Timken Co. v. United States

Slip Op. 16- 47

UNITED STATES COURT OF INTERNATIONAL TRADE

THE TIMKEN COMPANY,

Plaintiff,
Before: Jane A. Restani, Judge
v.
Consol. Court No. 14-00155
UNITED STATES,

Defendant,

NSK LTD., NSK CORPORATION, NSK
PRECISION AMERICA, INC., NTN
BEARING CORP. OF AMERICA, NTN
CORPORATION, NTN BOWER, INC., NTN
DRIVESHAFT, INC., AMERICAN NTN
BEARING MANUFACTURING CORP.,
JTEKT CORPORATION, JTEKT NORTH
AMERICA, INC., NACHI-FUJIKOSHI
CORPORATION, NACHI AMERICA, INC.,
NACHI TECHNOLOGY, INC., NSK
BEARINGS EUROPE, LTD., NSK EUROPE
LTD.,

Defendant-Intervenors.

OPINION

[Commerce’s Final Results of Remand Redetermination in antidumping duty administrative
review sustained.]

Dated: May 10, 2016

Geert M. De Prest, Stewart and Stewart, of Washington, DC, argued for plaintiff. With
him on the brief was Terence P. Stewart.

L. Misha Preheim, Senior Trial Counsel, Commercial Litigation Branch, Civil Division,
U.S. Department of Justice, of Washington, DC, argued for defendant. With him on the brief
were Benjamin C. Mizer, Principal Deputy Assistant Attorney General, Jeanne E. Davidson,
Director, and Claudia Burke, Assistant Director. Of counsel on the brief was Scott D. McBride,
Consol. Court No. 14-00155 Page 2

Senior Attorney, Office of the Chief Counsel for Trade Enforcement & Compliance, U.S.
Department of Commerce, of Washington, DC.

Alexander H. Schaefer and Hea Jin Koh, Crowell & Moring, LLP, of Washington, DC,
for defendant-intervenors NSK Ltd., NSK Corporation, NSK Precision America, Inc., NSK
Bearings Europe, Ltd., and NSK Europe Ltd.

Diane A. MacDonald, Riggle and Craven, of Chicago, IL, argued for defendant-
intervenors NTN Bearing Corp. of America, NTN Corporation, NTN Bower, Inc., NTN
Driveshaft, Inc., and American NTN Bearing Manufacturing Corp. With her on the brief was
David J. Craven.

Neil R. Ellis, Justin R. Becker, Rajib Pal, Sidley Austin, LLP, of Washington, DC, for
defendant-intervenors JTEKT Corporation and JTEKT North America, Inc.

Greyson L. Bryan, Jr., David J. Ribner, McAllister M. Jimbo, O’Melveny & Myers, LLP,
of Washington, DC, for defendant-intervenors Nachi-Fujikoshi Corporation, Nachi America,
Inc., and Nachi Technology, Inc.

Restani, Judge: This matter is before the court following the U.S. Department of

Commerce’s (“Commerce”) Final Results of Remand Redetermination, ECF No. 85 (“Remand

Results”). The court remanded to Commerce to apply its differential pricing (“DP”) analysis in

the 2009–2010 annual antidumping duty (“AD”) administrative review of imports of ball

bearings and parts thereof from Japan and the United Kingdom. Timken Co. v. United States, 79

F. Supp. 3d 1350, 1352, 1361 (CIT 2015) (“Timken”). Commerce has complied with the court’s

remand order and, for the reasons stated below, Commerce’s Remand Results are sustained.

BACKGROUND

The facts of this case have been documented in the court’s previous opinion, and the

court presumes familiarity with that opinion. See id. at 1351–55. There, The Timken Company

(“Timken”) contested Commerce’s decision not to apply its DP analysis in the challenged

administrative reviews. See id. at 1352; see also Ball Bearings and Parts Thereof from Japan and
Consol. Court No. 14-00155 Page 3

the United Kingdom: Final Results of Antidumping Duty Administrative Reviews and

Rescission of Review in Part; 2009–2010, 79 Fed. Reg. 35,312 (Dep’t Commerce June 20, 2014)

(“Final Results”). The court agreed that Commerce had abused its discretion by departing from

its routine practice of applying the DP analysis rather than the Nails test to address potential

targeted dumping. Timken, 79 F. Supp. 3d at 1352, 1361. Thus, on October 8, 2015, Commerce

published, under protest, its Remand Results, in which it applied its DP analysis. Remand

Results at 1, 2. The court’s reasoning in its prior opinion fully addresses the issue and will not

be addressed further here.

Commerce uses its DP analysis to “determine whether an alternative comparison

methodology is appropriate.” Id. at 3. Typically, in administrative reviews, Commerce uses the

average-to-average (“A-A”) methodology, without zeroing, 1 as the default methodology to

calculate weighted-average dumping margins. Antidumping Proceedings: Calculation of the

Weighted-Average Dumping Margin and Assessment Rate in Certain Antidumping Duty

Proceedings; Final Modification, 77 Fed. Reg. 8,101, 8,101–03 (Dep’t Commerce Feb. 14, 2012)

(“Final Modification”). Relevant to this case, Commerce also sometimes uses the average-to-

1
Under the average-to-average (“A-A”) methodology without zeroing, Commerce compares
“monthly weighted-average export prices with monthly weighted-average normal values, and . . .
grant[s] an offset for all such comparisons that show export price exceeds normal value in the
calculation of the weighted-average margin of dumping.” Antidumping Proceedings:
Calculation of the Weighted-Average Dumping Margin and Assessment Rate in Certain
Antidumping Duty Proceedings; Final Modification, 77 Fed. Reg. 8,101, 8,102 (Dep’t
Commerce Feb. 14, 2012) (“Final Modification”).
Consol. Court No. 14-00155 Page 4

transaction (“A-T”) alternative comparison methodology, with zeroing, 2 to calculate a weighted-

average dumping margin. Id. at 8,101.

