Saturday, June 28, 2014

Halliburton II – An Unexpected Gift to Plaintiffs

So, Halliburton Co. v. Erica P. John Fund, Inc. (2014) (“Halliburton II”)  came down, and to call it a change in the law is too generous – at best, it might qualify as a “clarification.”  After all of the angst  over the possibility that the Court might give plaintiffs the burden of proving the price impact of a particular misstatement, the Court soundly rejected that argument, reaffirmed Basic, Inc. v. Levinson (1988), and instead merely allowed defendants to rebut the fraud on the market presumption.  Because demonstrating a lack of price impact is as difficult as showing price impact in the first place, I don’t expect Halliburton II to change much in existing law – if anything, some of the rhetoric may make matters easier for plaintiffs.

[More under the cut]

As I’ve previously explained, fraud on the market theory – which the Supreme Court reaffirmed in Halliburton II – is actually two separate presumptions in the plaintiffs’ favor.  The first is that in an open and well-developed market, material information affects stock prices.  Therefore, if the defendant issues false material information, the relevant stock prices will be distorted.  The second presumption is that investors who trade in such a market do so with subjective reliance on the market price.  Together, the two presumptions satisfy the element of reliance in a 10b-5 fraud claim.  If the plaintiff subjectively trusts the market to supply her with the “right” price, and the price is distorted by the defendant’s fraud, the plaintiff has “relied” on the defendant's fraud.

In Halliburton II, the defendants mounted a number of challenges to the fraud on the market presumptions.  Among other things, the defendants argued that the market simply is not “efficient enough” to justify the presumptions.  As a fallback, the defendants claimed that, at minimum, the plaintiffs should have the initial burden of demonstrating that a particular misstatement affected stock prices, rather than obtaining the benefit of a presumption to that effect.

As I previously posted, this would have been an enormous challenge for plaintiffs.  This is because most of the time, a false statement isn’t designed to push stock prices up, but to keep stock prices level – or even to slow their descent.  The defendants might falsely claim, for example, that earnings met analyst expectations for the quarter, or came in just slightly under expectations when in fact there had been a precipitous drop.  In such circumstances, there will not be any detectable price movement in response to the lie.

Usually, then, plaintiffs might try to show “price impact” by demonstrating that the market reacted to disclosure of the truth.  The problem with that, however, is that a disclosure at Time 2 is never quite the same thing as a misstatement at Time 1.  For example, if a company claims to comply with federal safety standards, and then there’s an explosion that destroys one of its plants, what does the resulting stock price drop tell you about how the market treated the original misstatements about safety?  Nothing, really – whether or not the market reacted to the original misstatements, you’d still expect to see a drop upon catastrophic news.

But the Supreme Court didn’t go that route.  Instead, the Court simply held that the defendants may rebut the presumption of price impact.  This possibility was raised in Basic itself, which explicitly held that the presumption might be rebutted if the defendants showed that, for example, the truth was known, or became known, to a sufficient number of market actors.  Basic, however, suggested that this rebuttal showing might only be appropriate at the trial stage, leaving some degree of uncertainty about whether rebuttal evidence could be offered at class certification.

Despite that uncertainty, however, two circuits – the Second and the Third – both definitively held years ago that defendants had the right to rebut price impact at class certification.  See In re Salomon Analyst Metromedia Litig., 544 F.3d 474 (2d Cir. 2008); In re DVI, Inc. Sec. Litig., 639 F.3d 623 (3d Cir. 2011).  No circuit, until the Fifth Circuit in Halliburton II, had ever explicitly held to the contrary.  This is why I call the Supreme Court’s holding a “clarification” rather than a change in law.

That said, defendants have a heavy burden here, for the same reason that the plaintiffs would have had a heavy burden if the Supreme Court had forced them to prove price impact in the first instance.  There might not be any stock price movement when the initial lie is uttered, but that only tells you that the market was expecting the news – not that the news had no impact on prices.  Or, to put it another way, had the defendants told the truth (earnings were below expectations) or remained silent (failed to issue a required earnings report), the stock price would have fallen. The lie kept the stock price level.

