cannot be extended, consistently with the intent of Congress, to actions premised on negligent wrongdoing. Such extension would allow causes of action covered by §§ 11, 12 (2), and 15 to be brought instead under § 10 (b) and thereby nullify the effectiveness of the carefully drawn procedural restrictions on these express actions.30 See, e. g., Fisch*211 man v. Raytheon Mfg. Co., 188 F. 2d 783, 786-787 (CA2 1951); SEC v. Texas Gulf Sulphur Co., 401 F. 2d, at 867-868 (Friendly, J., concurring); Rosenberg v. Globe Aircraft Corp., 80 F. Supp. 123, 124 (ED Pa. 1948); 3 Loss, supra, n. 17, at 1787-1788; R. Jennings & H. Marsh, Securities Regulation 1070-1074 (3d ed. 1972). We would be unwilling to bring about this result absent substantial support in the legislative history, and there is none.31

*212D

We have addressed, to this point, primarily the language and history of § 10(b). The Commission contends, however, that subsections (b) and (c) of Rule 10b-5 are cast in language which — if standing alone— could encompass both intentional and negligent behavior. These subsections respectively provide that it is unlawful “[t]o 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 . . and “[t]o engage in any act, practice, or course of business which operates or would operate as a fraud or deceit upon any person . . . Viewed in isolation the language of subsection (b), and arguably that of subsection (c), could be read as proscribing, respectively, any type of material misstatement or omission, and any course of conduct, that has the effect of defrauding investors, whether the wrongdoing was intentional or not.

We note first that such a reading cannot be harmonized with the administrative history of the Rule, a history making clear that when the Commission adopted the Rule it was intended to apply only to activities that involved scienter.32 More importantly, Rule 10b-5 was *213adopted pursuant to authority granted the Commission under § 10 (b). The rulemaking power granted to an administrative agency charged with the administration of a federal statute is not the power to make law. Rather, *214it is “ ‘the power to adopt regulations to carry into effect the will of Congress as expressed by the statute.’ ” Dixon v. United States, 381 U. S. 68, 74 (1965), quoting Manhattan General Equipment Co. v. Commissioner, 297 U. S. 129, 134 (1936). Thus, despite the broad view of the Rule advanced by the Commission in this case, its scope cannot exceed the power granted the Commission by Congress under § 10 (b). For the reasons stated above, we think the Commission’s original interpretation of Rule 10b-5 was compelled by the language and history of § 10 (b) and related sections of the Acts. See, e. g., Gerstle v. Gamble-Skogmo, Inc., 478 F. 2d 1281, 1299 (CA2 1973); Lanza v. Drexel & Co., 479 F. 2d 1277, 1304-1305 (CA2 1973); SEC v. Texas Gulf Sulphur Co., 401 F. 2d, at 868 (Friendly, J., concurring); 3 Loss, supra, n. 17, at 1766; 6 id., at 3883-3885. When a statute speaks so specifically in terms of manipulation and deception, and of implementing devices and contrivances — the commonly understood terminology of intentional wrongdoing — and when its history reflects no more expansive intent, we are quite unwilling to extend the scope of the statute to negligent conduct.33

*215III

Recognizing that § 10 (b) and Rule 10b-5 might be held to require proof of more than negligent nonfeasance by Ernst & Ernst as a precondition to the imposition of civil liability, respondents further contend that the case should be remanded for trial under whatever standard is adopted. Throughout the lengthy history of this case respondents have proceeded on a theory of liability premised on negligence, specifically disclaiming that Ernst & Ernst had engaged in fraud or intentional misconduct.34 In these circumstances, we think it inappropriate to remand the action for further proceedings.

The judgment of the Court of Appeals is

Reversed.

Mr. Justice Stevens took no part in the consideration or decision of this case.

Mr. Justice Blackmun,

with whom Mr. Justice Brennan joins, dissenting.

Once again — see Blue Chip Stamps v. Manor Drug Stores, 421 U. S. 723, 730 (1975) — the Court interprets *216§ 10 (b) of the Securities Exchange Act of 1934, 15 U. S. C. § 78j (b), and the Securities and Exchange Commission’s Rule 10b-5, 17 CFR § 240.10b-5 (1975), restrictively and narrowly and thereby stultifies recovery for the victim. This time the Court does so by confining the statute and the Rule to situations where the defendant has “scienter,” that is, the “intent to deceive, manipulate, or defraud.” Sheer negligence, the Court says, is not within the reach of the statute and the Rule, and was not contemplated when the great reforms of 1933, 1934, and 1942 were effectuated by Congress and the Commission.

Perhaps the Court is right, but I doubt it. The Government and the Commission doubt it too, as is evidenced by the thrust of the brief filed by the Solicitor General on behalf of the Commission as amicus curiae. The Court’s opinion, to be sure, has a certain technical consistency about it. It seems to me, however, that an investor can be victimized just as much by negligent conduct as by positive deception, and that it is not logical to drive a wedge between the two, saying that Congress clearly intended the one but certainly not the other.

No one questions the fact that the respondents here were the victims of an intentional securities fraud practiced by Leston B. Nay. What is at issue, of course, is the petitioner accountant firm’s involvement and that firm’s responsibility under Rule 10b-5. The language of the Rule, making it unlawful for any person “in connection with the purchase or sale of any security”

“(b) To make any untrue statement of a material *217fact 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,”

seems to me, clearly and succinctly, to prohibit negligent as well as intentional conduct of the kind proscribed, to extend beyond common-law fraud, and to apply to negligent' omission and commission. This is consistent with Congress' intent, repeatedly recognized by the Court, that securities legislation enacted for the purpose of avoiding frauds be construed “not technically and restrictively, but flexibly to effectuate its remedial purposes.” SEC v. Capital Gains Research Bureau, 375 U. S. 180, 195 (1963); Superintendent of Insurance v. Bankers Life & Cas. Co., 404 U. S. 6, 12 (1971); Affiliated Ute Citizens v. United States, 406 U. S. 128, 151 (1972).

On motion for summary judgment, therefore, the respondents' allegations, in my view, were sufficient, and the District Court’s dismissal of the action was improper to the extent that the dismissal rested on the proposition that suit could not be maintained under § 10 (b) and Rule 10b-5 for mere negligence. The opposite appears to be true, at least in the Second Circuit, with respect to suits by the SEC to enjoin a violation of the Rule. SEC v. Management Dynamics, Inc., 515 F. 2d 801 (1975); SEC v. Spectrum, Ltd., 489 F. 2d 535, 541 (1973); SEC v. Texas Gulf Sulphur Co., 401 F. 2d 833, 854-855 (1968), cert. denied sub nom. Coates v. SEC, 394 U. S. 976 (1969). I see no real distinction between that situation and this one, for surely the question whether negligent conduct violates the Rule should not depend upon the plaintiff’s identity. If negligence is a violation factor *218when the SEC sues, it must be a violation factor when a private party sues. And, in its present posture, this case is concerned with the issue of violation, not with the secondary issue of a private party’s judicially created entitlement to damages or other specific relief. See Rondeau v. Mosinee Paper Corp., 422 U. S. 49 (1975).

The critical importance of the auditing accountant’s role in insuring full disclosure cannot be overestimated. The SEC has emphasized that in certifying statements the accountant’s duty “is to safeguard the public interest, not that of his client.” In re Touche, Niven, Bailey & Smart, 37 S. E. C. 629, 670-671 (1957). “In our complex society the accountant’s certificate and the lawyer’s opinion can be instruments for inflicting pecuniary loss more potent than the chisel or the crowbar.” United States v. Benjamin, 328 F. 2d 854, 863 (CA2), cert. denied sub nom. Howard v. United States, 377 U. S. 953 (1964). In this light, the initial inquiry into whether Ernst & Ernst’s preparation and certification of the financial statements of First Securities Company of Chicago were negligent, because of the failure to perceive Nay’s extraordinary mail rule, and in other alleged respects, and thus whether Rule 10b-5 was violated, should not be thwarted.

But the Court today decides that it is to be thwarted, and so once again it rests with Congress to rephrase and to re-enact, if investor victims, such as these, are ever to have relief under the federal securities laws that I thought had been enacted for their broad, needed, and deserving benefit.*

*313Justice Ginsburg

delivered the opinion of the Court.

This Court has long recognized that meritorious private actions to enforce federal antifraud securities laws are an essential supplement to criminal prosecutions and civil enforcement actions brought, respectively, by the Department of Justice and the Securities and Exchange Commission (SEC). See, e. g., Dura Pharmaceuticals, Inc. v. Broudo, 544 U. S. 336, 345 (2005); J. I. Case Co. v. Borak, 377 U. S. 426, 432 (1964). Private securities fraud actions, however, if not adequately contained, can be employed abusively to impose substantial costs on companies and individuals whose conduct conforms to the law. See Merrill Lynch, Pierce, Fenner & Smith Inc. v. Dabit, 547 U. S. 71, 81 (2006). As a check against abusive litigation by private parties, Congress enacted the Private Securities Litigation Reform Act of 1995 (PSLRA), 109 Stat. 737.

Exacting pleading requirements are among the control measures Congress included in the PSLRA. The PSLRA requires plaintiffs to state with particularity both the facts constituting the alleged violation, and the facts evidencing scienter, i. e., the defendant’s intention “to deceive, manipulate, or defraud.” Ernst & Ernst v. Hochfelder, 425 U. S. 185, 194, and n. 12 (1976); see 15 U.S.C. § 78u-4(b)(1), (2). *314This case concerns the latter requirement. As set out in §21D(b)(2) of the PSLRA, plaintiffs must “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind.” 15 U. S. C. § 78u-4(b)(2).

Congress left the key term “strong inference” undefined, and Courts of Appeals have divided on its meaning. In the case before us, the Court of Appeals for the Seventh Circuit held that the “strong inference” standard would be met if the complaint “allege[d] facts from which, if true, a reasonable person could infer that the defendant acted with the required intent.” 437 F. 3d 588, 602 (2006). That formulation, we conclude, does not capture the stricter demand Congress sought to convey in §21D(b)(2). It does not suffice that a reasonable factfinder plausibly could infer from the complaint’s allegations the requisite state of mind. Rather, to determine whether a complaint’s scienter allegations can survive threshold inspection for sufficiency, a court governed by §21D(b)(2) must engage in a comparative evaluation; it must consider, not only inferences urged by the plaintiff, as the Seventh Circuit did, but also competing inferences rationally drawn from the facts alleged. An inference of fraudulent intent may be plausible, yet less cogent than other, nonculpable explanations for the defendant’s conduct. To qualify as “strong” within the intendment of §21D(b)(2), we hold, an inference of scienter must be more than merely plausible or reasonable — it must be cogent and at least as compelling as any opposing inference of nonfraudulent intent.

I

Petitioner Tellabs, Inc., manufactures specialized equipment used in fiber optic networks. During the time period relevant to this case, petitioner Richard Notebaert was Tel-labs’ chief executive officer and president. Respondents (Shareholders) are persons who purchased Tellabs stock between December 11, 2000, and June 19, 2001. They accuse *315Tellabs and Notebaert (as well as several other Tellabs executives) of engaging in a scheme to deceive the investing public about the true value of Tellabs’ stock. See 437 F. 3d, at 591; App. 94-98.1

Beginning on December 11,2000, the Shareholders allege, Notebaert (and by imputation Tellabs) “falsely reassured public investors, in a series of statements . . . that Tellabs was continuing to enjoy strong demand for its products and earning record revenues,” when, in fact, Notebaert knew the opposite was true. Id., at 94-95, 98. From December 2000 until the spring of 2001, the Shareholders claim, Notebaert knowingly misled the public in four ways. 437 F. 3d, at 596. First, he made statements indicating that demand for Tel-labs’ flagship networking device, the TITAN 5500, was continuing to grow, when, in fact, demand for that product was waning. Id., at 596, 597. Second, Notebaert made statements indicating that the TITAN 6500, Tellabs’ next-generation networking device, was available for delivery, and that demand for that product was strong and growing, when in truth the product was not ready for delivery and demand was weak. Id., at 596, 597-598. Third, he falsely represented Tellabs’ financial results for the fourth quarter of 2000 (and, in connection with those results, condoned the practice of “channel stuffing,” under which Tellabs flooded its customers with unwanted products). Id., at 596, 598. Fourth, Notebaert made a series of overstated revenue projections, when demand for the TITAN 5500 was drying up and production of the TITAN 6500 was behind schedule. Id., at 596, 598-599. Based on Notebaert’s sunny assessments, the *316Shareholders contend, market analysts recommended that investors buy Tellabs’ stock. See id., at 592.

