CIT Upholds Use of Simple Average in 'd' Test After 2 Prior CAFC Rejections

The Court of International Trade on Feb. 12 sustained the Commerce Department's decision to use a simple average of standard deviations in the denominator of the Cohen's d test in detecting "masked" dumping as part of the antidumping investigation on steel nails from Taiwan. Despite a pair of decisions from the U.S. Court of Appeals for the Federal Circuit rejecting the use of simple averages in this case, Judge Claire Kelly said she could find no fault with the logic Commerce employed.

Ned Marshak, counsel for exporters led by PT Enterprise Inc., said in an email that he is "disappointed with the CIT’s decision," noting that an appeal is forthcoming.

"We will soon find out whether the Court of Appeals agrees with us that the DOC’s fourth bite at the apple is no better than its first three attempts to justify a methodology which is contrary to academic literature, basic principles of statistics and mathematics, judicial precedent and common sense," Marshak said.

To detect masked dumping, Commerce conducts three different types of tests, the first of which, the Cohen's d test, finds whether a company sells its goods across a "pattern of prices that differ significantly." The d test sees the agency compare the prices of "test groups" of an exporter's sales to a "comparison group" set by "region, purchaser, and time period."

Commerce creates groups from a respondent's sales, then computes the means and standard deviations of the test and comparison groups. The agency then creates a d coefficient by "dividing the difference in the groups’ means by the square root of the average of the squared standard deviations of each group." Commerce gets the average by adding them together and dividing by two -- a simple average -- as opposed to accounting for the size differences for each group by using a weighted average.

The Federal Circuit rejected this approach twice before in this case, once in 2019 and again in 2022. The appellate court found fault with Commerce's reasoning that the equally rational and equally genuine pricing choices warranted simple weighting in the d denominator, adding that Commerce needs reasoning for departing from the academic literature, which uses a weighted average.

On remand, Commerce said the literature uses a simple average when the sample sizes are equal and that the standard deviation of a full population is "in fact the actual standard deviation." As a result, since the agency used the full population of data in using the Cohen's d test, using a simple average is supported. Kelly said she "cannot find fault with Commerce's logic here."

Commerce gave an explanation which "logically connects" the relevance of full populations of data and the use of simple averaging, Kelly said. While there are other "reasonable alternatives" out there, the court found Commerce's reliability analysis as reasonable.

Exporters, led by PT Enterprise, didn't challenge the premise that using a simple average for equal sample sizes was valid, but instead argued that reliability is "dependent on a number of factors, at least with respect to samples." Kelly said that this claim falls apart since "Commerce is not comparing the reliability of a sample to the reliability of a full population," but that full populations have "equal reliability."

In its remand results, Commerce also addressed the Federal Circuit's suggestion that the agency could use the whole population's standard deviation. The agency said using this one number for both groups would "eliminate the relevancy of each individual standard deviation much in the same way that weighting the standard deviations would diminish the relevancy of one of the standard deviations," Kelly said.

Marshak said the Federal Circuit's forthcoming ruling in Stupp Corp. v. U.S., in which the court will rule on whether the use of the full population precludes Commerce from meeting basic statistical assumptions when using the d test (see 2401110037), will not affect this case. He said the exporters should win regardless of how Stupp is decided since the "manner in which the DOC calculated the Cohen’s d denominator remains unreasonable."

(Mid Continent Steel & Wire v. United States, Slip Op. 24-15, CIT # 15-00213, dated 02/12/24; Judge: Claire Kelly; Attorneys: Adam Gordon of The Bristol Group for plaintiff Mid Continent Steel & Wire; Bruce Mitchell of Grunfeld Desiderio for consolidated plaintiffs led by PT Enterprise Inc.; Mikki Cottet for defendant U.S. government)