Securities Litigation Should Not Be Based on Junk Science
The fraud-on-the-market presumption has made federal securities litigation a hotbed of junk science about capital market efficiency. This state of affairs traces back to the work of a single district court in the 1989 decision Cammer v. Bloom, a case that continues to be influential in securities cases 30 years later. Just two weeks ago, for example, the court in Signet Jewelers Limited Securities Litigation applied the Cammer factors on its own initiative in a case where no party challenged market efficiency.
When the United States Supreme Court endorsed the fraud-on-the-market presumption in its 1988 decision, Basic v. Levinson, it left district courts in a bit of a quandary. The fraud-on-the-market presumption “says that all traders who purchase stock in an efficient market are presumed to have relied on the accuracy of a company’s public statements,” but the high court was unclear about how the lower courts were to determine when a stock was purchased “in an efficient market” for purposes of the fraud-on-the-market presumption.
The Cammer court was the first to set out a list of facts that, it asserted, would indicate an efficient market for the security at issue. Key facts to allege were:
- Large trading volume;
- A significant number of reports by securities analysts;
- The presence of market-makers and arbitrageurs in the security;
- Eligibility of the issuer to file an S-3 registration statement; and
- A history of immediate stock price movements in response to unexpected corporate events and financial releases.
After these so-called Cammer factors came into play, another federal district court added three more factors: (1) the capitalization of the company; (2) the bid-ask spread of the stock; and (3) the percentage of stock not held by insiders. The Cammer factors have become ubiquitous in federal securities fraud litigation.
In a new research article, I explain why the Cammer factors are junk science. Put simply, none of the Cammer factors (or the additional factors added since) help establish whether the market for a security at issue is efficient in the sense of financial economics.
Consider the Cammer “fifth factor”: examining the history of price reactions or non-reactions to news for statistical significance. The fifth Cammer factor is a history of immediate movement of the stock price caused by unexpected corporate events or financial releases:
Finally, it would be helpful to a plaintiff seeking to allege an efficient market to allege empirical facts showing a cause and effect relationship between unexpected corporate events or financial releases and an immediate response in the stock price. This, after all, is the essence of an efficient market and the foundation for the fraud on the market theory.
In practice, the application of this factor has meant the generation of statistically-significant price reactions, that is, identifying events and then concluding that the market is inefficient if the resulting price reaction is not statistically significant. This test assumes that prices are not efficient unless price reactions are large enough to be statistically significant.
But that assumption reflects a serious misunderstanding of both efficiency and statistics. Prices can react efficiently to information even though the price reactions themselves are not so large in size as to approach statistical significance. Statistical significance is just a measure of the size of a reaction relative to its average and its variability, but nothing says that the corporate event must have a certain size to be efficiently reflected in the security’s price, especially since part of the information may already be in the price before the news released identified by the litigant’s expert.
Thus, while a statistically significant reaction to a firm-specific news event might be evidence that information was reflected in the price (absent confounding effects), the converse is not true — the failure of the price to react so extremely as to be two standard deviations from average does not establish that the market is inefficient. It may mean only that the correctly-sized value impact that occurred was less than 1.96 standard deviations from the mean. More courts are now refusing to fall for this bad argument.
As a result of the Cammer factors, there are court opinions in nine-figure securities cases that would fail a college freshman’s statistics exam on such reasoning — for example, In re American International Group Inc., a case where the court determined that defendants had rebutted the fraud-on-the-market presumption on certain dates because price moves were statistically significant only at the 10% level (size) but not the 5% level.
It used to be rare but refreshing when courts got it right. As courts have become more aware of the games litigants play in such situations, they have begun rejecting more frequently the junk science requirement of statistical significance proffered by (mainly defense) experts. Securities litigation should return to the commonsense approach of the earliest fraud-on-the-market cases: If a security trades in a free and open public market, then it is, in the sense of controlling Supreme Court precedent, “efficient,” and plaintiffs should be able to invoke the (rebuttable) fraud-on-the-market presumption.
The Cammer factors remain unhelpful in making this determination. What is important to the determination of a “free and open public market” is whether active investors — investors other than passive funds that will buy or sell according to index inclusion or the like — are able to buy and sell in the market, and whether there are substantial restrictions on participation, like substantial lockups of potential sellers or bans on short selling.
This approach is not anachronistic. Public markets that are free and open are necessarily subject to the profit-seeking of speculators — so-called arbitrageurs. An enormous amount of capital chases profit opportunities in the smallest of crevices, from life insurance policy acquisition to litigation funding to gold and silver coins to, of course, securities. The idea that a public market for a security is not subject to the scrutiny and trading of speculators is usually facially implausible, absent evidence of substantial limits on their trading.
The longevity of the Cammer factors reflects an interesting and more general phenomenon: Where courts come to misapply social science insights and methodology, it is sometimes because both lawyers and, as importantly, their retained experts, are repeat players in certain types of litigation. And because even fatally-flawed methodologies can deliver favorable outcomes to either side in a particular case, there is an incentive in a given single case not to challenge adequately the reliability of the methodology, even when its application in a given single case is detrimental to the lawyer’s client.
This is troubling, because it not only fails to give litigants the vigorous advocacy they deserve, but also facilitates the introduction and spread of junk science and attendant inaccuracy in adjudication. The Cammer factors have been a textbook example of this more general phenomenon, one that I am exploring in future research.
 Cammer v. Bloom , 711 F. Supp. 1264 (D.N.J. 1989).
 In re Signet Jewelers Limited Securities Litigation , No. 16CIV6728CMRWL, 2019 U.S. Dist. LEXIS 114695, at *31 (S.D.N.Y. July 10, 2019). The court stated: “And as to market efficiency — a factor that is most often disputed, but Defendants expressly do not dispute here — the Court looks to the Cammer and Krogman factors, the prevailing tests for market efficiency, which are so named after the district court cases expounding them.”
 Basic v. Levinson , 485 U.S. 224 (1988).
 Wal-Mart Stores Inc. v. Dukes , 564 U.S. 338, 352 n.6 (2011).
 Krogman v. Sterritt , 202 F.R.D. 467, 474 (N.D. Tex. 2001).
 Kill Cammer: Securities Litigation Without Junk Science, forthcoming, William & Mary Business Law Review, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3410446.
 Id. at 1287.
 See, for example, In re Petrobras Sec. Litig. , 862 F.3d 250, 278-79 (2d Cir. 2017) (“Notably, small sample sizes may limit statistical power, meaning that only very large-impact events will be detectable.”) (citing Alon Brav & J.B. Heaton, Event Studies in Securities Litigation: Low Power, Confounding Effects, and Bias, 93 Wash. U.L. Rev. 583 (2015))
 In re American International Group Inc. , 265 F.R.D. 157, 187 (S.D.N.Y. 2010), vacated and remanded, 689 F.3d 229 (2d Cir. 2012).
[This post originally appeared on Law360.com, July 23, 2019]