In February I discussed the Ninth Circuit’s curious new two-step method for analyzing scienter under the Private Securities Litigation Reform Act of 1995, which the court believed had been mandated by the Supreme Court in Tellabs, Inc. v. Makor Issues & Rights, Ltd., 551 U.S. 308, 127 S. Ct. 2499 (2007). It turns out that another court of appeals reads the same Supreme Court decision to require not two, but three steps.

The Fifth Circuit adopted its competing “three step approach” last July in Indiana Elec. Workers’ Pension Trust Fund IBEW v. Shaw Group, Inc., 537 F.3d 527 (5th Cir. 2008), and reaffirmed it last week in Flaherty & Crumrine Preferred Income Fund, Inc. v. TXU Corp.:

“First, the allegations must, as in federal pleadings generally, be taken as true. Second, courts may consider documents incorporated in the complaint by reference and matters subject to judicial notice. The facts must be evaluated collectively, not in isolation, to determine whether a strong inference of scienter has been pled. Third, a court must take into account plausible inferences opposing as well as supporting a strong inference of scienter. The inference of scienter must ultimately be ‘cogent and compelling,’ not merely ‘reasonable’ or ‘permissible.’”

— F.3d —, 2009 WL 930055, at *5 (5th Cir. Apr. 8, 2009) (quoting Ind. Elec., 537 F.3d at 533 (citing Tellabs, 127 S.Ct. at 2509-10)).

It’s always nice to be able to break a difficult task down into steps.  If you’re writing for a federal circuit court, it would be terrific if a recent Supreme Court decision already broke it down for you.  Tellabs, perhaps unfortunately, didn’t do that.  Rather, it “establish[ed]” three “prescriptions”:  1) “accept all factual allegations in the complaint as true“; 2) “consider the complaint in its entirety, as well as other sources courts ordinarily examine when ruling on Rule 12(b)(6) motions to dismiss”; and 3) when determining whether the facts give rise to a strong inference of scienter, “take into account plausible opposing inferences.” Tellabs, 127 S.Ct. at 2509-10.  The first of these prescriptions must be applied concurrently with, not prior to, the second, so they certainly aren’t “steps.”

The Fifth Circuit version leaves out the first part of prescription number two—”consider the complaint”—thus allowing a possible sequential interpretation:  look at allegations, look at documents, draw inferences.  Even if that reading were possible, though, it could as well apply to any area of federal civil litigation.    Everything special to the PSLRA pleading standard would be contained within a subset of the third step.  (Indeed, the first two prescriptions of Tellabs just affirm that, as to those two issues, the PSLRA did not alter the law.)

Judge Benavides, who authored last week’s decision in Flaherty & Crumrine, seemed a bit half-hearted about the three step approach, pointing out that the circuit had “recently interpreted Tellabs” this way—i.e., don’t blame me—and never mentioning any of the steps after their first dutiful recitation.  Like the Ninth Circuit’s “dual method, ” it’s not clear whether this approach has the potential to affect outcomes of cases, or if so, how.  If not, it’s not likely to be undone by an en banc court or by the Supreme Court, even through there is, in some sense, a circuit split.  If it can’t affect outcomes, perhaps it will be viewed as non-binding precedent by subsequent panels and will fade away.

The Fifth Circuit showed yesterday that it is possible to apply the PSLRA pleading standards without filling up dozens of pages of F.3d.  The plaintiff-appellant in Brunig v. Clark, — F.3d. — No. 08-20059, 2009 WL 376907 (Feb. 17, 2009), had been the defendant-respondent’s lawyer.  He had agreed to accept as his fee an interest in mineral leases, and later expected sale proceeds, from property owned by his client.  The dispute arose when the client sent the plaintiff a bill for operating expenses on the property and, when the bill wasn’t paid, a notice of default.

Brunig isn’t a class action, it has nothing to do with a drop in the price of any stock, and it isn’t the sort of case that Congress was was worried about when it enacted the PSLRA.  Although Brunig doesn’t say anything about congressional purpose, it does manage to keep a simple case simple, despite the statute.  The complaint, according to the court, is “prolix,” despite being, at fifty-six pages, on the slim side for a securities fraud complaint, at least in my experience.   By comparison with the Brunig court’s handiwork, however, the charge is justified.  In five paragraphs, the court states the elements of the Rule 10b-5 claim, assesses the complaint’s sufficiency as to both the PSLRA particularity requirements (misstatement/omission and scienter), and reverses the district court’s dismissal of the claim.  While it doesn’t break new legal ground, to my mind it doesn’t exclude anything important, either.  With equal efficiency, the court affirms the dismissal of a Securities Act claim and a RICO claim, and reverses a sanctions order.  Compliements to Judge Higginbotham, who wrote for the court.

