Hyundai Electronics Industries Co. v. United States

Slip Op. 05-105

UNITED STATES COURT OF INTERNATIONAL TRADE

BEFORE: RICHARD W. GOLDBERG, SENIOR JUDGE

HYUNDAI ELECTRONICS INDUSTRIES
CO., LTD. and HYUNDAI
ELECTRONICS AMERICA, INC.,

Plaintiffs,

v. PUBLIC VERSION

UNITED STATES, Cons. Court No. 00-01-00027

Defendant,

and

MICRON TECHNOLOGY, INC.,

Defendant-
Intervenor.

[Commerce antidumping duty remand determination sustained in part
and remanded in part.]

Dated: August 25, 2005

Willkie, Farr & Gallagher LLP (James P. Durling and Daniel L.
Porter) for Plaintiffs Hyundai Electronics Industries Co., Ltd.
and Hyundai Electronics America, Inc.

Peter D. Keisler, Assistant Attorney General, David M. Cohen,
Director, Jeanne E. Davidson, Deputy Director, Commercial
Litigation Branch, Civil Division, United States Department of
Justice (Kenneth S. Kessler); Patrick V. Gallagher, Jr., Of
Counsel, Office of the Chief Counsel for Import Administration,
United States Department of Commerce, for Defendant United
States.

King & Spalding LLP (Gilbert B. Kaplan and Daniel L.
Schneiderman) for Defendant-Intervenor Micron Technology, Inc.
Cons. Court No. 00-01-00027 Page 2

OPINION

GOLDBERG, Senior Judge: This case is before the Court following

remand to the United States Department of Commerce (“Commerce”).

In Hyundai Electronics Industries Co. v. United States, 28 CIT

__, 342 F. Supp. 2d 1141 (2004) (“Hyundai I”), familiarity with

which is presumed, the Court sustained in part and remanded in

part Commerce’s determination in the fifth administrative review

regarding Dynamic Random Access Memory semiconductors of one

megabit or above (“DRAMs”) from the Republic of Korea produced by

Hyundai Electronics Industries Co., Ltd. and Hyundai Electronics

America, Inc. (collectively “Hyundai”) and LG Semicon Co., Ltd.

(“LG Semicon”).1 See Dynamic Random Access Memory Semiconductors

of One Megabit or Above From the Republic of Korea: Final Results

of Antidumping Duty Administrative Review and Determination Not

To Revoke the Order in Part, 64 Fed. Reg. 69694 (Dec. 14, 1999)

(“Final Results”).

In Hyundai I, the Court found that Commerce was justified in

applying only partial adverse facts available (“AFA”) against LG

Semicon in determining its dumping margin. See Hyundai I, 28 CIT

at __, 342 F. Supp. 2d at 1155. The Court concluded that while

Commerce was correct in applying AFA against LG Semicon for its

1
After the fifth administrative review was completed,
respondent Hyundai acquired respondents LG Semicon Co., Ltd. and
LG Semicon America, Inc. (collectively “LG Semicon”). In this
opinion, Hyundai-as-successor-in-interest-to-LG Semicon is
referred to as LG Semicon.
Cons. Court No. 00-01-00027 Page 3

German sales to [ ] (“the customer”) because LG

Semicon knew or should have known that DRAMs sold to the customer

were destined for the United States, the use of total AFA was not

warranted because Commerce erred in using AFA for LG Semicon’s

Mexican sales to [ ]. Id. With

respect to Plaintiffs’ research and development (“R&D”) costs,

the Court held that Commerce had not adduced substantial evidence

to support its theory of cross-fertilization, which allowed the

inclusion of R&D expenditures for non-subject merchandise in

calculating the cost of producing the subject merchandise. See

id. at __, 342 F. Supp. 2d at 1157. Additionally, the Court

found that Commerce had not provided specific evidence

demonstrating why Plaintiffs’ amortized R&D costs did not

reasonably account for their actual R&D costs during the period

of review, or how Plaintiffs’ currently deferred R&D costs

affected production and revenue for the review period. Id. at

__, 342 F. Supp. 2d at 1159.2

The Court remanded the matter to Commerce with instructions

to: (1) recalculate LG Semicon’s dumping margin using the data

2
In addition to these holdings relevant to the issues
considered here, the Court in Hyundai I also found that: (1)
Commerce did not violate LG Semicon’s right to a fair and honest
proceeding; (2) Commerce’s calculation of Hyundai’s R&D cost
allocation ratio was reasonable; (3) Hyundai did not provide
sufficient evidence of double counting by Commerce; and (4)
Commerce’s treatment of Hyundai’s interest earned on severance
deposits was reasonable. See Hyundai I, 28 CIT at __, __, __,
__, 342 F. Supp. 2d at 1149-53, 1159-60, 1160, 1161-62.
Cons. Court No. 00-01-00027 Page 4

provided by LG Semicon for its Mexican sales, and applying AFA

only for LG Semicon’s sales to the customer’s German subsidiary;

(2) provide additional information specifically pointing to the

effect of non-subject merchandise R&D on the R&D for the subject

merchandise, or in the alternative, recalculate R&D costs on the

most product-specific basis possible; (3) provide specific

evidence explaining how Plaintiffs’ actual R&D costs for the

review period are not reasonably accounted for in their amortized

R&D costs, or in the alternative, accept Plaintiffs’ amortization

methodology; and (4) present substantial evidence demonstrating

how R&D costs for Plaintiffs’ long-term projects affect their

current projects for the period of review, or in the alternative,

accept Plaintiffs’ deferral methodology. See id. at __, __, __,

__, 342 F. Supp. 2d at 1155, 1157, 1159, 1159.

Commerce duly complied with the Court’s order. Commerce

issued draft Redetermination Results (Aug. 12, 2004) (“Draft

Remand Results”) and then, after receiving comments from

Plaintiffs and Defendant-Intervenor Micron Technology, Inc.

