GREEBEL v. FTP SOFTWARE, INC.

LYNCH, Circuit Judge.

This case requires us for the first time to interpret the provisions of the Private Securities Litigation Reform Act of 1995, 15 U.S.C. § 78u-4. Plaintiffs, purchasers of FTP Software stock from July 14, 1995 to January 3, 1996, brought suit under sections 10(b) and 20(a) of the Securities Exchange Act of 1934, 15 U.S.C. §§ 78j(b), 78t(a). During the period of plaintiffs' purchases, the stock reached a high of $38.875 per share. On January 4, 1996, the company announced that sales growth had declined and that it would have lower earnings. That same day, the stock price fell 52% on heavy trading, from $25.25 to $11.875 per share. By August 9, 1996, the stock price was $8 per share. Plaintiffs' suit was filed on March 3, 1996. It was dismissed on September 24, 1998. See Greebel v. FTP Software, Inc., 182 F.R.D. 370, 376 (D.Mass.1998).

We affirm the dismissal of the complaint under the standards we now adopt:

1. The PSLRA imposes requirements for pleading with particularity that are consistent with this circuit's prior rigorous requirements for pleading fraud with particularity under Fed.R.Civ.P. 9(b).

2. The PSLRA mandates neither the adoption nor the rejection of particular patterns of evidence to prove fraud and scienter, and thus does not alter this circuit's prior law on these points.

3. The PSLRA does, significantly, impose a requirement that pleadings raise a "strong" inference of scienter rather than a merely "reasonable" inference of scienter.

4. The PSLRA does not alter the previous definition of scienter, one that in this circuit includes a narrowly defined concept of recklessness which does not include ordinary negligence, but is closer to being a lesser form of intent.

I

The district court denied defendants' first motion to dismiss largely on the basis of the complaint's allegations that defendants had routinely "whited out" the contingency terms inserted by customers into purchase orders; this was allegedly done in furtherance of a scheme to inflate revenues by improperly booking contingent transactions as final sales. After limited discovery, the district court concluded that plaintiffs could not prove the white-out claims and entered judgment on those claims. The defendants renewed their motion to dismiss the complaint,2 and the plaintiffs, in response, sought to make their allegations of fraud more specific by referring to discovered documents, but did not formally move to amend. The district court dismissed the complaint with prejudice, thus effectively denying the plaintiffs an opportunity to amend their complaint. The court did so without deciding whether, in light of the new evidence and allegations, the complaint was adequate to survive.

Plaintiffs appeal saying that summary judgment on the white-out allegations was inappropriate; that they are given refuge by Rule 56(f); that the dismissal of the remaining allegations was improper; and that they were entitled to amend their complaint.

II

The complaint alleges the following. FTP Software, Inc. develops, markets, and supports Internet and Intranet software for personal computers and networks. By the beginning of the Class Period3 (from July 14, 1995 to January 3, 1996), the demand for FTP's software was diminishing because many of FTP's clients were either developing the technology themselves or acquiring competing systems from other manufacturers, such as Microsoft and Netscape. Microsoft, for example, was incorporating networking capabilities into its new Windows 95 software, free of additional charge. In addition, FTP was struggling to keep pace with "revolutionary" technological developments that threatened to render its software obsolete. In response, FTP and several of its directors and officers4 through fraudulent schemes inflated FTP's stock price and then made various false statements and material omissions.

Plaintiffs allege that FTP failed to disclose the threats to its continued success, as well as several "questionable" sales practices. These included the making of "warehouse shipments"—that is, booking a fictitious sale of a product to a non-existent buyer, shipping that product to a warehouse for storage, and then eventually returning it to FTP. According to plaintiffs, one FTP employee who complained about these shipments, and who refused (in at least one instance) to sign for the product return, was dismissed as a result of his protest, all before the Class Period. Other objectionable sales practices included excessively discounted sales (as high as 90%) and "channel stuffing" activity that compressed sales and orders into the final weeks of a fiscal quarter, with the intention of "cosmetically" improving the reported results for that quarter. Finally, plaintiffs say that FTP failed to disclose its practice of inducing distributors to purchase more product than they needed by promising that the distributors could return the unsold product. Distributors would send their orders to FTP with a notation that they were entitled to return any unsold product. FTP then booked these sales as revenue, but because FTP understood that recognizing such sales as revenue was improper (because of a right of return existed), it allegedly instructed the sales force to white-out these right-of-return notations on the distributors' order forms.

FTP also made several statements that the plaintiffs characterize as false or materially misleading. On July 14, 1995, the first day of the Class Period, David Zirkle, FTP's President and Chief Executive Officer, reported FTP's financial performance results for the second fiscal quarter of 1995. Zirkle declared: "We are pleased with our performance for the second quarter. Sales continue to be strong in both our U.S. and international channels." Zirkle also touted the release of several new products, stating that "[t]hese products should help us achieve our revenue objective for the second half of 1995." Plaintiffs argue that these comments "falsely convey[ed] the impression that sales were, and would continue to be, healthy and strong" and that this false impression was deliberately aided by FTP's failure to disclose that in or around January 1995, the French Post Office canceled its planned purchase of $10 million of FTP products "due to the impending release of `Windows '95.'"

On the same day, Zirkle discussed FTP's impending corporate "reconfiguration" into two business units. He predicted that "FTP Software [would] lead the market in providing applications and support that make it possible to share information and access resources across workgroups, LAN's, enterprise networks and the global Internet." After this announcement, FTP's stock fell from $31.75 to $28.25. Zirkle dismissed this decline as merely "a `knee-jerk' reaction to the short-term impact of the restructuring on earnings," and on the next trading day, the stock recovered, closing at $30.875. Plaintiffs argue that these comments were misleading because Zirkle did not disclose that FTP's costly investments in its reorganization would have to be continued over the long term.

FTP's management team next met with "the investment community and with securities analysts" to promote the company's products and stock. One securities firm rated FTP as a "long-term buy." Plaintiffs assert that this report "and the estimates contained therein were based upon communications with the management of FTP and were of a nature that could only have been provided (or be based on specific information provided) by [FTP] and its management."

Meanwhile, several of the individual defendants sold some of their FTP stock. In total, the six individual defendants sold over $23 million in stock during the Class Period.

Zirkle made another false statement, plaintiffs say, on October 25, 1995, when he reported FTP's financial results for the third quarter of fiscal year 1995:

Plaintiffs assert that the "new ventures" Zirkle referred to were failing to generate the expected new business.

On November 15, 1995, FTP filed its Form 10-Q report for the third quarter of 1995 with the SEC. The report revealed a dramatic increase in accounts receivable for the fiscal year ending on December 31, 1994. FTP explained that:

Plaintiffs claim that the 10-Q report, and these statements, were false and misleading because the increased unit sales were subject to the purchasers' right to return unsold merchandise and were not the result of FTP's growth.