Having misread the Supreme Court’s decision in Tellabs, Inc. v. Makor Issues & Rights, Ltd., 551 U.S. 308, 127 S. Ct. 2499 (2007), the Ninth Circuit has recently developed a new and nonsensical framework for analyzing securities fraud complaints.

The central issue in Tellabs was the meaning of the word “strong” in the provision of the Private Securities Litigation Reform Act of 1995 that requires plaintiffs 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 Tellabs Court also cautioned judges applying the provision “not to scrutinize each allegation in isolation but to assess all the allegations holistically.”  127 S. Ct. at 2511.  “[T]he 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?”  Id.

You might suppose that this “holistic” approach would be pretty uncontroversial, given the way human beings normally draw inferences.  An individual fact (an alleged fact) might be quite specific without strongly suggesting culpability, and yet, when viewed in conjunction with other specific and yet equally unsuggestive facts, might be strong evidence of guilt.  For example:

• Fact A:  The victim was run down by an orange Toyota at 9 pm on Feb. 2, three miles from Defendant’s home.
• Fact B:  Defendant’s neighbor owns an orange Toyota.
• Fact C:  Defendant does not own a car.
• Fact C:  Defendant’s neighbor’s orange Toyota was stolen at 8 pm on Feb. 2.
• Fact D:  Two of Defendant’s friends report that they were with Defendant at Defendant’s home in the late afternoon of Feb. 2, and that Defendant was drinking heavily and complaining about how much he hated his neighbor.
• Fact E:  At 9:15 pm on Feb. 2, Defendant was seen entering a bar, drunk and weeping, two blocks from where the victim was run down.

I could go on, but you get the idea.  Each fact individually is specific but innocuous, whereas taken together they give rise to an inference of guilt.  Although the facts don’t always fit together so neatly as this, it’s the norm that guilt is shown by an accumulation of evidence, not by a single fact.  This goes equally for allegations of scienter in a securities case.  A conversation might suggest that the defendant knew something particular about the operational aspect of the alleged fraud, an e-mail might show he was concealing something from the accountant, inconsistencies between a public statement and a private statement might suggest he was dissembling, and so on.

That looking at the whole complaint is the only approach that makes sense does not imply that the Court was wasting space saying it.  Divide and conquer is the defendant’s strategy in these cases.  Defendants address each allegation individually, explaining why it doesn’t give rise to a strong inference.  And they’re right:  merely alleging that the fraud was very large doesn’t raise a strong inference of scienter, merely alleging a SOx certification doesn’t support a strong inference, merely alleging that the defendant was present at such and such a meeting doesn’t support a strong inference, and so on.  The plaintiffs’ response is, “But we didn’t allege merely A or merely B or merely C.  We alleged A and B and C “!  For one reason or another, judges do sometimes seem to follow the defendant on this, although not expressly, perhaps in part because it’s easier to structure an opinion in this way.

Given the importance of this issue to the actual litigation of securities fraud cases, it’s not surprising that, even before Tellabs, most of the U.S. Courts of Appeals, including the Ninth Circuit, had expressly adopted the rule that a the court must look to the complaint as a whole in determining whether the plaintiff has satisfied  the PSLRA pleading requirement.  See, e.g., Gompper v. VISX, Inc., 298 F.3d 893, 896 (9th Cir. 2002) (“Under the PSLRA, the court ultimately reviews the complaint in its entirety to determine whether the totality of facts and inferences demonstrate a strong inference of scienter.”). Six such courts were identified in 2004 by the Eleventh Circuit, which joined them, summarizing the state of the law as follows:

