Saturday, October 11, 2014

Loss causation by hindsight

One of the most complex issues in Section 10(b) litigation concerns loss causation, i.e., the question whether the fraud ultimately resulted in a loss to the plaintiffs.

The reason loss causation is so complex is because companies rarely simply admit to wrongdoing, out of the blue.  Most of the time, the "truth" behind the fraud - whatever that truth may be - is revealed gradually or indirectly.   The first revelations concerning an accounting fraud, for example, might simply be a drop in earnings, as the company tries to "make up" for past premature revenue recognition without admitting to wrongdoing.  A company might announce a slowdown in product sales without ever admitting that it had previously lied about the product's features.  A key officer might resign without explanation.  And very often, the first rumblings of a problem come from the announcement of a government investigation - without any further details - that may or may not ultimately culminate in an enforcement action.

In response to any of these announcements, the company might experience a stock price drop, even though the market either is unaware of the possibility of fraud or uncertain as to whether a fraud exists and/or its scope.  In such situations, can the fraud be said to have "caused" a loss?

In a pair of decisions by the Fifth and Ninth Circuits, it appears that whether such early warning signals constitute "loss causation" depends very much on what happened later.

[More under the cut]

In Loos v. Immersion Corp., 2014 U.S. App. LEXIS 17813 (9th Cir. Aug. 7, 2014), Immersion announced several profitable quarters before it began to report losses.  Eventually, it announced an internal investigation into its own billing practices, and then it admitted that its financial statements could no longer be relied upon.  Each disclosure of losses, as well as the disclosure of the internal investigation, resulted in a stock price drop, but the disclosure that the financial statements were unreliable did not - presumably because by then, the market had already figured that out.

The plaintiffs filed a lawsuit alleging that Immersion had "cooked its books," and defined the class period to end upon the announcement of the internal investigation.  After the lawsuit was filed, Immersion formally announced the results of its internal investigation - it would in fact be required to restate its financials.  Again, the stock price barely reacted (hardly surprising, since Immersion had already announced the previous year that its financial statements were unreliable). 

The district court held, and the Ninth Circuit agreed, that neither the disclosure of the losses, nor the disclosure of the internal investigation, satisifed the plaintiffs' burden of showing loss causation, because none of these announcements necessarily revealed to the market that Immersion's financial statements were false.   The Ninth Circuit elaborated that an investigation, standing alone, only gives rise to "speculation" regarding fraud, which is not sufficient to show that the fraud "caused" a loss.  In so doing, the Ninth Circuit quoted Meyer v. Greene, 710 F.3d 1189 (11th Cir. 2013), that "the announcement of an investigation, 'standing alone and without any subsequent disclosure of actual wrongdoing, does not reveal to the market the pertinent truth of anything, and therefore does not qualify as a corrective disclosure.'"

The plaintiffs argued that in this case, there was subsequent disclosure of wrongdoing - the announcement that the financial statements were unreliable, and the restatement, both of which occurred after the class period.  But the Ninth Circuit held that this argument was waived, because the plaintiffs had not raised it below, nor had the plaintiffs discussed the impact of these disclosures in their complaint.

In Public Emples. Ret. Sys. of Miss. v. Amedisys, Inc., 2014 U.S. App. LEXIS 18894 (5th Cir. La. Oct. 2, 2014), the plaintiffs alleged that Amedisys committed Medicare fraud by pressuring employees to provide unnecessary home healthcare services, and that this underlying fraud rendered a number of the company's public statements false.  The plaintiffs claimed that losses were caused by a series of partial disclosures of the fraud.  These disclosures were: a report published by Citron Research raising questions about Amedisys's billing; the resignation of the COO and CIO; a Wall Street Journal article that analyzed Amedisys's data and concluded that Amedisys was "taking advantage of the Medicare reimbursement system"; and the announcement of investigations by the DOJ, SEC, and the Senate Finance Committee.  Each new disclosure was accompanied by a stock price drop.

The Fifth Circuit held that many of these disclosures, standing alone, would not satisfy the plaintiffs' burden to demonstrate loss causation - for example, the Citron report merely raised questions about Amedisys, but did not provide any definitive conclusions.  Nonetheless, the Fifth Circuit held that the complaint as a whole was sufficient to allege loss causation, because the pattern of disclosures gradually revealed the fraud to the market.

Taken together, these cases - and the earlier Meyer decision on which the Ninth Circuit relied - hold that whether a disclosure "revealed" the fraud is gauged not by what the market understood at the time of the disclosure, but is instead gauged by what disclosures come after.   So, for example, the announcement of an investigation does not "reveal" the fraud - unless, at some later time, there is an enforcement action, at which point, apparently, the earlier announcement becomes a revelation of fraud.  Fraud has not "caused" a loss if the company's stock price drops in reaction to a report raising suspicions about fraudulent billing - unless the report is later confirmed by subsequent disclosures, in which case, it has.

These holdings do not make sense on their own terms.  There is no way that a disclosure at Time 2 can change how the market understood, and reacted to, an earlier disclosure at Time 1.

The only way that these holdings can be understood is as an attempt to pack into loss causation a more substantive merits inquiry regarding the plaintiffs' claims.  I think these courts fear that if the announcement of an investigation - without a subsequent definitive revelation of wrongdoing - qualifies as loss causation, plaintiffs will find it too easy to bring meritless claims where there wasn't any wrongdoing at all.  Courts, I think, fear that there may be lots of reasons to launch an investigation that turns up empty, and they don't want companies that are taking reasonable steps to "clean house" to end up as targets of frivolous lawsuits.