In performing the DP analysis, Commerce uses two tests to determine whether there is “a

pattern of [export prices] or [constructed export prices] for comparable merchandise that differs

significantly among purchasers, regions, or time periods.” Remand Results at 3. “This pattern is

commonly referred to as ‘targeted dumping.’” Timken, 79 F. Supp. 3d at 1352. First,

Commerce uses the “Cohen’s d test” to measure “the extent of the difference between the mean

of a test group and the mean of a comparison group.” Remand Results at 4. Commerce may

calculate a Cohen’s d coefficient only when “the test and comparison groups of data each have at

least two observations, and when the sales quantity for the comparison group accounts for at

least five percent of the total sales quantity of the comparable merchandise.” Id. When

evaluating “the extent to which the net prices to a particular purchaser, region or time period

differ significantly from the net prices of all other sales of comparable merchandise,” Commerce

considers the difference significant if the calculated Cohen’s d coefficient “is equal to or exceeds

the large (i.e., 0.8) threshold.” Id. Second, Commerce uses the “ratio test” to determine “the

extent of the significant price differences for all sales.” Id. Under this test, Commerce

determines if the value of sales to purchasers, regions, and time periods that pass Cohen’s d falls

between the 0 to 33 percent range, 33 to 66 percent range, or the 66 to 100 percent range of total

2
Under the average-to-transaction (“A-T”) methodology with zeroing, Commerce compares “the
weighted-average normal value to the export price of individual transactions for comparable
merchandise . . . [and does not] offset the results of the comparisons for which export price was
less than normal value by the results of comparisons for which export price exceeded normal
value.” Id. at 8,101 (footnote omitted).
Consol. Court No. 14-00155 Page 5

sales. Id. at 4–5. Respectively to the percent range calculated, Commerce will either apply A-A

to all sales, A-A to the sales that don’t pass Cohen’s d and A-T to the sales that do pass Cohen’s

d (i.e., “mixed methodology”), or A-T to all sales. See id. at 5.

Next, Commerce “examine[s] whether using only the [A-A] method can appropriately

account for such differences” by determining “whether using an alternative method . . . yields a

meaningful difference in the weighted-average dumping margin as compared to” A-A without

zeroing. Id. Commerce considers a difference “meaningful” if “there is a 25 percent relative

change in the weighted-average dumping margin between the [A-A] method and the appropriate

alternative method when both rates are above the de minimis threshold, or . . . the resulting

weighted-average dumping margin moves across the de minimis threshold.” Id. If the difference

is meaningful as defined by Commerce, then Commerce determines that the default A-A without

zeroing methodology cannot account for the price differences and that use of an alternative

method is appropriate. Id.

Relevant to this appeal, after applying the DP analysis, Commerce calculated weighted-

average dumping margins for NTN Corporation and NTN Kongo Corporation of 6.37 percent,

for NSK Japan of 2.79 percent, and for NSK UK of 6.47 percent. 3 Remand Results at 41, 42.

3
For the antidumping duty (“AD”) order for Japan, Commerce assigned to the non-selected
respondents a rate of 4.58 percent, which was “the simple-average of the weighted average
dumping margin of NSK Japan and NTN, the two mandatory respondents[.]” Final Results of
Remand Redetermination 9, ECF No. 85 (“Remand Results”). In addition, Commerce
individually reviewed Asahi Seiko Co., Ltd. (“Asahi”) and Mori Seiki Co., Ltd. (“Mori Seiki”),
applying the differential pricing (“DP”) analysis to calculate a margin. Id. at 5–6. After
determining that each company had a pattern of prices for comparable merchandise that differed
significantly and that under the ratio test the value of sales that passed the Cohen’s d test fell
between 33 percent and 66 percent of total value, Commerce assigned a weighted-average
(continued . . . )
Consol. Court No. 14-00155 Page 6

For each of these companies, Commerce determined that under the Cohen’s d test, prices

differed significantly, and it determined that under the ratio test, the value of sales that passed the

Cohen’s d test fell between 33 percent and 66 percent of total value. Id. at 7–8. Commerce then

concluded that A-A without zeroing could not appropriately account for such price differences

for each of the companies. Id. Thus, the margins increased from the Final Results, where

Commerce had assigned zero margins to all companies from Japan and the United Kingdom.

Final Results, 79 Fed. Reg. at 35,313–14.

NTN Precision America, Inc., NTN Bearing Corp. of America, NTN Corporation, NTN

Bower, Inc., NTN Driveshaft, Inc., and American NTN Bearing Manufacturing Corp.

(collectively, “NTN” 4) challenge Commerce’s Remand Results. NTN argues that Commerce

deprived it of a meaningful opportunity to comment on the Draft Results of Remand

Redetermination, bar code 3296505-01 (Aug. 6, 2015) (“Draft Remand Results”). Def.-Intrvnr.

NTN’s Cmts. on Final Results of Remand Redetermination 2–7, ECF No. 118 (“NTN Cmts.”).

NTN also argues that Commerce abused its discretion by applying the DP analysis because the

methodology itself is erroneous, id. at 22–24, 28–29, and Commerce’s specific application of its

dumping margin of 1.33 percent to Asahi and of 0.65 percent to Mori Seiki. Id. at 40, 41. For
the AD order for the United Kingdom, Commerce assigned to “responding companies which [it]
did not select for individual examination” a margin of 6.47 percent, which was the margin
assigned to NSK UK—the only company that was individually examined. Id. at 9.
4
The court includes NTN Kongo Corporation, for which Commerce calculated an AD rate, when
it refers to “NTN” as there is no meaningful difference between the entities for the purposes of
this opinion.
Consol. Court No. 14-00155 Page 7

DP analysis in this case to NTN’s sampled sales database 5 was unlawful, id. at 7–22. NTN

further contends that Commerce failed to adequately explain why the default methodology could

not take into account price differences, id. at 24–28. 6

The government and Timken respond that NTN was provided sufficient notice of

Commerce’s application of its DP analysis and therefore had a meaningful opportunity to

respond to the Draft Remand Results. Def.’s Resp. to Cmts. Regarding the Remand

Redetermination 4–6, ECF No. 109 (“Gov. Resp.”); Pl.’s Cmts. in Supp. of the Final Results of

Remand Redetermination 7–12, ECF No. 110 (“Timken Resp.”). They maintain that

Commerce’s DP analysis is lawful, refuting each of NTN’s general challenges, Gov. Resp. at 6–

9, 13–15; Timken Resp. at 6–7, 23–27, and arguing that application of the DP analysis to NTN’s

sampled database was proper, Gov. Resp. at 16–25; Timken Resp. at 16–22, 27–30. They add

