TRADE REMEDIES | 2026-11-04

Targeted Dumping and the Cohen’s d Test: How Commerce Uses Differential Pricing to Raise Antidumping Margins

When an exporter’s prices differ sharply across customers, regions, or time periods, the Commerce Department can switch comparison methods and apply zeroing — often producing a much larger antidumping margin. California importers need to understand how the targeted-dumping analysis works.

By Anatolii Trembach, Esq. (California State Bar No. 349304; LL.M., USC Gould School of Law; admitted to the U.S. Court of International Trade), Trembach Law Firm, APC · Published

In most antidumping cases, the Commerce Department calculates a dumping margin by comparing the weighted-average price an exporter charges in the United States to the weighted-average price it charges at home — an average-to-average comparison that lets high U.S. prices offset low ones. But a statut

Antidumping duties are meant to offset the margin by which a foreign producer sells merchandise in the United States below its "normal value" — usually the price in its home market. How that margin is calculated can matter as much as the underlying prices, because the comparison methodology determines whether high-priced U.S. sales are allowed to offset low-priced ones. The targeted-dumping exception, implemented through Commerce’s differential pricing analysis, is the single most consequential methodological lever in modern antidumping practice. For California importers sourcing from countries under antidumping orders, the difference between the default method and the targeted-dumping method can be the difference between a workable duty and one that ends the business case for importing at all.

The Default Rule: Average-to-Average Comparison

Under 19 U.S.C. § 1677f-1(d)(1)(A), in an antidumping investigation Commerce normally determines the dumping margin by comparing the weighted-average normal value to the weighted-average of the export prices (or constructed export prices) of comparable merchandise. This average-to-average method groups sales and compares averages, which means that sales made at prices above normal value offset sales made below it. The practical effect is that an exporter whose prices vary around normal value may show little or no dumping, because the high and low sales cancel out.

This offsetting is not a loophole — it is the ordinary consequence of comparing averages, and after years of litigation at the World Trade Organization, Commerce abandoned "zeroing" (the denial of offsets) in the average-to-average context for investigations. The result is a methodology that treats an exporter’s U.S. pricing as a whole. The targeted-dumping exception exists precisely to address the situation where that whole-of-market picture is said to mask dumping aimed at particular customers, regions, or periods.

The Statutory Exception for Targeted Dumping

The exception appears in 19 U.S.C. § 1677f-1(d)(1)(B). Commerce may compare the weighted-average normal value to the export prices of individual U.S. transactions (an average-to-transaction comparison) if two conditions are met: first, there is a pattern of export prices that differ significantly among purchasers, regions, or periods of time; and second, Commerce explains why those differences cannot be taken into account using the ordinary average-to-average method. The statute is the vehicle Congress provided to reach "targeted" dumping — selling low to a particular segment of the U.S. market while keeping overall averages benign.

The significance of invoking § 1677f-1(d)(1)(B) is that, in the average-to-transaction setting, Commerce applies zeroing: it treats sales made above normal value as showing zero dumping rather than as offsets, and sums only the positive dumping amounts. The combination of switching to average-to-transaction and denying offsets is what can transform a margin. A respondent whose average-to-average margin would be zero or de minimis can face a substantial margin once the targeted-dumping method is applied to the portion of sales that fail the pattern test.

Differential Pricing Analysis and the Cohen’s d Test

To decide whether the statutory "pattern of prices that differ significantly" exists, Commerce developed the differential pricing analysis, which has three steps. The first and most contested is the Cohen’s d test. Cohen’s d is a statistical measure of "effect size" — it compares the mean price to one purchaser, region, or period (the "test group") against the mean price to all other purchasers, regions, or periods (the "comparison group"), and expresses the gap in units of pooled standard deviation. Commerce treats a Cohen’s d coefficient of 0.8 or greater — the threshold the statistical literature labels a "large" effect — as showing that prices to that group differ significantly.

The second step is the ratio test, which measures what share of the respondent’s U.S. sales (by value) passes the Cohen’s d test. If more than 66% of value passes, Commerce applies the average-to-transaction method to all sales; if between 33% and 66% passes, it applies a "mixed" approach — average-to-transaction for the passing sales and average-to-average for the rest; if less than 33% passes, the pattern is not meaningful and the default method stands. The third step is the "meaningful difference" test, which asks whether using the alternative method produces a margin that differs meaningfully from the default method — the explanation the statute requires for why the average-to-average method could not account for the differences.

Why the Method Can Multiply a Margin

The power of the targeted-dumping method comes from the interaction of two features. First, switching to average-to-transaction isolates individual low-priced sales. Second, zeroing then prevents those low sales from being offset by high ones. A respondent with a mix of sales scattered around normal value — the ordinary pattern of a competitive business — can find that the sales below normal value are aggregated into a positive dumping amount while the sales above normal value contribute nothing to reduce it. The arithmetic is asymmetric by design.

