The Obama Administration’s focus on criminal trade secret prosecutions under the Economic Espionage Act (EEA) highlights the legal complexities at the murky intersection between criminal and civil jurisprudence in trade secrets cases. As we previously discussed, when it comes time for sentencing, determining the “value” of the stolen trade secrets is often difficult—and courts have applied different valuation models.

In the latest example of the critical role that valuation can play in EEA sentencing, the Seventh Circuit recently vacated the three-year sentence of a criminal defendant when the district court erred by relying upon a $12 million loss figure that lacked factual support in the record.

The defendant, Yihao Pu, worked as a quantitative finance professional at two companies engaged in high frequency trading (“HFT”) in the stock market. Both companies had developed proprietary HFT platforms that performed the tasks necessary to execute stock trades at lightning speeds. While working at these companies, Pu illegally copied trade-secret-protected files and transferred them to his personal electronic storage devices. The government indicted Pu on multiple counts, including wire fraud and unlawful transmission and possession of trade secrets. Pu pled guilty to one count of unlawful possession of a trade secret and one count of unlawful transmission of a trade secret. The district court sentenced Pu to a three-year prison sentence.

EEA sentences are largely determined by the “loss” demonstrated in a particular case. Under federal guidelines, the district court has discretion to make a reasonable estimate of the loss amount, often relying on the greater of the actual or intended loss calculation for sentencing. As we previously noted, a lack of actual loss is not uncommon in EEA prosecutions. And in Pu, the parties agreed there was no actual loss. Instead, the district court relied upon the cost model (i.e., the development costs for the “trade secret”) to calculate the intended amount of loss used to determine the defendant’s sentence. Relying entirely on R&D costs submitted by the victim companies, the district court arrived at an intended loss amount of approximately $12 million.

On appeal, the Seventh Circuit vacated this finding, noting that while development costs for trade secrets are an “easy figure to use,” that was not the appropriate inquiry. The district court had to analyze whether the government had proved by a preponderance of the evidence that Pu actually intended to cause a loss to the victims that equaled the cost of development. The Court of Appeals found that here was no support in the record that Pu intended the victims to suffer that amount of loss. In fact, there was no evidence at all from which the court could infer how much loss Pu intended to cause. The Seventh Circuit vacated both Pu’s sentence and the district court’s restitution order that required Pu to pay one of the victim companies $750,000 due to a lack of evidentiary support.

The Pu case is therefore illustrative of the perils of relying upon the “cost model” for sentencing. The district court’s reliance on the R&D costs provided by the victim companies resulted in a 20-point increase to the base offense level for the defendant, but the increase was not able to withstand scrutiny by the Court of Appeals. It goes without saying that prosecutors would be wise to carefully select the appropriate model, and victims of trade secret theft should be sure to submit adequate factual evidence to support any “loss” amount for the stolen trade secrets at issue.