5
In this case and in most of the previous reviews under the AD order for imports of ball
bearings, Commerce has employed a sampling methodology for U.S. sales where it randomly
selects a total of six-weeks over the 52-week period of review, selecting one week from each
two-month period of the period of review. See Remand Results at 14 n.25, 18; NTN’s Cmts. on
Draft Remand Redetermination at Ex. 1, bar code 3299960-01 (Aug. 20, 2015) (“NTN Draft
Remand Cmts.”). The respondents then report U.S. sales information for the selected six weeks,
only.
6
NSK Ltd., NSK Corporation, NSK Precision America, Inc., NSK Bearings Europe, Ltd., and
NSK Europe Ltd. (collectively, “NSK”) and NTN continue to disagree with the court’s remand
order, and they argue that Commerce was not required to apply its DP analysis in this case.
Cmts. on the Dep’t’s Results of Remand Redetermination on Behalf of NSK 1–2, ECF No. 90
(“NSK Cmts.”); Def.-Intrvnr. NTN’s Cmts. on Final Results of Remand Redetermination 1–2,
ECF No. 118 (“NTN Cmts.”). As indicated, the court has already held that Commerce acted
arbitrarily by not applying its DP analysis and finds no reason to revisit that decision. See
Timken, 79 F. Supp. 3d at 1357–60. NSK and NTN have not provided the court with any
additional reasons as to why its prior holding is incorrect. Further, NSK acknowledges that
Commerce’s “Remand Results are consonant with this Court’s Opinion and order.” NSK Cmts.
at 2.
Consol. Court No. 14-00155 Page 8

that certain of NTN’s arguments were not properly exhausted. Gov. Resp. at 22; Timken Resp.

at 12–16. The government also responds that Commerce properly explained why the default

methodology could not account for the pattern of prices that differed significantly. Gov. Resp. at

9–11.

JURISDICTION AND STANDARD OF REVIEW

The court has jurisdiction pursuant to 28 U.S.C. § 1581(c) (2012). The court upholds

Commerce’s determination in an administrative review unless it is “unsupported by substantial

evidence on the record, or otherwise not in accordance with law.” 19 U.S.C.

§ 1516a(b)(1)(B)(i).

DISCUSSION

I. NTN’s Extension Request

NTN argues that it was deprived of a meaningful opportunity to comment on the Draft

Remand Results because it did not receive electronic notification of the filing of the Draft

Remand Results and, NTN should have been provided more than seven days to comment

because Commerce’s DP methodology is a complicated scientific technique not previously

applied in administrative reviews of this AD order. NTN Cmts. at 2–7. Thus, NTN requests

remand and for the court to order reopening of the record so that it may further develop its

arguments. Id. at 7. The government and Timken respond that Commerce corrected for the lack

of electronic notification by restarting the original seven-day comment period, and NTN

provided substantive comments. Gov. Resp. at 6; Timken Resp. at 9–12. The government also

contends that NTN had notice that Commerce would apply the then well-known DP analysis,

pursuant to the court’s remand order. Gov. Resp. at 4–5; see also Timken Resp. at 7–9.
Consol. Court No. 14-00155 Page 9

Commerce’s discretion in setting time limits to comment on draft results of

redetermination is broad, as there is no statute or regulation governing its conduct in this

situation. 7 When Commerce, however, provides interested parties an opportunity to comment, as

it did here, Commerce’s discretion is limited because “[w]here a right to be heard exists, due

process requires that right be accommodated at a meaningful time and in a meaningful manner.”

See Mid Continent Nail Corp. v. United States, 34 CIT 512, 517, 712 F. Supp. 2d 1370, 1375

(2010) (quoting Barnhart v. U.S. Treasury Dep’t, 588 F. Supp. 1432, 1438 (CIT 1984) (alteration

in original)).

Commerce provided NTN with a meaningful opportunity to comment on the Draft

Remand Results. Based on the facts of this case, NTN’s argument, regarding the failure of

Commerce’s electronic filing system, ACCESS, to send automated notification to NTN’s

counsel, is irrelevant. When Commerce filed its Draft Remand Results on ACCESS on August

6, 2015, it provided interested parties seven days to comment; and when Commerce learned of

the electronic notification issue after receiving extension requests from NTN and Asahi Seiko

Co., Ltd. (“Asahi”), it granted a seven-day extension until August 20, 2015 to file comments.

Draft Remand Results at 12; Letter from Commerce to All Interested Parties at 1, bar code

3298225-01 (Aug. 13, 2015). Commerce provided a remedy that fully addressed the lack of

earlier electronic notification. Moreover, although Commerce provided NTN with only seven

7
Neither the court’s order nor the statute expressly required Commerce to issue its Draft
Remand Results: Commerce was required to file only its final Remand Results with the court.
See Golden Dragon Precise Copper Tube Grp., Inc. v. United States, Slip Op. 16-17, 2016 WL
720659, at *3 (CIT Feb. 22, 2016) (“Commerce is permitted to establish remand procedures
encouraging parties to meaningfully participate in the administrative proceeding.”) (citing Vt.
Yankee Nuclear Power Corp. v. Nat. Res. Def. Council, Inc., 435 U.S. 519, 543–44 (1978)).
Consol. Court No. 14-00155 Page 10

days to comment on the Draft Remand Results, this amount of time was not unreasonable. See,

e.g., Sichuan Changhong Elec. Co. v. United States, 30 CIT 1886, 1892, 466 F. Supp. 2d 1323,

1329 (2006) (upholding a four-day comment period). NTN knew more than a month before

comments were initially due that the court had ordered Commerce to apply the DP analysis. See

Timken, 79 F. Supp. 3d at 1350, 1361. NTN, therefore, had an opportunity to familiarize itself

with that methodology. Indeed, NTN submitted thoughtful comments regarding the DP analysis

in general and specifically as to its application to NTN’s sampled database. 8 See NTN’s

Comments on Draft Remand Redetermination at 3–14, bar code 3299960-01 (Aug. 20, 2015)

(“NTN Draft Remand Cmts.”); see also Mid Continent Nail, 34 CIT at 517, 712 F. Supp. 2d at

1375 (“If, however, a [party] makes thoughtful comments that Commerce addresses in its

determination, then, ‘as a practical matter [the party] was not substantially deprived of an

opportunity to be heard before the agency.’” (quoting Borden Inc. v. United States, 23 CIT 372,

375 n.3 (1999))). Accordingly, NTN was provided an opportunity and actually did meaningfully

comment on Commerce’s Draft Remand Results; therefore, its request to remand again and order

reopening of the record is denied.