FTP continued the "drumbeat" of misleading positive statements, according to plaintiffs, when Zirkle, speaking in an interview published in the November 27—December 3, 1995 issue of Mass High Tech, said that "[t]he networking business (TCP/IP) is a cash cow that is feeding the development of other businesses, which are feeding back new technology that makes the core business even better." In December 1995, two other securities firms issued positive reports on FTP; after these statements, FTP's stock rose.

Finally, the complaint alleges, the "truth [began] to emerge" on January 4, 1996, when FTP announced that its earnings for the fourth fiscal quarter of 1995 would be less than the same period in 1994. FTP stated that this decline reflected, in part, the company's investment in its New Ventures Business Unit, but, nonetheless, the company's stock fell $13.375 per share to close at $11.875 per share (a one-day decline of 52%). Plaintiffs emphasize that this decline represented a $27 drop in market value (an approximately 70% decrease) from a Class Period high of $38.875 per share.

III

On March 3, 1996, plaintiffs brought suit against FTP and the individual defendants for violations of sections 10(b) and 20(a) of the Securities Exchange Act of 1934. The defendants moved to dismiss. The procedural history was recited above. The district court ultimately granted summary judgment on the white-out allegations because plaintiffs' only witness, Ms. Trudy Nichols, was unavailable and her testimony was potentially inadmissible (as hearsay). Furthermore, the plaintiffs did not establish that the defendants had ordered the alteration of any documents.

The district court also granted the defendants' renewed motion to dismiss. The court found that the complaint failed to plead the circumstances of fraud with specificity, and could not even meet the pleading standards required to establish scienter under Fed.R.Civ.P. 9(b).

IV

Interpretation of the PSLRA

The enactment of the PSLRA in 1995 marked a bipartisan effort to curb abuse in private securities lawsuits, particularly the filing of strike suits.5 See H.R. Conf. Rep. No. 104-369, at 32 (1995), reprinted in 1995 U.S.C.C.A.N. 730, 731.

The PSLRA restated the requirements for securities fraud actions at subsections 21D(b)(1) and (2), codified at 15 U.S.C. § 78u-4(b)(1)-(2). Those provisions, involved here, read as follows:

15 U.S.C. § 78u-4(b)(1)-(2).

The parties and the SEC as amicus have framed different possible interpretations of these provisions. Essentially, these questions are raised:

First, did the PSLRA alter the standards for pleading fraud with particularity previously adhered to by this circuit?

Second, did the PSLRA restrict the characteristic patterns of facts that may be pleaded in order to establish a "strong inference" of scienter? Specifically, are the two methods of showing scienter endorsed earlier by the Second Circuit—motive and opportunity or circumstantial evidence of reckless or conscious behavior sufficient to raise a "`strong inference' of fraudulent intent," see, e.g., In re Time Warner Inc. Sec. Litig., 9 F.3d 259, 268-69 (2d Cir.1993) (quoting O'Brien v. National Property Analysts Partners, 936 F.2d 674, 676 (2d Cir.1991) (internal quotation marks omitted))—now available?

Third, did the PSLRA alter the scienter requirement for actions under section 10(b) and Rule 10b-5, 17 C.F.R. § 240.10b-5? Specifically, is some form of recklessness sufficient to satisfy the scienter requirement?

The parties and amicus6 all rely heavily on competing excerpts from the congressional history of the Act and on the pre-Act case law from the Second Circuit.

The words of the statute are the first guide to any interpretation of the meaning of the statute. The usual maxim is that courts do not go beyond the text of the statute if the meaning is plain. See United Food & Commercial Workers Union, Local 328 v. Almac's Inc., 90 F.3d 1, 5 (1st Cir.1996). But that maxim has inherent flexibility. Even seemingly straightforward text should be informed by the purpose and context of the statute. See Stafford v. Briggs, 444 U.S. 527, 535, 100 S.Ct. 774, 63 L.Ed.2d 1 (1980); Puerto Rico Tel. Co. v. Telecommunications Regulatory Bd., 189 F.3d 1, 8-10 (1st Cir. 1999). Both this court and the Supreme Court have checked a sense of a statute's plain meaning against undisputed legislative history as a guard against judicial error. See, e.g., Bob Jones Univ. v. United States, 461 U.S. 574, 586, 103 S.Ct. 2017, 76 L.Ed.2d 157 (1983) ("It is a well-established canon of statutory construction that a court should go beyond the literal language of a statute if reliance on that language would defeat the plain purpose of the statute. . . ."); Cablevision of Boston v. Public Improvement Comm'n, 184 F.3d 88, 101 (1st Cir.1999). If the meaning is not plain from the words of the statute, then resort to legislative history is required. See Akins v. Penobscot Nation, 130 F.3d 482, 488 (1st Cir.1997).

On some of the points neither text nor history is indisputably clear. The legislative history is irretrievably conflicted as to the second issue—which characteristic patterns of facts may be pleaded in order to establish a "strong inference" of scienter—with all sides finding some support for their positions. About all that can be said with confidence on that issue is that Congress agreed on the need to curb abuses, that it attempted to do so in the guise of what are articulated as procedural requirements, and that there was agreement on the words of the statute and on little else. And so we return to the text of the statute and its purpose.

A. Pleading Standards for Fraud Allegations

The text of the Act requires now that any complaint alleging that a statement or omission is misleading must:

15 U.S.C. § 78u-4(b)(1). The effect of this is to embody in the Act itself at least the standards of Rule 9(b), Fed.R.Civ.P.

Before the PSLRA, a securities fraud claim had to meet the standards set by Rule 9(b). See Simcox v. San Juan Shipyard, Inc., 754 F.2d 430, 439 (1st Cir.1985). Rule 9(b) provides that "[i]n all averments of fraud or mistake, the circumstances constituting fraud or mistake shall be stated with particularity. Malice, intent, knowledge, and other condition of mind of a person may be averred generally." Fed. R.Civ.P. 9(b). This circuit has interpreted Rule 9(b) to require "specification of the time, place, and content of an alleged false representation." McGinty v. Beranger Volkswagen, Inc., 633 F.2d 226, 228 (1st Cir.1980).7 "Even where allegations are based on information and belief, supporting facts on which the belief is founded must be set forth in the complaint. And this holds true even when the fraud relates to matters peculiarly within the knowledge of the opposing party." Hayduk v. Lanna, 775 F.2d 441, 444 (1st Cir.1985) (internal quotation marks and citations omitted).

The PSLRA's pleading standard is congruent and consistent with the pre-existing standards of this circuit. This circuit has been notably strict and rigorous in applying the Rule 9(b) standard in securities fraud actions. See Maldonado v. Dominguez, 137 F.3d 1, 9 (1st Cir.1998) ("This court has been especially rigorous in applying Rule 9(b) in securities fraud actions. . . .") (quotation marks omitted); Shaw v. Digital Equip. Corp., 82 F.3d 1194, 1223 (1st Cir.1996) (similar); Romani v. Shearson Lehman Hutton, 929 F.2d 875, 878 (1st Cir.1991) ("We have been especially rigorous in demanding . . . factual support in the securities context. . . .").