The first public glimmer that business was not so healthy came in March 2001 when Tellabs modestly reduced its first quarter sales projections. Ibid. In the next months, Tel-labs made progressively more cautious statements about its projected sales. On June 19, 2001, the last day of the class period, Tellabs disclosed that demand for the TITAN 5500 had significantly dropped. Id., at 593. Simultaneously, the company substantially lowered its revenue projections for the second quarter of 2001. The next day, the price of Tellabs stock, which had reached a high of $67 during the period, plunged to a low of $15.87. Ibid.

On December 3,2002, the Shareholders filed a class action in the District Court for the Northern District of Illinois. Ibid. Their complaint stated, inter alia, that Tellabs and Notebaert had engaged in securities fraud in violation of § 10(b) of the Securities Exchange Act of 1934, 48 Stat. 891, 15 U. S. C. §78j(b), and SEC Rule 10b-5,17 CFR §240.10b-5 (2006), also that Notebaert was a “controlling person” under § 20(a) of the 1934 Act, 15 U. S. C. §78t(a), and therefore derivatively liable for the company’s fraudulent acts. See App. 98-101, 167-171. Tellabs moved to dismiss the complaint on the ground that the Shareholders had failed to plead their case with the particularity the PSLRA requires. The District Court agreed, and therefore dismissed the complaint without prejudice. App. to Pet. for Cert. 80a-117a; see Johnson v. Tellabs, Inc., 303 F. Supp. 2d 941, 945 (ND Ill. 2004).

The Shareholders then amended their complaint, adding references to 27 confidential sources and making further, more specific, allegations concerning Notebaert’s mental state. See 437 F. 3d, at 594; App. 91-93,152-160. The District Court again dismissed, this time with prejudice. 303 F. Supp. 2d, at 971. The Shareholders had sufficiently, pleaded that Notebaert’s statements were misleading, the *317court determined, id., at 955-961, but they had insufficiently alleged that he acted with scienter, id., at 954-955, 961-969.

The Court of Appeals for the Seventh Circuit reversed in relevant part. 437 F. 3d, at 591. Like the District Court, the Court of Appeals found that the Shareholders had pleaded the misleading character of Notebaert’s statements with sufficient particularity. Id., at 595-600. Unlike the District Court, however, the Seventh Circuit concluded that the Shareholders had sufficiently alleged that Notebaert acted with the requisite state of mind. Id., at 603-605.

The Court of Appeals recognized that the PSLRA “unequivocally raise[d] the bar for pleading scienter” by requiring plaintiffs to “plea[d] sufficient facts to create a strong inference of scienter.” Id., at 601 (internal quotation marks omitted). In evaluating whether that pleading standard is met, the Seventh Circuit said, “courts [should] examine all of the allegations in the complaint and then ... decide whether collectively they establish such an inference.” Ibid. “[W]e will allow the complaint to survive,” the court next and critically stated, “if it alleges facts from which, if true, a reasonable person could infer that the defendant acted with the required intent.... If a reasonable person could not draw such an inference from the alleged facts, the defendants are entitled to dismissal.” Id., at 602.

In adopting its standard for the survival of a complaint, the Seventh Circuit explicitly rejected a stiffer standard adopted by the Sixth Circuit, i. e., that “plaintiff's are entitled only to the most plausible of competing inferences.” Id., at 601, 602 (quoting Fidel v. Farley, 392 F. 3d 220, 227 (2004)). The Sixth Circuit’s standard, the court observed, because it involved an assessment of competing inferences, "could potentially infringe upon plaintiffs’ Seventh Amendment rights.” 437 F. 3d, at 602. We granted certiorari to resolve the disagreement among the Circuits on whether, and to what extent, a court must consider competing inferences in determining whether a securities fraud complaint *318gives rise to a “strong inference” of scienter.2 549 U. S. 1105 (2007).

II

Section 10(b) of the Securities Exchange Act of 1934 forbids the “use or employ, in connection with'the purchase or sale of any security ... , [of] any manipulative or deceptive device or contrivance in contravention of such rules and regulations as the [SEC] may prescribe as necessary or appropriate in the public interest or for the protection of investors.” 15 U. S. C. §78j(b). SEC Rule 10b-5 implements § 10(b) by declaring it unlawful:

“(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 ... 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 CFR §240.10b-5.

Section 10(b), this Court has implied from the statute’s text and purpose, affords a right of action to purchasers or sellers of securities injured by its violation. See, e. g., Dura Pharmaceuticals, 544 U. S., at 341. See also id., at 345 (“The securities statutes seek to maintain public confidence in the marketplace ... by deterring fraud, in part, through the availability of private securities fraud actions.”); Borak, 377 U. S., at 432 (private securities fraud actions provide “a most effective weapon in the enforcement” of securities laws and *319are “a necessary supplement to Commission action”). To establish liability under § 10(b) and Rule 10b-5, a private plaintiff must prove that the defendant acted with scienter, “a mental state embracing intent to deceive, manipulate, or defraud.” Ernst & Ernst, 425 U. S., at 193-194, and n. 12.3 ****8

In an ordinary civil action, the Federal Rules of Civil Procedure require only “a short and plain statement of the claim showing that the pleader is entitled to relief.” Fed. Rule Civ. Proc. 8(a)(2). Although the rule encourages brevity, the complaint must say enough to give the defendant “fair notice of what the plaintiff’s claim is and the grounds upon which it rests.” Dura Pharmaceuticals, 544 U. S., at 346 (internal quotation marks omitted). Prior to the enactment of the PSLRA, the sufficiency of a complaint for securities fraud was governed not by Rule 8, but by the heightened pleading standard set forth in Rule 9(b). See Greenstone v. Cambex Corp., 975 F. 2d 22, 25 (CA1 1992) (Breyer, J.) (collecting cases). Rule 9(b) applies to “all averments of fraud or mistake”; it requires that “the circumstances constituting fraud ... be stated with particularity” but provides that “[mjalice, intent, knowledge, and other condition of mind of a person may be averred generally.”

Courts of Appeals diverged on the character of the Rule 9(b) inquiry in § 10(b) cases: Could securities fraud plaintiffs allege the requisite mental state “simply by saying that sci-enter existed,” In re GlenFed, Inc. Securities Litigation, 42 F. 3d 1541, 1546-1547 (CA9 1994) (en bane), or were they required to allege with particularity facts giving rise to an *320inference of scienter? Compare id., at 1546 (“We are not permitted to add new requirements to Rule 9(b) simply because we like the effects of doing so.”), with, e. g., Greenstone, 975 F. 2d, at 25 (were the law to permit a securities fraud complaint simply to allege scienter without supporting facts, “a complaint could evade too easily the ‘particularity’ requirement in Rule 9(b)’s first sentence”). Circuits requiring plaintiffs to allege specific facts indicating scienter expressed that requirement variously. See 5A C. Wright & A. Miller, Federal Practice and Procedure § 1301.1, pp. 300-302 (3d ed. 2004) (hereinafter Wright & Miller). The Second Circuit’s formulation was the most stringent. Securities fraud plaintiffs in that Circuit were required to “specifically plead those [facts] which they assert give rise to a strong inference that the defendants had” the requisite state of mind. Ross v. A. H. Robins Co., 607 F. 2d 545, 558 (1979) (emphasis added). The “strong inference” formulation was appropriate, the Second Circuit said, to ward off allegations of “fraud by hindsight.” See, e.g., Shields v. Citytrust Bancorp, Inc., 25 F. 3d 1124, 1129 (1994) (quoting Denny v. Barber, 576 F. 2d 465, 470 (CA2 1978) (Friendly, J.)).

Setting a uniform pleading standard for § 10(b) actions was among Congress’ objectives when it enacted the PSLRA. Designed to curb perceived abuses of the § 10(b) private action — “nuisance filings, targeting of deep-pocket defendants, vexatious discovery requests and manipulation by class action lawyers,” Dabit, 547 U. S., at 81 (quoting H. R. Conf. Rep. No. 104-369, p. 31 (1995) (hereinafter H. R. Conf. Rep.)) — the PSLRA installed both substantive and procedural controls.4 Notably, Congress prescribed new proce*321dures for the appointment of lead plaintiffs and lead counsel. This innovation aimed to increase the likelihood that institutional investors — parties more likely to balance the interests of the class with the long-term interests of the company— would serve as lead plaintiffs. See id., at 33-34; S. Rep. No. 104-98, p. 11 (1995). Congress also “limit[ed] recoverable damages and attorney’s fees, provide[d] a ‘safe harbor’ for forward-looking statements,... mandate[d] imposition of sanctions for frivolous litigation, and authorize^] a stay of discovery pending resolution of any motion to dismiss.” Dabit, 547 U. S., at 81. And in §21D(b) of the PSLRA, Congress “impose[d] heightened pleading requirements in actions brought pursuant to § 10(b) and Rule 10b-5.” Ibid.

Under the PSLRA’s heightened pleading instructions, any private securities complaint alleging that the defendant made a false or misleading statement must: (1) “specify each statement alleged to have been misleading [and] the reason or reasons why the statement is misleading,” 15 U. S. C. § 78u-4(b)(l); and (2) “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind,” § 78u-4(b)(2). In the instant case, as earlier stated, see supra, at 317, the District Court and the Seventh Circuit agreed that the Shareholders met the first of the two requirements: The complaint sufficiently specified Notebaert’s alleged misleading statements and the reasons why the statements were misleading. 303 F. Supp. 2d, at 955-961; 437 F. 3d, at 596-600. But those courts disagreed on whether the Shareholders, as required by §21D(b)(2), “state[d] with particularity facts giving rise to a strong inference that [Notebaert] acted with [scienter],” § 78u-4(b)(2). See supra, at 317.

The “strong inference” standard “unequivocally raise[d] the bar for pleading scienter,” 437 F. 3d, at 601, and signaled Congress’ purpose to promote greater uniformity among the Circuits, see H. R. Conf. Rep., p. 41. But “Congress did not. .. throw much light on what facts . .. suffice to create *322[a strong] inference,” or on what “degree of imagination courts can use in divining whether” the requisite inference exists. 437 F. 3d, at 601. While adopting the Second Circuit’s “strong inference” standard, Congress did not codify that Circuit’s case law interpreting the standard. See § 78u-4(b)(2). See also Brief for United States as Amicus Curiae 18. With no clear guide from Congress other than its “in-tension] to strengthen existing pleading requirements,” H. R. Conf. Rep., p. 41, Courts of Appeals have diverged again, this time in construing the term “strong inference.” Among the uncertainties, should courts consider competing inferences in determining whether an inference of scienter is “strong”? See 437 F. 3d, at 601-602 (collecting cases). Our task is to prescribe a workable construction of the “strong inference” standard, a reading geared to the PSLRA’s twin goals: to curb frivolous, lawyer-driven litigation, while preserving investors’ ability to recover on meritorious claims.

Ill

A

We establish the following prescriptions: First, faced with a Rule 12(b)(6) motion to dismiss a § 10(b) action, courts must, as with any motion to dismiss for failure to plead a claim on which relief can be granted, accept all factual allegations in the complaint as true. See Leatherman v. Tarrant County Narcotics Intelligence and Coordination Unit, 507 U. S. 163, 164 (1993). On this point, the parties agree. See Reply Brief 8; Brief for Respondents 26; Brief for United States as Amicus Curiae 8, 20, 21.