(“Micron”), the Final Results of Redetermination Pursuant to

Court Remand (Aug. 31, 2004) (“Remand Results”). In the Remand

Results, Commerce recalculated LG Semicon’s dumping margin and

applied a new rate of 89.10 percent, which Commerce concluded was

“the highest non-aberrational margin calculated for any U.S.

transaction for LG [Semicon] in the period of review[.]” Remand
Cons. Court No. 00-01-00027 Page 5

Results at 4. Commerce also complied with the Court’s request

for more information regarding its theory of cross-fertilization

by providing scientific articles, new expert testimony, and the

titles of some of Hyundai’s development projects. Id. at 4-5,

11-14. In addition, although it expressed disagreement with the

Court’s findings regarding amortization in Hyundai I, Commerce

stated that it could not provide specific evidence showing how

amortization did not reasonably account for Plaintiffs’ actual

R&D costs incurred during the period of review. Id. at 5. Thus,

Commerce recalculated Plaintiffs’ R&D costs to allow for

amortization. Id. Finally, Commerce continued to find that

Plaintiffs’ deferred R&D costs should be expensed in the period

incurred because Plaintiffs did not offer any reasonable evidence

demonstrating how their deferred costs would have discernible

future benefits. Id. at 6, 22.

Plaintiffs submitted Comments on the Final Results of

Redetermination (“Pls.’ Br.”), and Micron submitted a Memorandum

Addressing the Final Results of Redetermination Pursuant to Court

Remand (“Def.-Intvr.’s Br.”). Commerce then submitted its

Response to Plaintiffs’ and Defendant-Intervenor’s Comments.

Plaintiffs subsequently submitted Response Comments on the Final

Results of Redetermination, and Micron submitted a Response Brief

Addressing Plaintiffs’ Comments.

The Court has jurisdiction under 28 U.S.C. § 1581(c). The
Cons. Court No. 00-01-00027 Page 6

Court must uphold Commerce’s determination if it is supported by

substantial evidence and otherwise in accordance with law. 19

U.S.C. § 1516a(b)(1)(B)(i). After due consideration of the

parties’ submissions, the administrative record, and all other

papers had herein, and for the reasons that follow, the Court

sustains in part and reverses and remands in part.

I. DISCUSSION

A. Commerce’s Decision to Apply a Margin of 89.10 Percent as
Partial AFA Is Supported by Substantial Evidence and
Otherwise in Accordance with Law.

In Hyundai I, the Court held that Commerce was justified in

applying partial AFA to LG Semicon’s German sales, but not in

applying total AFA to LG Semicon’s entire U.S. sales database.

See Hyundai I, 28 CIT at __, 342 F. Supp. 2d at 1153-55. With

respect to LG Semicon’s German sales, the Court sustained

Commerce’s finding that LG Semicon knew or should have known that

the DRAMs it sold to the customer’s German subsidiary were

destined for the U.S. market, and that its failure to submit

these German sales as U.S. sales justified the use of AFA under

19 U.S.C. § 1677e(b). Id. at __, 342 F. Supp. 2d at 1155.

However, with respect to LG Semicon’s Mexican sales, the Court

found that Commerce did not meet the requisite standard for

applying AFA. Id. Thus, the Court ordered Commerce to

recalculate LG Semicon’s dumping margin using AFA only for LG

Semicon’s sales to the customer’s German subsidiary. Id.
Cons. Court No. 00-01-00027 Page 7

Commerce complied with the Court’s instructions and

calculated a new AFA rate of 89.10 percent. See Remand Results

at 2-4. In choosing this rate, Commerce stated that it selected

the highest non-aberrational margin calculated for any of LG

Semicon’s U.S. transactions during the period of review.3 See

id. at 4. Commerce also noted that the 89.10 percent rate fell

within “a range of margins for a large portion of LG [Semicon]’s

review period transactions that decrease[d] steadily by small

amounts.” Id. Finally, Commerce observed that the new rate was

sufficiently adverse to ensure that LG Semicon would not have

obtained a more favorable result by failing to cooperate than if

it had cooperated fully, and also furthered the statutory purpose

underlying the AFA rule to induce respondents to provide Commerce

with complete and accurate information in a timely manner. See

id.

1. The 89.10 Percent Rate Is Non-Aberrational.

Plaintiffs argue that Commerce failed to explain why the

89.10 percent rate is non-aberrational, while the other margins

above it are aberrational. Pls.’ Br. at 3. According to

Plaintiffs, Commerce’s discretion in applying AFA is not

unlimited; rather, Commerce must demonstrate why a particular AFA

rate is indicative of a respondent’s selling practices and

3
Calculated rates ranging as high as 223 percent were
actually available for LG Semicon on remand. See Remand Results
at 4.
Cons. Court No. 00-01-00027 Page 8

rationally related to its sales. Id. Here, Plaintiffs contend,

Commerce failed to demonstrate in the Remand Results that the

89.10 percent rate is indicative of LG Semicon’s sales, and

therefore non-aberrational. Id.