The requirements of the PSLRA's new pleading standard in § 78u-4(b)(1) were largely imposed under First Circuit law, although this court has not used the same precise terminology. First, this court had already required a fraud plaintiff to specify each allegedly misleading statement or omission. See, e.g., Romani, 929 F.2d at 878 (plaintiffs' isolation of the offering materials as the source of the alleged fraud was "sufficient to identify the time and place of the alleged misrepresentations"); New England Data Servs., Inc. v. Becher, 829 F.2d 286, 292 (1st Cir.1987) (rejecting plaintiffs' claims as merely "conclusory allegations of mail and wire fraud . . . with no description of any time, place or content of the communication").

Second, this court has required a securities fraud plaintiff to explain why the challenged statement or omission is misleading by requiring that "the complaint . . . provide some factual support for the allegations of fraud." Romani, 929 F.2d at 878 (citation omitted). This means that the plaintiff must not only allege the time, place, and content of the alleged misrepresentations with specificity, but also the "factual allegations that would support a reasonable inference that adverse circumstances existed at the time of the offering, and were known and deliberately or recklessly disregarded by defendants." Id.

Finally, this court has required plaintiffs who bring their claims on information and belief to "set forth the source of the information and the reasons for the belief." Romani, 929 F.2d at 878; see also New England Data Servs., 829 F.2d at 288; Hayduk, 775 F.2d at 444-45; Wayne Inv., Inc. v. Gulf Oil Corp., 739 F.2d 11, 13 (1st Cir.1984).

Our previous strict pleading requirements under Rule 9(b) are, in our view, consistent with the PSLRA. See Maldonado, 137 F.3d at 10 n. 6.

B. Pleading Required State of Mind: Characteristic Fact Patterns

Where a plaintiff can recover money damages on proof that a defendant acted with a particular state of mind, the PSLRA now requires a complaint to "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). The "required state of mind" for liability under section 10(b) and Rule 10b-5 is referred to as scienter, which the Supreme Court has defined as "a mental state embracing intent to deceive, manipulate, or defraud." Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193 n. 12, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976).

The debate between the plaintiffs and the defendants is largely over whether Congress intended to embody, as the SEC says, the prior Second Circuit methods for proving scienter (i.e., by showing motive and opportunity or evidence of reckless or conscious behavior sufficient to raise a strong inference) or, as the defendants say, to prohibit use of at least the motive and opportunity method. This focus of the parties is not surprising, as there was much debate in Congress on these points.

The plaintiffs and the SEC argue that the PSLRA does not prohibit use of the Second Circuit's methods for proving scienter. They refer to the bill reported out of the Senate Committee on Banking, Housing, and Urban Affairs, see S. 240, 104th Cong. § 104(b) (1995), reprinted in 141 Cong. Rec. S9222 (daily ed. June 28, 1995), and to the Senate Report, which states, in part, that:

S. Res. 98, 104th Cong., at 15 (1995), reprinted in 1995 U.S.C.C.A.N. 679, 694 (footnotes omitted). They also rely on the comments of Senator Dodd, co-sponsor of the PSLRA, explaining that Congress intended to codify the Second Circuit's "pleading standards." 141 Cong. Rec. S17960 (daily ed. Dec. 5, 1995). Finally, they argue that their position is bolstered by the Statement of Managers of the Securities Litigation Uniform Standards Act of 1998, Pub.L. No. 105-353, 112 Stat. 3227, which declared that "the managers again emphasize that the clear intent in 1995 and our continuing intent in this legislation is that neither the Reform Act nor [the Standards Act] in any way alters the scienter standard in Federal securities fraud suits." Joint Explanatory Statement of the Committee of Conference, Conference Report to Accompany S. 1260, H.R. Conf. Rep. No. 105-803 ("1998 Conf. Rep."), at 15 (1998).

The defendants argue that allegations of the existence of motive and opportunity to commit fraud (or simple recklessness) do not satisfy the scienter requirement. To support their view, they note Congress' statement that "[t]he Conference Committee language is based in part on the pleading standard of the Second Circuit," (emphasis added) and that "[b]ecause the Conference Committee intends to strengthen existing pleading requirements, it does not intend to codify the Second Circuit's case law interpreting this pleading standard." H.R. Conf. Rep. 104-369, at 41 (1995), reprinted in 1995 U.S.C.C.A.N. 730, 740. Defendants also rely heavily on the footnote associated with this sentence, which states: "For this reason, the Conference Report chose not to include in the pleading standard certain language relating to motive, opportunity, or recklessness." Id. at 41 n. 23, reprinted in 1995 U.S.C.C.A.N. at 747. Further, the defendants emphasize that although the Senate Bill (S.240) included an amendment that would codify Second Circuit law, the Conference Committee eliminated that amendment. See Amend. 1485, S. 240, 104th Cong., 1st Sess. (1995), 141 Cong. Rec. S9170 (daily ed. June 27, 1995).

Finally, the defendants place considerable weight on Congress' decision to override President Clinton's veto, in light of the President's statement that in the Act Congress "intended to `strengthen' the existing pleading requirements of the Second Circuit . . . [and] to erect a higher barrier to bringing suit than any now existing[.]" H.R. Doc. No. 104-150, 104th Cong., 1st Sess. (1995), 141 Cong. Rec. H15214 (Dec. 20, 1995).8 Counsel for defendant FTP candidly admitted before the district court that the legislative history standing alone could be read either way, but argued that the President's veto—based on his reading of the Act as overruling the Second Circuit motive and opportunity test—administered the coup de grace. While the President's view of what Congress meant has some informational value, we give that view little weight: the real issue is what the intent was of the Congress, not the President.

The legislative history is inconclusive on whether the Act was meant to either embody or to reject the Second Circuit's pleading standards. As the Third Circuit has noted, "[t]he Reform Act's legislative history on this point is ambiguous and even contradictory." In re Advanta Corp. Sec. Litig., 180 F.3d 525, 531 (3d Cir.1999). The history and text show no agreement to restrict the types of evidence which may be used to show a strong inference of scienter. Indeed, it would be unusual for Congress to legislate on what fact patterns could or could not prove fraud or scienter. At best, there appears to have been an agreement to disagree on the issue of Second Circuit standards (other than the strong inference standard), and perhaps, as is common, to leave such matters for courts to resolve. See, e.g., Burlington Indus., Inc. v. Ellerth, 524 U.S. 742, 118 S.Ct. 2257, 2264, 141 L.Ed.2d 633 (1998).

