US, Petitioner Defend Use of Simple Average in Cohen's 'd' Coefficient Denominator

The U.S. and antidumping duty petitioner Mid Continent Steel & Wire defended the Commerce Department's use of the Cohen's d test to detect "masked" dumping, in a pair of reply briefs at the U.S. Court of Appeals for the Federal Circuit. Taiwanese steel nail exporters, led by PT Enterprise, challenged Commerce's use of a simple average for the denominator of the Cohen's d coefficient instead of a weighted average (Mid Continent Steel & Wire v. United States, Fed. Cir. # 24-1556).

The government claimed that since the academic literature doesn't speak directly to the d test, the agency reasonably used the "general principle of reliability presented in the academic literature" when invoking the simple average in the AD investigation on steel nails from Taiwan. Dr. Jacob Cohen, the creator of the test, called for a simple average when populations' standard deviations aren't equal.

Understanding that this equation applies when sample sizes are equal, "a simple average is reasonable because sample size is an indicator of reliability, and, thus, the averaged estimated standard deviations are equally reliable," the brief said.

The U.S. also said Commerce addressed the Federal Circuit's previous suggestion of using a single standard deviation commingling the data from the test and comparison groups, claiming that the single standard deviation of the commingled groups isn't an "appropriate" way to contextualize the differences in the means of the two populations. The single standard deviation of the commingled has the variance within each group and also the difference between the groups, "and, thus, single standard deviation will vary with the difference in the means between the two groups," the brief said.

Mid Continent took a more direct approach to the argument, telling the court that the Taiwanese exporters want the weighted average since it would "eliminate" the dumping margin and revoke the AD order. The petitioner has said that while it has taken years to get to this point, Commerce has now "articulated a clear and reasonable rationale and justification for its methodology."

Mid Continent claimed that it's reasonable to use a simple average "because Commerce’s goal is to measure an abstract effect, i.e., pricing behavior in the test group vis-à-vis pricing behavior in the comparison group." The statute refers only to identifying a "pattern of export prices," and doesn't contemplate the size of a sale, making the use of a weighted average, at best, not prescribed by law and, at worst, barred under the statute.

The brief said it's irrelevant why a pattern of prices exists, "as is the size of the sale associated with the price being examined; it is the presence of a pattern of prices, on its own, that the statute directs Commerce to assess."