If defendants cannot prove a lack of price impact by showing that prices remained steady when the lie was first uttered,they may instead try to prove lack of price impact by showing lack of price movement upon disclosure of the truth.  The problem is that the defendants then run smack into Erica P. John Fund, Inc. v. Halliburton Co. (2011) (“Halliburton I”) which laid out all the reasons why a lack of loss causation has nothing to do with whether the stock price was inflated initially.  Maybe, for example, the truth leaked out in some way that caused the stock price to fall in gradual, undetectable increments.  Maybe some horrible catastrophe intervened and brought the stock price so low that new information had no further effect.   All of these possibilities bear on plaintiffs’ ability to satisfy the element of loss causation in a 10b-5 action, but they tell you nothing about whether the price was impacted initially.

It is for this reason that in the past, defendants have not usually fought for the right to rebut price impact per se – instead, they have tried to get courts to accept proxies for price impact that are easier to prove, namely, loss causation and materiality.  But the Supreme Court rejected those proxies in Halliburton I and Amgen Inc. v. Conn. Ret. Plans & Trust Funds (2013).

What does make Halliburton II interesting, however, was the manner in which the Court described market efficiency.

One ongoing debate in 10b-5 litigation concerns how efficient is efficient.  Plaintiffs are only permitted the benefit of the fraud on the market presumptions if they can first show that the market was open and well-developed – sometimes described by courts as “efficient,” even though the concept of efficiency in the legal context bears little resemblance to how that term is used by economists.

Courts have developed a rather strict test for 10b-5 efficiency, generally requiring the plaintiffs to show that the market rapidly and fully reflects available information.  How rapidly?  What is fully?  There are no bright line rules, but courts have often wanted to see information fully incorporated into stock prices within hours of its announcement. 

The problem, of course, is that no market is perfect – which is exactly the argument that the Halliburton defendants made.  Studies show, for example, that new information may take weeks to be fully impounded into a stock’s price, even for stocks that are widely followed and traded.

As a result, scholars have often argued that courts’ tests for efficiency are not only too open-ended, but they are also too demanding.  So, maybe a stock shows abnormal returns for a prolonged period of time after new information is announced, suggesting that it is continuing to react to the information – so what?  Why should that mean that a plaintiff whose purchase was only partly affected by the false information – because the information was not yet fully reflected in the stock price – should be denied any recovery at all?  Doesn’t it make more sense to sort out the precise degree to which prices were affected in the damages phase, rather than in the class certification phase?  See, e.g., Donald Langevoort, Basic at Twenty: Rethinking Fraud on the Market, Wis. L. Rev. 151 (2009); Bradford Cornell & James Rutten, Market Efficiency, Crashes, and Securities Litigation, 81 Tul. L. Rev. 443 (2006).

Interestingly enough, the Supreme Court’s language in Halliburton II opens a path to a looser definition of efficiency that more closely comports with the relevant legal question, i.e., whether the market has characteristics that make it reasonable to assume that false statements affect stock prices in some way.  In his opinion for the Court, Chief Justice Roberts described the Basic presumption as “modest,” and emphasized that the presumption does not depend economists’ definitions of efficiency, or on a theory of how quickly or completely market prices are affected by material information.  He even recognized that an investor “relies” on market efficiency when she purchases stock in the belief that the market has failed to recognize its true value – because she is necessarily expecting that the market will correct itself “within a reasonable period.”  This is an extremely generous concept of “market efficiency,” apparently allowing for even a substantial delay between the issuance of information and its reflection in stock prices.  If lower courts read the Court’s opinion this way, plaintiffs may find that they have an easier time proving efficiency in borderline cases.

Ann Lipton | Permalink


Thanks for this post, Ann. You've given us lots to chew on here.

You ask: "Why should that mean that a plaintiff whose purchase was only partly affected by the false information – because the information was not yet fully reflected in the stock price – should be denied any recovery at all?" I assume you mean to refer here to a person who cannot prove reliance, since that always is an option. Of course, it's also possible for a court to deny class certification in a class action because of differences in the plaintiffs' claims (meaning that more distinct classes of plaintiffs may need to be developed or that some plaintiffs would have to bring individual actions). So, it's not clear to me that timing differences result in a failure of plaintiffs to recover.