The problem with this reasoning, of course, is that there are other elements to a Section 10(b) claim, and it is these other elements that are meant to protect against such a scenario.  Plaintiffs still have to show that there were false statements; they still have to show that the statements were material; and they still have to show that the defendants acted intentionally.  There is no need for loss causation to carry this entire burden, and forcing it to do so will only result in incoherence.

Ann Lipton | Permalink


You sure know how to wake me up ion a Sunday morning, Ann!

Without disagreeing with your general premise (that courts are placing too much of a burden on loss causation, as construed by the court decisions that you cite, in Section 10(b)/Rule 10b-5 private actions), I would note that the matter is more complex than you suggest at the end of your post. Yes, there are other elements of a Section 10(b)/Rule 10b-5 claim that may in certain cases carry this weight. But some of these others elements also have significant defects as sorting mechanisms for frivolous litigation. My favorites, as many who know me would guess, are materiality (for claims emanating from alleged misstatements and omissions) and scienter. The analysis under these elements also can involve, in many factual scenarios, very difficult analyses at the motion to dismiss and summary judgment phases of a case. Loss causation analyses seem a lot less squirrelly—easier to analyze with hard, cold facts—than materiality and scienter. So, perhaps that’s why courts are working with loss causation.

Again, this does not discredit your observation in the slightest. In fact, it is quite important that you are raising the issue. But maybe we ought to be fixing—better concretizing—the elements of proof needed under each of these elusive elements in order to “fix” Section 10(b)/Rule 10b-5 litigation, since litigants and courts will always look for the next place to push on uncertainties for the benefit of their clients. In other words, the problem you identify is really just a piece of a bigger mess that we need to clean up. Agreed?

Posted by: joanheminway | Oct 12, 2014 6:09:51 AM

"In other words, the problem you identify is really just a piece of a bigger mess that we need to clean up. Agreed?"

Yes and no? I think there are a lot of really unsettled boundaries regarding 10b, that have a lot to do with what its purpose is as a cause of action. But I think that given the PSLRA's pleading burdens - which are applied by courts with increasing strictness - scienter and falsity in particular do more than eliminate frivolous litigation, they also eliminate lots of meritorious litigation. And at this point, loss causation has become a crapshoot - when is a revelation "speculative" and therefore doesn't count, versus certain? Moreover, why should we say fraud hasn't "caused" a loss just because the company reports a decrease in earnings (that is in fact due to an underlying fraud) but attributes that loss to some other factor, so that the market does not yet realize there was fraud? In several circuits, that is the law - so long as the company declares bankruptcy before it admits to accounting fraud, there is no 10b liability. Which is actually the 6th Circuit's unpublished holding in the K-Mart litigation.

So basically - and hey, this is part of the thesis of my article that will be forthcoming in April! - my feeling is, 10b-5 is carrying an awful lot of weight in terms of enforcing very broad disclosure requirements, and courts keep trying to cabin it in fairly broad brush, unstructured ways.

Posted by: Ann Lipton | Oct 12, 2014 6:30:33 AM

Right. I think we do agree on general principles. And I am not suggesting that the cases you point to are correctly decided. You have done significantly more work on that issue than I.

Of course, causation is problematic in many other areas of law, too. Folks have been writing about it for years. Proximate cause seems to be what's at issue here. Standard approaches are "but for" and "substantial factor" tests. I am sure there are others . . . . I teach the AUSA v. E&Y case for its discussion of these issues, which I find helpful.

Does any of this general causation literature (and the case law that contributes to and uses it) help with the issue you raise? I would think that resolving the loss causation element in the context of the broader elements of a cause of action under Section 10(b)/Rule 10b-5 could benefit from a deep dive into that scholarship and decisional law. Perhaps you've already done that and found it unhelpful . . . ?

Posted by: joanheminway | Oct 12, 2014 8:34:20 AM

AUSA isn't a fraud on the market case, though, right? See, that's what makes it all so complicated - in a fraud on the market case, courts are very obsessed with disclosures and what they reveal, such that a stock price reaction can be taken to mean that artificial inflation has dissipated. My favorite example is the FDA case - company has a drug with a 50/50 shot of approval, but lies and makes it sound like chances of approval of 80/20. Stock price trades higher than it would have otherwise traded. Then the FDA rejects the drug, so the stock price drops - but since the chances of approval are now zero, the price drops a lot, well below what it would have been if the truth had been told initially. Is the denial of the drug application a "corrective disclosure"? It depends on what counts as "revealing" the truth, and the cases are all over the map. There's a clear circuit split but even within the circuits the courts are wildly inconsistent. There has been a lot of writing on the subject but it's really a big mess.

Posted by: Ann Lipton | Oct 12, 2014 9:12:53 AM

I know what you mean.

Nevertheless, I find myself reacting negatively to your example. It reminds me of a comment that a litigator friend of mine once made. He said (effectively, not literally): "I am always right in guesstimating chances of success in a case if I am not gauging them at 0% or 100% (since either outcome is possible under all probabilities in between)." I think you can see why I have concern about your example in that context . . . . I would need to know a lot more before I would be sympathetic to a plaintiff's claim in those circumstances.

Anyway, thanks for the great thoughts.

Posted by: joanheminway | Oct 12, 2014 10:46:42 AM

Post a comment