For a California importer, the lesson is that a supplier’s margin in a prior segment of a proceeding is not a reliable guide to its margin in the next one. A producer that varies its U.S. pricing — offering volume discounts to large customers, regional promotions, or end-of-season pricing — may generate exactly the kind of price dispersion that triggers the Cohen’s d test, even without any intent to target the U.S. market. The methodology keys on the pattern of prices, not on motive.

The Federal Circuit’s Scrutiny: Stupp and Apex

The courts have both upheld the framework and pressed Commerce to justify it. In Apex Frozen Foods Private Ltd. v. United States, 862 F.3d 1322 (Fed. Cir. 2017), the Federal Circuit sustained Commerce’s authority to use the differential pricing analysis and to apply zeroing in the average-to-transaction context, rejecting the argument that zeroing was impermissible there. Apex confirmed that the targeted-dumping exception and its associated denial of offsets rest on a reasonable reading of § 1677f-1(d)(1)(B).

In Stupp Corp. v. United States, 5 F.4th 1341 (Fed. Cir. 2021), however, the Federal Circuit questioned whether Commerce could apply the Cohen’s d test without satisfying the statistical assumptions that underlie it — in particular, assumptions about normal distribution, sufficient sample size, and roughly equal variances between the test and comparison groups. The court remanded for Commerce to explain how the test remains reliable when those assumptions are not met, as is common with the small, non-normal datasets in antidumping proceedings. The Stupp remand did not abolish the differential pricing analysis, but it signaled that Commerce’s statistical shortcuts are not immune from review and must be defensible on their own terms.

How the Analysis Plays Out in a Proceeding

In practice, differential pricing is litigated through the record. Commerce runs its programs on the respondent’s reported U.S. sales database, defines the test and comparison groups by purchaser, region, and quarter, and reports the Cohen’s d and ratio-test results in its preliminary and final determinations. Respondents and petitioners submit case briefs contesting how groups were defined, whether the data satisfy the test’s assumptions, and whether the "meaningful difference" standard was correctly applied. Because the analysis is mechanical once the inputs are set, the most effective advocacy often targets the inputs — how regions are grouped, how time periods are divided, and whether outliers distort the standard deviation.

California importers are rarely the respondents themselves — the respondents are the foreign producers and exporters — but importers bear the duty consequences and have a direct stake in the outcome. An importer that understands which of its suppliers are vulnerable to a targeted-dumping finding can diversify sourcing, seek suppliers with their own low (or zero) rates, or participate in reviews to protect a favorable rate. Because Commerce sets cash-deposit rates prospectively and assesses final duties on entries after each administrative review, methodology changes ripple directly into an importer’s landed cost.

Managing the Risk for California Importers

Several strategies reduce exposure. Supplier selection matters: a producer with a long record of de minimis margins, or one not subject to the order at all, carries less methodological risk. Review participation matters: because assessment rates are set in administrative reviews, an importer whose supplier declines to participate may face the adverse all-others or country-wide rate rather than a calculated one. Documentation matters: accurate, well-organized sales data help a cooperating respondent rebut an unfavorable grouping or defend the statistical reliability of its pricing.

Importers should also plan for volatility. Because a supplier’s margin can shift between review periods based on methodology alone, the prudent importer treats antidumping duty as a variable cost subject to annual recalculation, not a fixed line item. Bonding, duty-drawback planning where available, and contractual allocation of antidumping risk between buyer and seller all help absorb the uncertainty that the differential pricing analysis injects into trade-remedy cases. For a broader overview of trade-remedy proceedings and customs enforcement, see our international trade and customs practice.

Administrative Reviews and the Shifting Margin

Antidumping duties in the United States operate on a retrospective assessment system, which magnifies the importance of methodology. When merchandise enters, the importer posts a cash deposit at the current rate, but the final duty is not fixed until Commerce completes an administrative review covering that entry’s period under 19 U.S.C. § 1675(a). In each annual review, Commerce recalculates the respondent’s margin on fresh sales data — and reruns the differential pricing analysis. A supplier that passed the ratio test one year may fail it the next, or vice versa, because the pricing pattern changes with the respondent’s commercial behavior during the review period.

The consequence for a California importer is direct exposure to retroactive duty liability. If a review sets an assessment rate higher than the deposits posted, the importer owes the difference plus interest; if lower, it receives a refund. Because the differential pricing method can move a rate sharply, importers cannot treat the cash-deposit rate as the true cost of goods. Prudent importers reserve for the possibility that a review will raise the rate, monitor whether their suppliers are mandatory respondents or are receiving the all-others rate, and consider whether to request a review themselves to obtain a calculated rate rather than inherit an adverse one. The retrospective system means that the methodology questions discussed above are not academic — they determine a cash liability that can arrive years after the entry.