Second, courts must consider the complaint in its entirety, as well as other sources courts ordinarily examine when ruling on Rule 12(b)(6) motions to dismiss, in particular, documents incorporated into the complaint by reference, and matters of which a court may take judicial notice. See 5B Wright & Miller § 1357 (3d ed. 2004 and Supp. 2007). The inquiry, as several Courts of Appeals have recognized, is *323whether all of the facts alleged, taken collectively, give rise to a strong inference of scienter, not whether any individual allegation, scrutinized in isolation, meets that standard. See, e.g., Abrams v. Baker Hughes Inc., 292 F. 3d 424, 431 (CA5 2002); Gompper v. VISX, Inc., 298 F. 3d 893, 897 (CA9 2002). See also Brief for United States as Amicus Curiae 25.

Third, in determining whether the pleaded facts give rise to a “strong” inference of scienter, the court must take into account plausible opposing inferences. The Seventh Circuit expressly declined to engage in such a comparative inquiry. A complaint could survive, that court said, as long as it “alleges facts from which, if true, a reasonable person could infer that the defendant acted with the required intent”; in other words, only “[i]f a reasonable person could not draw such an inference from the alleged facts” would the defendant prevail on a motion to dismiss. 437 F. 3d, at 602. But in § 21D(b)(2), Congress did not merely require plaintiffs to “provide a factual basis for [their] scienter allegations,” ibid. (quoting In re Cerner Corp. Securities Litigation, 425 F. 3d 1079, 1084, 1085 (CA8 2005)), i. e., to allege facts from which an inference of scienter rationally could be drawn. Instead, Congress required plaintiffs to plead with particularity facts that give rise to a “strong” — i. e., a powerful or cogent— inference. See American Heritage Dictionary 1717 (4th ed. 2000) (defining “strong” as “[p]ersuasive, effective, and cogent”); 16 Oxford English Dictionary 949 (2d ed. 1989) (defining “strong” as “[powerful to demonstrate or convince” (definition 16b)); cf. 7 id., at 924 (defining “inference” as “a conclusion [drawn] from known or assumed facts or statements”; “reasoning from something known or assumed to something else which follows from it”).

The strength of an inference cannot be decided in a vacuum. The inquiry is inherently comparative: How likely is it that one conclusion, as compared to others, follows from the underlying facts? To determine whether the plaintiff *324has alleged facts that give rise to the requisite “strong inference” of scienter, a court must consider plausible, non-culpable explanations for the defendant’s conduct, as well as inferences favoring the plaintiff. The inference that the defendant acted with scienter need not be irrefutable, i. e., of the “smoking-gun” genre, or even the “most plausible of competing inferences,” Fidel, 392 F. 3d, at 227 (quoting Helwig v. Vencor, Inc., 251 F. 3d 540, 553 (CA6 2001) (en banc)). Recall in this regard that §21D(b)’s pleading requirements are but one constraint among many the PSLRA installed to screen out frivolous suits, while allowing meritorious actions to move forward. See supra, at 320-321, and n. 4. Yet the inference of scienter must be more than merely “reasonable” or “permissible” — it must be cogent and compelling, thus strong in light of other explanations. A complaint will survive, we hold, only if- a reasonable person would deem the inference of scienter cogent and at least as compelling as any opposing inference one could draw from the facts alleged.5

*325B

Tellabs contends that when competing inferences are considered, Notebaert’s evident lack of pecuniary motive will be dispositive. The Shareholders, Tellabs stresses, did not allege that Notebaert sold any shares during the class period. See Brief for Petitioners 50 (“The absence of any allegations of motive color all the other allegations putatively giving rise to an inference of scienter.”). While it is true that motive can be a relevant consideration, and personal financial gain may weigh heavily in favor of a scienter inference, we agree with the Seventh Circuit that the absence of a motive allegation is not fatal. See 437 F. 3d, at 601. As earlier stated, supra, at 322-323, allegations must be considered collectively; the significance that can be ascribed to an allegation of motive, or lack thereof, depends on the entirety of the complaint.

Tellabs also maintains that several of the Shareholders’ allegations are too vague or ambiguous to contribute to a strong inference of scienter. For example, the Shareholders alleged that Tellabs flooded its customers with unwanted products, a practice known as “channel stuffing.” See supra, at 315. But they failed, Tellabs argues, to specify whether the channel stuffing allegedly known to Notebaert was the illegitimate kind (e. g., writing orders for products customers had not requested) or the legitimate kind (e. g., offering customers discounts as an incentive to buy). Brief for Petitioners 44-46; Reply Brief 8. See also id., at 8-9 (complaint lacks precise dates of reports critical to distinguish legitimate conduct from culpable conduct). But see 437 F. 3d, at 598, 603-604 (pointing to multiple particulars *326alleged by the Shareholders, including specifications as to timing). We agree that omissions and ambiguities count against inferring scienter, for plaintiffs must “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind.” §78u-4(b)(2). We reiterate, however, that the court’s job is not to scrutinize each allegation in isolation but to assess all the allegations holistically. See supra, at 322-323; 437 F. 3d, at 601. In sum, the reviewing court must ask: When the allegations are accepted as true and taken collectively, would a reasonable person deem the inference of scienter at least as strong as any opposing inference?6

IV

Accounting for its construction of §21D(b)(2), the Seventh Circuit explained’that the court “th[ought] it wis[e] to adopt an approach that [could not] be misunderstood as a usurpation of the jury’s role.” 437 F. 3d, at 602. In our view, the Seventh Circuit’s concern was undue.7 A court’s comparative assessment of plausible inferences, while constantly as*327suming the plaintiff’s allegations to be true, we think it plain, does not impinge upon the Seventh Amendment right to jury trial.8

Congress, as creator of federal statutory claims, has power to prescribe what must be pleaded to state the claim, just as it has power to determine what must be proved to prevail on the merits. It is the federal lawmaker’s prerogative, therefore, to allow, disallow, or shape the contours of— including the pleading and proof requirements for — § 10(b) private actions. No decision of this Court questions that authority in general, or suggests, in particular, that the Seventh Amendment inhibits Congress from establishing whatever pleading requirements it finds appropriate for federal statutory claims. Cf. Swierkiewicz v. Sorema N. A., 534 U. S. 506, 512-513 (2002); Leatherman, 507 U. S., at 168 (both recognizing that heightened pleading requirements can be established by Federal Rule, citing Fed. Rule Civ. Proe. 9(b), which requires that fraud or mistake be pleaded with particularity).9

Our decision in Fidelity & Deposit Co. of Md. v. United States, 187 U. S. 315 (1902), is instructive. That case concerned a rule adopted by the Supreme Court of the District of Columbia in 1879 pursuant to rulemaking power delegated by Congress. The rule required defendants, in certain con*328tract actions, to file an affidavit “specifically stating ... , in precise and distinct terms, the grounds of his defen[s]e.” Id., at 318 (internal quotation marks omitted). The defendant’s affidavit was found insufficient, and judgment was entered for the plaintiff, whose declaration and supporting affidavit had been found satisfactory. Ibid. This Court upheld the District’s rule against the contention that it violated the Seventh Amendment. Id., at 320. Just as the purpose of § 21D(b) is to screen out frivolous complaints, the purpose of the prescription at issue in Fidelity & Deposit Co. was to “preserve the court from frivolous defen[s]es,” ibid. Explaining why the Seventh Amendment was not implicated, this Court said that the heightened pleading rule simply “prescribes the means of making an issue,” and that, when “[t]he issue [was] made as prescribed, the right of trial by jury accrues.” Ibid.; accord Ex parte Peterson, 253 U. S. 300, 310 (1920) (Brandéis, J.) (citing Fidelity & Deposit Co., and reiterating: “It does not infringe the constitutional right to a trial by jury [in a civil case], to require, with a view to formulating the issues, an oath by each party to the facts relied upon.”). See also Walker v. New Mexico & Southern Pacific R. Co., 165 U. S. 593, 596 (1897) (Seventh Amendment “does not attempt to regulate matters of pleading”).

In the instant case, provided that the Shareholders have satisfied the congressionally “prescribe^]... means of making an issue,” Fidelity & Deposit Co., 187 U. S., at 320, the case will fall within the jury’s authority to assess the credibility of witnesses, resolve any genuine issues of fact, and make the ultimate determination whether Notebaert and, by imputation, Tellabs acted with scienter. We emphasize, as well, that under our construction of the “strong inference” standard, a plaintiff is not forced to plead more than she would be required to prove at trial. A plaintiff alleging fraud in a § 10(b) action, we hold today, must plead facts rendering an inference of scienter at least as likely as any plausible opposing inference. At trial, she must then prove her *329case by a “preponderance of the evidence.” Stated otherwise, she must demonstrate that it is more likely than not that the defendant acted with scienter. See Herman & MacLean v. Huddleston, 459 U. S. 375, 390 (1983).

* * *

While we reject the Seventh Circuit’s approach to §21D(b)(2), we do not decide whether, under the standard we have described, see supra, at 322-326, the Shareholders’ allegations warrant “a strong inference that [Notebaert and Tellabs] acted with the required state of mind,” 15 U. S. C. § 78u-4(b)(2). Neither the District Court nor the Court of Appeals had the opportunity to consider the matter in light of the prescriptions we announce today. We therefore vacate the Seventh Circuit’s judgment so that the case may be reexamined in accord with our construction of §21D(b)(2).

The judgment of the Court of Appeals is vacated, and the case is remanded for further proceedings consistent with this opinion.

It is so ordered.

Justice Scalia,

concurring in the judgment.

I fail to see how an inference that is merely “at least as compelling as any opposing inference,” ante, at 314, can conceivably be called what the statute here at issue requires: a “strong inference,” 15 U. S. C. § 78u-4(b)(2). If a jade falcon were stolen from a room to which only A and B had access, could it possibly be said there was a “strong inference” that B was the thief? I think not, and I therefore think that the Court’s test must fail. In my view, the test should be whether the inference of scienter (if any) is more plausible than the inference of innocence.*

*330The Court’s explicit rejection of this reading, ante, at 323-324, and n. 5, rests on two assertions. The first (doubtless true) is that the statute does not require that “[t]he inference that the defendant acted with scienter ... be irrefutable, i. e., of the ‘smoking-gun’ genre,” ante, at 324. It is up to Congress, however, and not to us, to determine what pleading standard would avoid those extremities while yet effectively deterring baseless actions. Congress has expressed its determination in the phrase “strong inference”; it is our job to give that phrase its normal meaning. And if we are to abandon text in favor of unexpressed purpose, as the Court does, it is inconceivable that Congress’s enactment of stringent pleading requirements in the Private Securities Litigation Reform Act of 1995 somehow manifests the purpose of giving plaintiffs the edge in close cases.

The Court’s second assertion (also true) is that “an inference at least as likely as competing inferences can, in some cases, warrant recovery.” Ante, at 324, n. 5 (citing Summers v. Tice, 33 Cal. 2d 80, 84-87, 199 P. 2d 1, 3-5 (1948)). Summers is a famous case, however, because it sticks out of the ordinary body of tort law like a sore thumb. It represented “a relaxation” of “such proof as is ordinarily required” to succeed in a negligence action. Id., at 86,199 P. 2d, at 4 (internal quotation marks omitted). There is no indication that the statute at issue here was meant to relax the ordinary rule under which a tie goes to the defendant. To the contrary, it explicitly strengthens that rule by extending it to the pleading stage of a case.

