A Case for Statistical Evidence in Insider Trading Claims
POMERANTZ MONITOR | MAY JUNE 2022
Insider trading is tough to prove, and it can take years to gather enough evidence to successfully prosecute it. Preet Bharara, the former U.S. Attorney for the Southern District of New York, has called insider trading, undertaken by company insiders and hedge funds, “rampant.” The SEC has developed data analytics tools to aid the investigation of insider trading cases with statistical evidence. However, in SEC v. Clark and Wright, the court rejected the SEC’s statistical evidence as “just a matter of speculation.” On occasion, evidence of insider trading is used to determine whether scienter exists in Rule 10b-5 or securities fraud cases. This article will describe the value of statistical evidence and explain what steps plaintiffs’ lawyers can take to leverage such evidence in prosecuting insider trading claims.
What is Insider Trading?
Insider trading involves trading in a public company’s stock by someone who has material nonpublic information (“MNPI”) about that stock. MNPI is data relating to a company that has not been made public but would have an impact on its share price when made public. Insider trading is often carried out by a group of people rather than by an individual acting alone. Frequently, the person who receives inside information from within an organization (“a tipper”) does not trade on that information, but rather passes information on to others (“tippees”), who ultimately trade based on that information. This type of insider trading is illegal because insiders cannot “misappropriate” information for their own benefit.
Investors may be deterred from participating in the market if they know that others are trading on nonpublic information. However, insider trading violations are difficult both to prosecute and to prove. Often, defendants in insider trading cases will deny their awareness of MNPI at the time of a securities trade, or defendants will claim that their reason for trading was completely unrelated to the information in their possession. Because insider trading involves secret information and communications, it is rare to find a smoking gun proving that a trader was tipped and by whom (i.e., what they knew, when they knew it, and how they found out). Therefore, the government and the plaintiffs’ bar often rely––at least initially––on circumstantial evidence to draw the strong inference that the defendant was aware of MNPI and used that information for personal profit when trading.
SEC v. Clark and Wright and the SEC’s Data Analytics Tools
To aid the investigation of insider trading cases, the SEC has developed numerous data analytics tools for analyzing massive amounts of data to identify suspicious trading, such as improbably successful trading across different securities over time. In SEC v. Clark and Wright, the SEC argued that the trades in question were suspicious because they had “an improbable success rate.” The SEC alleged that Christopher Clark and his brother-in-law, William Wright, the former controller for CEB, Inc., engaged in insider trading in advance of CEB’s acquisition. Based on the information tipped by Wright, Clark allegedly purchased highly speculative, out-of-the money call options. The SEC further alleged that, after the public announcement of the acquisition of CEB for $2.6 billion, Clark liquidated his CEB options and made a profit of over $240,000. The cornerstone of the SEC’s case was “suspicious trading”: Clark and his son purchased CEB’s highly risky options before the merger announcement after Clark maxed out his family’s credit line, took out a loan on his car and liquidated his wife’s IRA account to finance these trades. According to the SEC, Clark and his son “were the only investors in the entire world willing to buy such risky options” and accounted for 100% of the buy-side volume on the days in question. The SEC took the position that Clark’s too-good-to-be-true trades, combined with the fact that Clark borrowed money to make the trades, spent ample time with Wright, and traded after communicating with Wright, clearly pointed to insider trading. Instead of initially presenting testimony or other direct evidence, which is hard to obtain at the outset of insider trading cases, the SEC’s case largely hinged on its statistical surveillance tools’ identification of the trades as “highly suspicious.
Wright settled with the SEC in October 2021. Clark proceeded to trial. Courts have repeatedly held that evidence of suspicious trading that coincides with communications between the alleged tippee and tipper should go to the jury. Judge Hilton, however, ended the SEC’s trial against Clark without hearing Clark’s arguments or allowing the jury to weigh in, finding that the SEC’s statistical evidence was “just a matter of speculation” and that the “improbable success rate” of Clark’s trades was not evidence of anything at all: “the government can speculate that he made a little too much money, he was a little successful or more successful than he ought to be, so therefore he’s getting insider information, but there’s no evidence of it.” Judge Hilton added that “[t]here’s just simply no circumstantial evidence here that gives rise to an inference that he received the insider information.”