The Court disagrees for two reasons. First, the Court finds

that the 89.10 percent rate is inherently indicative of LG

Semicon’s selling practices because it was derived from LG

Semicon’s “own sales data from the instant review segment.”

Remand Results at 8. When Commerce utilizes a respondent’s own

sales data, it is afforded broad discretion in the selection of

the adverse rate, and this is true even if the selected rate is

reflective of only a small proportion of the respondent’s sales.

See Ta Chen Stainless Steel Pipe, Inc. v. United States, 298 F.3d

1330, 1339 (Fed. Cir. 2002). Thus, although the 89.10 percent

rate chosen by Commerce may be higher than other calculated

margins, this fact alone does not render the AFA rate

aberrational or unrelated to LG Semicon’s sales practices.

Second, the Court finds that Commerce did adequately explain

in the Remand Results why the 89.10 percent rate is non-

aberrational: “[T]he margin selected falls in a range of margins

for a large portion of LG [Semicon]’s review period transactions

that decrease steadily by small amounts.” Remand Results at 8.

The record evidence clearly shows that the selected 89.10 percent

margin is within a grouping of sales whose margins differ by very
Cons. Court No. 00-01-00027 Page 9

small amounts (e.g., 89.10 percent, [ ] percent, [ ]

percent, and [ ] percent), while several sales margins ([

]) above 89.10 percent range from [ ] percent to 223 percent

and do not form a cluster. See Confidential Administrative

Record (“Conf. Admin. R.”) at Ex. 8 (SAS Margin Program Log and

Output dated Aug. 28, 2004) at 54. In contrast to this record

evidence supporting Commerce’s determination, Plaintiffs did not

provide any evidence showing how the 89.10 percent rate is

aberrational, or otherwise unreflective of LG Semicon’s selling

practices. As a result, the Court finds that the 89.10 percent

rate selected by Commerce on remand is non-aberrational.

2. The 89.10 Percent Rate Is Not Unduly Punitive.

Plaintiffs also argue that the 89.10 percent rate is unduly

punitive and excessive. Pls.’ Br. at 3. Plaintiffs assert that

this new rate is sixteen times greater than Hyundai’s previous

dumping margin and almost eighteen times greater than LG

Semicon’s previous margin. See id. at 4. Additionally,

Plaintiffs note that while LG Semicon’s weighted-average dumping

margin was 10.44 percent under total AFA, this margin rose to

15.87 percent upon Commerce’s application of only partial AFA to

LG Semicon’s German sales. See id. at 3-4. Concluding that LG

Semicon has been made worse off with partial AFA than with total

AFA, Plaintiffs argue that Commerce has failed to create an

incentive for foreign companies to cooperate in administrative
Cons. Court No. 00-01-00027 Page 10

reviews, and that Congress’s intention to strike “a balance

between deterrence for non-compliance and assuring a reasonable

margin” has not been achieved. Id. at 4 (quoting F.lli De Cecco

di Filippo Fara S. Martino S.p.A. v. United States, 216 F.3d

1027, 1032 (Fed. Cir. 2000)).

The Court, however, finds that the 89.10 percent rate

selected by Commerce upon remand is neither unduly punitive nor

excessive. First, Commerce acts within its discretion and does

not “overreach reality” so long as the “selected . . . dumping

margin [is] within the range of [the respondent’s] actual sales

data[.]” See Ta Chen, 298 F.3d at 1340. Here, because Commerce

selected an AFA rate from a range of margins calculated using LG

Semicon’s own sales data, Commerce did not abuse its discretion.

Moreover, the new rate of 89.10 percent is in fact lower than

several ([ ]) other margins Commerce calculated for LG

Semicon, and therefore is not excessive.

Second, the 89.10 percent rate is not rendered unduly

punitive simply because it is higher than the previous rate

calculated using total AFA. As the Court has already noted, new

margins ranging as high as 223 percent became available for LG

Semicon’s sales practices after the Court ordered the

recalculation in Hyundai I. See Remand Results at 4. Since 19

U.S.C. § 1677e(b) does not prohibit Commerce from including such

new sales in its remand calculations, Commerce was free to select
Cons. Court No. 00-01-00027 Page 11

an AFA rate incorporating these new margins, even though that

meant assigning a higher rate to LG Semicon under partial AFA

than under total AFA.

Finally, the Court rejects Plaintiffs’ assertion that

Commerce failed to create an incentive to cooperate in

administrative reviews, and also failed to strike a balance

between deterrence and assuring a reasonable margin. If LG

Semicon had correctly reported certain sales it knew or should

have known were destined for the United States, Commerce would

not have applied AFA at all, and the mere fact that LG Semicon

received a higher rate on remand does not diminish its incentive

to cooperate more fully in future reviews in an effort to avoid

the application of AFA. In addition, the issue of whether

Commerce properly balanced deterrence with providing a reasonable

margin is inapposite, because the Court has already determined

that Commerce did provide LG Semicon with a reasonable margin.

