In a change from the constitutional issues that have plagued the SEC in court of late, in SEC v. World Tree Financial, LLC, the Fifth Circuit affirmed a securities-fraud judgment based on “a fraudulent ‘cherry-picking’ scheme, in which [defendants] allocated favorable trades to themselves and favored clients and unfavorable trades to disfavored clients.” Proof of this scheme required sophisticated statistical analysis, summarized as follows:
To analyze World Tree’s allocation data, [the SEC’s expert] divided the client accounts into three categories: (1) accounts controlled by Perkins, Gilmore, or both (“Favored-Perkins accounts”); (2) accounts owned by World Tree clients other than Matthew LeBlanc and his business Delcambre Cellular (“Favored-Client accounts”); and (3) accounts owned by LeBlanc and Delcambre (“Disfavored accounts”). She then measured several performance measures and subsets of trades: most and least profitable trades, day trades, average first-day returns, earnings-day trades, overlapping stocks, and trades after LeBlanc complained to Perkins about his accounts’ poor performance. According to her analysis, from July 2012 to July 2015, Perkins methodically allocated trades with favorable first-day returns to the FavoredPerkins and Favored-Client accounts, while allocating trades with unfavorable first-day returns to the Disfavored accounts.
[She] opined that the “evidence overwhelmingly indicates that Perkins engaged in cherry-picking.” Though she acknowledged at trial that the data reflected only a pattern and that she did not “have the ability to identify individual trades that may or may not be improper,” the data in the aggregate showed a “one in one million chance that these patterns could have occurred if allocations were being made without regard to first-day return.”
No. 21-30063 (Aug. 4, 2022).