8
In addition, Commerce’s regulations provide that “[u]nless expressly precluded by statute,
[Commerce] may, for good cause, extend any time limit established by this part.” 19 C.F.R.
§ 351.302(b) (2013). As discussed above, however, time limits to comment on draft results of
redetermination are not provided for in Commerce’s regulations, meaning they are not
“established by this part.” Therefore, in the present situation, it is not clear, and the parties have
not argued, whether it is Commerce’s practice to grant extensions for “good cause.”
Nevertheless, Commerce did grant a seven-day extension, for what is presumably good cause
due to the failure of Commerce’s electronic notification system. See Letter from Commerce to
All Interested Parties at 1, bar code 3298225-01 (Aug. 13, 2015). NTN has not established that
there was good cause to extend the deadline for an additional seven days, given that NTN knew
Commerce would be applying its DP methodology as ordered by the court.
Consol. Court No. 14-00155 Page 11

II. Differential Pricing

Commerce’s regulations provide that in an administrative review it will use the A-A

method “unless [it] determines another method is appropriate in a particular case.” 19 C.F.R.

§ 351.414(c)(1) (2013). To determine whether an alternative comparison methodology is

appropriate, Commerce has decided that its “examination of this question in the context of

administrative reviews, . . . is, in fact, analogous to the issue in antidumping investigations.”

Remand Results at 3; see also Final Modification, 77 Fed. Reg. at 8,102 (“[W]hen conducting

reviews under the modified methodology, [Commerce] will determine on a case-by-case basis

whether it is appropriate to use an alternative comparison methodology by examining the same

criteria that [Commerce] examines in original investigations pursuant to section 777A(d)(1)(A)

and (B) of the Act.”). Therefore, the statutory language governing investigations is instructive.

The statute provides that Commerce may determine whether dumping is occurring—that

is “the subject merchandise is being sold in the United States at less than fair value”—by using

the A-T method if: “(i) there is a pattern of export prices (or constructed export prices) for

comparable merchandise that differ significantly among purchasers, regions, or periods of time,

and (ii) [Commerce] explains why such differences cannot be taken into account using” A-A or

transaction-to-transaction (“T-T”). 9 19 U.S.C. § 1677f-1(d)(1)(B). Because the statute is silent

with regard to which comparison methodology Commerce is to use in administrative reviews and

how Commerce should determine which alternative methodology to use, see JBF RAK LLC v.

9
Under the transaction-to-transaction (“T-T”) methodology, which Commerce uses “in unusual
situations,” Commerce “compar[es] normal values of individual transactions to the export prices
of individual transactions for comparable merchandise.” Final Modification, 77 Fed. Reg. at
8,101 n.7.
Consol. Court No. 14-00155 Page 12

United States, 790 F.3d 1358, 1368 (Fed. Cir. 2015), Commerce’s approach is deemed lawful if

its “methodology and procedures are reasonable means of effectuating the statutory purpose, and

there is substantial evidence in the record supporting [its] conclusions,” Ceramica

Regiomontana, S.A. v. United States, 10 CIT 399, 404–05, 636 F. Supp. 961, 966 (1986), aff’d

810 F.2d 1137 (Fed. Cir. 1987).

A. Commerce’s Methodology

1. General Challenges

NTN argues that Commerce’s DP analysis is “fatally flawed” because it double counts

the sales that “pass” the Cohen’s d test, NTN Cmts. at 22–23; it captures tiny differences such as

exchange rate fluctuations that are not the result of targeted dumping, id. at 23–24; and it

compares prices between CONNUMs rather than comparing prices to particular customers,

regions, or time periods, id. at 24. NTN also challenges the methodology as unlawfully

incorporating zeroing in administrative reviews, arguing that zeroing is World Trade

Organization (“WTO”) inconsistent. Id. at 28–29.

Timken argues that the validity of the DP analysis is not properly before the court, and, in

any event, the court has already sustained Commerce’s use of DP. Timken Resp. at 23–24. The

government and Timken also both disagree with NTN’s challenges and respond that the DP

analysis does not double count because the value of the sales that pass are only included once in

the numerator of the ratio test. Gov. Resp. at 8; Timken Resp. at 25–26. They also contend that

even if the DP analysis accounts for tiny differences, the statute does not require Commerce to

consider the reason why prices differ significantly. Gov. Resp. at 8–9; Timken Resp. at 24–25.

They further dispute NTN’s argument relating to CONNUMs as failing to acknowledge that the
Consol. Court No. 14-00155 Page 13

statutory language, adopted in reviews, guides Commerce to consider a pattern of export prices

“for comparable merchandise.” Gov. Resp. at 9–10; Timken Resp. at 26–27. Both parties argue

that the use of zeroing has been upheld in the DP context and other contexts, with the

government arguing that the court is not in a position to determine whether Commerce is

violating the United States’ WTO obligations. Gov. Resp. at 13–14; Timken Resp. at 6–7.

Timken is incorrect that the validity of the DP analysis is not subject to review by the

court. Although Timken is correct that an intervening party is not allowed to bring claims

“clearly beyond the scope of the original litigation,” see Torrington Co. v. United States, 14 CIT

56, 59, 731 F. Supp. 1073, 1076 (1990), the scope of the original litigation centered on whether

Commerce should have applied its DP analysis. The validity of Commerce’s DP analysis is

plainly within the scope of Timken’s complaint, which requests that Commerce apply its DP

analysis. See Compl. ¶ 30, ECF No. 7. If Commerce has constructed the methodology in a way

that is invalid and therefore unlawful, Commerce may not apply it. It follows then that NTN, for

whom Commerce originally assigned a zero margin when it did not apply the DP analysis, but on

remand received a non-de minimis margin when Commerce did apply DP, could challenge the

validity and application of the DP analysis employed. Thus, NTN has not impermissibly

enlarged the scope of the original litigation, as framed by the parties. Further, the validity of the

DP methodology as applied to this case could not be addressed by the court until it actually was

applied to the facts of this case. 10

10
It was appropriate, however, for NTN to have raised some general, non-case specific
arguments prior to remand, most particularly its zeroing arguments. Such arguments, which are
(continued . . . )
Consol. Court No. 14-00155 Page 14