*331One of petitioners’ amici suggests that my reading of the statute would transform the text from requiring a “strong” inference to requiring the “strongest” inference. See Brief for American Association for Justice as Amicus Curiae 27. The point might have some force if Congress could have more clearly adopted my standard by using the word “strongest” instead of the word “strong.” But the use of the superlative would not have made any sense given the provision’s structure: What does it mean to require a plaintiff to plead “facts giving rise to the strongest inference that the defendant acted with the required state of mind”? It is certainly true that, if Congress had wanted to adopt my standard with even greater clarity, it could have restructured the entire provision — to require, for example, that the plaintiff plead “facts giving rise to an inference of scienter that is more compelling than the inference that the defendant acted with a nonculpable state of mind.” But if one is to consider the possibility of total restructuring, it is equally true that, to express the Court’s standard, Congress could have demanded “an inference of scienter that is at least as compelling as the inference that the defendant acted with a nonculpable state of mind” Argument from the possibility of saying it differently is clearly a draw. We must be content to give “strong inference” its normal meaning. I hasten to add that, while precision of interpretation should always be pursued for its own sake, I doubt that in this instance what I deem to be the correct test will produce results much different from the Court’s. How often is it that inferences are precisely in equipoise? All the more reason, I think, to read the language for what it says.

The Court and the dissent criticize me for suggesting that there is only one reading of the text. Ante, at 324-325, n. 5; post, at 336, n. 1 (Stevens, J., dissenting). They are both mistaken. I assert only that mine is the natural reading of the statute (i. e., the normal reading), not that it is the only *332conceivable one. The Court has no standing to object to this approach, since it concludes that, in another respect, the statute admits of only one natural reading, namely, that competing inferences must be weighed because the strong-inference requirement “is inherently comparative,” ante, at 323. As for the dissent, it asserts that the statute cannot possibly have a natural and discernible meaning, since “Courts of Appeals” and “Members of this Court” “have divided” over the question. Post, at 336, n. 1. It was just weeks ago, however, that the author of the dissent, joined by the author of today’s opinion for the Court, concluded that a statute’s meaning was “plain,” Rockwell Int’l Corp. v. United States, 549 U. S. 457, 479 (2007) (STEVENS, J., dissenting), even though the Courts of Appeals and Members of this Court divided over the question, id., at 470, n. 5. Was plain meaning then, as the dissent claims it is today, post, at 336, n. 1, “in the eye of the beholder”?

It is unremarkable that various Justices in this cáse reach different conclusions about the correct interpretation of the statutory text. It is remarkable, however, that the dissent believes that Congress “implicitly delegated significant lawmaking authority to the Judiciary in determining how th[e] [strong-inference] standard should operate in practice.” Post, at 335. This is language usually employed to describe the discretion conferred upon administrative agencies, which need not adopt what courts would consider the interpretation most faithful to the text of the statute, but may choose some other interpretation, so long as it is within the bounds of the reasonable, and may later change to some other interpretation that is within the bounds of the reasonable. See Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984). Courts, by contrast, must give the statute its single, most plausible, reading. To describe this as an exercise of “delegated lawmaking authority” seems to me peculiar — unless one believes in lawmakers who have no discretion. Courts must apply judgment, to be sure. But judgment is not discretion.

*333Even if I agreed with the Court’s interpretation of “strong inference,” I would not join the Court’s opinion because of its frequent indulgence in the last remaining legal fiction of the West: that the report of a single committee of a single House expresses the will of Congress. The Court says, for example, that “Congresses] purpose” was “to promote greater uniformity among the Circuits,” ante, at 321, relying for that certitude upon the statement of managers accompanying a House Conference Committee Report whose text was never adopted by the House, much less by the Senate, and as far as we know was read by almost no one. The Court is sure that Congress “ *inten[ded] to strengthen existing pleading requirements,”’ ante, at 322, because — again— the statement of managers said so. I come to the same conclusion for the much safer reason that the law which Congress adopted (and which the Members of both Houses actually voted on) so indicates. And had the legislation not done so, the statement of managers assuredly could not have remedied the deficiency.

With the above exceptions, I am generally in agreement with the Court’s analysis, and so concur in its judgment.

Justice Alito,

concurring in the judgment.

I agree with the Court that the Seventh Circuit used an erroneously low standard for determining whether the plaintiffs in this case satisfied their burden of pleading “with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind.” 15 U. S. C. § 78u-4(b)(2). I further agree that the case should be remanded to allow the lower courts to decide in the first instance whether the allegations survive under the correct standard. In two respects, however, I disagree with the opinion of the Court. First, the best interpretation of the statute is that only those facts that are alleged “with particularity” may properly be considered in determining whether the allegations of scienter are sufficient. Second, I agree with Justice Scalia that a “strong inference” of scienter, *334in the present context, means an inference that is more likely than not correct.

I

On the first point, the statutory language is quite clear. Section 78u-4(b)(2) states that “the complaint shall, with respect to each act or omission alleged to violate this chapter, state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind.” Thus, “a strong inference” of scienter must arise from those facts that are stated “with particularity.” It follows that facts not stated with the requisite particularity cannot be considered in determining whether the strong-inference test is met.

In dicta, however, the Court states that “omissions and ambiguities” merely “count against” inferring scienter, and that a court should consider all allegations of scienter, even nonparticularized ones, when considering whether a complaint meets the “strong inference” requirement. Ante, at 326. Not only does this interpretation contradict the clear statutory language on this point, but it undermines the particularity requirement’s purpose of preventing a plaintiff from using vague or general allegations in order to get by a motion to dismiss for failure to state a claim. Allowing a plaintiff to derive benefit from such allegations would permit him to circumvent this important provision.

Furthermore, the Court’s interpretation of the particularity requirement in no way distinguishes it from normal pleading review, under which a court naturally gives less weight to allegations containing “omissions and ambiguities” and more weight to allegations stating particularized facts. The particularity requirement is thus stripped of all meaning.

Questions certainly may arise as to whether certain allegations meet the statutory particularity requirement, but where that requirement is violated, the offending allegations cannot be taken into account.

*335II

I would also hold that a “strong inference that the defendant acted with the required state of mind” is an inference that is stronger than the inference that the defendant lacked the required state of mind. Congress has provided very little guidance regarding the meaning of “strong inference,” and the difference between the Court’s interpretation (the inference of scienter must be at least as strong as the inference of no scienter) and Justice Scalia’s (the inference of scienter must be at least marginally stronger than the inference of no scienter) is unlikely to make any practical difference. The two approaches are similar in that they both regard the critical question as posing a binary choice (either the facts give rise to a “strong inference” of scienter or they do not). But Justice Scalia’s interpretation would align the pleading test under §78u-4(b)(2) with the test that is used at the summary-judgment and judgment-as-a-matter-of-law stages, whereas the Court’s test would introduce a test previously unknown in civil litigation. It seems more likely that Congress meant to adopt a known quantity and thus to adopt Justice Scalia’s approach.

Justice Stevens,

dissenting.

As the Court explains, when Congress enacted a heightened pleading requirement for private actions to enforce the federal securities laws, it “left the key term ‘strong inference’ undefined.” Ante, at 314. It thus implicitly delegated significant lawmaking authority to the Judiciary in determining how that standard should operate in practice. Today the majority crafts a perfectly workable definition of the term, but I am persuaded that a different interpretation would be both easier to apply and more consistent with the statute.

The basic purpose of the heightened pleading requirement in the context of securities fraud litigation is to protect defendants from the costs of discovery and trial in unmeritori*336ous cases. Because of its intrusive nature, discovery may also invade the privacy interests of the defendants and their executives. Like citizens suspected of having engaged in criminal activity, those defendants should not be required to produce their private effects unless there is probable cause to believe them guilty of misconduct. Admittedly, the probable-cause standard is not capable of precise measurement, but it is a concept that is familiar to judges. As a matter of normal English usage, its meaning is roughly the same as “strong inference.” Moreover, it is most unlikely that Congress intended us to adopt a standard that makes it more difficult to commence a civil case than a criminal case.1

In addition to the benefit of its grounding in an already familiar legal concept, using a probable-cause standard would avoid the unnecessary conclusion that “in determining whether the pleaded facts give rise to a ‘strong’ inference of scienter, the court must take into account plausible opposing inferences.” Ante, at 323 (emphasis added). There are times when an inference can easily be deemed strong without any need to weigh competing inferences. For example, if a known drug dealer exits a building immediately after a *337confirmed drug transaction, carrying a suspicious looking package, a judge could draw a strong inference that the individual was involved in the aforementioned drug transaction without debating whether the suspect might have been leaving the building at that exact time for another unrelated reason.

If, using that same methodology, we assume (as we must, see ante, at 322, 326) the truth of the detailed factual allegations attributed to 27 different confidential informants described in the complaint, App. 91-93, and view those allegations collectively, I think it clear that they establish probable cause to believe that Tellabs’ chief executive officer “acted with the required intent,” as the Seventh Circuit held.2 437 P. 3d 588, 602 (2006).

Accordingly, I would affirm the judgment of the Court of Appeals.

EASTERBROOK, Circuit Judge.

Baxter International, a manufacturer of medical products, released its second-quarter financial results for 2002 on July 18 of that year. Sales and profits did not match analysts’ expectations. Shares swiftly fell from $43 to. $32. This litigation followed; plaintiffs contend that the $43 price was the result of materially misleading projections on November 5, 2001, projections that Baxter reiterated until the bad news came out on July 18, 2002. Plaintiffs want to represent a class of all investors who. purchased during that time either in the open market or by exchanging their shares of Fusion Medical Technologies. (Baxter acquired Fusion in a stock-for-stock transaction; plaintiffs think that Baxter juiced up the market price so that it could secure Fusion in exchange for fewer of its own shares.) Bypassing the question whether the suit could proceed as a class action, but see Fed.R.Civ.P. 23(c)(1)(A), the district court dismissed the complaint for failure to state a claim on which relief may be granted. 2003 WL 21825498, 2003 U.S. Dist. Lexis 12905 (N.D.I11. July 17, 2003). The court did not doubt that the allegations ordinarily would defeat a motion under Fed.R.Civ.P. 12(b)(6). Still, it held, Baxter’s forecasts come within the safe harbor created by the Private Securities Litigation Reform Act of 1995,15 U.S.C. §§ 77z-2(c), 78u-5(c). The PSLRA creates rules that judges must enforce at the outset of the litigation; plaintiffs do not question the statute’s application before discovery but do dispute the district court’s substantive decision.

Baxter’s projection, repeated many times (sometimes in documents filed with the SEC, sometimes in press releases, sometimes in executives’ oral statements), was that during 2002 the business would yield revenue growth in the “low teens” compared with the prior year, earnings-per-share growth in the “mid teens,” and “operational cash flow of at least $500 million.” Baxter often referred to these forecasts as “our 2002 full-year commitments,” which is a strange locution. No firm can make “commitments” about the future — Baxter can’t compel its customers to buy more of its products — unless it plans to engage in accounting shenanigans to make the numbers come out right no matter what happens to the business. But nothing turns on the word; the district court took these “commitments” as “forward-looking statements,” see, 15 U.S.C. §§ 77z-2(a), 78u-5(a), and plaintiffs,do not quarrel with that understanding. What they do say is that the projections were too rosy, and that Baxter knew it. That charges the defendants with stupidity as much as with knavery, for the truth was bound to come out quickly, but the securities laws forbid foolish frauds along with clever ones.

According to the. complaint, Baxter’s projections were materially false because: (1) its Renal Division had not met its internal budgets in years; (2) economic instability in Latin America adversely affected Baxter’s sales, in that part of the world; (3) Baxter closed plants in Ronneby, Sweden, *729and Miami Lakes, Florida, that had been its principal source of low-cost dialysis products; (4) the . market for albumin (blood-plasma) products was “over-saturated,” resulting in lower prices and revenue for the BioSciences Division; (5) sales of that division’s IGIV immunoglobin' products had fallen short of internal predictions; and (6) in March 2002 the BioSeience Division had experienced a sterility failure in the manufacture of a major product, resulting in the destruction of multiple lots and a loss exceeding $10 million. The district court assumed, as shall we, that failure to disclose these facts would create problems but for the statutory safe harbor — though items (2) and (4) at least are general business matters rather than Baxter’s secrets, and the securities laws do not require issuers to disclose the state of the world, as opposed to facts about the firm. See Wielgos v. Commonwealth Edison Co., 892 F.2d 509 (7th Cir.1989). Item (3) also was public knowledge (Baxter issued a press release announcing the closings and a substantial charge against earnings)— though the cost of products that had been made at these plants may have been secret. Whether all firm-specific non-disclosures add up to a material, non-disclosure — and whether Baxter had some nonpublic information about those matters that seem to be general information — are topics we need not tackle.