The 89.10 percent rate is a non-aberrational rate that was

derived from LG Semicon’s own sales data. Consequently, the

margin was within Commerce’s discretion to select and must be

considered reasonable. See Ta Chen, 298 F.3d at 1339.

Accordingly, Commerce’s decision on remand to apply an AFA

rate of 89.10 percent is supported by substantial evidence and

otherwise in accordance with law.
Cons. Court No. 00-01-00027 Page 12

B. Commerce’s Treatment of Plaintiffs’ R&D Costs Is Sustained
in Part and Reversed and Remanded in Part.

In Hyundai I, the Court found that Commerce did not offer

substantial evidence in support of its cross-fertilization

theory. Hyundai I, 28 CIT at __, 342 F. Supp. 2d at 1157. The

Court also held that Commerce’s decision to reject Plaintiffs’

amortization of R&D costs was not supported by substantial

evidence. See id. at __, 342 F. Supp. 2d at 1157-59. In

addition, the Court found that Commerce failed to provide

specific record evidence showing how deferred R&D costs actually

affect production and revenue for the period of review. Id. at

__, 342 F. Supp. 2d at 1159.

On remand, Commerce (1) provided additional factual

information in support of its cross-fertilization theory; (2)

recalculated Plaintiffs’ R&D costs to allow for amortization; and

(3) continued to find that Plaintiffs’ R&D costs should be

expensed in the period incurred, instead of being deferred. See

Remand Results at 4-6. Plaintiffs challenge Commerce’s findings

with respect to cross-fertilization and deferral of R&D expenses,

while Micron challenges Commerce’s finding with respect to

amortization of R&D expenses.

1. Commerce’s Decision Not to Calculate Costs on a
Product-Specific Basis Is Not Supported by Substantial
Evidence.

In Hyundai I, the Court rejected Commerce’s suggestion that

“intrinsic benefits . . . occur between R&D expenditures on non-
Cons. Court No. 00-01-00027 Page 13

subject merchandise and production of subject merchandise,” and

that “R&D costs for non-subject merchandise [therefore] should be

included in the cost of production analyses.” Hyundai I, 28 CIT

at __, 342 F. Supp. 2d at 1156. Specifically, the Court found

that Commerce failed to offer substantial evidence in support of

its cross-fertilization theory because the findings of its

expert, Dr. Murzy Jhabvala, were based on a prior antidumping

investigation that concerned “different products and different

parties” than those at issue in Hyundai I. Id. at __, 342 F.

Supp. 2d at 1156-57. Thus, the Court remanded for Commerce “to

provide additional information specifically pointing to the

effect of non-subject merchandise R&D on the R&D for the subject

merchandise, or alternatively, [to] recalculat[e] R&D costs on

the most product-specific basis possible for both LG Semicon and

Hyundai.” Id. at __, 342 F. Supp. 2d at 1157.

On remand, Commerce complied with the Court’s request for

additional information by providing three exhibits that were

originally submitted for the record of the fifth administrative

review by Micron. See Def.-Intvr.’s Br. at Ex. 2 (Micron’s

Submission of Factual Information dated Dec. 18, 1998) at Doc.

1(a) (Memorandum from Dr. Murzy Jhabvala dated Sept. 8, 1997)

(“Jhabvala Letter 1”); id. at Doc. 1(b) (World Technology

Evaluation Center Report on the Korean Electronics Industry dated

Oct. 1, 1997) (“WTEC Report”); id. at Doc. 2 (Memorandum to the
Cons. Court No. 00-01-00027 Page 14

File from Dr. Murzy Jhabvala dated Dec. 18, 1997) (“Jhabvala

Letter 2”); id. at Doc. 3 (Letter from Eugene Cloud dated Oct.

15, 1997) (“Cloud Letter”); id. at Doc. 3(a) (Attachment A

Articles dated Sept.-Oct. 1997) (“Attachment A Articles”); id. at

Doc. 3(b) (Attachment B Articles dated Oct. 1997) (“Attachment B

Articles”). Commerce also presented the names of certain

research projects that were conducted by Hyundai in non-memory IC

labs but featured the word “DRAM” in their titles. See Remand

Results at 14.

Plaintiffs maintain that the additional factual information

supplied by Commerce suffers from the same deficiencies as the

information on which Commerce initially relied – namely, the new

evidence does not relate to non-subject merchandise, and it does

not specifically address Plaintiffs’ operations. Pls.’ Br. at 5.

Moreover, Plaintiffs contend, simply because the word “DRAM” is

in a project does not provide substantial evidence that the R&D

actually relates to DRAM development. Id. at 7.