NTN’s general challenges to Commerce’s DP analysis, however, fail. First, NTN’s

double counting claim is unpersuasive. The purpose of Commerce’s DP analysis is to find a

pattern of prices that differ significantly; therefore, in NTN’s example situation 11 where there are

only two purchasers, A and B, if the prices to purchaser A differ significantly from prices to

purchaser B and vice-versa, then it is reasonable that sales to both purchasers will be counted as

passing Cohen’s d. As Commerce explains, these sales are then included in both the numerator

and the denominator of the ratio, and, importantly, the value of these sales is included only once

in the numerator and once in the denominator. See Remand Results at 26–27. Under

Commerce’s methodology, even if some sales are included in a test group and later in a

comparison group, their value is counted only once in the numerator of the ratio if they pass

Cohen’s d. Moreover, NTN’s concern that the number of sales that pass Cohen’s d will be

artificially inflated cannot be taken as a truism because, as Commerce correctly explains, if in the

hypothetical above, sales to purchasers A and B both do not pass Cohen’s d, then the value of

essentially facial challenges, were known and are deemed waived, but they are addressed here
for completeness.
11
NTN’s example provides:

If there are two customers (A and B) for a CONNUM, [Commerce] compares
sales to each customer. If sales to A differ significantly from sales to B, sales to
A pass the Cohen’s d test. Then, however, [Commerce] examines sales to B,
which, because they differ significantly from sales to A, will also pass the test.
This effectively applies the test to the same sales twice (one with a positive
Cohen’s d and one with the negative Cohen’s d of the same value) . . . [and]
unfairly inflates the number of sales that “pass” the Cohen's d test.

NTN Cmts. at 23.
Consol. Court No. 14-00155 Page 15

both will be included in the denominator of the ratio test. See id. at 27. Therefore, in the second

situation, sales passing Cohen’s d will be reduced.

Second, NTN has not established that Commerce’s DP analysis, which may capture

differences in prices that are not the result of targeted dumping, is unlawful. Despite guidance

from the Statement of Administrative Action (“SAA”), which indicates that A-T was typically

used to avoid potentially concealing targeted dumping under an A-A methodology, see Uruguay

Round Agreements Act, Statement of Administrative Action, H.R. Doc. No. 103-316, vol. 1, at

843, reprinted in 1994 U.S.C.C.A.N. 4040, 4177–78, the U.S. Court of Appeals for the Federal

Circuit (“Federal Circuit”) has made clear that the statute does not require Commerce “to

determine the reasons why there is a pattern of export prices for comparable merchandise that

differs significantly,” JBF RAK, 790 F.3d at 1368. Therefore, in the light of the Federal

Circuit’s decision, whether NTN’s U.S. sales were not intentionally targeted or whether the price

differences were due to an external factor, such as shifting exchange rates, Commerce’s

methodology lawfully identifies a pattern of export prices that differ significantly. See Remand

Results at 24–25 (“The court has already found that the purpose or intent behind an exporter’s

pricing behavior in the U.S. market is not relevant to [Commerce’s] analysis.”). 12

12
NTN only provides a hypothetical and does not provide evidence that the price differences in
its sales were solely due to an external factor totally beyond an exporter’s control, such as
exchange rates, and not targeting. The Federal Circuit’s opinion in JBF RAK does not appear to
speak to a situation where a respondent actually demonstrates that the price differences are not
the result of targeting. Although the Federal Circuit expressed concern that “requiring
Commerce to determine the intent of a targeted dumping respondent would create a tremendous
burden on Commerce that is not required or suggested by the statute,” JBF RAK, 790 F.3d at
1368 (internal quotation marks omitted) (quoting JBF RAK LLC v. United States, 991 F. Supp.
2d 1343, 1355 (CIT 2014)), such a burden might not exist where a respondent itself provides that
(continued . . . )
Consol. Court No. 14-00155 Page 16

Third, Commerce’s decision to compare purchasers, regions, and time periods by

analyzing CONNUMs appears reasonable. The government correctly recognizes that the statute

governing investigations is helpful because it directs Commerce to use A-T when there is a

“pattern of export prices (or constructed export prices) for comparable merchandise that differ

significantly among purchasers, regions, or periods of time.” See Gov. Resp. at 9 (quoting 19

U.S.C. § 1677f-1(d)(1)(B)(ii)). Thus, in reviews as well as in investigations, Commerce looks

for a pattern of prices for comparable merchandise, a term that Commerce identifies by “product

control number[, i.e. CONNUM,] and any characteristics of the sales, other than purchaser,

region and time period, that [Commerce] uses in making comparisons between [export price] or

[constructed export price] and [normal value] for the individual dumping margins.” Remand

Results at 3. Then, Commerce compares the prices for comparable merchandise for a particular

purchaser, region, or time period to prices for all other purchasers, regions, or time periods. Id.

at 29. NTN fails to demonstrate that using CONNUMs as a basis for establishing “comparable

merchandise” is unreasonable; instead, it incorrectly believes that Commerce cannot limit its

price comparisons to comparable merchandise. See NTN Cmts. at 24. But, as discussed, NTN’s

information. Indeed, faced with the Statement of Administrative Action’s (“SAA”) clear
expression that the alternative comparison methodology is typically employed to address
targeted dumping, the substantial evidence standard might require a different result. Uruguay
Round Agreements Act, Statement of Administrative Action, H.R. Doc. No. 103-316, vol. 1, at
843, reprinted in 1994 U.S.C.C.A.N. 4040, 4177–78 (“New section [19 U.S.C. § 1677f-
1(d)(1)(B)] provides for a comparison of average normal values to individual export prices or
constructed export prices in situations where an average-to-average or transaction-to-transaction
methodology cannot account for a pattern of prices that differ significantly among purchasers,
regions, or time periods, i.e., where targeted dumping may be occurring.”). This is not the case
before the court.
Consol. Court No. 14-00155 Page 17

argument cannot be reconciled with the statutory language for investigations, which guides

Commerce’s practice here.