Section 77z-2, which deals with statements covered by the Securities Act of 1933 (here, those in the registration statement and prospectus for the. stock that Baxter exchanged for Fusion’s shares) and § 78u-5, which deals with statements covered by the Securities Exchange Act of 1934 (here, the statements in Baxter’s press releases, press conferences, and periodic filings) are identical in all significant respects, so from now on we mention only .the former statute. The statutory safe harbor forecloses liability if a forward-looking statement “is accompanied by meaningful cautionary statements identifying important factors that could cause 'actual results to differ materially from those in the forward-looking statement” (§ 77z-2(c)(l)(A)(i)). The fundamental problem is that the statutory requirement of “meaningful cautionary statements” is not itself meaningful. What must the firm say? Unless it is possible to give a concrete and reliable answer, the harbor is not “safe”; yet a word such as “meaningful” resists a concrete rendition and thus makes administration of the safe harbor difficult if not impossible. It rules out a caution such as: “This is a forward-looking statement: caveat emptor.” But it does not rule in any particular caution, which always may be challenged as not sufficiently “meaningful” or not pinning down the “important factors that could cause actual results to differ materially”— for if it had identified all of those factors, it would not be possible to describe the forward-looking statement itself as materially misleading. A safe harbor matters only when the firm’s disclosures (including the accompanying cautionary statements) are false or misleadingly incomplete; yet whenever that condition is satisfied, one can complain that the cautionary statement must have been inadequate. The safe harbor loses its function. Yet it would be unsound to read the statute so that the safe harbor never works; then one might as well treat § 77z-2 and § 78u-5 as defunct.

Baxter provided a number of cautionary statements throughout the class period. This one, from its 2001 Form 10-K filing— a document to which many of the firm’s press releases and other statements referred — is the best illustration:

Statements throughout this report that are not historical facts are forward-looking statements. These statements are *730based on the company’s current expectations and involve numerous risks and uncertainties. Some of these risks and uncertainties are factors that affect all international businesses, while some are specific to the company and the health care arenas in which it operates.

Many factors could affect the company’s actual results, causing results to differ materially, from those expressed in any such forward-looking statements. These factors include, but are not limited to, interest rates; technological advances in the medical field; economic conditions; demand and market acceptance risks for new and existing products, technologies and health care services; the impact of competitive products and pricing; manufacturing capacity; new plant start-ups; global regulatory, trade and tax policies; regulatory, legal or other developments relating to the company’s Series A, AF, and AX dialyzers; continued price competition; product development risks, including technological difficulties; ability to enforce patents; actions of regulatory bodies and other government authorities; reimbursement policies of government agencies; commercialization factors; results of product testing; and other factors described elsewhere in this report or in the company’s other filings with the Securities and Exchange Commission. Additionally, as discussed in Item 3 — “Legal Proceedings,” upon the resolution of certain legal matters, the company may incur charges in excess of presently established reserves. Any such change could have a material adverse effect on the company’s results of operations or cash flows in the period in which it is recorded.

Currency fluctuations are also a significant variable for global companies, especially fluctuations in local currencies where hedging opportunities are unreasonably expensive or unavailable. If the United States dollar strengthens significantly against mist foreign currencies, the company’s ability to realize projected growth rates in its sales and net earnings outside the United States could be negatively impacted.

The company believes that its expectations with respect to forward-looking statements are based upon reasonable assumptions within the bounds of its knowledge of its business operations, but there can be no assurance that the actual results or performance of the company will conform to any future results or performance expressed or implied by such forward-looking statements.

The district court concluded that these are “meaningful cautionary statements identifying important factors that could cause actual results to differ materially from those in the forward-looking statement”. They deal with Baxter’s business specifically, mentioning risks and product lines. Plaintiffs offer two responses. First they contend that the cautionary statements did not cover any of the six matters that (in plaintiffs’ view) Baxter had withheld. That can’t be dispositive; otherwise the statute would demand prescience. As long as the firm reveals the principal risks, the fact that some other event caused problems cannot be dispositive. Indeed, an unexpected turn of events cannot demonstrate a securities problem at all, as there cannot be “fraud by hindsight.” Denny v. Barber, 576 F.2d 465, 470 (2d Cir.1978) (Friendly, J.). See also Murray v. Abt Associates Inc., 18 F.3d 1376 (7th Cir.1994); DiLeo v. Ernst & Young, 901 F.2d 624 (7th Cir.1990). The other response is that the cautionary statement did not follow the firm’s fortunes: plants closed but the cautionary statement remained the *731same; sterilization failures occurred but the cautionary statement remained the same; and bad news that (plaintiffs contend) Baxter well knew in November 2001 did not east even a shadow in the cautionary statement.

Before considering whether plaintiffs’ objections defeat the safe harbor, we ask whether the cautionary statements have any bearing on Baxter’s potential liability for statements in its press releases, and those its managers made orally. The press releases referred to, but did not repeat verbatim, the cautionary statements in the Form 10-K and other documents filed with the Securities and Exchange Commission. The oral statements did not do even that much. Plaintiffs say that this is fatal, because § 77z — 2(c)(1)(A)(i) provides a safe harbor only if a written statement is “accompanied by” the meaningful caution; a statement published elsewhere differs from one that accompanies the press release. As for the oral statements: § 77z — 2(c)(2)(A)(ii), a special rule for oral statements, provides a safe harbor only if the statement includes “that the actual results could differ materially from those projected in the forward-looking statement” and in addition:

(i) the oral forward-looking statement is accompanied by an oral statement that additional information concerning factors that could cause actual results to differ materially from those in the forward-looking statement is contained in a readily available written document, or portion thereof;

(ii) the accompanying oral statement referred to in clause (i) identifies the document, or portion thereof, that contains the additional information about those factors relating to the forward-looking statement; and

(iii)the information contained in that written document is a cautionary statement that satisfies the standard established in paragraph ' (1)(A).

15 U.S.C. § 77z-2(c)(2)(B). When speaking with analysts Baxter’s executives did not provide them with all of this information, such as directions to look in the 10-K report for the full cautionary statement. It follows, plaintiffs maintain, that this suit must proceed with respect to the press releases and oral statements even if the cautionary language filed with the SEC in registration' statements and other documents meets the statutory standard.

If this were a traditional securities suit — if, in other words, an investor claimed to have read or heard the statement and, not having access to the truth, relied to his detriment on the falsehood— then plaintiffs’ argument would be correct. But this is not a traditional securities claim. It is a fraud-on-the-market claim. None of the plaintiffs asserts that.he read any of Baxter’s press releases or listened to an executive’s oral statement. Instead the theory is that other people (professional traders, mutual fund managers, securities analysts) did the reading, and that they made trades or recommendations that influenced the price. In an efficient capital market, all information known to the public affects the price and thus affects every investor. Basic Inc. v. Levinson, 485 U.S. 224, 241-47, 108 S.Ct.978, 99 L.Ed,2d 194 (1988), holds that reliance on the accuracy of the price can substitute for reliance on the accuracy of particular written or oral statements, when the statements affect the price — as they do for large and well-followed firms such as Baxter, for which there is a liquid public market. This works only to the extent that markets efficiently reflect (and thus convey to investors the economic equivalent *732of) all public information. See Daniel R. Fischel, Efficient Capital Markets, the Crash, and the Fraud on the Market Theory, 74 Cornell L.Rev. 907, 917-22 (1989); Jonathan R. Macey & Geoffrey P. Miller, Good Finance, Bad Economics: An Analysis of the Fraud-on-the-Market Theory, 42 Stan. L.Rev. 1059 (1990).

When markets are informationally efficient, it is impossible to segment information as plaintiffs propose. They ask us to say that they received (through the price) the false oral statements but not the cautionary disclosures. That can’t be; only if the market is inefficient is partial transmission likely, and if the market for Baxter’s stock is inefficient then this suit collapses because a fraud-on-the-market claim won’t fly. An investor who invokes the fraud-on-the-market theory must acknowledge that all public information is reflected in the price, just as the Supreme Court said in Basic. See 485 U.S. at 246, 108 S.Ct. 978. See Flamm v. Eberstadt, 814 F.2d 1169, 1179-80 (7th Cir.1987); In re Apple Computer Securities Litigation, 886 F.2d 1109, 1115-16 (9th Cir.1989); Grossman v. Novell, Inc., 120 F.3d 1112, 1122-28 (10th Cir.1997). Thus if the truth or the nature of a business risk is widely known, an incorrect statement can have no deleterious effect, and if a cautionary statement has been widely disseminated, that news too affects the price just as if that statement had been handed to each investor. If the executives’ oral statements came to plaintiffs through professional traders (or analysts) and hence the price, then the cautions reached plaintiffs via the same route; market professionals are savvy enough to discount projections appropriately. Then § 77z-2(c)(2)(B) has been satisfied for the oral statements (and so too § 77z — 2(c)(1)(A)(i) for the press releases). And if the cautions did not affect the price, then the market must be inefficient and the suit fails for that reason. So we take the claim as the pleadings framed it: the market for Baxter’s stock is efficient, which means that Baxter’s cautionary language must be treated as if attached to every one of its oral and written statements. That leaves the question whether these statements satisfy the statutory requirement that they adequately “identify! ] important factors that could cause actual results to differ materially from those in the forward-looking statement”.

The parties agree on two propositions, each with support in decisions of other circuits. First, “boilerplate” warnings won’t do; cautions must be tailored to the risks that accompany the particular projections. Second, the cautions need not identify what actually goes wrong and causes the projections to be inaccurate; prevision is not required. See Halperin v. eBanker USA.COM, Inc., 295 F.3d 352, 359 (2d Cir.2002); Helwig v. Vencor, Inc., 251 F.3d 540, 558-59 (6th Cir.2001) (en banc); Ehlert v. Singer, 245 F.3d 1313, 1320 (11th Cir.2001) (discussing a judicially developed defense, the bespeaks-caution doctrine, that applies to statements such as those made in tender offers to which the statutory safe harbor does not apply); Semerenko v. Cendant Corp., 223 F.3d 165, 182-83 (3d Cir.2000) (same); Harris v. Ivax Corp., 182 F.3d 799, 807 (11th Cir.1999) (same). Unfortunately, these principles don’t decide any concrete case — for that matter, the statutory language itself does not decide any concrete case. It is the result of a compromise between legislators who did not want any safe harbor (or, indeed any new legislation), and those who wanted a safe harbor along the lines of the old Rule 175 (discussed in our Wielgos decision) that did not require any cautionary statements but just required the projection to have a reasonable basis. Rule 175 was limited to statements in certain documents *733filed with the SEC; proponents of the PSLRA wanted to extend this to all statements, including oral declarations and press releases. As is often the situation, a compromise enabled the bill to pass but lacks much content; it does not encode a principle on which political forces agreed as much as it signifies conflict about both the scope and the wisdom of the safe harbor. Compromises of this kind lack spirit. Still, the language was enacted, and we must make something of it.

Plaintiffs say that Baxter’s cautions were boilerplate, but they aren’t. Statements along the lines of “all businesses are risky” or “the future lies ahead” come to nothing other than caveat emptor (which isn’t enough); these statements, by contrast, at least included Baxter-specific information and highlighted some parts of the business that might cause problems. For its part, Baxter says that mentioning these business segments demonstrates that the caution is sufficient; but this also is wrong, because then any issuer could list its lines of business, say “we could have problems in any of these,” and avoid liability for statements implying that no such problems were on the horizon even if a precipice was in sight.

