Choi v. Tower Research Capital LLC

Background

Myun-Uk Choi, Jin-Ho Jung, Sung-Hun Jung, Sung-Hee Lee, and Kyung-Sub Lee sued Tower Research Capital LLC and Mark Gorton on behalf of themselves and a proposed class of investors who traded KOSPI 200 futures during the 2012 overnight market. The only remaining claim was for unjust enrichment. The plaintiffs alleged that Tower used a two-part trading strategy: it placed non-marketable orders that allegedly attracted other traders to the market, then used large marketable orders to make favorable trades. The plaintiffs characterized this as a two-sided form of spoofing.

The plaintiffs sought certification under Federal Rule of Civil Procedure 23(b)(3), which requires common questions to predominate over questions affecting only individual class members and requires a class action to be the superior method of resolving the dispute. They also sought appointment of class representatives and class counsel. Defendants opposed certification and moved to seal portions of the parties’ briefs and exhibits because they contained confidential business information.

Choice of Law

The court held that the plaintiffs had not shown that the laws potentially applicable to the unjust-enrichment claims had no meaningful differences. The proposed class could include members from different states and countries, and the laws of unjust enrichment could differ on issues such as whether an indirect benefit is sufficient, whether actual loss is required, whether dishonesty or fraud must be shown, and whether unjust enrichment is an independent claim. The court therefore did not decide which choice-of-law test would apply. For purposes of analyzing the motion, it assumed without deciding that New York law governed.

Why the Court Found Common Issues Did Not Predominate

Under the assumed New York law, unjust enrichment requires proof that the defendant received a benefit, that the benefit came at the plaintiff’s expense, and that fairness requires repayment. The court concluded that the plaintiffs had not shown that any of these elements could be proved for the proposed class through generalized evidence.

First, the plaintiffs had not established a reliable classwide method for identifying artificial prices caused by Tower’s conduct. The court found that the model developed by the plaintiffs’ expert, Haim Bodek, could identify the trading pattern that he called an “Artificial Price State,” but could not establish that Tower’s conduct actually changed the price. The model also treated all trades during such a state as affected, even though some trades could have occurred at the same price without Tower’s alleged manipulation. In addition, the model used later prices as a benchmark without adequately distinguishing price changes caused by Tower from later changes caused by other market events.

Second, for similar reasons, the plaintiffs could not use common evidence to show that Tower’s benefit came at each plaintiff’s expense. The court stated that the plaintiffs needed to identify the price at which each trade should have occurred rather than relying on the actual execution price, and they had not provided a valid method for doing so.

Third, the court found that deciding whether fairness required repayment would require individualized evidence. The plaintiffs’ theory depended on identifying traders whose orders were actually lured into the market by Tower’s alleged Supporting Orders. Some orders were placed before Tower’s first Supporting Order, while others would have been placed even without Tower’s conduct. The plaintiffs’ expert had not provided a method for separating lured orders from non-lured orders. The court also noted that some proposed class members might have been sophisticated traders who knowingly accepted the risk that faster firms could trade against their resting orders. Although Tower’s alleged intent not to execute its Supporting Orders could present a common question, the court found that this issue did not predominate over the individualized questions.

Damages Methodology

The court separately found that the plaintiffs’ damages model did not satisfy the requirement that a classwide model measure damages resulting from the asserted theory of injury. The model lacked a justified benchmark price, did not determine which orders had been lured by Tower, included some orders that the plaintiffs’ expert acknowledged should be excluded, and could not distinguish price changes caused by delayed movement toward a supposed true price from later changes caused by independent market information. The court found these to be conceptual problems rather than defects that could be fixed with a simple change to the model.

Rulings

The court denied the plaintiffs’ motion for class certification. It denied as moot the motions to appoint class representatives and class counsel, and denied as moot the parties’ joint request for oral argument. The court provisionally granted defendants’ motions to seal portions of the briefs and exhibits, except for passages quoted in the opinion, for no more than ten weeks through December 6, 2022, absent further order. It required defendants by October 25, 2022, to file unredacted copies showing the proposed redactions and to explain the reasons for each requested redaction. The court also directed the parties to submit a joint letter proposing the next steps in the litigation by that date.