From the words of the Act, certain conclusions can be drawn. First, Congress plainly contemplated that scienter could be proven by inference, thus acknowledging the role of indirect and circumstantial evidence. See 15 U.S.C. § 78u-4(b)(2) (requiring that "the complaint . . . state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind") (emphasis added). Second, the words of the Act neither mandate nor prohibit the use of any particular method to establish an inference of scienter. Third, Congress has effectively mandated a special standard for measuring whether allegations of scienter survive a motion to dismiss. While under Rule 12(b)(6) all inferences must be drawn in plaintiffs' favor, inferences of scienter do not survive if they are merely reasonable, as is true when pleadings for other causes of action are tested by motion to dismiss under Rule 12(b)(6). See Conley v. Gibson, 355 U.S. 41, 45-46, 78 S.Ct. 99, 2 L.Ed.2d 80 (1957). Rather, inferences of scienter survive a motion to dismiss only if they are both reasonable and "strong" inferences.9

Indeed, the debate about adoption or rejection of prior Second Circuit standards strikes us as somewhat beside the point. The categorization of patterns of facts as acceptable or unacceptable to prove scienter or to prove fraud has never been the approach this circuit has taken to securities fraud. As stated in Maldonado, 137 F.3d at 10 n. 6, this court has never adopted the Second Circuit test. Instead we have analyzed the particular facts alleged in each individual case to determine whether the allegations were sufficient to support scienter. See, e.g., Shaw v. Digital Equip. Corp., 82 F.3d 1194, 1209 (1st Cir.1996). In this, the approach of this circuit has been like that taken by the Supreme Court as to the issue of materiality in Basic Inc. v. Levinson, 485 U.S. 224, 108 S.Ct. 978, 99 L.Ed.2d 194 (1988).

This court has considered many different types of evidence as relevant to show scienter. Examples include: insider trading (discussed below); divergence between internal reports and external statements on the same subject (see Serabian v. Amoskeag Bank Shares, Inc., 24 F.3d 357, 361 (1st Cir.1994)); closeness in time of an allegedly fraudulent statement or omission and the later disclosure of inconsistent information (see Shaw, 82 F.3d at 1224-25); evidence of bribery by a top company official (see Greenstone v. Cambex Corp., 975 F.2d 22, 26 (1st Cir.1992)); existence of an ancillary lawsuit charging fraud by a company and the company's quick settlement of that suit (see id.); disregard of the most current factual information before making statements (see Glassman v. Computervision Corp., 90 F.3d 617, 627 (1st Cir. 1996)); disclosure of accrual basis information in a way which could only be understood by a sophisticated person with a high degree of accounting skill (see Holmes v. Bateson, 583 F.2d 542, 552 (1st Cir.1978)); the personal interest of certain directors in not informing disinterested directors of impending sale of stock (see Estate of Soler v. Rodriguez, 63 F.3d 45, 54 (1st Cir. 1995)); and the self-interested motivation of defendants in the form of saving their salaries or jobs (see Serabian, 24 F.3d at 368). While a number of these cases could be thought of as falling into motive and opportunity patterns, this court continues to prefer a more fact-specific inquiry. See, e.g., Glassman, 90 F.3d at 624 (fact that lead underwriter may have had incentive to inflate the offering price was significant, but overall, complaint failed to state a claim on which relief could be granted).

The most salient feature of the PSLRA is that whatever the characteristic pattern of the facts alleged, those facts must now present a strong inference of scienter. A mere reasonable inference is insufficient to survive a motion to dismiss. Our pre-Act case law had used both the language of "strong" inference and of "reasonable" inference in various contexts. For example, "strong inference" is used in Maldonado, 137 F.3d at 9 and Suna v. Bailey Corp., 107 F.3d 64, 68 (1st Cir. 1997). "Reasonable inference" language was used in Gross v. Summa Four Inc., 93 F.3d 987, 996 (1st Cir.1996); Shaw, 82 F.3d at 1224; Serabian, 24 F.3d at 368; Greenstone, 975 F.2d at 25; and Romani, 929 F.2d at 878. It is clear that scienter allegations now must be judged under the "strong inference" standard at the motion to dismiss stage.

Our view of the Act is thus close to that articulated by the Sixth Circuit. That court held that a plaintiff could survive a motion to dismiss by "pleading facts that give rise to a strong inference of [scienter]." In re Comshare, Inc. Sec. Litig., 183 F.3d 542, 550 (6th Cir.1999) (internal quotation marks omitted). The Sixth Circuit found that evidence of motive and opportunity to commit fraud did not, of itself, constitute scienter for purposes of section 10(b) and Rule 10b-5. See id. at 551. "Indeed, those courts addressing motive and opportunity in Securities Act cases have held only that facts showing a motive and opportunity may adequately allege scienter, not that the existence of motive and opportunity may support, as scienter itself, liability under § 10b or Rule 10b-5." Id. The court held that evidence of motive and opportunity was "relevant" to pleading facts that could establish scienter, and, on occasion, could "rise to the level of creating a strong inference of reckless or knowing conduct." Id. Nevertheless, such evidence, standing alone, could not "constitute the pleading of a strong inference of scienter." Id.; accord Bryant v. Avado Brands, Inc., 187 F.3d 1271, 1282-83 (11th Cir.1999).

Without adopting any pleading litany of motive and opportunity, we reject defendants' argument that facts showing motive and opportunity can never be enough to permit the drawing of a strong inference of scienter. But, as we cautioned in Maldonado, 137 F.3d at 10 n. 6, merely pleading motive and opportunity, regardless of the strength of the inferences to be drawn of scienter, is not enough. Three circuits have interpreted the PSLRA as permitting use of motive and opportunity type pleading if it raises a strong inference. See In re Advanta Corp. Sec. Litig., 180 F.3d 525, 534-35 (3d Cir.1999); Press v. Chemical Inv. Servs. Corp., 166 F.3d 529, 537-38 (2d Cir.1999); Williams v. WMX Techs., Inc., 112 F.3d 175, 178 (5th Cir.1997) (dicta). Like the Third Circuit, we caution that "catch-all allegations that defendants stood to benefit from wrongdoing and had the opportunity to implement a fraudulent scheme are [not] sufficient." In re Advanta Corp., 180 F.3d at 535.

Similarly, the PSLRA neither prohibits nor endorses the pleading of insider trading as evidence of scienter, but requires that the evidence meet the "strong inference" standard.10 Unusual trading or trading at suspicious times or in suspicious amounts by corporate insiders has long been recognized as probative of scienter. See Shaw, 82 F.3d at 1204; Rubinstein v. Collins, 20 F.3d 160, 169-70 (5th Cir.1994); Greenstone, 975 F.2d at 26. The vitality of the inference to be drawn depends on the facts, and can range from marginal, see Shaw, 82 F.3d at 1204, to strong, see Rubinstein, 20 F.3d at 169-70. This continues to be true in litigation after the effective date of the PSLRA. Indeed, in Greenstone we noted, and still think today, that allegations of unusual insider trading by a defendant with access to material non-public information can support a strong inference of scienter. See Greenstone, 975 F.2d at 26. We similarly caution that mere pleading of insider trading, without regard to either context or the strength of the inferences to be drawn, is not enough. See Maldonado, 137 F.3d at 9-10. At a minimum, the trading must be in a context where defendants have incentives to withhold material, non-public information, and it must be unusual, well beyond the normal patterns of trading by those defendants.11

C. Substantive Scienter Standards

The parties disagree about the effect, if any, of the PSLRA on the substantive standard for proving scienter. We start with the state of the law in this circuit before the enactment of the PSLRA.