The Court agrees that the additional information provided by

Commerce on remand does not constitute substantial evidence of

cross-fertilization between DRAMs and non-DRAM merchandise with

respect to R&D costs in this case. First, as in Hyundai I, the

new evidence on which Commerce relies does not demonstrate

“direct contact or experience with Plaintiffs’ practices during

this review[,]” but rather concerns “different products and
Cons. Court No. 00-01-00027 Page 15

different parties to that of the current review[.]” Hyundai I,

28 CIT at __, 342 F. Supp. 2d at 1157. In fact, Jhabvala Letter

1 is the same document rejected by the Court in Hyundai I, while

Jhabvala Letter 2 is based on similar research from a prior

antidumping investigation that involved different parties and

products. See Jhabvala Letter 1 at 1-2; Jhabvala Letter 2 at 1-

2. Moreover, the WTEC Report referenced by Commerce only refers

to Plaintiffs’ DRAM market shares, and only mentions DRAM R&D in

the context of industry-wide research goals – it does not address

Plaintiffs’ specific R&D operations or production practices. See

WTEC Report at 4-9. Finally, the Cloud Letter, Attachment A

Articles, and Attachment B Articles all fail to mention

Plaintiffs by name, and none demonstrates direct contact with

Plaintiffs’ R&D practices during the review.4 See Cloud Letter

at 1-3; Attachment A Articles at 1-3; Attachment B Articles at 1-

14.

Second, Commerce’s additional information fails to

“specifically point[ ] to the effect of non-subject merchandise

R&D on the R&D for the subject merchandise[.]” Hyundai I, 28 CIT

4
The Attachment A Articles fail to mention either
Plaintiffs or the subject merchandise, while the Attachment B
Articles only discuss how DRAM technology has been applied to
Application Specific Integrated Circuits (“ASICs”) by other
firms. See Attachment A Articles at 1-3; Attachment B Articles
at 3, 6-7, 9-10. Likewise, the Cloud Letter does not name
Plaintiffs once; instead, it was written in response to “letters
submitted by Dr. Bruce Wooley on behalf of ISSI and Dr. David
Angel on behalf of Samsung.” Cloud Letter at 1.
Cons. Court No. 00-01-00027 Page 16

at , 342 F. Supp. 2d at 1157. Both Jhabvala Letter 1 and

Jhabvala Letter 2 argue that cross-fertilization occurs within

the semiconductor industry, but the evidence and research

contained therein deal exclusively with SRAMs rather than the

subject merchandise DRAMs. See Jhabvala Letter 1 at 1-2;

Jhabvala Letter 2 at 1-2. Similarly, the WTEC Report discusses

only an industry-wide goal of utilizing DRAM R&D to improve SRAM

production; it does not establish that SRAM R&D actually benefits

DRAM R&D.5 See WTEC Report at 8-9. Moreover, Mr. Cloud’s focus

and expertise, like that of Dr. Jhabvala, is on SRAMs rather than

DRAMs, and the various technological developments discussed in

the Attachment A and B Articles fail to identify any specific

effect on R&D for the subject merchandise.6 See Cloud Letter at

1-3; Attachment A Articles at 1-3; Attachment B Articles at 1-14.

Finally, Commerce’s apparent reliance on the names of

Hyundai’s R&D projects is misplaced. According to Commerce, the

names of three research projects “[c]learly [show] Hyundai was

5
Furthermore, the WTEC Report does not indicate that
Plaintiffs have achieved, or even share, the industry-wide goal
of using DRAM technology to increase SRAM capacity. See WTEC
Report at 8-9.
6
The Attachment A Articles deal with copper metallization
technology and do not mention DRAMs. See Attachment A Articles
at 1-3. The Attachment B Articles merely note that DRAM
technology has been introduced into the design of ASICs. See
Attachment B Articles at 1-14. There is no specific explanation
of how these various advancements directly impact DRAM R&D. See
id.
Cons. Court No. 00-01-00027 Page 17

conducting [memory] research during the POR in its [non-memory]

IC Lab[, which] clearly indicate[s] that in the semiconductor

industry, there is enough similarity among semiconductor products

and process technology objectives, that advances from R&D for one

type of semiconductor product can benefit other semiconductor

products.” Remand Results at 14. However, as the Court has

previously explained, the simple recitation of R&D project titles

does not constitute substantial evidence of cross-fertilization.

Hynix Semiconductor, Inc. v. United States, 28 CIT __, __, 295 F.

Supp. 2d 1365, 1372 (2003). Thus, the Court holds that

Commerce’s presentation of the names of Hyundai’s R&D projects

does not establish that non-subject merchandise R&D benefits

subject merchandise R&D.

Accordingly, the Court finds that Commerce failed to provide

substantial evidence to support its theory of cross-

fertilization. The Court therefore remands this issue to

Commerce with instructions to recalculate Plaintiffs’ R&D costs

on the most product-specific basis possible.

2. Commerce’s Decision to Accept Plaintiffs’ Amortization
of R&D Expenses Is Supported by Substantial Evidence
and Otherwise in Accordance with Law.

In disapproving Commerce’s initial refusal to accept

Plaintiffs’ amortization of R&D costs, the Court in Hyundai I

rejected Commerce’s finding that “Plaintiffs’ practice of

‘continually changing’ [accounting] methodologies produces
Cons. Court No. 00-01-00027 Page 18

‘aberrationally high amounts of R&D expense in some years, and

aberrationally low amounts of R&D expense in other years, that do

not reasonably reflect [production] costs.’” Hyundai I, 28 CIT

at __, 342 F. Supp. 2d at 1158 (quoting Final Results, 64 Fed.

Reg. at 69699). The Court stated that “Plaintiffs’ previous

changes in accounting methodology are not relevant in this case

as the Court is concerned with the actions of the parties with

respect to their R&D costs only for this period of review.” Id.

