The recent revamp of the Department of Commerce’s targeted dumping test has brought significant changes in the use of zeroing in the calculation of margins in anti-dumping cases. The former “Nails Test” method for identifying whether a company engaged in targeted dumping was wholly replaced by a two-staged test known as the Differential Pricing Analysis. Beginning in March, 2013 with the Xanthan Gum cases from China and Austria, the Department has applied this new methodology over 50 times with interesting results.
Identifying a Pattern of Price Differences
Changing the targeting methodology required a reasonable method for establishing whether a significant pattern of price difference exists. The Department adopted a statistic little known in the antidumping world:the Cohen’s d test, a ratio of the difference between the means of two comparison groups divided by the standard deviation. In our case, the comparisons are U.S. net selling prices and the standard deviation is the average dispersion of prices for the two groups.
Stage one of the Differential Pricing Analysis calculates the Cohen d ratio numerous times, comparing prices for groups of sales by purchaser, geographic region and quarter. If a company has a large percentage of its sales with high Cohen d ratios, a possible pattern of price difference might be indicated, and the Department may be justified in applying zeroing (no offsets for negative margins).
Three Levels of Differential Pricing and their Methods of Zeroing
A company’s overall score in the Cohen d tests is the percentage of US sales with high Cohen d ratios. The Department defines three thresholds of differential pricing depending on this score and applies corresponding averaging and zeroing methods as shown in the following table.[table id=13 /]
In Stage two of the Differential Pricing Analysis the Department assesses the difference between the margins using the various methods and asks the question: Is there a meaningful difference in the weighted-average dumping margins when calculated using the average-to-average method versus the mixed or alternative method. A relative change in the margin greater than 25% or a move from a de minimis margin to above is considered meaningful.
Results Thus Far
Our analysis to date shows that the DOC has employed the test in over 50 decisions. Differential pricing was found over 75% of the time. Of the companies with a finding of differential pricing, 25 decisions showed no meaningful difference between the zeroed, and the non-zeroed outcome. Despite the large occurrence of differential pricing, zeroing of all sales in the average-to-transaction method was applied 6 times and mixed zeroing 8 times.Our early assessment of the new targeting methodology indicates that while most companies are found to be differentially pricing at some level, Commerce nevertheless refrains from applying zeroing because its applications yields no meaningful difference in anti-dumping margins. The efficacy of the Cohen’s d test has been argued in numerous cases. We suspect that the debate over targeting or differential pricing is far from over, and the Cohen’s d test is not the final word on how differential pricing is to be measured. Indeed, in the recent final decision for Plywood Lumber the Department states that “it is important for us to gain practical experience with this new analytical approach, on a case-by-case basis before turning this methodology into a hard and fast rule to be applied broadly in all anti-dumping proceedings.” [table id=30 /]