The Supreme Court has defined scienter as "a mental state embracing intent to deceive, manipulate, or defraud." Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193 n. 12, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976). The Court explicitly reserved the issue of whether recklessness sufficed, saying "[i]n certain areas of the law recklessness is considered to be a form of intentional conduct for purposes of imposing liability for some act." Id. Before enactment of the PSLRA, most circuits had held that scienter in civil securities fraud actions could be shown by showing recklessness.12 It is accepted that recklessness may establish intent to defraud in criminal prosecutions under § 17(a)(1) of the Exchange Act and under the mail and wire fraud statutes. See 8 L. Loss and J. Seligman, Securities Regulation 3656-57 (3d ed.1991). The question became whether the standard for criminal liability should also apply to civil liability. That, in turn, raised the question of what was meant by recklessness.

We have used a definition of recklessness articulated by the Seventh Circuit:

Sundstrand Corp. v. Sun Chem. Corp., 553 F.2d 1033, 1045 (7th Cir.1977) (quoting Franke v. Midwestern Okla. Dev. Auth., 428 F.Supp. 719, 725 (W.D.Okla.1976)). Without at first explicitly adopting that standard, this court assumed that it applied and cited it with approval in Cook v. Avien, Inc., 573 F.2d 685, 692 (1st Cir. 1978). In Hoffman v. Estabrook & Co., 587 F.2d 509, 515-16 (1st Cir.1978), this court again assumed that recklessness could ground a Rule 10b-5 action and, again, quoted with approval a Seventh Circuit definition:

Sanders v. John Nuveen & Co., 554 F.2d 790, 793 (7th Cir.1977). This court explicitly rejected a formulation of recklessness as mere negligence, finding it had to come closer to being a lesser form of intent than merely a greater degree of ordinary negligence. See Hoffman, 587 F.2d at 516 & n. 10.

In Serabian this court held that it was error to dismiss a securities fraud complaint that alleged sufficient facts to draw "an inference that the [defendant] knew, or should have known, that its public statements were inconsistent with the actual conditions then being reported to [it]." Serabian, 24 F.3d at 365 (emphasis added and original emphasis omitted). We understand Serabian to have used "should have known" in the reckless disregard sense used in Cook and Hoffman. This court has explicitly tested factual allegations to see whether they supported an inference that defendants acted with reckless disregard. See Romani, 929 F.2d at 878. Since then, there have been encapsulated references to this rule in dicta in other cases. See, e.g., Maldonado, 137 F.3d at 9 n. 4. The rule in this circuit has been to accept recklessness, as narrowly defined in the two Seventh Circuit cases (Sundstrand and Sanders), as a method of proving scienter. See Serabian, 24 F.3d at 365. That definition of recklessness does not encompass ordinary negligence and is closer to a lesser form of intent.

The effect of the PSLRA on the standard for scienter has been much disputed. The Act itself is silent on the general scienter requirements for 10b-5 actions, referring only to scienter as "the required state of mind." 15 U.S.C. § 78u-4(b)(2). Plaintiffs and the SEC maintain that the PSLRA did not intend or purport to change whatever the preexisting standard was, and that standard included some form of recklessness. The defendants say that only the most heightened recklessness standard should be used, and that the SEC's views are not entitled to any weight.

The circuit courts have reached different results. A panel in the Ninth Circuit has held that the PSLRA elevated the standard to one of "deliberate recklessness," In re Silicon Graphics Inc. Sec. Litig., 183 F.3d 970, 974 (9th Cir.1999). One district court within this circuit has concluded that recklessness is insufficient and that only conscious conduct would suffice under the Act. See Friedberg v. Discreet Logic Inc., 959 F.Supp. 42, 48 (D.Mass.1997).

In contrast, the Sixth Circuit has concluded that the PSLRA did not alter the state of mind requirement and thus a complaint pleading facts that give rise to a "`strong inference of recklessness' of the kind required for securities fraud liability" suffices. In re Comshare, Inc. Sec. Litig., 183 F.3d 542, 552-53 (6th Cir.1999). The Sixth Circuit had previously adopted, as had this circuit, the Sundstrand definition of recklessness, see id. at 550, and concluded that same standard applied after the PSLRA was enacted. The Eleventh Circuit has expressed its "basic agreement" with the Sixth Circuit, Bryant v. Avado Brands, Inc., 187 F.3d 1271, 1282-83 (11th Cir.1999), holding that the PSLRA did not "substantively change the actionable level of scienter," id. at 1284-85. The Third Circuit has concluded that "[a]lthough the Reform Act established a uniform pleading standard, it did not purport to alter the substantive contours of scienter," In re Advanta Corp. Sec. Litig., 180 F.3d 525, 534 (3d Cir.1999), and reiterated its prior law that recklessness, as defined in Sundstrand, suffices, see id. Without much discussion of the issue, the Second Circuit has adhered to its previous acceptance of recklessness. See Press v. Chemical Inv. Servs. Corp., 166 F.3d 529, 537-38 (2d Cir.1999). The Fourth Circuit has also concluded that the PSLRA did not change the pre-existing scienter standard. See Phillips v. LCI Int'l, Inc., 190 F.3d 609, 620 (4th Cir.1999) ("[T]o establish scienter, a plaintiff must still prove that the defendant acted intentionally, which may perhaps be shown by recklessness."). Two district courts in this circuit have reached the conclusion that some form of recklessness is available. See In re PLC Sys., Inc. Sec. Litig., 41 F.Supp.2d 106, 115 (D.Mass. 1999); Lirette v. Shiva Corp., 27 F.Supp.2d 268, 282 (D.Mass.1998).

We agree with those courts that hold that the PSLRA did not address the substantive definition of scienter. The legislative history shows no such intent and the language of the Act itself does not address the topic. See generally Dunn, Note, Pleading Scienter After the Private Securities Litigation Reform Act, 84 Cornell L.Rev. 193 (1998). The opinions of the Sixth and Third Circuits discuss these points ably and there is no reason to repeat the discussion here.

We add additional reasons that buttress this conclusion. The PSLRA does in fact discuss the role of "knowing" violations, and thus an aspect of scienter, in two different respects. The first concerns contribution; the second, "safe harbors" for defendants. As to contribution,13 the Act, through a new § 21D(g), added a section "to preserve joint and several liability for persons who knowingly commit securities fraud, but otherwise to proportionately limit liability to the `portion of the judgment that corresponds to the percentage of responsibility of that covered person.'" 10 Loss & Seligman, at 4687 (quoting 15 U.S.C. § 78u-4(f)(2)(B)(i)). The PSLRA specifies that:

15 U.S.C. § 78u-4(f)(2)(A). In turn "knowingly commits a violation of the securities laws" is defined (for 10b-5 purposes) as requiring "actual knowledge" that a representation is false or an omission renders a representation false. Id. § 78u-4(f)(10)(A). The definition specifically excludes "reckless conduct" as a basis for construing a knowing commission of a violation. Id. § 78-u(f)(10)(B).