The Court also dismissed Commerce’s concern that the inadvertent

result of the change in accounting practice would allow

Plaintiffs to recognize less than one-fifth of the current year’s

R&D costs. Id. (citing Final Results, 64 Fed. Reg. at 69699).

According to the Court, “in switching from expensing to

amortization, a difference in costs will likely occur, as

amortization by definition permits the allocation of costs over

the market life of the product, while expensing costs during the

period incurred necessarily implies a one-time charge.” Id.

Therefore, the Court remanded to Commerce “to provide specific

evidence regarding how Plaintiffs’ actual R&D costs for this

period of review are not reasonably accounted for in its

amortized R&D costs.” Id. at __, 342 F. Supp. 2d at 1159.

On remand, Commerce stated:

We believe that in the Final Results, we fully
explained, and supported with substantial evidence, our
positions regarding the amortization of LG [Semicon]’s
and Hyundai’s R&D costs. Nevertheless, the Court has
Cons. Court No. 00-01-00027 Page 19

found that the information cited by the Department does
not constitute substantial evidence supporting this
determination. Therefore, although we disagree with
the Court’s findings, we have recalculated LG
[Semicon]’s and Hyundai’s R&D costs to allow for
amortization.

Remand Results at 5.

Micron urges that, although Commerce conceded it was unable

to comply with the Court’s request that it provide additional

evidence in support of its determination regarding amortization,

Commerce actually did point to such evidence in the Remand

Results.7 See Def.-Intvr.’s Br. at 1. The Court, however, has

carefully reviewed Commerce’s discussion of the amortization

issue in the Remand Results, see Remand Results at 5, 17-19, and

finds that it is noticeably void of any “specific evidence

regarding how Plaintiffs’ actual R&D costs for this period of

review are not reasonably accounted for in its amortized R&D

costs[,]” which is what the Court instructed Commerce to provide

7
In addition to this argument, the Court notes that
Micron’s Memorandum Addressing the Final Results of
Redetermination Pursuant to Court Remand, which was filed on
September 30, 2004, is largely comprised of arguments regarding
the alleged incompatibility between the Court’s decision in the
instant proceeding and the Court’s decision in Micron Technology,
Inc. v. United States, 23 CIT 380 (1999), as well as a section
explaining why portions of the Court’s reasoning in Hyundai I
are, in Micron’s view, “irrelevant.” See Def.-Intvr.’s Br. at
10-15, 16. As such, Micron’s brief is more akin to a motion for
rehearing and reconsideration of the Court’s April 16, 2004
decision in Hyundai I – something the Court declines to entertain
at this late stage. See USCIT R. 59(b) (“A motion for a . . .
rehearing shall be served and filed not later than 30 days after
the entry of the judgment or order.”).
Cons. Court No. 00-01-00027 Page 20

on remand. Hyundai I, 28 CIT at __, 342 F. Supp. 2d at 1159

(emphasis added). Moreover, the concerns expressed by Commerce

in the Remand Results were already considered and rejected by the

Court in Hyundai I.

Accordingly, the Court sustains Commerce’s decision to

accept Plaintiffs’ amortization of R&D expenses for purposes of

calculating the cost of production.

3. Commerce’s Rejection of Plaintiffs’ Deferral of R&D
Expenses Is Not Supported by Substantial Evidence.

In Hyundai I, the Court held that Commerce failed to provide

specific evidence demonstrating why Plaintiffs’ deferred R&D

costs should be allocated into the cost of production

calculation. Hyundai I, 28 CIT at __, 342 F. Supp. 2d at 1159.

In particular, the Court was unconvinced by Commerce’s argument

that “the practice of indefinite deferral of R&D costs is

inconsistent with the conservatism principle in accounting[,]”

since “Plaintiffs point[ed] out that their [deferral]

methodology, which is in accordance with Korean GAAP, does follow

the principle of conservatism in accounting.”8 Id. Thus, after

observing that “[o]nly R&D costs that are related to the

production and revenue of the subject merchandise for the review

8
Conservatism in accounting calls for the recognition of
expenses when incurred if the probability of associated revenue
is remote or uncertain. Hyundai I, 28 CIT at __, 342 F. Supp. 2d
at 1159.
Cons. Court No. 00-01-00027 Page 21

period should be included in Commerce’s calculations[,]” see 19

U.S.C. § 1677b(f)(1)(A), the Court remanded to Commerce to

provide “specific evidence on the record to show that R&D costs

that are currently deferred actually affect production and

revenue for this review period.” Hyundai I, 28 CIT at __, 342 F.

Supp. 2d at 1159.

On remand, Commerce continues to rely upon general

accounting principles in support of its determination that R&D

expenditures for long-term projects should be included in the

current cost of production calculation.9 See Remand Results at

5, 22-23. In addition, Commerce faults Plaintiffs for failing to

offer any “reasonable evidence to indicate that their deferred

[R&D] costs will benefit future periods.” Id. at 6.

