Background
The dispute arose from the Commerce Department’s 2015 less-than-fair-value investigation of welded line pipe from South Korea. Commerce used a differential-pricing analysis to determine whether export prices differed significantly among purchasers, regions, or time periods and whether it could use an average-to-transaction comparison method to address targeted dumping. SeAH Steel Corporation challenged Commerce’s use of Cohen’s d as the price-difference test.
After earlier proceedings, the Federal Circuit held that Commerce’s use of Cohen’s d was unreasonable because the relevant sales data did not satisfy that test’s statistical assumptions. It directed Commerce to re-perform its differential-pricing analysis without Cohen’s d. On the fourth remand, Commerce adopted a two-percent price-difference test and discontinued its mixed methodology, applying average-to-transaction comparisons to all sales when significantly different sales exceeded 33% of total sales value. SeAH challenged both changes.
The Court’s Holding
The Court of International Trade sustained Commerce’s fourth remand redetermination. It held that Commerce reasonably used a two-percent threshold to determine whether prices differed significantly. A percentage-based threshold accounted for relative differences among products and industries, and Commerce adequately supported its selection by pointing to other antidumping contexts in which a two-percent difference carries legal or administrative significance. The statute did not require Commerce to use a recognized statistical effect-size test.
The court also held that Commerce acted within the scope of the Federal Circuit’s remand when it abandoned the mixed methodology. The mandate broadly permitted Commerce to re-perform its differential-pricing analysis without Cohen’s d, allowing reconsideration of the analysis’s interrelated components. Commerce reasonably retained the 33% ratio-test threshold for identifying a pattern but applied the average-to-transaction method to all sales once that threshold was met, explaining that the change better aligned with the statutory text and improved its ability to address masked dumping.
Key Takeaways
- Commerce may use a two-percent price-difference test because the antidumping statute does not require a particular statistical test for deciding whether prices differ significantly.
- A remand directing Commerce to re-perform its differential-pricing analysis permitted the agency to reconsider its ratio-test methodology, not merely replace Cohen’s d.
- Commerce reasonably discontinued its mixed methodology and applied average-to-transaction comparisons to all sales when significantly different sales exceeded 33% of total sales value.
Why It Matters
The decision approves Commerce’s replacement for the Cohen’s d test after the Federal Circuit rejected that test’s application to sales data lacking its required statistical characteristics. It also confirms Commerce’s substantial methodological discretion when implementing the statutory exception for patterns of significantly different export prices.
For antidumping litigants, the ruling indicates that Commerce may rely on administratively selected thresholds rather than conventional statistical effect-size measures and may revise connected parts of its differential-pricing framework when an appellate remand requires the agency to reconsider the analysis as a whole.