These special contribution provisions lead to several conclusions. Congress, having explicitly eliminated recklessness as a basis for imposing joint and several liability, should not be taken as implicitly having eliminated recklessness as a basis for any liability. See In re Silicon Graphics, 183 F.3d at 995 (Browning, J., concurring in part and dissenting in part). Because joint and several liability is more onerous than individual liability, the exclusion of recklessness as the basis for imposing joint and several liability constitutes a recognition that some form of recklessness may suffice for individual liability. Furthermore, Congress took great care to insure that the actual knowledge requirement was restricted to the joint and several liability provisions of the Act. Section 78u-4(f)(1) provides that "nothing in this subsection shall be construed to create, affect, or in any manner modify, the standard for liability associated with any action arising under the securities laws." 15 U.S.C. § 78u-4(f)(1). Including this language would not make sense if Congress had altered the general scienter requirement to restrict it to actual knowledge.

The "safe harbor" provisions of the Act similarly buttress the conclusion that the Act did not alter pre-existing law defining scienter. The PSLRA adopted a statutory "safe harbor" by adding a new section 27A to the 1933 Act, 15 U.S.C. § 77z-2, and a new section 21E to the 1934 Act, 15 U.S.C. § 78u-5. The safe harbor has two alternative inlets: the first shelters forward-looking statements that are accompanied by meaningful cautionary statements. See 15 U.S.C. § 78u-5(c)(1)(A)(i). The second inlet is of importance here. It focuses on the state of mind of the defendant and precludes liability for a forward-looking statement unless the maker of the statement had actual knowledge it was false or misleading. See § 78u-5(c)(1)(B); H.R. Conf. Rep. No. 104-369 (1995), reprinted in 1995 U.S.C.C.A.N. 730, 743. Again, this new section 21E to the 1934 Act is explicit, in contrast with the lack of such language in the definition of scienter in § 21D(b)(2). See Bryant, 187 F.3d at 1284-85; In re Silicon Graphics, 183 F.3d at 995 (Browning, J., concurring in part and dissenting in part).

Concluding that the Act does not alter the pre-existing definition of scienter adopted by this circuit, we accordingly test the complaint against that definition.

V

Application of Standards to Plaintiffs' Complaint

We review the dismissal of plaintiffs' complaint de novo, giving plaintiffs the benefit of all reasonable inferences, but holding plaintiffs to the standard of showing a strong inference of scienter. See Gross v. Summa Four, Inc., 93 F.3d 987, 991 (1st Cir.1996). We think the claims are either insufficiently particularized or, where particularized, do not permit a strong inference of scienter.14

At the heart of plaintiffs' case is the allegation that defendants consistently overstated the earnings of the company by improperly booking as revenue (and inadequately reserving) "sales" that were actually contingent transactions. This was improper, plaintiffs say, under generally accepted accounting principles ("GAAP"), specifically Statement of Financial Accounting Standards No. 48 ("FAS 48"). Plaintiffs claim that the sales were contingent because there were unlimited return rights. Plaintiffs say that they have evidence both tending directly to show conscious wrongdoing on the part of defendants and circumstantial evidence from which such wrongdoing may be inferred, including that the defendants had both motive and opportunity.

A. Direct Evidence of Scienter

Two types of allegations tend to show conscious wrongdoing: the "white-out" and the warehousing allegations.

1. The "White-Out" Allegations and the Rule 56(f) Motion

The "white-out" allegations claimed that FTP personnel "whited out" (in a manner undetectable to company auditors) the customers' additions to standard purchase orders; those additions made orders contingent on the customers' unlimited right to return the goods to FTP. The white-out allegations were powerful, as the district court recognized in initially denying the motion to dismiss. If adequately supported, claims that management deliberately altered company records to hide material information from company auditors could well create strong inferences of scienter. But, as the district court correctly ruled on summary judgment,15 plaintiffs could not produce admissible evidence to support the white-out allegations, and so we disregard these allegations.

2. The Warehousing Allegations

The warehousing allegations remain. In essence, the complaint asserts that at some time before the Class Period, the company made a phony sale or sales and caused to be booked as goods sold certain product that was shipped to a warehouse and not to customers; the company then recognized the revenue from such phony sales. After a period, the product was sent back from the warehouse as "returned" goods. The allegations state that Robert Casa, an employee who refused to sign for the "returned" product and complained about the practice, was fired. If true, such practices by a company are very serious. See United States v. Bradstreet, 135 F.3d 46, 48 (1st Cir.1998) (affirming a criminal conviction for, inter alia, "knowingly falsifying [a company's] books and records in an attempt to conceal [securities] fraud").

The complaint is deficient in not identifying when this took place. The complaint is specific only in saying this occurred before the Class Period. The complaint alleges, on information and belief, that the practice continued into the Class Period but provides no specifics about why the practice is thought to have occurred during the Class Period or why it caused harm to plaintiffs. The defendants say the temporal lag means the allegations are irrelevant and should be disregarded. The allegations are not irrelevant—evidence of past practice may indeed be probative of present practice. But there is scant else from which to infer that this was the company's practice at any pertinent time, and the allegations are not enough to support a strong inference of scienter. See Lefkowitz v. Smith Barney, Harris Upham & Co., 804 F.2d 154, 155-56 (1st Cir.1986).

B. Indirect Evidence of Scienter

There is also indirect evidence from which plaintiffs say scienter can be inferred. Defendants, or so the complaint alleges, knew the company was in trouble because they knew that Microsoft would have, as an integral part of its Windows 95 package, a product that would compete with FTP's wares. Purchasers of Windows 95 would therefore have no reason to purchase FTP's product. Before and during the Class Period (July 14, 1995, to January 3, 1996), FTP engineers repeatedly warned management of this competitive threat. Indeed, certain large orders, such as a $10 million order by the French Post Office, were cancelled. Plaintiffs allege that FTP's management responded through two strategies, "channel stuffing" and contingent sales, which artificially inflated the company's revenues.

1. Channel Stuffing

"Channel stuffing" means inducing purchasers to increase substantially their purchases before they would, in the normal course, otherwise purchase products from the company. It has the result of shifting earnings into earlier quarters, quite likely to the detriment of earnings in later quarters. There is nothing inherently improper in pressing for sales to be made earlier than in the normal course, and we do not understand plaintiffs' complaint to make any such claim. Plaintiffs make use of the channel stuffing allegations in a different way. They say evidence of channel stuffing supports their contention that management knew that revenues during the Class Period would be low and attempted to hide that fact by shifting income through channel stuffing (which remained undisclosed) and by artificially inflating income through improper revenue recognition. In this context, the channel stuffing evidence has some probative value.16 But that value is weak. Unlike altering company documents, there may be any number of legitimate reasons for attempting to achieve sales earlier. Thus, it does not support a strong inference of scienter.

