Price Impact in Securities Class Actions Post-Halliburton II

Jorge Baez and Dr. Renzo Comolli are Senior Consultants at NERA Economic Consulting. This post is based on a NERA publication authored by Mr. Baez and Dr. Comolli. Related research from the Program on Corporate Governance includes Rethinking Basic by Lucian Bebchuk and Allen Ferrell (discussed on the Forum here).

On July 25, 2015, the United States District Court for the Northern District of Texas issued the much-anticipated ruling on class certification in Erica P. John Fund, Inc. v. Halliburton Co. The economic analysis of price impact was front and center in the Court’s ruling.

This ruling follows the Supreme Court’s decision on price impact that is widely known as Halliburton II. Although this ruling involves facts that are unique to Halliburton’s particular disclosures, attorneys may look at it as a roadmap for guiding economic analysis of price impact in future cases in the post-Halliburton II world.

The case initially involved three categories of alleged fraud:

  • Alleged overstatement of cost savings from a merger;
  • Alleged accounting issues related to certain revenues; and
  • Alleged misrepresentations regarding asbestos liabilities.

According to Plaintiff the alleged truth was disclosed to the market by 13 separate alleged corrective disclosures.

Halliburton, has now reduced Plaintiff’s claim to one alleged corrective disclosure related solely to the allegations regarding asbestos liabilities.

In response to Halliburton and its expert’s price impact analysis, Plaintiff first narrowed their focus from 13 to six alleged corrective disclosures. Then, Halliburton, after a full-day evidentiary hearing before the Court, successfully defeated five of the six remaining corrective disclosures, leaving a single corrective disclosure.

The analysis by Lucy Allen, Halliburton’s expert, helped demonstrate to the Court that there was no price impact stemming from alleged corrective disclosures by using economic principles that generally apply to securities class action cases, as well as complex statistical analyses that may apply to certain cases. Some of the economic and statistical principles included:

  • Level of statistical significance: Statistical analysis can provide information on the likelihood that a price movement of the magnitude observed could be seen solely by chance (i.e., whether the price movement was statistically significant). The Court found that, in the analysis of a particular date, Plaintiff’s expert used a level of statistical significance less stringent than what is “necessary” and thus could not show that there was price impact on that date. [1]
  • Length of event window: Event windows are the periods over which stock price movements are analyzed. The Court concluded that, in this case, Plaintiff’s expert’s “use of a two-day window is inappropriate to measure price impact in an efficient market,” and thus did not support Plaintiff’s claim of price impact. [2]
  • Market efficiency and new information: The Court concluded that information that was previously disclosed to the market should not have been used in this case to show price impact in an efficient market. In particular, the Court found that on one alleged corrective disclosure, Halliburton “demonstrated that the disclosure that allegedly caused the price reaction was already disclosed to the market on July 25th [weeks before the alleged corrective disclosure], to no price reaction.” [3]
  • Internal inconsistencies in expert model: Halliburton argued that Plaintiff’s expert’s event study methodology was internally inconsistent. The Court was persuaded “that Allen’s adjustments to [Plaintiff’s expert’s] model were appropriate to achieve internal consistency.” [4] After making these internal consistency adjustments to Plaintiff’s expert’s model, several of the alleged corrective disclosures were no longer statistically significant and thus did not support Plaintiff’s claim of price impact on those dates.
  • Proper selection and testing of dates: The Court questioned the reliability of Plaintiff’s first expert’s method for selecting and testing dates. [5] According to Halliburton’s expert, Plaintiff’s expert inappropriately cherry-picked the alleged corrective disclosure dates by first finding dates that had statistically significant price reactions and then trying to find news that was related to the alleged misrepresentations.
  • Statistical adjustment for the number of tests performed (multiple comparisons adjustment): An issue arises when multiple price reactions are tested for statistical significance: the more price reactions that are tested, the greater the probability of finding statistical significance simply due to chance (i.e., when there is no company-specific news of importance to the market). This issue is known in statistics as “multiple comparisons.” The Court found “that the use of a multiple comparison adjustment is proper in this case because of the substantial number of comparisons.” [6] After making a multiple comparisons adjustment, several of the alleged corrective disclosures were no longer statistically significant.

With regard to the one remaining alleged corrective disclosure, Halliburton argued that there was no price impact because the information was not corrective. The Court held that it cannot make a determination at the class certification stage as to whether a disclosure is corrective (“the Court will not determine as a matter of law whether the verdict announcement was corrective”). [7] Thus, the Court concluded that Halliburton had not shown that there was no price impact as to this date and certified the class. The Court noted that “the burdens of production and persuasion to show lack of price impact are properly placed on [the defendant].” [8]

While legal questions raised by Halliburton II may well be subject to further judicial interpretation in the months and years to come, the District Court’s decision underscores the importance of sound economic analysis at class certification in the post-Halliburton II world.

Endnotes:

[1] Memorandum Opinion and Order, Case No. 3:02-CV-1152-M, 2015 WL 4522863 (N.D. Tex. July 25, 2015), at 32.
(go back)

[2] Id. at 29.
(go back)

[3] Id. at 37
(go back)

[4] Id. at 22.
(go back)

[5] Id. at 25.
(go back)

[6] Id. at 25.
(go back)

[7] Id. at 46.
(go back)

[8] Id. at 7.
(go back)

Both comments and trackbacks are currently closed.