In response, Plaintiffs claim they did provide substantial

evidence demonstrating that “the future benefits of [DRAM]

research are both readily discernible and imminent at the time of

the expenditures.” Pls.’ Br. at 12 (citing Supplement to

Hyundai’s Appendix of the Administrative Record at CR 42 (Hyundai

Cost Verification Exhibit 10 and Report dated June 11, 1999) at

Ex. 10(a), 10(c)). Plaintiffs also note that Commerce failed on

remand to show even one example of how R&D for long-term projects

9
Specifically, Commerce cites to both International
Accounting Standard 9 and U.S. GAAP to assert that “R&D should
not be deferred because the future economic benefits cannot be
quantified and measured with a reasonable degree of certainty.”
Remand Results at 22.
Cons. Court No. 00-01-00027 Page 22

impacted projects for the current review period. Id. at 9.

Finally, Plaintiffs contend that Commerce erroneously shifted its

burden of providing additional information on remand to

Plaintiffs. Id.

The Court agrees that Commerce failed to provide specific

evidence in the Remand Results showing how Plaintiffs’ “R&D costs

that are currently deferred actually affect production and

revenue for this review period.” Hyundai I, 28 CIT at __, 342 F.

Supp. 2d at 1159 (emphasis added). By continuing to cite to

general accounting principles on remand, Commerce is merely

attempting to resurrect its previous argument, which the Court

already rejected in Hyundai I, that the future benefits of

deferred R&D costs are unquantifiable.10 The general accounting

principles on which Commerce relies do not directly address any

of Plaintiffs’ expenditures for long-term R&D projects, and they

do not explain how such R&D costs might affect revenue and

production for the current review period, which is what the Court

ordered Commerce to consider on remand. Moreover, since it was

the burden of Commerce – not Plaintiffs – to provide additional

10
In Hyundai I, Commerce claimed that Plaintiffs’ practice
of indefinite deferral of R&D costs was inconsistent with the
conservatism principle in accounting because the probability of
associated revenue was remote or uncertain. Hyundai I, 28 CIT at
__, 342 F. Supp. 2d at 1159. In the Remand Results, Commerce
again “conclude[s] that R&D should not be deferred because the
future economic benefits cannot be quantified and measured with a
reasonable degree of certainty.” Remand Results at 22.
Cons. Court No. 00-01-00027 Page 23

information on remand, Commerce’s assertion that Plaintiffs

failed to offer “reasonable evidence” in defense of R&D deferral

improperly shifts the burden of providing additional information

to Plaintiffs. See Remand Results at 6.

Accordingly, because Commerce entirely failed to comply with

the Court’s instructions in Hyundai I to provide specific

evidence showing how deferred R&D costs actually affect

production and revenue for the current review period, the Court

remands to Commerce with instructions to accept Plaintiffs’

deferral methodology in calculating R&D expenses for the cost of

production.

C. Commerce Properly Decided to Correct an Error in the
Calculation of Hyundai’s Entered Value.

After Commerce issued the Draft Remand Results, Micron

raised a challenge to Commerce’s calculation of Hyundai’s total

entered value. See Conf. Admin. R. at Ex. 2 (Micron’s Comments

on the Draft Final Results of Redetermination dated Aug. 18,

2004) at 11-13. According to Micron, the margin program used by

Commerce to calculate Hyundai’s entered value improperly

multiplied quantity and value variables that were stated in

inconsistent units of measure, which led to an erroneously low

importer-specific assessment rate. See id. at 12. In response,

Plaintiffs agreed with Micron that Commerce should address the

miscalculation issue, but Plaintiffs further argued that

Hyundai’s entered value should also include all of the sales made
Cons. Court No. 00-01-00027 Page 24

by Hyundai Electronics America, Inc. (“HEA”) – not just the sales

that HEA resold to U.S. customers. See Conf. Admin. R. at Ex. 3

(Hynix’s Rebuttal to Micron’s Comments on the Draft Final Results

of Redetermination dated Aug. 23, 2004) at 7-8 (“Pls.’ Draft

Reply Br.”). In the Remand Results, Commerce made the

programming changes suggested by Micron, but did not address the

issue raised by Plaintiffs. Remand Results at 32.

Plaintiffs now contend that Commerce and Micron were time

barred from revisiting the calculation methodology for Hyundai’s

entered value. Pls.’ Br. at 15. In addition, Plaintiffs argue

that Commerce’s change to the entered value calculation did not

involve correcting “a mere clerical error.” Id. at 16. Finally,

Plaintiffs assert that if Commerce is permitted to change its

calculation of the entered value at all, then it should include

all of HEA’s sales transactions in the calculation. Id. at 17-

19.

1. Commerce Properly Corrected the Ministerial Error in
the Calculation of Hyundai’s Entered Value Identified
by Micron.

With respect to Plaintiffs’ first issue, the Court finds

that although Plaintiffs allege Commerce and Micron were time

barred from recalculating Hyundai’s entered value, in reality it

is Plaintiffs who are barred from objecting to the recalculated

entered value, because Plaintiffs have reversed their agency

position on appeal. After Commerce released the Draft Remand
Cons. Court No. 00-01-00027 Page 25

Results, Plaintiffs did not object to Micron’s argument regarding

the improper multiplication of the entered value, and in fact

“agree[d] that the Department should address this issue.” Pls.’

Draft Reply Br. at 7 (emphasis added). On appeal, Plaintiffs now

contend that Commerce and Micron were time barred from addressing

the miscalculation at all. See Pls.’ Br. at 15.

