Mid Continent Steel & Wire v. United States
940 F.3d 662
Fed. Cir.2019Background
- Commerce investigated antidumping of certain steel nails from Taiwan; PT Enterprise and affiliated Pro-Team were named mandatory respondents.
- Commerce compared constructed value (home-market side) to export price (U.S. side); constructed value relied on costs paid to numerous "toll" manufacturers.
- Mid Continent argued some tollers were affiliated with PT and their transactions should be disregarded when computing constructed value; Commerce found no affiliation.
- In the final determination Commerce found "differential pricing," applied a mixed average-to-transaction (A-to-T) approach using a Cohen’s d test, and calculated a dumping margin slightly above Commerce’s de minimis threshold.
- Mid Continent appealed the affiliation finding; PT cross-appealed three aspects of Commerce’s differential-pricing methodology (zeroing at aggregation, the fixed d≥0.8 threshold, and use of a simple vs. weighted pooled variance).
Issues
| Issue | Plaintiff's Argument | Defendant's Argument | Held |
|---|---|---|---|
| Whether PT was "affiliated" with tollers (ability to control) | Mid Continent: tollers effectively controlled/dependent on PT and transactions should be treated as between affiliates | Commerce: no control—tollers sold to others, were profitable, no common ownership/management or long-term ties | Court: Affirmed Commerce; substantial evidence supports finding of no affiliation |
| Use of "zeroing" negative margins from A-to-A group when aggregating overall margin | PT: zeroing at aggregation stage conflicts with statute | Commerce/US: statute silent on this implementation; zeroing preserves remedial effect of A-to-T and is reasonable | Court: Rejected PT’s statutory challenge; upheld zeroing as reasonable exercise of discretion |
| Rigid Cohen’s d cutoff of 0.8 to identify significant pricing differences | PT: threshold is inflexible; small absolute $ differences can be meaningless even if d≥0.8 | Commerce/US: 0.8 is widely used effect-size standard; relative-to-dispersion measure is appropriate and objective | Court: Upheld Commerce’s use of 0.8 as reasonable |
| Using simple average (vs weighted) to compute pooled variance in Cohen’s d | PT: pooled variance should be weighted (e.g., by quantity) as statistics sources recommend; simple average can distort results | Commerce/US: simple averaging promotes predictability and avoids manipulation of how respondents report sales; statute silent | Held: Court vacated and remanded—Commerce must better justify choice and address weighted alternatives |
Key Cases Cited
- SKF USA Inc. v. United States, 630 F.3d 1365 (Fed. Cir. 2011) (affiliate transactions may be disregarded if not reflecting usual market amounts)
- Apex Frozen Foods Private Ltd. v. United States, 862 F.3d 1322 (Fed. Cir. 2017) (interpreting §1677f-1(d)(1)(B) and upholding broad A-to-T application)
- U.S. Steel Corp. v. United States, 621 F.3d 1351 (Fed. Cir. 2010) (antidumping statute silent/ambiguous as to zeroing methodology)
- CS Wind Vietnam Co., Ltd. v. United States, 832 F.3d 1367 (Fed. Cir. 2016) (agency must provide adequate explanation of methodology for judicial review)
- Consolo v. Federal Maritime Comm’n, 383 U.S. 607 (U.S. 1966) (agency findings may be supported even where evidence permits contrary inferences)
- Diamond Sawblades Mfrs. Coal. v. United States, 866 F.3d 1304 (Fed. Cir. 2017) (standard of review for Commerce decisions)