What investors would like to have is a full disclosure of the assumptions and calculations behind the projections; then they could apply their own discount factors. For reasons covered at length in Wielgos, however, this is not a sensible requirement. Many of the assumptions and calculations would be more useful to a firm’s rivals than to its investors. Suppose, for example, that Baxter had revealed its sterility failure in the BioSciences Division, the steps it had taken to restore production, and the costs and prospects of each. Rivals could have used that information to avoid costs and hazards that had befallen Baxter, or to find solutions more quickly, and as Baxter could not have charged the rivals for this information they would have been able to undercut Baxter’s price in future transactions. Baxter’s shareholders would have been worse off. Similarly Baxter might have added verisimilitude to its projections by describing its sales policies and the lowest prices it would accept from major customers, but disclosing reservation prices would do more to help the customers than to assist the investors.

Another form a helpful caution might take would be the disclosure of confidence intervals. After saying that it expected growth in the low teens, Baxter might have added that events could deviate 5% in either direction (so the real projection was that growth would fall someplace between 8% and 18%); disclosure of the probability that growth will be under 10% (or over 16%) would have done much to avoid the hit stock prices took when the results for the first half of 2002 proved to be unexpectedly low. Baxter surely had developed internally some estimate of likely variance. Revealing the mean, median, and standard deviation of these internal estimates, and pinpointing the principal matters that could cause results to differ from the more likely outcome, could help to generate an accurate price for the stock. Knowledge that the mean is above the median, or that the standard deviation is substantial, would be particularly helpful to those professional investors whose trades determine the market price. (It might imply, for example, that as in Wiel-gos the firm was projecting what would happen if nothing unexpected happened; because some things always go wrong, investors could apply discounts.) Perhaps, *734however, a firm’s data do not permit estimates to be stated in probabilities. If, for example, a major source of uncertainty for Baxter’s business was how Congress would resolve the debate about Medicare coverage for prescription drugs, or whether a rival would manage to win the FDA’s approval for a product that would compete with one of Baxter’s most profitable items, it would be hard to reduce these chances to probabilities. Events such as these are discrete rather than continuous variables, so standard confidence intervals would be meaningless even if probabilities could be attached to the likely outcomes.

Whether or not Baxter could have made the cautions more helpful by disclosing assumptions, methods, or confidence intervals, none of these is required. The PSLRA does not require the most helpful caution; it is enough to “identify! ] important factors that could cause actual results to differ materially from those in the forward-looking statement”. This means that it is enough to point to the principal contingencies that could cause actual results to depart from the projection. The statute calls for issuers to reveal the “important factors” but not to attach probabilities to each potential bad outcome, or to reveal in detail what could go wrong; as we have said, that level of detail might hurt investors (by helping rivals) even as it improved the accuracy of stock prices. (Requiring cautions to contain elaborate detail also would defeat the goal of facilitating projections, by turning each into a form of registration statement. Undue complexity would lead issuers to shut up, and stock prices could become even less accurate. Incomplete information usually is better than none, because market professionals know other tidbits that put the news in context.) Moreover, “[i]f enterprises cannot make predictions about themselves, then securities analysts, newspaper columnists, and charlatans have protected turf. There will be predictions aplenty outside the domain of the securities acts, predictions by persons whose access to information is not as good as the issuer’s. When the issuer adds its information and analysis to that assembled by outsiders, the collective assessment will be more accurate even though a given projection will be off the mark.” Wielgos, 892 F.2d at 514 (emphasis in original).

Yet Baxter’s chosen language may fall short. There is no reason to think' — -at least, no reason that a court can accept at the pleading stage, before plaintiffs have access to discovery — that the items mentioned in Baxter’s cautionary language were those that at the time were the (or any of the) “important” sources of variance. The problem is not that what actually happened went unmentioned; issuers need not anticipate all sources of deviations from expectations. Rather, the problem is that there is no reason (on this record) to conclude that Baxter mentioned those sources of variance that (at the time of the projection) were the principal or important risks. For all we can tell, the major risks Baxter objectively faced when it made its forecasts were exactly those that, according to the complaint, came to pass, yet the cautionary statement mentioned none of them. Moreover, the cautionary language remained fixed even as the risks changed. When the sterility failure occurred in spring 2002, Baxter left both its forecasts and cautions as is. When Baxter closed the plants that (according to the complaint) were its least-cost sources of production, the forecasts and cautions continued without amendment. This raises the possibility — no greater confidence is possible before discovery — that Baxter omitted important variables from the cautionary language *735and so made projections more certain than its internal estimates at the time warranted. Thus this complaint could not be dismissed under the safe harbor, though we cannot exclude the possibility that if after discovery Baxter establishes that the cautions did reveal what were, ex ante, the major risks, the safe harbor may yet carry the day.

Baxter urges us to affirm the judgment immediately, contending that the full truth had reached the market despite any shortcomings in its cautionary statements. If this is so, however, it is hard to understand the sharp drop in the price of its stock. A “truth-on-the-market” defense is available in principle, as we discussed in Flamm, but not at the pleading stage. Likewise one must consider the possibility that investors looked at all of the projections as fluff and responded only to the hard numbers; on this view it was a reduction in Baxter’s growth rate, not the embarrassment of a projection, that caused the price decline in July 2002; again it is too early in the litigation to reach such a conclusion. It would be necessary to ask, for example, whether the price rose relative to the rest of the market when Baxter made its projections; if not, that might support. an inference that the projections were so much noise.

Nor has the time arrived to evaluate Baxter’s contention that its projections panned out, so there was no material error. Baxter insists that all of the projections dealt with the entire calendar year 2002, and that by year-end performance was up to snuff — close enough to the projections that any difference was immaterial. Once again, it is inappropriate to entertain such an argument at the pleading stage. The district court will need to determine whether all of the forward-looking statements referenced calendar 2002 as a whole, rather than anticipated improvements quarter-by-quarter over the preceding year. It will be necessary to evaluate whether differences between the projections and the outcome were material under the standard of Basic. Finally it may be necessary to explore what Baxter’s full-year results actually were; plaintiffs’ reply brief accuses Baxter of using gimmicks to report extra revenue in 2002 at the expense of later years. The implication is that Baxter may have overstated its 2002 results. Whether 'that is so cannot be determined on the pleadings, even when supplemented with the documents that Baxter has filed with the SEC.

ReveRsed And Remanded.

LOHIER, Circuit Judge:

Plaintiffs-Appellants appeal from a September 2010 judgment of the United States District Court for the Southern District of New York (Daniels, J.) granting the summary judgment motion of Defendant-Appellee Grant Thornton LLP (“GT”) and dismissing the Plaintiffs’ claims *151arising from GT’s audit of the financial statements of its client, Winstar Communications, Inc. (“Winstar”). The Plaintiffs claimed that GT committed securities fraud in violation of Section 10(b) of the Securities Exchange Act of 1934, 15 U.S.C. § 78j(b) (the “Act” or the “Exchange Act”), and 17 C.F.R. § 240.10b-5, and made false and misleading statements in an audit opinion letter in violation of Section 18 of the Act, 15 U.S.C. § 78r. We conclude that genuine issues of material fact exist as to each of these claims. We therefore VACATE the District Court’s grant of summary judgment and REMAND for further proceedings.

BACKGROUND

1. Facts

Reviewing the District Court’s grant of summary judgment in favor of GT, “we construe the evidence in the light most favorable to the [Plaintiffs], drawing all reasonable inferences and resolving all ambiguities in [their] favor.”1 In re Omnicom Grp., Inc. Sec. Litig., 597 F.3d 501, 504 (2d Cir.2010) (quotation marks omitted).

Winstar was a broadband communications company whose core business was to provide wireless Internet connectivity to various businesses. GT served as Wins-tar’s independent auditor from 1994 until Winstar filed for bankruptcy in April 2001, and GT regarded Winstar as “one of [its] largest and most important clients.”2

In 1999, however, the relationship deteriorated. Winstar warned GT that it would likely terminate the relationship if GT’s performance on unrelated international tax planning and other accounting matters proved unsatisfactory. In March 1999 at least one member of Winstar’s board of directors openly urged during a board meeting that the GT partner overseeing the audit of Winstar be removed from the Winstar account. GT eventually re-staffed the Winstar account so that the 1999 audit was managed by a partner, Gary Goldman, and a senior manager, Patricia Cummings, neither of whom had previously reviewed or audited the financial records of a telecommunications company.

As relevant to this appeal, GT’s audit for 1999 included several “large account” transactions that Winstar consummated in an attempt to conceal a decrease in revenue associated with Winstar’s core business. Most of the large account transactions involved Lucent Technologies, Inc. (“Lucent”), Winstar’s strategic partner, and all of them were consummated at the end of Winstar’s fiscal quarters in 1999. Together, the transactions accounted for $114.5 million in revenue, or approximately 26 percent of Winstar’s reported 1999 operating revenues and 32 percent of its “core” revenues that year. At the time, GT considered these transactions to be “red flags,” warranting the accounting firm’s “heightened scrutiny.**3 However, *152GT ultimately approved Winstar’s recognition of revenue in connection with each of these transactions.

We discuss the evidence relating to each category of transaction in turn.

A. Questionable Sales

The first category of questionable transactions involved a series of six end-of-quarter and end-of-year transactions, primarily reported as equipment sales, for which there was little evidence that any goods or services were ordered and delivered. For example, for the third quarter of 1999 Winstar recognized $15 million in revenue for the sale of Lucent equipment to Anixter Brothers, Inc. (“Anixter”), a wire and cable distributor. There were several unusual aspects of this sale. First, Anixter ordinarily purchased equipment directly from Lucent, not Winstar. Second, equipment sales were not part of Winstar’s core business of creating and operating wireless networks. Third, during GT’s audit Cummings noted that the Anixter transaction was “apparently” completed on September 30, 1999, the last day of Winstar’s fiscal quarter, but GT’s work papers included no documents reflecting the sale’s completion beyond a purchase order from Winstar to Lucent and an invoice from Lucent to Winstar. Moreover, neither the' purchase order nor the invoice included an itemized list of the goods sold or indicated the shipping terms, even though the items were to be shipped on September 30, 1999, and delivered on October 4, 1999.4 Absent too was any document evidencing Anixter’s agreement to purchase the items. Lastly, not a single employee of Lucent, Winstar, or Anixter who was asked about the equipment sale could recall it.

Five other transactions that were not part of Winstar’s core business were consummated at the end of one of Winstar’s fiscal quarters and were barely documented. Winstar nevertheless recognized a total of $49.7 million in revenue associated with these five transactions. First, Wins-tar recognized $5 million in revenue in the first quarter of 1999 for a “feasibility study” that Winstar was scheduled to conduct for Lucent, but which had not been delivered by at least 2000. Second, Wins-tar recognized $21.1 million in revenue in the first and second quarters of 1999 in connection with the sale of Lucent equipment to Williams Communications, Inc. (“Williams”). The equipment was shipped by Lucent, not Winstar, on the last business day of the first and second quarters (March 31, 1999 and June 30, 1999, respectively), with no written agreement. Third, Winstar recognized $9.1 million in revenue in the second quarter of 1999 in connection with the sale of Lucent equipment to Vo*153Call Communications Corporation (“VoCall”) on June 30, 1999. Although the sale was referenced in a series of non-numbered purchase orders, it was not referenced in any executed, final agreement or shipping document. Fourth, Winstar recognized $4.5 million in revenue in the third quarter of 1999 in connection with the sale of unspecified “WinStar Equipment” to Cignal Global Communications (“Cignal”), which was contracted for on September 30, 1999, the last day of that quarter. However, GT was unable to produce a document evidencing that the equipment had been shipped to Cignal during that quarter. Fifth, Winstar recognized $10 million in revenue in the fourth quarter of 1999 in connection with the sale of wireless radio equipment (“radios”) to Lucent under an agreement dated December 30, 1999. GT endorsed the recognition of revenue even though its work papers included shipping documents with conflicting dates, no document specified the goods purchased, and Lucent, not Winstar, was in the business of manufacturing and selling radios. The same agreement also involved a $2 million “promotional credit” purchased by Lucent for services that had not yet been rendered by Winstar. Although GT specifically advised Winstar that recognizing and recording the amount of the credit as revenue was improper and in violation of generally accepted accounting principles (“GAAP”),5 Winstar nevertheless recognized the full $2 million in revenue.

