Judge Admits Statistical Expert Testimony in Western Asset CIO Fraud Case

NEW YORK (LN) — U.S. District Judge Gregory H. Woods denied cross-motions to exclude expert testimony on Thursday case against S. Kenneth Leech, II, a former Chief Investment Officer at Western Asset Management Company, ruling that neither side’s statistical analyses were unreliable enough to warrant exclusion under Federal Rule of Evidence 702.

Leech faces 5 counts of fraud and false statements, including allegations that he engaged in a “cherry-picking” scheme from January 2021 to October 2023. The government alleges Leech would execute trades in Treasury futures and options, wait to see how they performed, and then allocate profitable trades to favored accounts and underperforming trades to disfavored accounts.

The government’s experts, Harvard Business School Professor Lauren Cohen and New York University Professor Petter Kolm, propose to testify that their analyses of trade blotters show a statistically significant difference in first-day returns between accounts Leech allocated to the “Macro Opps” strategy versus the “Core Strategies.”

Professor Cohen proposes to testify that trades allocated to Macro Opps had an average first-day return of $0.24 million, while trades allocated to the Core Strategies had an average first-day return of -$0.31 million. He plans to testify that this pattern is “consistent with Mr. Leech using information about the first-day performance of his Treasury F&O trades in making allocation decisions between Macro Opps and the Core Strategies.”

The defense moved to exclude the government’s experts, arguing their methods were flawed because they relied on proxies for allocation times and that their testimony impermissibly opined on Leech’s state of mind under Rule 704(b). Woods rejected those arguments, ruling the experts’ methodologies sufficiently reliable and noting that their testimony describes conduct rather than explicitly stating Leech’s intent.

Conversely, the government sought to exclude testimony from defense experts Aaron Dolgoff of Charles River Associates and New York University Professor Bruce Tuckman. The defense experts conducted a “blind allocation exercise” in which Professor Tuckman allocated trades without knowing their first-day performance, resulting in a 56.8% match rate with Leech’s actual allocations.

Woods found the exercise reliable and relevant, noting it was designed to test whether lawful allocation rationales could produce performance differences similar to those alleged by the government. The judge also allowed testimony from defense rebuttal expert Professor Kevin Maloney, whose regression analyses challenge the government’s conclusions about duration management.

While denying the motions, Woods ordered the parties to meet and confer on mutually agreeable language for the experts to use at trial to avoid confusing the jury regarding Leech’s mental state. The judge specifically cautioned against the defense using the term “null hypothesis” to describe the government’s case, calling it an “inapposite use of a statistical term of art.”

The parties must file a letter with the court within 2 weeks detailing the status of their discussions on expert language. The defense also has 2 weeks to propose a limiting instruction regarding evidence of Leech’s trading activity after October 2023, when he no longer had discretion to allocate trades to the Core Strategies.