Nothing in th[e] language [of the PSLRA] suggests that scienter may only be inferred from individual facts, each of which alone gives rise to a strong inference of scienter, rather than from an aggregation of particularized facts. We readily join the courts that have interpreted the PSLRA to permit the aggregation of facts to infer scienter. See  Broudo v. Dura Pharms., Inc., 339 F.3d 933, 940 (9th Cir.2003) (“This court has made clear that allegations of scienter must be collectively considered.”); In Re Cabletron Sys., 311 F.3d 11, 39 (1st Cir.2002) (“ ‘The plaintiff may combine various facts and circumstances indicating fraudulent intent’ … to satisfy the scienter requirement.”) (quoting Aldridge v. A.T. Cross Corp., 284 F.3d 72, 82 (1st Cir.2002)); Abrams v. Baker Hughes, Inc., 292 F.3d 424, 431 (5th Cir.2002) (“The appropriate analysis … is to consider whether all facts and circumstances ‘taken together’ are sufficient to support the necessary strong inference of scienter on the part of the plaintiffs.”); Fla. State Bd. of Admin. v. Green Tree Fin. Corp., 270 F.3d 645, 660 (8th Cir.2001) (“[U]nder the Reform Act, a securities fraud case cannot survive unless its allegations collectively add up to a strong inference of the required state of mind.”); Rothman v. Gregor, 220 F.3d 81, 92 (2nd Cir.2000) (“Taken together with the allegations of poor sales and the pleadings in various lawsuits filed by GT, the Appellants have alleged sufficient facts to support a strong inference of recklessness.”) . . . .

Phillips v. Scientific-Atlanta, Inc., 374 F.3d 1015, 1016-17 (11th Cir. 2004).  At least two more circuits had joined the group prior to Tellabs.  See Teachers’ Retirement System v. Hunter, 477 F.3d 162, 174 (4th Cir. 2007) (“Determining whether the complaint satisfies this standard necessarily entails a case-by-case assessment of the complaint as a whole.”); Fidel v. Farley, 392 F.3d 220, 233 (6th Cir. 2004) (“[T]his court employs a “totality of the circumstances analysis whereby the facts argued collectively must give rise to a strong inference of at least recklessness.” (internal quotation marks omitted)). As far as I know, no circuit has come out the opposite way, ruling that a plaintiff can satisfy the PSLRA standard only by alleging at least one fact that, all by itself, gives rise to a strong inference of scienter.

The Ninth Circuit, however, in two opinions authored by Judge Bybee, interpreted Tellabs‘ repetition of this well established principle as a substantial reformulation of the analysis under the PSLRA.  As a result of Tellabs, the court held in Rubke v. Capitol Bancorp, Inc., “we can no longer summarily dismiss a complaint whose individual allegations are insufficient under the PSLRA.”  551 F.3d 1156, 1165 (9th Cir. 2009).  Thus, according to Zucco Partners, LLC v. Digimarc Corp., Tellabs “calls into question” the Circuit’s prior “methodology that relies exclusively on a segmented analysis of scienter.” – F.3d –, No. 06-35758, 2009 WL 311070, at *6, (9th Cir. Feb. 10, 2009), amending 552 F.3d 981 (9th Cir. 2009).

To be fair, an intervening case, South Ferry LP, No. 2 v. Killinger, did rely on the “holistically” language in Tellabs to modify somewhat the Circuit’s analysis under the PSLRA. 542 F.3d 776, 785-86 (9th Cir. 2008). South Ferry, however, merely broadened the Circuit’s acceptance of “vague” allegations as support for specific allegations elsewhere in the complaint. 542 F.3d 776, 785-86 (9th Cir. 2008). Rubke and Zucco went much further, suggesting that the prior approach of the Circuit was simply to consider allegations in isolation.

Most perplexing, Rubke and Zucco imposed a new two-step analysis of scienter within the Circuit:

Thus, following Tellabs, we will conduct a dual inquiry: first, we will determine whether any of the plaintiff’s allegations, standing alone, are sufficient to create a strong inference of scienter; second, if no individual allegations are sufficient, we will conduct a “holistic” review of the same allegations to determine whether the insufficient allegations combine to create a strong inference of intentional conduct or deliberate recklessness.

Zucco, 2009 WL 311070, at *6; see also Rubke, 551 F.3d at 1165 (“[W]e must perform a second holistic analysis to determine whether the complaint contains an inference of scienter that is greater than the sum of its parts.”).  Contrary to the court’s assumption, it is the first of these two stages that is the innovation, not the second.  And, indeed, the first stage is without purpose, because allegations are generally meaningless out of context (and, in any event, are never more meaningful out of context than they are in context).