Each of these transactions appears to have violated the provisions of Staff Accounting Bulletin No. 101 (“SAB 101”), issued by the Securities and Exchange Commission (“SEC”), which states that four conditions must be satisfied before revenue can be recognized: (1) “Persuasive evidence of an arrangement [for the sale of goods or services] exists,” (2) “Delivery has occurred or services have been rendered,” (3) “The seller’s price to the buyer is fixed or determinable,” and (4) “Collectibility is reasonably assured.” SAB 101 at 3.6

*154GT requested that Winstar’s counterparties provide additional documentary evidence of the relevant sales underlying each questionable transaction. By doing so, consistent with SAB 101, GT sought to obtain independent support for Winstar’s recognition of revenue for each transaction — in other words, support from documents that were not generated by Winstar itself. As of February 10, 2000, GT still had not received responsive documents from four of these customers. Nonetheless, it issued an audit opinion letter opining that Winstar’s 1999 financial statements accurately reflected its financial condition and complied with GAAP.

B. Bifurcated Accounting

In connection with at least three other transactions, Winstar employed a bifurcated accounting scheme that GT ultimately approved prior to its audit of Winstar’s financial statements. Two of these transactions involved leasing or subleasing fiber optic network capacity in units called indefeasible rights of use (“IRUs”). Winstar accounted for these IRUs using a dubious bifurcated accounting method, pursuant to which it recognized as much as 94 percent of the revenue from the leases upon execution of the lease, reflecting the cost of optical equipment (“optronics”) that transmitted data over fiber optic cable (“cable” or “fiber”). Winstar then would recognize the balance of the revenue in later quarters, as payments were received over the span of the lease, representing the customer’s actual use of the network. In other words, Winstar split the value of the leases so that the revenue associated with the optronics was reported separately from revenue associated with the cable. By employing this accounting method, Winstar was able to recognize $30.9 million in revenue up front in 1999.

During discovery, a forensic accountant retained by the Plaintiffs opined that the rules of the Financial Accounting Standards Board and interpretive rules of the SEC prohibited the division of leases for fiber optic networks because both the cable and the optronics were essential to the network. Winstar conceded that the fiber and the optronics were not separable, and that no other company previously had employed this bifurcated method in accounting for IRUs. Indeed, Winstar specifically advised GT that the bifurcated approach had been criticized by the accounting firm Deloitte & Touche LLP (“Deloitte”).

GT was in any event aware that revenue associated with an IRU contract could be recognized only if the leased circuit was operational, or “lit,” in the language of the fiber optics field. In the third and fourth quarters of 1999, however, Winstar had failed to “light” many of its IRU circuits, a fact that should have precluded the company from recognizing revenue associated with those IRUs.

In the midst of GT’s audit, Winstar sent form letters dated December 30, 1999, and December 31, 1999, to the counterparties to the two IRU transactions, Wam!Net, Inc. and Cignal, respectively, to confirm that the IRU circuits had been “deliver[ed]” and “accept[ed].” A representative of Cignal signed a form letter confirming delivery and acceptance of the circuits. By contrast, Wam!Net’s Senior Vice President of Finance responded to a letter from Winstar as follows: “To our knowledge [the circuits] are not currently in place.” A subsequent amended letter from a different Wam!Net employee, which does not appear in the record but which Cummings referenced in an email, purported to “accept” the circuits but did not address the earlier letter response. After receiving the amended letter, GT did not further review whether the circuits were installed *155and operational. Even though it appears not to have received the amended letter until after February 11, 2000, GT approved Winstar’s recognition of revenue for the WamlNet IRU circuit sale on February 10, 2000.

While GT appears to have neglected to verify that WamlNet’s IRU circuits were operational, there was evidence that GT actually knew that the leased Cignal IRU circuits were inoperative. GT nevertheless approved Winstar’s recognition of revenue for the Cignal IRUs in the third quarter of 1999.

Winstar employed a similar bifurcated accounting method in the fourth quarter of 1999 for its sale of radios to Lucent. The agreement between the two companies provided for Winstar to install the radios, but Winstar recognized revenue for the transaction immediately, upon delivering them to Lucent.7 GT expressed doubt that the radios and installation services were separable,8 but it nevertheless approved of Winstar’s recognition of $10 million in revenue in connection with the transaction.

C. Round-Trip Transactions

Several of the transactions discussed above involved “round-trip” transactions with Cignal and WamlNet at a time when the two customers were struggling financially. “ ‘Round-tripping’ typically refers to reciprocal agreements, involving the same products or services, that lack economic substance but permit [both] parties to book revenue and improve their financial statements.” Teachers’ Ret. Sys. of LA v. Hunter, 477 F.3d 162, 169 (4th Cir.2007). The two principal round-trip transactions that were the focus of discovery involved a scheme pursuant to which Winstar overpaid both companies for purportedly unrelated goods and services in exchange for their purchase of IRUs, equipment, and services from Winstar by the end of Winstar’s third fiscal quarter of 1999 (Cignal), and the end of the fourth quarter of 1999 (WamlNet).

Both round-trip transactions were material to Winstar’s financial performance in 1999. In the larger of the two transactions, Winstar recognized approximately $39 million in revenue in the third and fourth quarters of 1999 in connection with sales to Cignal while it paid Cignal $29.5 million under a separate agreement during the same time period and an additional $4.8 million in the first quarter of 2000. Similarly, Winstar recognized $19.6 million in revenue in connection with sales to WamlNet in the fourth quarter of 1999, even as it paid WamlNet $25 million in the same quarter for “prepaid marketing” and the lease of a “data service center.” In February 2000 GT questioned the “[a]rms length nature of the [round-trip] transactions.” It later acknowledged that the transactions “raise[d] a concern” within GT “as to whether or not absent Winstar’s payment ... collectability was reasonably assured.” Nevertheless, GT approved Winstar’s recognition of the full amount of revenue for both transactions.

D. The Audit Opinion Letter

By letter dated February 10, 2000, GT issued an unqualified audit opinion letter *156stating that Winstar’s annual Form 10-K report for fiscal- year 1999 complied with GAAP and fairly represented Winstar’s financial condition at the end of that year:

We have audited the accompanying consolidated balance sheets of Winstar Communications, Inc. ...

We conducted our audits in accordance with auditing standards generally accepted in the United States. Those standards require that we plan and perform the audit to obtain reasonable assurance about whether the financial statements are free of material misstatement. ... We believe that our audits provide a reasonable basis for our opinion.

In our opinion, the consolidated financial statements referred to above present fairly, in all'material respects, the consolidated financial position of Winstar Communications, Inc. and Subsidiaries as of December 31, 1999 ... in conformity with accounting principles generally accepted in the United States.

Joint App. at 529.

E. Investment by the Jefferson Plaintiffs

From December 1998 to February 2001, the Jéfferson Plaintiffs purchased over $200 million worth of Winstar stock. The investment portfolios of most of the Jefferson Plaintiffs were managed by Ronald Clark, the Chief Investment Officer for Allianz of America (“Allianz”). The remaining entities deferred to Clark to select their investments in United States securities. Although Clark enjoyed ultimate authority for these investment decisions, he relied on recommendations from a team of analysts, including Livia Asher, who recommended that Allianz purchase Winstar stock. Based on Asher’s recommendation, Clark caused Allianz and the other Jefferson Plaintiffs to invest in Winstar.

Clark, it appears, did not personally review Winstar’s financial statements prior to making the decision to invest in Wins-tar. Instead, he relied on Asher to review the statements as part of her job. During discovery, however, Asher acknowledged that she was uncertain of the date of, or reason for, her recommendation that Allianz purchase Winstar stock. Nor could she specifically recall reading Winstar’s 1999 Form 10-K report. However, Asher testified that she “probably flipped through every single page” of the report, based on her practice. She explained, “I can’t imagine any reason why I would not have looked at this, ... given our position in the stock and given what I would normally do.” Asher added that she habitually read auditors’ opinion letters included in Forms 10-K to make sure that auditors believed that an issuer’s reports were “kosher,” but she did not specifically recall reviewing GT’s audit report.

F. Winstar’s Decline

Winstar’s stock reached a price per share of over $60 in March 2000. In May 2000 Lucent provided Winstar with financing in the form of a renewed credit facility in the aggregate amount of $2 billion. Almost a year later, however, in March 2001, Asensio & Company (“Asensio”), an investment firm, issued a press release criticizing Winstar’s ability to pay its debts and explaining that certain measurements of Winstar’s financial performance were “questionable” due in part to Winstar’s “revenue recognition from non-core businesses and sales of equipment and services to related parties.” Joint App. at 2412. The same press release warned that “[a]ny adjustment to Winstar’s aggressive revenue accounting and capitalization of cash expenditures would negatively and materially impact Winstar’s reported [earnings] and analyst’s [sic] opinions of its opera*157tions.” Id. at 2413. On March 19, 2001, Asensio issued another press release reporting on Winstar’s “debt collapse,” again criticizing its revenue accounting practices and noting, “Winstar has recognized revenues that created a slew of uncollected items.... Its revenues include sales to related parties and non-core items such as equipment sales and installation services.” Id. at 2414.

The Asensio press releases preceded a significant downgrade in Winstar’s credit rating on April 3, 2001. Moreover, on April 16, 2001, Winstar announced that Lucent was cancelling Winstar’s credit facility, that Winstar would delay filing its Form 10-K report for 2000, and that Wins-tar was considering a reorganization under Chapter 11 of the Bankruptcy Code.

By the time of the Asensio press releases, Winstar’s revenues and its stock price had decreased significantly during a marketwide decline in the prices of technology stocks. However, the press releases, coupled with the subsequent announcements of Winstar’s financial troubles, were followed almost immediately by an additional steep decline in Winstar’s stock price, from over $10 per share before March 2001 to $0.14 per share by mid-April 2001. On April 18, 2001, Winstar filed for bankruptcy.

During discovery, an economist retained by the Plaintiffs as an expert witness attributed the decline in Winstar’s stock price to, among other causes, the partial disclosure of Winstar’s alleged fraud by Asensio. The economist calculated that members of the putative class who had purchased Winstar common stock lost as much as $974 million, that class members who purchased Winstar notes lost up to $197 million, and that the Jefferson Plaintiffs lost over $96 million by investing in Winstar stock.

2. Procedural History

On April 10, 2001, the Lead Plaintiffs filed a putative class action complaint in the Southern District of New York against Winstar, GT, Lucent, and certain Winstar executives alleging securities fraud under Section 10(b) of the Exchange Act, 15 U.S.C. § 78j(b). In December 2001 the Jefferson Plaintiffs filed a separate complaint in the District Court against GT, claiming violations of both Section 10(b) and Section 18 of the Exchange Act, 15 U.S.C. § 78r(a).9 As a result of various settlements among the parties, the claims against GT were all that remained by 2008. After discovery, GT moved for summary judgment to dismiss the Plaintiffs’ claims and also moved unsuccessfully to preclude the Plaintiffs’ expert loss causation analysis.

In September 2010 the District Court granted GT’s motion for summary judgment, concluding that (1) the Plaintiffs had failed to demonstrate that a genuine dispute of material fact existed as to whether GT acted intentionally or recklessly, as required under Section 10(b) of the Exchange Act, and (2) the Jefferson Plaintiffs had failed to demonstrate that such a dispute existed as to whether they actually relied on GT’s audit opinion letter, as required under Section 18 of the Exchange Act. In re Winstar Commc’ns Sec. Litig., No. 01 CV 11522(GBD), 2010 WL 3910322, at *5-6 (S.D.N.Y. Sept. 29, 2010).