The Role of Statistical Evidence in Civil Cases
Is there a qualitative distinction between statistical and non-statistical evidence? According to Judge Posner, “[t]he probabilities that are derived from statistical studies are no less reliable in general than the probabilities that are derived from direct observation, from intuition, or from case studies of a single person or event.” The ambiguity of “statistical” evidence does not differ in kind from the ambiguity of “non-statistical” evidence. The “real problem” of statistical evidence is not the explicit characteristics of the evidence itself; it is instead the interpretation given to that evidence. At its core, evidence takes on meaning for trials only through the process of being considered by an individual. And Judge Hilton’s interpretation of the SEC’s statistical evidence of “suspicious trading” could have been very different from the jury’s interpretation of the same evidence. Moreover, while the use of statistical analysis to identify insider trading is novel, the use of statistical analysis in other fields to provide legal proof is not. Many courts have permitted proof of causation through statistically-based evidence in toxic tort cases, even when the evidence is thin and attenuated, and stronger and better evidence is unavailable. The courts and regulators also rely on statistical analysis to help prove systemic employment discrimination and the efficacy of treatments in clinical trials. This is because the approach employed in statistical analysis of arriving at a conclusion by ruling out plausible alternative explanations is consistent with judicial fact-finding.
The SEC has been very successful in litigating insider trading claims solely on the basis of statistical evidence. For example, in SEC v. Ieremenko, et. al., a case brought against a hacker and several traders who traded on nonpublic information stolen from the SEC’s EDGAR database, the SEC successfully argued that the defendants’ trading was correlated with the EDGAR hacks: “[i]t is virtually impossible that [the suspicious trading] could have occurred by random chance. Statistical analysis shows that for each of the Trader Defendants, the odds of that trader trading so disproportionately in hacked events by random chance ranged from less than 7 in 10 million to less than 1 in 1 trillion. This means that for each of the Trader Defendants, it is nearly impossible that their trading is uncorrelated with the hack of the EDGAR system.”
Thus, it is surprising that the SEC’s case against Clark was dismissed so early. The SEC appealed and, if it prevails, the ultimate legal impact of Judge Hilton’s outlier decision should be minimal. Judge Hilton did not give the SEC a chance to establish evidence of scienter based on strong circumstantial statistical evidence.
An Expert’s Advice for the Plaintiffs’ Bar for Successful Prosecution of Insider Trading Claims
What steps can plaintiffs’ lawyers take to leverage the value of statistical evidence in identifying and deterring wrongdoing in insider trading cases? To find out, I interviewed Daniel Taylor, a professor at The Wharton School of the University of Pennsylvania, who leads the Wharton Forensic Analytics Lab and has done extensive research on insider trading. According to Taylor, a weakness of the plaintiffs’ bar is that it tends to rely on experts who are credentialed law professors not trained in data analysis and visualization. Lawyers don’t understand the power of data and often do not know how to convincingly present statistical evidence to a fact-finder. Taylor says that the plaintiffs’ bar needs a broader expert network for insider trading that includes experts in data analysis and the tools for presenting that analysis to laymen, such as charts and graphs. According to Taylor, there is a lack of visualization within many legal briefs that allege suspicious trading. Taylor says that in an insider trading case, persuasive and compelling evidence would consist of probability calculations and an analysis of counter-factuals. For example, one can highlight an extreme outlier trade by comparing the trade to the distribution of normal trades. Another helpful technique would be to explain the counter-factual: for example, as it was with Clark, it is exceptionally rare for an individual to liquidate their retirement accounts, max out their credit lines, take out an auto loan, and risk those proceeds investing in out-of-the-money options unless the individual knew they had MNPI.