2. Contingent Sales

Plaintiffs allege that the financial statements included in FTP's Form 10-Q report for the third quarter of 1995 were prepared in violation of GAAP and contained improperly inflated revenues and earnings. Specifically, plaintiffs claim that FTP recognized revenues from sales that included a right of return, which did not meet the requirements for revenue recognition set forth in FAS 48. When a buyer has the right to return a product, FAS 48 prohibits the seller from recognizing income from the sale unless six conditions are met. See Statement of Financial Accounting Standards No. 48, ¶ 6 (Fin. Accounting Standards Bd., June 1981).

Plaintiffs focus on three of the six FAS 48 conditions: that the buyer's obligation to pay the seller is not contingent on resale of the product; that the seller does not have significant obligations for future performance directly to bring about resale of the product by the buyer; and that the amount of future returns can be reasonably estimated. See id. If any of the conditions are not met at the time of the sale, sales revenue cannot be immediately recognized. See id.

Plaintiffs allege that during the Class Period FTP recorded as "sales" transactions including a right of return that violated all three of the above conditions. Plaintiffs claim FTP "induc[ed] distributors to purchase more product than they needed with the promise that they could return the product if it were unsold. . . . [S]ubsequently, a material portion of these `sales' were either returned in the fourth fiscal quarter of 1995 or remained with distributors, but were unpaid for." Plaintiffs contend that "[u]nder certain circumstances, the distributor could defer payment until FTP's or its own sales force had booked a sale to an ultimate customer for the products. FTP, however, would not receive payment until the distributor was in a position to book the sale and ship the product and receive payment from the end user. . . ." Furthermore, "due to the constantly changing competitive environment and the release of Microsoft's `Windows '95', FTP had no way to reasonably estimate returns." Finally, even if all the conditions for immediate revenue recognition are met, FAS 48 requires that the seller reduce sales revenue and cost of sales reported in the income statement to reflect estimated returns. See id. ¶ 7. Plaintiffs allege that FTP violated this by failing to adequately reserve for returns.

Violations of GAAP standards such as FAS 48 could provide evidence of scienter. See Malone v. Microdyne Corp., 26 F.3d 471, 478-79 (4th Cir.1994). To support even a reasonable inference of scienter, however, the complaint must describe the violations with sufficient particularity; "a general allegation that the practices at issue resulted in a false report of company earnings is not a sufficiently particular claim of misrepresentation." Gross v. Summa Four, Inc., 93 F.3d 987, 996 (1st Cir.1996) (quoting Serabian v. Amoskeag Bank Shares, Inc., 24 F.3d 357, 362 n. 5 (1st Cir.1994)). Here, as the district court correctly concluded, the complaint clearly falls short. The allegations in the complaint do not include such basic details as the approximate amount by which revenues and earnings were overstated, see Gross, 93 F.3d at 996; the products involved in the contingent transactions, cf. Malone, 26 F.3d at 476-77 (products that were "sold" with rights of return specifically identified); the dates of any of the transactions; or the identities of any of the customers or FTP employees involved in the transactions. We do not say that each of these particulars must appear in a complaint, but their complete absence in this case is indicative of the excessive generality of these allegations.

As part of their opposition to defendants' renewed motion to dismiss, plaintiffs attempted to introduce evidence of four alleged contingent transactions drawn from defendants' automatic disclosure. The district court, however, refused to take this new evidence into consideration in its ruling. The court reasoned that there would have been no disclosure but for the white-out allegations, because the complaint would have been dismissed at the outset; the white-out allegations proved insubstantial, even with the benefit of disclosure; and therefore the remainder of the complaint should be judged without the additional evidence obtained through disclosure. We need not decide whether the district court's refusal to consider the additional evidence was correct, because we believe that plaintiffs' additional evidence would not have sufficed to prevent dismissal.

Plaintiffs identify four sets of transactions from the third quarter of 1995 that allegedly involve improperly booked revenue. Plaintiffs present invoices, purchase orders, and other documentation that show, they contend: (a) a set of transactions totaling $678,000 with a distributor, Merisel, in which Merisel was not obliged to pay for the product; (b) a $705,250 transaction with reseller CC-OPS in which CC-OPS was given an unlimited right to return the product; (c) a $1.14 million transaction with reseller Afina Sistemas that involved the "sale" of an FTP product that did not yet exist; and (d) a $416,325 transaction with reseller Force 3 that was contingent upon Force 3 receiving a government contract.

a. The Merisel Transactions

The Merisel allegations involve one $130,078 transaction in August 1995 and two transactions in late September 1995 totaling $548,192. Plaintiffs claim that the August transaction, for which FTP issued an invoice, was not a true sale but a "stock rotation," in which FTP replaced outdated products in Merisel's inventory at no charge. Plaintiffs point to a credit issued to Merisel by FTP in mid-October for the full amount. Less detail is provided concerning the September transactions. To support their contention that those transactions were improperly booked contingent sales, plaintiffs proffer three items: the $548,192 posted to accounts receivable on September 29; the fact that $494,872 remained unpaid as of December 31; and the agreement between Merisel and FTP, which states that Merisel will pay FTP for all copies of FTP's products sub-licensed by Merisel and its dealers.

A possible—though far from necessary—conclusion is that the August transaction was an exchange of new products for old improperly booked as a sale, as the original Merisel purchase order contains the notation "[o]ffsetting order f. stockrotation" (sic). The September transactions, on the other hand, are described in insufficient detail to support plaintiffs' allegations. The mere existence of an overdue receivable does not support an inference that the original transaction was booked as a sale in violation of GAAP.

b. The CC-OPS Transaction

The CC-OPS allegation concerns a September 29 order for $705,250 of FTP products, which FTP immediately booked as a sale. A letter from FTP's sales director for the Americas apparently accompanied the invoice. The letter stated that it was FTP's "policy" to "allow[ ] large or frequent customers to return product without contingency within 60 days of receipt of order." Large customers were defined as those generating over $100,000 per year. A copy of the original letter is not present in the invoice file. On November 27, CC-OPS faxed a copy of the letter back to FTP, and FTP extended the right of return from 60 to 90 days. This revised letter is present in the file. Shortly after the return period was lengthened, FTP issued a credit for the full amount of the invoice and authorized the return of the product. On the original invoice is written: "Credit per Jack Geraghty. Not recognizable revenue."17