Judicial Review at the Court of International Trade

A respondent or petitioner dissatisfied with Commerce’s determination may seek judicial review at the U.S. Court of International Trade (CIT) under 28 U.S.C. § 1581(c), with further appeal to the Federal Circuit. The CIT reviews Commerce’s antidumping determinations under the substantial evidence standard and asks whether the agency acted in accordance with law. Methodology challenges — including attacks on the grouping of test and comparison sales, the statistical reliability of the Cohen’s d test, and the application of zeroing — are litigated there on the administrative record.

The Stupp and Apex decisions arose from exactly this review path, and they illustrate its limits and its force. Courts give Commerce considerable deference on technical methodology, so a challenger must usually show that the agency failed to explain its choices or departed from the statute, not merely that a different method would have been reasonable. At the same time, Stupp shows that deference is not abdication: when Commerce invokes a statistical test, it must be prepared to defend the test’s reliability on the record. For importers, the practical takeaway is that preserving methodology arguments in case briefs before Commerce — the prerequisite to raising them in court — is an essential part of managing a supplier’s rate, since unexhausted arguments are generally forfeited.

Frequently Asked Questions

What is targeted dumping?

Targeted dumping refers to a pattern of U.S. export prices that differ significantly among particular purchasers, regions, or time periods — as opposed to uniform low pricing. Under 19 U.S.C. § 1677f-1(d)(1)(B), when Commerce finds such a pattern and explains why the ordinary averaging method cannot account for it, it may use an alternative comparison method that often yields a higher margin.

What is the Cohen’s d test?

Cohen’s d is a statistical measure of effect size that Commerce uses to decide whether prices to a test group differ significantly from prices to a comparison group. Commerce treats a coefficient of 0.8 or greater — the "large" effect threshold in the statistical literature — as showing a significant difference.

Why does the method raise antidumping margins?

Switching from an average-to-average comparison to an average-to-transaction comparison isolates individual low-priced sales, and Commerce then applies zeroing, which denies offsets for sales above normal value. The combination aggregates only positive dumping amounts, which can substantially increase the margin.

What is zeroing?

Zeroing is the practice of treating sales made above normal value as showing zero dumping rather than as offsets against sales below normal value. Commerce abandoned zeroing in the average-to-average context but continues to apply it in the average-to-transaction context used for targeted dumping, as upheld in Apex Frozen Foods v. United States, 862 F.3d 1322 (Fed. Cir. 2017).

Did the courts strike down the Cohen’s d test?

No. In Stupp Corp. v. United States, 5 F.4th 1341 (Fed. Cir. 2021), the Federal Circuit remanded for Commerce to explain how the Cohen’s d test remains reliable when the statistical assumptions behind it — normal distribution, adequate sample size, equal variances — are not met. The court did not abolish the differential pricing analysis.

Does a producer have to intend to target the U.S. market?

No. The analysis keys on the statistical pattern of prices, not on the exporter’s motive. Ordinary commercial pricing — volume discounts, regional promotions, seasonal pricing — can generate the price dispersion that triggers the Cohen’s d test even without any intent to target.

How does this affect a California importer rather than the exporter?

The foreign producer is the respondent, but the importer pays the antidumping duties. A targeted-dumping finding that raises a supplier’s margin directly increases the importer’s cash-deposit and assessed-duty burden, so importers have a strong interest in supplier selection and in participating in administrative reviews.

How can an importer reduce exposure to targeted-dumping margins?

Strategies include choosing suppliers with a record of low or de minimis margins, ensuring suppliers participate in administrative reviews so a calculated rate applies rather than an adverse rate, maintaining accurate sales documentation, and allocating antidumping risk contractually between buyer and seller. A trade attorney can assess a specific supply chain’s vulnerability.

Speak With a California Trade-Remedies Attorney

Antidumping methodology is technical, but its consequences are concrete: the choice between an average-to-average and an average-to-transaction comparison can move a margin from de minimis to prohibitive. If your business imports merchandise subject to an antidumping order, or sources from producers under investigation, Trembach Law Firm can help you assess your exposure and plan around it. Call (213) 645-2246 to discuss your supply chain and your options under the antidumping laws.

This article is for general informational purposes and is not legal advice. Antidumping determinations are fact-specific and depend on the record of each proceeding. Consult a qualified trade attorney about your particular circumstances.

Contact Trembach Law Firm at (818) 514-7680 for a confidential consultation.

Trembach Law Firm | 27001 Agoura Road, Suite 350, Calabasas, CA 91301