This month, ExxonMobil filed two court memos regarding the New York attorney general’s (NY AG) expert witnesses in the trial for the lawsuit alleging ExxonMobil misled investors. Based on glaring flaws in the witnesses’ reasoning and methodologies, ExxonMobil sought to exclude their testimonies or have the court otherwise resolve major issues in the witnesses’ assertations. Even though New York Supreme Court Justice Barry Ostrager, who is overseeing the trial, denied ExxonMobil’s motion to exclude the witness’ testimonies, the serious inadequacies in the witnesses’ qualifications and arguments underscore the deep-seated weaknesses of the NY AG’s case.
The NY AG’s expert witnesses Eli Bartov, an accounting professor at New York University, and Peter Boukouzis, an Assistant Professor at University of Saint Katherine, were tasked with clarifying issues in ExxonMobil’s argument that require specialized professional or technical knowledge, in an effort to help advance the NY AG’s case. At least, that is what expert witnesses are supposed to do. But Bartov’s testimony is unreliable, lacks proper foundation, and is riddled with gaps in conclusions and methodologies. Similarly, Boukouzis also lacks reliable methodology, and on top of that, the professor has no relevant experience or background to serve as an expert witness.
Bartov’s Testimony is Unreliable, Disconnected from the Allegations, and Devoid of Generally Accepted Methodology
Eli Bartov asserts that ExxonMobil misrepresented data and that impacted the company’s stock price to a degree that was statistically significant. He bases this conclusion on research that he did at the request of the NY AG’s office, specifically on “GHG Emission Proxy Costs,” impairment testing and the climate change risks to ExxonMobil’s business. Bartov’s “GHG Emission Proxy Costs” ignores the distinction between the proxy cost of carbon and GHG costs; the proxy cost of carbon explores the impact of potential climate policies on future energy demand, while GHG costs estimate direct financial impact of emissions regulations on specific projects. The NY AG’s complaint also erroneously conflates the two measurements.
Bartov claims his research shows that the company inappropriately omitted “GHG Emission Proxy Costs” from its cost projections in its impairment testing on Mobile Bay in 2015, and this omission ultimately inflated the company’s stock price. But this conclusion does not stand up in the face of serious methodological flaws, process errors and general shoddy research tactics. Let’s break this down.
Bartov’s study fails to follow best practices in statistical analysis and cherry-picks data. Bartov refers to three instances that he alleges shows artificial inflation in ExxonMobil’s stock price. The findings in only one of these instances is within a 5 percent threshold of statistical significance; the other two findings have a 10 percent chance that his observations on the stock price impact are totally random. 5 percent is the accepted threshold for statistical significance across the board – precedent, academic literature, and even Bartov himself agree on the 5 percent standard. Courts typically throw out testimony that disregards the well-established limits of statistical significance.
On top of this, Bartov cherry-picks the dates he used in his study to find data that fit within his preconceived conclusion. Mobile Bay in 2015 was the only asset and time Bartov reviewed, and as ExxonMobil explains, “His analysis boils down to a search for dates with a stock price movement, which he retrofits to support a preconceived conclusion about price inflation.” The stock price impacts based on this process are therefore impermissible and unreliable.
Bartov fails to complete a corrective disclosure. In a corrective disclosure, new information not previously available to the market must correct alleged false information. But Bartov’s corrective disclosures only serve to announce the existence of an investigation into ExxonMobil’s climate change disclosures, relying on a government investigation announcement as proof of new information. The only “facts” that an investigation announcement reveals is that there is an investigation. Such an announcement could only have a negative – not corrective – effect on stock prices.
The investigation wasn’t even new news during the date range to which Bartov limited his study. Bartov uses a Los Angeles Times article reporting that the California AG was investigating ExxonMobil (fact check: it didn’t go anywhere), which itself relied on a series of articles published months before by major outlets including The New York Times, The Wall Street Journal, U.S. News & World Report, and Washington Post.
In a videotaped examination played for the court yesterday, Roger Read, a senior energy analyst at Wells Fargo, testified that he did not publish any comment or notification related to the California AG’s investigation. In other words, the California AG’s investigation did not impact Wells Fargo’s analysis or rating of the company. Read continued:
Q: Do you have any recollection of adjusting your stock rating for ExxonMobil as a result of the New York Attorney General’s investigation?
A: We did not.
Q: Do you have a recollection of adjusting your target price for ExxonMobil as a result of the New York Attorney General’s investigation?