Fourth, Commerce lawfully employed zeroing as part of its DP analysis. The courts have

repeatedly held that zeroing is lawful in administrative reviews. See Union Steel v. United

States, 713 F.3d 1101, 1103, 1104 (Fed. Cir. 2013); Timken Co. v. United States, 354 F.3d 1334,

1342 (Fed. Cir. 2004). 13 The use of zeroing here is no different. Here, NTN’s concern derives

from the fact that when Commerce uses its “mixed methodology,” Commerce uses A-T with

zeroing on sales that pass Cohen’s d and then it allegedly applies zeroing again when it

determines the final AD margin by not allowing negative dumping margins to offset positive

margins for the targeted sales. NTN Cmts. at 28–29. But, as the court has recognized correctly,

the goal of measuring the effect of a pattern of prices that differ significantly is effectuated by

using A-T with zeroing at both stages and not by somehow diluting the A-T margin when it is

aggregated. See Apex Frozen Foods Private Ltd. v. United States, Slip Op. 16-9, 2016 WL

471948, at *20–21 (CIT Feb. 2, 2016). Commerce, in applying its mixed methodology,

13
These cases also address NTN’s argument that the application of zeroing is inconsistent with
the United States’ World Trade Organization (“WTO”) obligations. See NTN Cmts. at 29; Def.-
Intrvnr. NTN’s Reply Cmts. in Opp’n to the Final Results of Remand Redetermination 7, ECF
No. 116 (“NTN Reply”) (citing Panel Report, United States—Anti-Dumping and Countervailing
Measures on Large Residential Washers from Korea ¶ 8.1.a.xii, WT/DS464/T (Mar. 11, 2016)
(“United States—Washers from Korea”)). In the case cited by NTN, the WTO Panel simply
extends the same reasoning for why it believed zeroing to be WTO-inconsistent in the past. See
United States—Washers from Korea at ¶ 7.206 (“We consider that the use of zeroing in the
context of the [A-T] comparison methodology would not lead to a fair comparison, . . . [and]
therefore find that the use of zeroing in the context of the [A-T] comparison methodology is ‘as
such’ inconsistent with Article 2.4.”). The Federal Circuit has rejected this same reasoning as
unpersuasive to show that Commerce’s practice of zeroing in reviews is unreasonable under the
U.S. statute. Timken Co. v. United States, 354 F.3d 1334, 1343–44 (Fed. Cir. 2004).
Consol. Court No. 14-00155 Page 18

reasonably and “proportionately applies the remedy across the sales.” Id. Substantial record

evidence, such as the 54.7 percent of NTN’s sales that pass Cohen’s d, Remand Results at 8,

supports the proposition that zeroing should be applied to those 54.7 percent sales so that

Commerce may measure the effect of the concealed dumping. See Apex, 2016 WL 471948, at

*18 (“[W]ithout zeroing the A-A and A-T comparison methodologies would always be

mathematically equivalent, obviating any benefit derived from having an alternative comparison

methodology in the statute.” (internal quotation omitted)). Thus, NTN’s general challenges to

Commerce’s DP analysis in this case are without merit.

2. Application to NTN’s Sampled Sales Database

NTN challenges Commerce’s application of its DP analysis to NTN’s sampled sales

database, arguing that the sampled database is not representative of NTN’s overall sales. NTN

Cmts. at 7–13. NTN also argues that Commerce was required to test the validity of the sampled

sales database under general statistical principles before it could use the database in its DP

analysis. Id. at 13–16. It further contends that Commerce abused its discretion by applying the

DP analysis to a subset of sales because certain portions of the DP analysis requires

consideration of a respondent’s total sales, id. at 20–22, and that Commerce could not ascertain a

pattern from the data provided in the six-week sample, id. at 17–20.

The government and Timken respond that NTN failed to exhaust certain arguments by

failing to raise them before Commerce. Gov. Resp. at 17, 22; Timken Resp. at 12–16, 21. They

dispute NTN’s arguments pertaining to the statistical validity of the sample and the general

statistical principles. Gov. Resp. at 16, 21–25; Timken Resp. at 21–22, 27–30. They also argue

that Commerce did not abuse its discretion in applying its DP analysis to a sampled sales
Consol. Court No. 14-00155 Page 19

database and that the application in this case was lawful. Gov. Resp. at 18–21; Timken Resp. at

18–21.

As a preliminary matter, NTN did fail to exhaust certain arguments in its comments on

Commerce’s Draft Remand Results. The court “shall, where appropriate, require the exhaustion

of administrative remedies.” 28 U.S.C. § 2637(d). Exhaustion “protect[s] administrative agency

authority and promote[s] judicial efficiency.” Itochu Bldg. Prods. v. United States, 733 F.3d

1140, 1145 (Fed. Cir. 2013) (citing McCarthy v. Madigan, 503 U.S. 140, 145 (1992)). Before

the court, NTN asserts that Commerce’s methodology of randomly selecting six weeks of

sampled sales 14 is not statistically valid, NTN Cmts. at 7–13, and argues that Commerce’s use of

an unrepresentative sampled database in its DP analysis is unlawful because the database must

fulfill key statistical “assumptions of normality, independence of observations, [and] variances,”

id. at 13–16. Neither of these arguments were properly raised and developed before

Commerce. 15 See NTN Draft Remand Cmts. at 1–14. Issues pertaining to Commerce’s

14
Here, Commerce employed the same random six-week sampling methodology that it
employed in the 1990–1991 administrative review of the AD order on ball bearings. See Nachi-
Fujikoshi Corp. v. United States, 19 CIT 914, 916, 917, 890 F. Supp. 1106, 1108–09 (1995); see
also NTN Draft Remand Cmts. at Ex. 1 (providing Commerce’s current sampling methodology).
NTN attempts to discredit Commerce’s sampling methodology by pointing out that in the 2010–
2011 administrative review (i.e., the review immediately following the present review),
Commerce stopped sampling and used each respondents’ entire sales database. NTN Cmts. at
11–12. But, Commerce’s change in methodology is insufficient to show its present sampling
methodology is unlawful. See Apex Frozen Foods Private Ltd. v. United States, 37 F. Supp. 3d
1286, 1302 n.7 (CIT 2014) (“If an agency practice reasonably complies with statute, that practice
is not rendered invalid simply because the agency replaces it with an equally defensible
policy.”).
15
With regards to the statistical validity, the statute provides that Commerce may use
“statistically valid samples, if there is a significant volume of sales of the subject merchandise”
(continued . . . )
Consol. Court No. 14-00155 Page 20

sampling and DP methodologies are areas where requiring exhaustion is especially appropriate

because these issues relate to “complex economic and accounting decisions of a technical nature,

for which agencies possess far greater expertise than courts.” See Fujitsu Gen. Ltd. v. United

States, 88 F.3d 1034, 1039 (Fed. Cir. 1996). Therefore, the court declines to consider these

arguments.