The Court has repeatedly held that “a party may not reverse

the position it took before the agency and raise contrary

arguments on appeal” because this “den[ies Commerce] the

opportunity to review plaintiffs’ arguments” and “deprive[s] the

other parties of their right to respond to plaintiffs’ position.”

Calabrian Corp. v. United States Int’l Trade Comm’n, 16 CIT 342,

347, 794 F. Supp. 377, 383 (1992). Plaintiffs have clearly

reversed their agency position on appeal; therefore, the Court

will not entertain Plaintiffs’ new objection to the calculation

of Hyundai’s entered value.

Moreover, even if Plaintiffs were not barred from reversing

their agency position on appeal, their objection to Hyundai’s

recalculated entered value would nonetheless fail on the merits.

First, Commerce “may, with or without a party’s request, correct

errors that it reasonably regards as ministerial in final

determinations.” Shandong Huarong Gen. Corp. v. United States,

25 CIT 834, 848, 159 F. Supp. 2d 714, 727 (2001) (quoting Aramide

Maatschappij V.o.F. v. United States, 19 CIT 1094, 1103, 901 F.
Cons. Court No. 00-01-00027 Page 26

Supp. 353, 361 (1995)), aff’d sub nom. Shandong Huarong Gen.

Group Corp. v. United States, Appeal No. 02-1095 (Fed. Cir. Jan.

10, 2003). Here, because the calculation of Hyundai’s entered

value involved multiplying quantity and value variables that were

stated in inconsistent units of measure, the calculation was

clearly an “error in addition, subtraction, or other arithmetic

function,” and Commerce therefore correctly regarded the

miscalculation as a ministerial error. See 19 C.F.R. §

351.224(f) (defining “ministerial error”).

Second, the Court itself has a responsibility to “exercise

its discretion to prevent knowingly affirming a determination

with errors.” Torrington Co. v. United States, 21 CIT 1079, 1082

(1997); see also Fed.-Mogul Corp. v. United States, 18 CIT 1168,

1172, 872 F. Supp. 1011, 1014 (1994). At various stages of this

review, all parties have agreed that the Draft Remand Results

contained a miscalculation of Hyundai’s entered value. See

Remand Results at 31-32; Pls.’ Draft Reply Br. at 7. Therefore,

the Court would be “knowingly affirming a determination with

errors” if it did not sustain the correction made by Commerce in

the Remand Results. Torrington, 21 CIT at 1082.

Accordingly, Commerce’s recalculation of Hyundai’s entered

value in the Remand Results is sustained.
Cons. Court No. 00-01-00027 Page 27

2. Commerce Properly Refused to Address Plaintiffs’
Challenge to the Calculation Methodology for Hyundai’s
Entered Value.

The Court also affirms Commerce’s decision not to address

Plaintiffs’ argument concerning the inclusion of all HEA’s sales

in Hyundai’s entered value. “While Commerce is required to allow

respondents to correct clerical errors discovered late in the

administrative process, clerical errors are distinguished from

substantive errors and do not encompass methodological

modifications.” Tianjin Mach. Imp. & Exp. Corp. v. United

States, 28 CIT ___, ___, 353 F. Supp. 2d 1294, 1304 (2004).

Clerical, or ministerial, errors are defined as “error[s] in

addition, subtraction, or other arithmetic function, clerical

error[s] resulting from inaccurate copying, duplication, or the

like, and any other similar type of unintentional error which the

Secretary considers ministerial.” 19 C.F.R. § 351.224(f); see

also Maui Pineapple Co. v. United States, 27 CIT ___, ___, 264

F. Supp. 2d 1244, 1261 (2003) (quoting Certain Fresh Cut Flowers

From Colombia; Final Results of Antidumping Duty Administrative

Reviews, 61 Fed. Reg. 42833 (Aug. 19, 1996)).

Although Plaintiffs claim that Commerce made an incorrect

entered value calculation, Commerce’s decision to use only HEA’s

U.S. sales in calculating Hyundai’s entered value was not an

“error in addition, subtraction, or other arithmetic function,”

did not involve “inaccurate copying [or] duplication,” and was
Cons. Court No. 00-01-00027 Page 28

not an “unintentional error.” 19 C.F.R. § 351.224(f). Rather,

the decision to exclude HEA’s non-U.S. sales involved many issues

of methodology and fact,11 and Commerce intentionally rejected

Plaintiffs’ alternative methodology because “the record does not

appear to contain the data necessary to support Hyundai’s claim.”

Remand Results at 33. Thus, unlike the miscalculation identified

by Micron, Plaintiffs’ challenge does not involve a clerical

error, and Commerce therefore properly refused to address this

issue in the Remand Results.

II. CONCLUSION

For the aforementioned reasons, the Remand Results are

sustained in part and reversed and remanded in part. A separate

order will be issued accordingly.

/s/ Richard W. Goldberg
Richard W. Goldberg
Senior Judge

Dated: August 25, 2005
New York, New York

11
These issues included (1) the extent of HEA’s
transshipped sales to third countries; (2) whether Customs could
distinguish between entries destined for the United States and
those destined for third countries; and (3) whether transshipped
entries should have been included in the assessment rate
calculation. See Pls.’ Br. at 17-19; Def.-Intvr.’s Br. at 17-19.