A: We did not.
Q: Do you have a recollection of adjusting your stock rating price for ExxonMobil after media reports of a California Attorney General investigation of ExxonMobil?
A: We did not.
Q: Do you have a recollection of adjusting your target price for ExxonMobil after media reports of a California Attorney General Investigation?
A: We did not.
Bartov fails to follow best practices in his analysis of ExxonMobil’s impairment testing. When conducting an impairment test, there are three steps one must meet; and if step one is not satisfied, the testing does not proceed further. Bartov himself acknowledges this, but, even so, he also admits that he only assumed the crucial first step was satisfied. Bartov’s failure to follow the most basic rules in conducting his inquiry is inexcusable and mind-boggling, and calls into question his credibility in this case.
Data from ExxonMobil and its independent auditor, PriceWaterhouse Coopers (PwC), show that step one was actually not fulfilled; ExxonMobil states in its 2015 10-K that it found no trigger event in assessing Mobile Bay for impairment that would signal step one was fulfilled. PwC reviewed ExxonMobil’s impairment study and concluded the company was correct in its assessment. The very individual who developed and authored the guidance that governs impairment testing, Lisa MacDonald, has found several key flaws in Bartov’s reasoning, leading him to apply the guidance incorrectly. MacDonald will have the opportunity to speak in greater depth on these flaws during the trial, as she will appear as an expert witness for ExxonMobil.
Boukouzis’s Testimony is Unqualified and Unsupported
The NY AG seeks to use Peter Boukouzis’s testimony to establish that ExxonMobil’s reporting on proxy costs and greenhouse gas (GHG) costs were both misleading and material to investors. In reality ExxonMobil has found that his testimony is “nothing more than NYAG’s factual and legal argument masquerading as expert opinion,” featuring more narrative and his own thoughts and opinions than analysis backed by facts and data.
Boukouzis is not qualified to stand as an expert witness on this issue. Boukouzis is new to academia by his own admission – he began his assistant professorship barely a year ago at an institution with 200 students that was only accredited in 2016. He has never published an academic article on any subject at all. His decades of experience in mergers and acquisitions investment banking do not automatically provide him experience with climate regulations or climate change disclosures, especially considering that he did not advise on stock investment decisions nor did he work with equity analysts. He even admitted in his deposition that he is “’Not an expert in climate change policy’ or in greenhouse gas ‘regulations.’” The Court doesn’t even need an expert witness to investigate whether ExxonMobil’s representations on proxy costs and GHG costs in its published disclosures misled investors, the task Boukouzis was relegated; all he did was read disclosures that are already provided for the general public. “By definition, no special expertise is needed to determine how a reasonable investor would interpret a public disclosure,” ExxonMobil explains in its Court filing on why Boukouzis’s testimony should be excluded.
Boukouzis’s testimony lacks reliable methodology. Boukouzis asserts that climate change regulatory risks are “important” to investors, but provides only anecdotal evidence that summarizes out-of-court documents, including a haphazard array of shareholder proposals, analyst reports, and statements by activists. This means that over a third of his expert report simply summarizes the contents of reports that the Court is able to read itself.
And he engaged in even more cherry-picking, relying on a keyword search of generic terms, including “climate change” and “GHG emissions” to identify 99 analyst reports he argued proved that the companies disclosures concerning proxy costs and GHG costs were “relevant” to investors, instead of using a random sample to determine if the keywords were present in a significant number of reports. He did not describe how he identified the collection of reports through which to search.
On top of that, he failed to define what constitutes relevancy, saying in his deposition, “I don’t remember exactly what triggered relevance versus irrelevance,” explaining that it was inconsequential whether a report only included a keyword buried in a footnote or if it discussed climate change as a broad issue.
His dataset lacked a statistical analysis – another glaring omission – so ExxonMobil expert witness Dr. Marc Zenner ran one of his own on 500 analyst reports. Zenner found that only 6% contained key terms “emission,” “greenhouse,” “GHG,” or “climate change.”
Conclusion
A trial court has the duty to confirm that expert witnesses’ methodology is scientifically reliable and that the evidence is admissible. But ExxonMobil’s lawyers have shown that both Bartov and Boukouzis are unqualified, unreliable, and lack sound methodology. As Justice Ostrager has nevertheless permitted their participation, we anticipate the trial will involve even more discussion on the weaknesses of their testimonies.