Further, Commerce did not abuse its discretion when it applied the DP analysis to a

subset of sales. Although NTN points to language from Commerce’s explanation of its DP

methodology to argue that Commerce can conduct a DP analysis only when it has all sales data,

see NTN Cmts. at 20–22, the sampled sales database is the universe of sales on the record and

and Commerce has the exclusive “authority to select . . . statistically valid samples.” 19 U.S.C.
§§ 1677f-1(a), (b). NTN did state in its comments before Commerce that Commerce

has never stated that these sample [six] weeks are in any way representative of the
entire year of bearing sales, or that they are statistically-valid samples, and it [sic]
unlikely that they ever could be. The sample weeks are chosen by [Commerce]
without knowledge of the entire universe of sales, are different each year, and are
not evenly spaced throughout the year.

NTN Draft Remand Cmts. at 5. NTN, however, in its comments before Commerce, does not cite
to 19 U.S.C. § 1677f-1(a) or (b), does not provide factual support from the record to explain why
Commerce’s approach is not statistically valid, and does not attempt to establish the limitations
imposed by the phrase “statistically valid.” Therefore, it does not appear that NTN’s statement
was sufficient to apprise Commerce of the fact that NTN was attempting to challenge the
sampling methodology as a violation of 19 U.S.C. § 1677f-1(a) or (b) and any such challenge is
waived. Regardless, absent any fact-based argument which would cause it to reach a different
conclusion, the court adheres to Nachi-Fujikoshi, which held that Commerce’s bearings
sampling methodology employed in this case, where Commerce randomly selects one week from
each two-month interval of the period of review, is a reasonable exercise of agency discretion
under a previous, but similarly-worded version of the statute. See 19 CIT at 918–19, 890 F.
Supp. at 1109–10.
Consol. Court No. 14-00155 Page 21

Commerce reasonably applied its DP analysis to the sales information that it had. 16 NTN knew

before remand that Commerce had a sampled database, even objecting to the DP analysis as

invalid based on that limitation, NTN Draft Remand Cmts. at 5, but it did not ask Commerce to

reopen the record to obtain full reporting of sales and it did not ask the court to require it. It

never offered its full database and cannot now complain about its absence. 17

Given that the court has already upheld Commerce’s sampling methodology employed in

this case as a basis for calculating a weighted-average dumping margin, it appears reasonable for

Commerce to use the same six-week sample for purposes of calculating a weighted-average

dumping margin when conducting a DP analysis. The relevant statutory language, covering

targeting in investigations, does not require a different conclusion as it does not limit the type or

representativeness of databases that Commerce may use in its pattern analysis. See 19 U.S.C.

§ 1677f-1(d)(1)(B).

NTN’s argument that Commerce cannot ascertain a “pattern” from the sampled database

in this case fails. As explained, Commerce’s methodology requires that the test group and the

comparison group each have two observations before Commerce can compare the two groups.

Remand Results at 4. If they do not have at least two observations each, then Commerce cannot

compare the two groups and the value of the sales in those groups is included in the denominator

of the ratio test. Thus, NTN’s argument that Commerce used “point estimates” and “discerned a

16
Thus, where the DP methodology calls for Commerce to consider “total sales” or “all sales,”
Commerce has interpreted that to mean total or all reported sales (i.e., all sales covered in the six
sampled weeks). See Remand Results at 4–5.
17
In fact, it has not established that it exists.
Consol. Court No. 14-00155 Page 22

‘pattern’ based on the comparison of one sale to one other sale” is incorrect. 18 See NTN Cmts. at

17–18. Instead, Commerce disqualified those sales and, by putting their value in the

denominator of the ratio test, decreased the likelihood that the DP analysis would show a pattern

of prices that differed significantly. 19 Commerce’s application of its DP analysis to NTN’s sales

was lawful. 20

18
NTN also argues that a pattern discerned from comparing two point estimates would be
“unreliable for projecting generalizations for the entire year of pricing behavior.” NTN Cmts. at
18. NTN fails to provide information on how often Commerce’s DP analysis relied on only two
observations for each group. Regardless, NTN’s argument also relies on the mistaken belief that
Commerce was required to determine a pattern for all sales. As the court has explained,
Commerce properly treated the sampled sales database as the universe of sales in conducting its
DP analysis.
19
This situation differs importantly from a challenge to Commerce’s DP analysis that required
remand by the court. There, a domestic producer successfully challenged Commerce’s
explanation of its ratio test thresholds because Commerce’s inclusion of disqualified sales in the
denominator of the ratio test decreased the benefit derived by the domestic producer by
decreasing the overall value of sales that might receive the A-T remedy. See U.S. Steel Corp. v.
United States, No. 14-00263, Slip Op. 16-44, at *14–17 (CIT May 5, 2016). NTN, on the other
hand, stands to benefit from Commerce’s disqualification of these sales, and Timken, the
domestic producer, does not challenge Commerce’s approach. In addition, although NTN argues
that significant quantity of sales CONNUMs had only one or two observations, NTN Cmts. at
17, sales value is what is important for the ratio test. As Commerce recognized, the sales value
for these disqualified sales that were included in the denominator was quite small as a percentage
of overall sales value, see Remand Results at 20, and is insufficient in this case to render
Commerce’s decision unlawful.
20
NTN couches many of its challenges to Commerce’s DP analysis using the terms
“representative” and “statistically valid.” These terms appear to derive from the portions of the
statute that discuss the method by which Commerce may select samples (or may average). See
19 U.S.C. §§ 1677f-1(a)–(b). In fact, the word “representative” does not appear in the current
version of the statute but instead derives from an old version. See 19 U.S.C. § 1677f-1(b)
(1984). Regardless, once Commerce appropriately selects a sample, these provisions do not
speak to how Commerce must use that sample (e.g., in a DP analysis). Therefore, these statutory
provisions do not further NTN’s argument.
Consol. Court No. 14-00155 Page 23

B. Commerce’s Explanation Requirement

NTN requests that the court remand the case to Commerce because it argues that

Commerce has not explained why the default methodology, A-A without zeroing, cannot account

for the price differences. NTN Cmts. at 24–27. NTN also argues that any difference in the

margin derived from each of the methods is mostly attributable to the effect of zeroing used in

A-T. Id. at 27–28. The government responds that its explanation, where it explained by

reference to an example, in this case was adequate. Gov. Resp. at 11–13.