This appeal followed.

DISCUSSION

We review a district court’s grant of summary judgment de novo, viewing the *158evidence in the light most favorable to the non-moving party and drawing all reasonable inferences and resolving all ambiguities in its favor. E.g., Grain Traders, Inc. v. Citibank, N.A., 160 F.3d 97, 100 (2d Cir.1998). We will affirm the judgment “only if there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed.R.Civ.P. 56(a). “There is no genuine issue of material fact where the record taken as a whole could not lead a rational trier of fact to find for the non-moving party.” Durakovic v. Bldg. Serv. 32 BJ Pension Fund, 609 F.3d 133, 137 (2d Cir.2010) (quotation marks and brackets omitted). Put another way, “[a]n issue of fact is genuine if the evidence is such that a reasonable jury could return a verdict for the nonmoving party.” Omnicom, 597 F.3d at 509 (quotation marks omitted).

To sustain a claim under Section 10(b) of the Exchange Act and Rule 10b-5, the Plaintiffs must show “(i) a material misrepresentation or omission; (ii) scienter; (iii) ‘a connection with the purchase or sale of a security!;]’ (iv) reliance by the plaintiffs); (v) economic loss; and (vi) loss causation.” Id. (alteration in original) (quoting Dura Pharms., Inc. v. Broudo, 544 U.S. 336, 341-42, 125 S.Ct. 1627, 161 L.Ed.2d 577 (2005)).

1. Scienter

Plaintiffs may satisfy the scienter requirement by producing “evidence of conscious misbehavior or recklessness.” See ECA Local 134 IBEW Joint Pension Trust of Chicago v. JP Morgan Chase Co., 553 F.3d 187, 198 (2d Cir.2009). Scienter based on conscious misbehavior, in turn, requires a showing of “deliberate illegal behavior,” Novak v. Kasaks, 216 F.3d 300, 308 (2d Cir.2000), a standard met “when it is clear that a scheme, viewed broadly, is necessarily going to injure,” AUSA Life Ins. Co. v. Ernst & Young, 206 F.3d 202, 221 (2d Cir.2000) (alterations omitted).

In AUSA Life, we applied this standard to representations by the accounting firm Ernst & Young (“E & Y”), which, as plaintiffs allege with respect to GT, “consistently noticed, protested, and then acquiesced in” the financial misrepresentations of an audit client under pressure from the client’s management. Id. at 205. We held that by issuing an unqualified audit report despite its knowledge of accounting improprieties by the client, E & Y “intentionally engaged in manipulative conduct,” id. at 221 (quotation marks omitted), in violation of Section 10(b). We explained that

E & Y is not an accounting dilettante. It knows well that its opinions and certifications are afforded great weight, and it must have known that its financial certifications with regard to [its client] would be compelling to the investors .... [I]t is sufficient [for the purposes of showing scienter] for a plaintiff to allege and prove that a defendant could have foreseen the consequences of his actions but forged ahead nonetheless ....

Id.; see also SEC v. KPMG LLP, 412 F.Supp.2d 349, 379 (S.D.N.Y.2006) (triable issue as to conscious fraud existed when accountant was “[a]ware of [the client’s] earnings pressures” but allowed aggressive accounting policies “without any serious study to determine that these unusual, indeed unique, accounting treatments would result in financial statements that fairly represented [the client’s] financial condition”).

Scienter based on recklessness may be demonstrated where a defendant has engaged in conduct that was “highly unreasonable, representing an extreme departure from the standards of ordinary care ... to the extent that the danger was *159either known to the defendant or so obvious that the defendant must have been aware of it.” Rothman v. Gregor, 220 F.3d 81, 90 (2d Cir.2000) (quotation marks omitted). Recklessness may be established where a defendant “failed to review or check information that [it] had a duty to monitor, or ignored obvious signs of fraud.” Novak, 216 F.3d at 308.

The District Court concluded that Winstar engaged in “dubious accounting practices,” that “[m]uch of the supporting documentation that Winstar supplied to GT was a mere contrivance meant to paper the transactions and create the appearance of legitimacy,” and that GT “failed to uncover the accounting fraud being perpetuated by the Winstar defendants.” Wins-tar, 2010 WL 3910322, at *3, *5. GT does not seriously dispute these conclusions, which are-in any event supported by the record before us. The District Court concluded, however, that the evidence demonstrated only that GT was “performing its role as [Winstar’s] independent auditor” and fell short of demonstrating scienter in the form of either conscious misbehavior or recklessness. Id. at *3. We disagree.

Some evidence supports the Plaintiffs’ contention that GT consciously ignored Winstar’s fraud when it approved Wins-tar’s recognition of revenue for the suspicious 1999 transactions. This evidence goes beyond a mere failure to uncover the accounting fraud and, in general, relates to (1) Winstar’s recognition of revenue for the sale of equipment or services without sufficient indicia of delivery, (2) its recognition of all revenue associated with the incompíete sale of telecommunications systems, and (3) its recognition of revenue for sales of IRUs, equipment, and services to financially unstable companies to whom Winstar paid back large sums under separate contractual obligations.

There is also evidence that GT failed to confirm Winstar’s representations regarding these transactions or to retain and review documents evidencing each transaction. Indeed, an expert forensic accountant retained by the Plaintiffs testified that GT’s failure to exercise due professional care, gather reliable documents, and issue an adverse opinion in this regard represented a violation of Generally Accepted Auditing Standards (“GAAS”)10 with regard to all of the challenged transactions. Furthermore, the record evidence supports a conclusion that GT knew, even under the terms of Winstar’s bifurcated accounting method — a method that had been criticized by Deloitte but reviewed and approved, with skepticism, by GT— that revenue could not be properly recognized for IRUs unless the leased circuits were operational. During discovery, Cummings acknowledged that GT knew the circuits for the Cignal IRUs were not operational. Under these circumstances, GT’s approval of Winstar’s recognition of revenue for the nonfunctional circuits presents a genuine issue of material fact as to whether GT acted with the requisite scienter under Section 10(b).

We point to the IRU transactions only as one example of a transaction that suggests that GT acted with scienter. Triable *160issues regarding GT’s scienter exist for the other questionable transactions as well. Broadly speaking, there was admissible evidence that in the course of its audit GT learned of and advised against the use of indisputably deceptive accounting schemes, but eventually acquiesced in the schemes by issuing an unqualified audit opinion. See AUSA Life, 206 F.3d at 221. Based on this record, we cannot conclude that this and the other evidence adduced by the Plaintiffs was insufficient to raise a genuine issue of material fact about whether, in issuing an unqualified audit opinion letter, GT “agreed to [Winstar’s] accounting abuses, knowing ... that [investors] and others would be receiving and relying upon the manipulated financial reports.” Id,'. At this stage, the Plaintiffs have proffered enough facts constituting evidence of conscious misbehavior or recklessness to survive summary judgment.

We note that in granting summary judgment in GT’s favor, the District Court placed particular emphasis on the magnitude of GT’s audit work, both in time spent and documents reviewed. For example, it noted that GT “spent 1,928 hours of professional time,” assembled working papers spanning 3,000 pages, and reviewed “copies of contracts, Winstar business plan[s], press releases, board minutes,” and memos prepared by Winstar and GT addressing accounting issues. Winstar, 2010 WL 3910322, at *2. The number of hours spent on an audit cannot, standing alone, immunize an accountant- from charges that it has violated the securities laws. Here, the Plaintiffs adduced evidence that, notwithstanding the magnitude of its audit, GT repeatedly failed to scrutinize serious signs of fraud by an important client, including: (1) significant end-of-quarter transactions;11 (2) the absence of documents confirming the goods or services ordered by Winstar customers, the fact of delivery, or the existence of an underlying agreement; (3) the repeated failure of Winstar or its clients timely to provide supporting documentation requested by GT; (4) Winstar’s transactions outside its core business; and (5) round-trip transactions in which revenues were subsequently offset by Winstar’s payments to financially unstable customers under unrelated contractual obligations.

In short, regardless of the hours GT spent or the number of documents it reviewed in the course of its 1999 audit of Winstar, a jury reasonably could determine that the audit was so deficient as to be “highly unreasonable, representing an extreme departure from the standards of ordinary care ... to the extent that the danger was either known to [GT] or so obvious that [GT] must have been aware of *161it.” Rothman, 220 F.3d at 90 (quotation marks omitted).

2. Reliance

Section 18 of the Exchange Act, 15 U.S.C. § 78r(a), requires actual rather than constructive reliance upon a materially false or misleading statement by one who has purchased or sold a security. Heit v. Weitzen, 402 F.2d 909, 916 (2d Cir.1968). The District Court concluded that the Jefferson Plaintiffs failed to show reliance because they could not demonstrate that they or their representatives “actually saw Winstar’s 1999 Form 10-K filing, much less read the included independent accountant report of GT.” Winstar, 2010 WL 3910322, at *6. Even assuming that such “eyeball” reliance is the sort of actual reliance required by our precedents, the District Court’s conclusion somewhat understates the record evidence on this score. Ronald Clark and Livia Asher worked on behalf of the Jefferson Plaintiffs. Although Asher was unable to recall specifically that she reviewed GT’s audit opinion letter, there was evidence that she actively reviewed such letters as a matter of practice in deciding whether to recommend certain stocks. At this stage of the proceedings, Asher’s testimony is enough; from that evidence, a jury reasonably could infer that she actually reviewed and relied on the relevant statements in the documents. See, e.g., Fed.R.Evid. 406. Accordingly, we conclude that the District Court erred in granting summary judgment in GT’s favor on the Jefferson Plaintiffs’ Section 18 claims.

3. Causation

GT argues in the alternative that the District Court’s grant of summary judgment should be affirmed because the Plaintiffs failed to show loss causation. We have described loss causation as “the causal link between the alleged misconduct and the economic harm ultimately suffered by the plaintiff.” Emergent Capital Inv. Mgmt., LLC v. Stonepath Grp., Inc., 343 F.3d 189, 197 (2d Cir.2003). Among other things, the misconduct must be a “substantial factor in the sequence of responsible causation.” AUSA Life, 206 F.3d at 215 (quotation marks omitted). With these principles in mind, however, we have also warned that “if the loss was caused by an intervening event, like a general fall in the price of Internet stocks, the chain of causation will not have been established!, b]ut such is a matter of proof at trial.” Emergent Capital, 343 F.3d at 197; see also AUSA Life, 206 F.3d at 214-15 (“[W]hen the plaintiffs loss coincides with a market-wide phenomenon causing comparable losses to other investors, the prospect that the plaintiffs loss was caused by the fraud decreases.”).

Relying largely on the deposition testimony of an expert witness economist, the Plaintiffs argue that they have adduced proof of loss causation in the form of the press releases from Asensio and Winstar’s April 2001 announcements, which publicly exposed Winstar’s substantial financial weaknesses and together suggested for the first time that Winstar had engaged in improper revenue recognition practices for a period of time that included 1999. Although a much closer call, a jury could reasonably infer based on the expert testimony and other evidence that the precipitous decline in Winstar’s stock price in 2001 was attributable in part to the alleged fraud.

GT counters that any decline in Wins-tar’s stock price that was not caused by broader market trends resulted not from the alleged fraud but from Lucent’s cancellation of its credit facility. This may be true for a portion of the collapse in Wins-tar’s stock price. There is support as well *162for the argument that the downgrade in Winstar’s credit rating resulted in a substantial decline in the stock price. But these facts, if established, hardly foreclose the reasonable inference that some part of the decline was substantially caused by the disclosures about the fraud itself. We therefore conclude that a jury reasonably could find the requisite “causal link between [Winstar’s] alleged misconduct and the economic harm ultimately suffered” by the Plaintiffs. See Emergent Capital, 343 F.3d at 197.

CONCLUSION

We have considered GT’s other arguments, and we conclude that they are without merit. For the foregoing reasons, we VACATE the District Court’s grant of summary judgment, and we REMAND for further proceedings.