Plaintiffs charge that the initial booking of revenue from this transaction violated GAAP; that the letter indicates that FTP had a policy of granting unlimited return rights to its customers; and that the absence of the original September 29 letter from the file indicates that FTP was attempting to conceal the existence of return rights. However, FAS 48 permits sellers to recognize sales that include a right of return, so long as the required conditions are met and the seller establishes a reasonable reserve for returns. The granting of a right of return in a particular transaction, or even a general policy of granting return rights, does not per se mean that revenue cannot be recognized at the time of sale. Plaintiffs merely make an allegation that FTP failed to adequately reserve and materially overstated FTP's revenues. Without any information on FTP's experience with past return rates, the size of its reserve for returns, or how the reserve changed over time, it is difficult to infer that FTP's revenue recognition decisions were unreasonable enough to violate GAAP, or that they give rise to a strong inference of scienter. "`Generally accepted accounting principles,' . . . tolerate a range of `reasonable' treatments, leaving the choice among alternatives to management." Thor Power Tool Co. v. Commissioner of Internal Revenue, 439 U.S. 522, 544, 99 S.Ct. 773, 58 L.Ed.2d 785 (1979).

c. The Afina Sistemas Transaction

The Afina Sistemas ("Afina") allegation involves two purchase orders totaling $1.14 million issued on September 28. The orders were for 200,000 copies of FTP's Internet browser, custom made for an Afina client. FTP booked the entire amount as revenue on September 29 and carried the amount as an account receivable throughout the fourth quarter. Plaintiffs claim that this revenue was improperly booked because the version of the browser ordered by Afina was then still under development. An internal FTP document indicates that the test version of the Spanish edition of the browser was not scheduled to be completed until late October. Plaintiffs also point to notations reading "DO NOT SHIP PRODUCTS" on documents attached to each invoice.

This transaction is difficult to classify. This appears to be one part of a larger undertaking (the "Telefonica project" referred to on the Afina purchase order) about which plaintiffs present only fragmentary information. It is not clear that the custom version of FTP's browser referred to in Afina's purchase order is the same as the Spanish version listed on FTP's development schedule. Furthermore, marking "DO NOT SHIP" prominently on documents seems an odd way to conceal an improperly booked sale from auditors. Finally, the fact that an overseas customer with 90 days to pay has not paid after 94 days is not highly suspicious. It is possible to infer, however—at least tentatively—that this transaction should not have been booked as a sale in September. It is a leap from there to a strong inference of scienter.

d. The Force 3 Transaction

The Force 3 allegations concern a purchase order for $416,325 issued on September 30, which stated on its face that it was contingent on Force 3's receipt of a government contract. FTP nonetheless immediately recorded the entire amount as an account receivable. On December 29, FTP issued Force 3 a credit for the full amount. The same day, Force 3 sent FTP a new purchase order for the same products it ordered on September 30, but this purchase order did not refer to any contingency. FTP issued a new invoice and again booked the amount as a sale. The original booking of this sale in September appears to have violated the requirement in FAS 48 that the buyer's obligation to pay the seller is not contingent on resale of the product.

At best, plaintiffs' additional evidence supports an inference that FTP improperly recognized from $416,000 to $1.55 million in revenue in the third quarter of 1995. Because FTP reported overall revenue during the quarter of $37.1 million, these transactions do not support a strong inference of scienter.18 It is equally possible to conclude that FTP made some incorrect accounting decisions regarding a limited number of transactions. Seeing fraud, however, requires too great of an inferential leap. In short, even when viewed in combination with plaintiffs' other allegations, plaintiffs' additional evidence does not support a strong inference of scienter, and thus the district court's decision not to consider the evidence could not have affected the outcome of the motion to dismiss.

3. Insider Trading

The allegations of insider trading do not, either alone or together with the other allegations, suffice. The individual defendants sold FTP common stock during the Class Period. For example, Zirkle sold 40,000 shares on July 21, 1995 at a price of $26.27 per share, for total proceeds of $1,050,800. Goodnow sold 40,000 shares on July 27 at a price of $29.00 per share, for a total price of $1,160,000. Last, on August 2, 1995, Charlotte Evans sold 200 shares at $28.50 per share, for a total of $57,000. All three defendants sold stock later in the Class Period as well.

We first look at context. The timing does not appear very suspicious. None of these three key players sold at the high points of the stock price. Each waited to sell until after FTP announced a corporate reorganization on July 14, an announcement which caused the price of the stock to fall. Each sold some stock before an allegedly manipulated analyst's report from Brookehill Equities recommended FTP stock as a long-term buy on August 3, 1995, and before a favorable Cowen & Co. report on December 1, 1995.

The total sum of sales involved—over $23 million during a six month period—could be suspicious, but a closer look provides ready explanations. Goodnow, as plaintiffs allege, retired on October 26, 1995, from his position as vice president, CFO, and treasurer. His sale of stock in July occurred not long before he left the company. He also sold a considerable number of shares after he left the company—690,000 shares accounting for $19 million out of his total of almost $20.2 million in sales during the Class Period. The vast majority of the $23 million in sales by the individual defendants, more than $20 million worth, were by one individual who was leaving the company, and more than $19 million was after that individual had left the company. It is not unusual for individuals leaving a company, like Goodnow, to sell shares. Indeed, they often have a limited period of time to exercise their company stock options. As to the others, the sales do not reflect either unusual sales or sales made before a big "event" unknown to the public. Selling after delivering news that causes a company's stock price to go down is not suggestive of withholding information. But that is what happened here. Plaintiffs provided no information on sales by corporate insiders at times outside the Class Period, so there is no comparison point.

Although the total sum involved was large, the district court correctly concluded that plaintiffs produced no evidence that the trading was out of the ordinary or suspicious. Absent additional evidence, it is not possible to draw a strong inference of scienter based on improper trading on material, non-public information.

C. Other Alleged False Statements and Material Omissions

1. Zirkle's Statements

Zirkle's upbeat statements of optimism and puffing about the company's prospects, described earlier, have each been reviewed and we conclude that they are not actionable. See Glassman, 90 F.3d at 635-36; Shaw, 82 F.3d at 1217-19.

2. Form 10-Q Report for Third Quarter 1995

The 10-Q Report stated there was an increase in accounts receivable and attributed it to increased sales. Plaintiffs do not say this statement was false; only that it was misleading because it did not say the increased sales were subject to return rights. As an independent ground, this is too slight; as a ground in service of the contingent sales/improper booking argument, it fails for the same reasons that argument fails.

D. Individual Defendants

Because the dismissal of the complaint is upheld, we do not reach the arguments of the individual defendants that the facts alleged do not make out a claim against them.

E. Section 20(a) Claim

Plaintiffs also assert claims against the individual defendants under Section 20(a) of the Exchange Act, which provides for derivative liability of persons who "control" others found to be primarily liable under the Exchange Act. See 15 U.S.C. § 78t(a). Because plaintiffs' complaint does not adequately allege an underlying violation of the securities laws, the district court was correct to dismiss the Section 20(a) claim. See Suna v. Bailey Corp., 107 F.3d 64, 72 (1st Cir.1997).

VI

The district court correctly refused to dismiss the complaint originally and was well within its discretion in limiting the discovery it afforded. The difficult and different balance the Act now requires—testing allegations before little or no discovery, but holding plaintiffs to a strong inference of scienter standard—has been honored in this case. Plaintiffs did not have enough weight on their side of the balance to meet the requirements of the Act, and so we affirm the dismissal. Costs to appellants.