It will be interesting to see whether this pointless two-step analysis becomes the standard method for assessing scienter allegations within the Circuit.  While these decisions are binding on subsequent panels, there are other post-Tellabs decisions that do not apply this method.  Arguably there is therefore a conflict within the Circuit, which is one criterion for hearing a case en banc.  However, because this method doesn’t have any clear benefit to either side in these cases (or at least I don’t see any), it is unlikely to attract enough interest among the active judges to go en banc.  The issue is likely to remain unresolved, therefore, unless and until it arises in a case that is heard en banc for another reason.  In the meantime, panels may rely on the conflict as a basis for making their own choice as to whether to follow a two-step method.  Alternatively they may view a methodological issue of this nature as dictum (a term that not everyone defines in the same way) and therefore outside of the binding aspects of these decisions.

Readers (if you’re out there and have lasted this long), may be interested in Lyle Roberts’s post on this at the 10b-5 Daily.  I’m a regular reader and generally find his analysis astute.  I think he misses on this one, though.

Plaintiffs in private securities fraud litigation frequently point to insider trading as support for a “strong inference” that the defendant acted with scienter, as required to survive a motion to dismiss under the Private Securities Litigation Reform Act of 1995.  The defendant sold her stock, the plaintiff argues, because she knew that the price was inflated as a result of false information in the market.  Generally this argument works only if the trading was somehow suspicious, such as where the defendant sold an especially large percentage of her holdings. 

Last August, in Metzler Investment GMBH v. Corinthian Colleges, Inc., 540 F.3d 1049 (9th Cir. 2008), the Ninth Circuit rejected a plaintiff’s insider trading argument as to one of three defendants, in part on the ground that the other defendants’ sales weren’t suspicious.  Id. at 1067 (“Moore sold only 37% of his total stock holdings during the Class Period. We typically require larger sales amounts—and corroborative sales by other defendants—to allow insider trading to support scienter.” (emphasis added)).  One of the other defendants had sold no stock during the relevant period, “suggesting,” the court reasoned, “that there was no insider information from which to benefit.”  Id. The court cited two cases in support of this inference, neither of which does support it, although they do generally support the other point the court made in the same sentence.  See In re Silicon Graphics Inc. Sec. Litig., 183 F.3d 970, 987-88 (9th Cir. 1999); Tripp v. Indymac Fin., Inc., 2007 WL 4591930 at *4-5 (C.D. Cal. Nov.29, 2007).

Regardless of support, though, can this be the right rule—that if one defendant’s sales were not suspicious, then sales by a different defendant weren’t either?  There is some sense to it.  An insider’s failure to capitalize on the fraud through insider trading undermines allegations of his scienter.  And if one defendant didn’t know, then it’s certainly more likely that another didn’t, either.  While this sort of group-scienter argument is not available to plaintiffs, no one ever claimed that the PSLRA was intended to create reciprocal benefits to plaintiffs and defendants.

My view, though, is that without some affirmative reason for thinking that defendant B wouldn’t have known unless defendant A did, too, it’s an unreasonable inference in favor of the defendant.  (Unlike on most motions to dismiss, the court is supposed to consider defendant-favoring inferences.)  The same principle, moreover, could as easily be applied to any allegations bearing on scienter, not just to insider trading.  If it were applied broadly, it could affect plaintiffs’ decisions about whom to name as individual defendants.  I suppose it’s an open question whether the Metzler court would have considered in defendants’ favor the failure of the plaintiffs to allege insider trading by insiders who were not named—and/or documents offered by the defendants at the pleading stage to show the innocence of unnamed insiders’ trades.  Either way, a plaintiff could lower the likelihood that insider A’s comparatively innocent behavior would rub off on insider B by not naming A.  That a plaintiff might leave a less guilty-looking defendant off the complaint for this reason could be viewed as either a positive or a negative consequence of the rule, depending probably on how one views the balance of interests in the current system of enforcement.

I used to be a plaintiff’s lawyer, which presumably shapes my perspective on this to some degree at least.  Also, I’m an admirer of Judge Betty Fletcher, who authored the Metzler court’s opinion, and usually agree with her.  For these reasons, among others, I’m quite open to the possibility that I’m not weighing this right.  If anyone with a view on this is reading, please share it.

(For more on Metzler, see the Corporate Securities Blog and the 10b-5 Daily.)