Leaving aside issues with the application of the teachings of the ECMH to a single firm's stock, my experience is that a firm's stock price reacts quite quickly to the release of material information released fully and on its own. But even assuming my intuition here is correct, there are at least four additional issues to contend with. First, information may not be released fully (i.e., all at once). Second, information may not be (most often is not) released alone (without the release of other information at the same time). Next, the observation depends on the materiality of information, which is a separate element of a Section 10(b)/Rule 10b-5 claim. Finally, market prices also move on immaterial information (but I guess that's OK since materiality must be separately pleaded and proven).

Moreover, omissions cases--where there is no lie from the outset, just a failure to speak when there is a duty to speak--raise additional questions about market price/timing analyses. And, imho, the Affiliate Ute reliance presumption does not resolve all of these issues. You may disagree.

Anyway, you address some of this in your post, but I am wondering if you have any additional observations about any of those factors . . . ?

Posted by: Joan Heminway | Jun 29, 2014 7:25:49 AM

Hi, Joan! Glad the post interested you!

As for your questions:

For your first question, I'm assuming that we're talking about a 10b-5 class action based on fraud on the market, in a situation where the plaintiffs have trouble meeting the efficiency standards currently demanded by most courts. I.e., the market reacts to new information, but the reaction is not necessarily immediate and complete - like, maybe the market continues to react to something new after a day or two, rather than an hour or two. In that situation, a lot of courts might deny certification - the question is, why? There would be differences among plaintiffs - some might have more distortion from a false statement, and thus more damages - but assuming the plaintiffs could tease that out at the damages phase of litigation, there's no reason why the fact that the fraudulent statement was not immediately and fully incorporated into the stock price should be a basis for denying certification. My point is, courts don't really have a standard of how efficient is efficient "enough" and as such, tend to demand very rapid and complete incorporation of new information. They might deny certification for bonds, for example, even when it's clear that the bond prices are reacting within days to new information. The current standard could be loosened, with appropriate adjustments made in the damages phase. Right now, courts aren't saying "well, there are too many differences among the plaintiffs for less-than-perfect markets" -they don't get that far. They simply say Basic doesn't apply if the market doesn't meet some high standard of efficiency, without tethering the high standard to the purposes that Basic is meant to serve, i.e., creating a reasonable presumption that false information will have some effect on prices.

As to your second point, yes, those are always issues in 10b-5 class actions. Typically, the plaintiff brings on an expert who calculates the degree to which each new release of information further inflated/affected stock prices, and who controls for information unrelated to the fraud. This means there's a formula - i.e., people who bought between dates x and y, and who held through the release of truthful information on date z, get such-and-such damages. And the formula is adjusted throughout the class period. Of course, the plaintiffs' expert battles it out with the defense expert - recently, for example, in the BP case, the court partly denied certification because it didn't believe the plaintiffs' expert had sufficiently shown that damages could be calculated for the class. But my argument is really simpler - the plaintiffs will always have the burden of teasing out confounding factors and determining varying levels of inflation throughout the class. But right now, before even asking whether that is possible, courts will deny certification if the plaintiffs can't meet some artificial - and somewhat undefined - level of efficiency. They shouldn't; they should just ask when the market reacts quickly enough to make the presumption of price impact a reasonable one.

I don't believe that omissions cases are really any different from misstatement cases in the context of fraud on the market. Typically, there's going to be a disclosure at the end of the class period, to which the stock price will react. Using that disclosure, and controlling for unrelated factors, the plaintiffs will try to show that the stock price at the end of the class period represents the "true" value of the stock at an earlier point in time, had the truthful information been disclosed. The issue is going to be how precisely to control for unrelated factors, but I don't think the mechanics of it are any different - the nub will be establishing that there was a duty to disclose in the first place.

Oh I should add - I had an earlier post on Affiliated Ute, but basically, I don't believe it does any work for impersonal market transactions - i.e., I believe it's only properly used where the plaintiffs directly received the relevant materials and allege they have read them. [] So in an impersonal market, omissions cases should play out like misstatement cases - the plaintiffs will first have to show that the market was sufficiently efficient to justify a presumption that information affects prices (and that the omitted information would have affected prices), etc.

Hope that answers your questions!

Posted by: Ann Lipton | Jun 29, 2014 8:06:41 AM

Post a comment