Recent CIT Decisions Upholding 'd' Test Don't Address Questions Raised by CAFC, Exporters Tell CIT
None of the Court of International Trade's conclusions upholding the use of the Cohen's d test to root out "masked" dumping address the Commerce Department's "core error" of using a 0.8 threshold "when the statistical assumptions of normality, variance, and size have not been proven," thermal paper exporters led by Koehler Paper argued. In a reply brief at the trade court, Koehler said CIT's recent decisions in Stupp Corp. v. U.S. and Marmen v. U.S. "do nothing to mitigate the fundamental flaws" of using the d test (Koehler Paper, et al. v. United States, CIT # 21-00632).
Sign up for a free preview to unlock the rest of this article
Timely, relevant coverage of court proceedings and agency rulings involving tariffs, classification, valuation, origin and antidumping and countervailing duties. Each day, Trade Law Daily subscribers receive a daily headline email, in-depth PDF edition and access to all relevant documents via our trade law source document library and website.
The U.S. Court of Appeals for the Federal Circuit originally raised questions about the use of the test, asking Commerce whether it could reasonably use the test even though the data used by the agency did not satisfy its basic statistical assumptions. In its defense, the government said it did not need to satisfy the assumptions since it used the entire population of the data and not a sample. In Marmen (see 2303200066) and Stupp (see 2302270049), the trade court upheld this defense.
For instance, in Stupp, CIT upheld use of the d test because (1) Commerce combines its Cohen's d analysis with the ratio and meaningful difference tests, which make up for inaccuracies in the d test, (2) small fluctuations in price will not lead to "false positives," and (3) use of the 0.8 threhsold leads to a reasonably infrequent use of alternative methodologies. While parties in Koehler's case said this opinion backs up its position, the exporter said it actually does nothing to settle the questions raised by the Federal Circuit.
Koehler said the Cohen's d analysis does not actually work with the raito test and meaningful difference test to compensate for the d test's inaccuracies. The tests operate sequentially and represent more of the axiom "garbage in, garbage out" since "less accurate results in d necessarily leads to less accurate results in subsequent tests."
Commerce also errs when it claims that small price fluctuations are less likely to lead to "false positives" when using the test, since the entire concept of false positives only works in the context of sample data, Koehler said. "It is nearly incoherent to defend a false positive rate, which can only stem from taking samples, when the Department has stated it is in control of the full universe of sales," the brief said. Koehler dubbed the agency's claim a "red herring," since the question of false positives is not relevant to the data being tested.