Commerce’s explanation of why A-A cannot account for the price differences here is

adequate. 21 The present case is different from other cases where the court determined that

Commerce provided a conclusion rather than an adequate explanation under 19 U.S.C. § 1677f-

1(d)(1)(B)(ii). See, e.g., Beijing Tianhai Indus. Co. v. United States, 106 F. Supp. 3d 1342, 1351

(CIT 2015) (rejecting Commerce’s explanation because it relied on “confirmation bias” where it

in essence reasoned that “because substantial dumping was not found using A-A, but [was]

found using A-T, it was permissible for [Commerce] to use the alternative A-T methodology.”);

Beijing Tianhai Indus. Co. v. United States, 7 F. Supp. 3d 1318, 1332 (CIT 2014) (holding that

Commerce failed to adequately explain where Commerce simply stated that A-T was necessary

because any pattern of price differences was hidden using A-A). Commerce has explained that

21
Although the statute for investigations also requires Commerce to explain why the transaction-
to-transaction (“T-T”) methodology cannot account for the pattern of prices that differ
significantly, see 19 U.S.C. § 1677f-1(d)(1)(B)(ii), Commerce, in the present administrative
review, has only provided a reason why A-A is not appropriate. No party has argued that
Commerce was required by its practice to explain why T-T is not an appropriate comparison
method in an administrative review and it is fairly obvious that this product is not the kind of
large-scale item for which T-T might be used.
Consol. Court No. 14-00155 Page 24

“there must be a significant and meaningful difference in U.S. prices in order to resort to an

alternative comparison method,” indicating that an alternative comparison is appropriate where,

in comparing the alternative methods, “the weighted-average dumping margin will change by at

least 25 percent or the weighted-average dumping margin will change [from non-de minimis] to

be de minimis.” 22 Remand Results at 33.

According to Commerce, in such instances, “the normal value is in the middle of the

range of individual U.S. prices such that there is both a significant amount of dumping and a

significant amount of offsets generated from non-dumped sales,” id. at 32, and, therefore,

Commerce will “find that the A-A method is not appropriate,” id. at 33. Commerce reasons that

in this situation A-T is appropriate because the significant differences in U.S. price are “large

enough” such that “not only is there a non-de-minimis amount of dumping, but that there also is

a meaningful amount of offsets to impact the identified amount of dumping.” Id. at 33.

Commerce’s rationale, therefore, recognizes that in this situation, A-T can account for this by not

allowing these offsets to dilute the weighted-average dumping margin in the same way that A-A

might. See id. at 31 (“The comparison of a dumping margin based on weighted-average U.S.

price . . . precisely examines the impact on the amount of dumping which is hidden or masked.”).

Such an explanation is reasonable and demonstrates why Commerce believes A-A, which allows

22
In reviews, Commerce treats “as de minimis any weighted-average dumping margin . . . that is
less than 0.5 percent ad valorem, or the equivalent specific rate.” 19 C.F.R. § 351.106(c).
Moreover, at oral argument, the government noted that NTN’s non-de minimis rate of 6.37
percent using A-T with zeroing constituted more than a 25 percent relative change in the
dumping margin. As discussed in the Remand Results, NTN’s rate also crossed over from de
minimis using A-A without zeroing to non-de minimis using A-T with zeroing. Remand Results
at 8. NTN does not refute that the change in its dumping margin was meaningfully or
significantly different, as defined by either of Commerce’s criteria.
Consol. Court No. 14-00155 Page 25

for offsets and might mask significant price differences, cannot account for these price

differences. See Apex, 2016 WL 471948, at *17 n.24 (“The court can discern from Commerce’s

explanation that A-A cannot account for the pattern of significant price differences because A-A

masked the dumping that was occurring as revealed by the A-T calculated margin.”).

NTN’s other arguments also lack merit. As previously discussed, although NTN argues

that Commerce’s use of zeroing with A-T creates artificially higher dumping margins, NTN

Cmts. at 27–28, NTN has not explained why the use of zeroing in this context is unlawful.

Indeed, Commerce’s methodology appears to be a lawful attempt to effectuate the goal of the

statute by measuring the effect of the pattern of prices that differ significantly. See Apex Frozen

Foods Private Ltd. v. United States, 37 F. Supp. 3d 1286, 1296 (CIT 2014) (“[B]y comparing

Apex’s nonzeroed A-A rate to its zeroed A-T rate, the agency found the precise amount of

dumping—including dumping from the targeted sales—that A-A masked. Commerce could then

decide whether that dumping was great enough to merit an exceptional remedy.”). Moreover,

NTN’s citation to United States—Anti-Dumping and Countervailing Measures on Large

Residential Washers from Korea, WT/DS464/T (Mar. 11, 2016) (“United States—Washers from

Korea”) is unconvincing. There, the WTO Panel interpreted a similarly-worded but ultimately

different provision, Article 2.4.2 of the Agreement on Implementation of Article VI of the

General Agreement on Tariffs and Trade 1994 (“AD Agreement”). United States—Washers

from Korea ¶ 7.73. The differences in the two provisions is evident: the WTO Panel clarified

that Article 2.4.2 requires “an investigating authority [to] analyse the prevailing factual

circumstances in order to consider the possibility that something other than targeted dumping is

responsible for these relevant price differences,” id., but the Federal Circuit has made clear that
Consol. Court No. 14-00155 Page 26

the U.S. statute does not require Commerce to investigate why a pattern of prices that differ

significantly exists, JBF RAK, 790 F.3d at 1368. In any event, whatever light a WTO decision

might eventually shed, the WTO Panel’s decision in United States—Washers from Korea is not

final as it has been appealed by the United States. U.S. Appeals Panel Report on Large

Residential Washers from Korea, World Trade Organization: 2016 News Items (Apr. 19, 2016),

available at https://www.wto.org/english/news_e/news16_e/ds464apl_19apr16_e.htm (last

visited Apr. 22, 2016). Thus, NTN’s reliance on the WTO’s interpretation of Article 2.4.2 is

misplaced.

CONCLUSION

For the reasons stated above, Commerce’s Remand Results are sustained. Judgment will

enter accordingly.

/s/ Jane A. Restani
Jane A. Restani
Judge

Dated: May 10, 2016
New York, New York