Korean Exporter Objects to Commerce's Defense of 'd' Test in Suit Returned to Federal Circuit
The Commerce Department improperly used Cohen's d test to root out masked dumping because the agency violated statistical assmptions inherent to the test, SeAH Steel told the U.S. Court of Appeals for the Federal Circuit in the opening brief of its appeal. While Commerce justified its use of the test because it used a whole population, not a sample, SeAH said the academic literature shows the d test was meant to be used as a measure of effect size only when the data comes from samples with "normal distributions, with roughly equal variance, and a sufficient number of data-points" (Stupp Corp. v. United States, Fed. Cir. # 23-1663).
Sign up for a free preview to unlock the rest of this article
Export Compliance Daily combines U.S. export control news, foreign border import regulation and policy developments into a single daily information service that reliably informs its trade professional readers about important current issues affecting their operations.
The Federal Circuit previously remanded the case, questioning Commerce's use of the d test without these statistical assumptions, among other reasons (see 2107150032). The agency said at the Court of International Trade that Professor Jacob Cohen, the founder of the test, only addresses the use of the test when analyzing samples, adding that none of the statistical sources prohibits the use of the test for entire populations.
"Tellingly, Commerce failed to identify any texts that support its proposed use of Cohen’s d when the assumptions described by Professor Cohen are not satisfied," SeAH countered. "Moreover, Commerce also failed to provide any independent logical, mathematical, or empirical justification for its use of Cohen’s d when the usual assumptions are not satisfied." If Commerce seeks to use the test as it did here in the antidumping duty investigation on welded line pipe from South Korea, "it must provide some other basis for concluding that" the thresholds it uses are reasonable.
When the case was first before the appellate court, the Federal Circuit saw that under Commerce's use of the d test, an "imperceptible price difference" could be found to be "significant," leading to a change in the way the agency compares home market and U.S. sales when setting the dumping margin. At CIT, Commerce said the separate "meaningful difference" test as part of its differential pricing analysis (DPA) alleviates this concern since insignificant differences would not lead to increased margins from zero to above the 2% de minimis threshold.
SeAH said this was wrong. While an incorrect significant price difference for a particular customer, region or time period for one good may not lead to a higher margin for that product, it will affect the number of transactions found to pass the test. Since the "ratio test" depends on the number of transactions that pass the test, an incorrect number of passing transactions "may lead to an alternate comparison methodology," which "will inflate the dumping margin," the brief said. The claim that the meaningful difference test resolves the problem is true, "if at all," for only single products but not for multiple products, as is the case in the present investigation.
The exporter also objected to Commerce's assertion that an analysis showing the DPA only altered the comparison methodology in 21% or 22% of cases "demonstrated" the analysis was reasonable. SeAH said the "logical error in that argument should be obvious," since unless a party knows how often Commerce should have altered its methodology, "one cannot assess whether Commerce's actual experience is reasonable or not."