La Molisana S.p.A. v. United States

Slip Op. 17 - 111

UNITED STATES COURT OF INTERNATIONAL TRADE

: LA MOLISANA S.p.A, : : Plaintiff, : : v. : Before: R. Kenton Musgrave, Senior Judge : Court No. 16-00047 : UNITED STATES, : : Defendant, : : and : : NEW WORLD PASTA CO. and : DAKATA GROWERS PASTA CO., : Defendant-Intervenors. : :

OPINION AND ORDER

[Remanding eighteenth administrative review of certain pasta from Italy.]

Dated: August 23, 2017

David J. Craven and Saichang Xu, Sandler, Travis & Rosenberg, of Chicago, IL, for the plaintiff.

Elizabeth Speck, Senior Trial Counsel, Commercial Litigation Branch, Civil Division, U.S. Department of Justice, of Washington, DC, for the defendant. On the brief were Chad A. Readler, Acting Assistant Attorney General, Jeanne E. Davidson, Director, and Reginald T. Blades, Jr., Assistant Director. Of Counsel was Mykhalo Gryzlov, Senior Counsel, Office of the Chief Counsel for Trade Enforcement and Compliance, U.S. Department of Commerce.

Paul C. Rosenthal and David C. Smith, Kelley Drye & Warren LLP, of Washington, DC, for the defendant-intervenors.

Court No. 16-00047 Page 2

Musgrave, Senior Judge: The plaintiff, La Molisana S.p.A., challenges two determinations from the eighteenth (2013-2014) antidumping duty administrative review of certain dry pasta from Italy1: (1) whether the U.S. Department of Commerce, International Trade Administration (“Commerce” or “Department”) erred in requiring La Molisana’s pasta sales product shapes to be reported without variance from the proceeding’s pasta shape classification list; and whether Commerce failed to provide meaningful opportunity for addressing the agency’s differential pricing analysis. This being a timely-filed matter described by 19 U.S.C. §1516a(a)(2), subsections (A)(i)(I) and (B)(iii), jurisdiction is proper under 28 U.S.C. §1581(c). The standard of review is substantial evidence on the record. 19 U.S.C. §1516a(b)(1)(B)(i). The case will be remanded per the following.

Discussion

I

A. Background

The pasta shape classification list provides for eight different types of shape categories: (1) long cuts (e.g., linguine, spaghetti), (2) specialty long cuts (e.g., capellini, fioccini), (3) nested/folded/coiled, (4) lasagna, (5) short cuts (e.g., fagiolini, medium shells), (6) specialty short cuts (e.g., mezzanelli, pasta mista), (7) soupettes (e.g., ditali, corallini), and (8) combination of shapes. E.g., Initial Questionnaire (“IQ”) (Oct. 3, 2014), PDoc 24, at B-8, C-7. Appendix III of the

1 Certain Pasta From Italy: Final Results of Antidumping Administrative Review; 2013-2014, 81 Fed. Reg. 8043 (Feb. 17, 2016) (“Final Results”); Certain Pasta From Italy: Amended Final Results of Antidumping Duty Administrative Review; 2013-2104, 8l Fed. Reg. 12690 (Mar. 10, 2016) (“Amended Final Results”) and the accompanying issues and decision memorandum (“IDM”), PDoc 228. Citations to other specific administrative reviews are herein abbreviated “AR” followed by its numeric sequence.

Court No. 16-00047 Page 3 IQ sent to all respondents lists some 256 different shapes and their corresponding control number category (i.e., CONNUMs) and provided therein in relevant part is as follows:

SHAPE CLASSIFICATION You are required to classify the pasta types reported in field 3.9 into one of the shape categories specified in field 3.1 in accordance with the questionnaire examples and the attached “Classification of Pasta Shapes.” If you sold pasta in shapes that do not appear on the attached list, please use the most similar pasta type on this list as a guide for determining the appropriate shape classification. Support any such classification with a description and picture of the pasta type, the production line on which it is produced, the standard production capacity of that line (e.g., pounds per hour), and the line speed (e.g., pounds per hour) for the pasta type in question. Please note that any revisions in shape classification must also be reflected in control numbers and variable manufacturing cost information.

IQ at 151 (or Appx. III at 4; italics added).

Long (and linear) cuts are made on long cut pasta machines and short cuts are made on short cut pasta machines. Once the determination of long or short is made, the question is then whether a cut is a “standard” or “special” cut, which for purposes of this matter is a critical distinction: the “dividing line” is not due to any physical differences in pasta cuts within the long or short shape categories but due to the higher production costs that would be associated with the slower line speeds of the latter. See, e.g., Certain Pasta From Italy, 78 Fed. Reg. 9364 (Feb. 8, 2013) (AR15 final) and accompanying issues and decision memorandum (all such memoranda hereinafter “I&D Memo”) at cmt 1; Certain Pasta From Italy, 64 Fed. Reg. 6615 (Feb. 10, 1999) (inter alia AR1 final) and accompanying I&D Memo at cmt 7.

Commerce attempts to frame La Molisana’s dispute as a disagreement with methodology, but the parties rather appear in agreement that the methodology is proper and long standing. Onward since the 1996 investigation, Commerce’s practice has been to use a 75 percent line-speed benchmark to distinguish “speciality” cuts from “regular” cuts produced on the same long Court No. 16-00047 Page 4 or short production line. See, e.g., id. La Molisana’s initial Section B and C questionnaire responses included detailed exhibits in which the production speeds of the various shapes which La Molisana believed were improperly categorized in Commerce’s shape list and which La Molisana believed should be re-classified. La Molisana also included as part of this exhibit a list of those shapes which it believed did not appear on Commerce list and, using the same methodology, placed these shapes into what it believed to be appropriate categories. See LM Response to Sections B and C of the IQ (Dec. 2, 2014), CDocs 38-54, at Exs B-1 and C-1. La Molisana had submitted pictures of these shapes as part of its Section A response. See LM Response to Section A of the IQ (Nov. 5, 2015), PDocs 34-39, at Ex A-13(a).

About a month later, Commerce issued a supplemental questionnaire (“SQ”) to La Molisana stating:

1. In your [Section B Questionnaire Response,] page 2, and at Ex B-2, you state that you classified shapes “by taking the production speed for the cut on one of the lines as compared to the production speed for a shape classified by the Department in a particular category.” However, the Department does not classify products by production characteristics, rather through sales characteristics. Please assign shape codes to each of your observations in the home market and U.S. market databases according to the shape classifications included in Appendix III of the initial questionnaire. If you believe a change in the model match is warranted, you must submit a formal request to the Department for a change to the model match methodology.

SQ (Jan. 6, 2015), PDoc 86, at 2. (bracketing and italics added; footnote omitted).

La Molisana requested clarification of this instruction on January 13, 2015, noting Commerce’s extant practice. Commerce’s reply three days later to this request was as follows:

A. Please report your shapes, and the appropriate shape code, according to the existing list of classifications as presented in Appendix III of the Initial Questionnaire.

Court No. 16-00047 Page 5

B. For shapes not listed in Appendix III, please report the characteristics based on the most similar shape listed in Appendix III.

Dep’t Resp. to Request for Clarification (Jan. 16, 2015), PDoc 92, at 2 (italics added).

Believing it had already placed on the record the correct classifications, La Molisana explains, in order to avoid confusion resulting from multiple competing databases it sent in a database of the shapes as Commerce apparently wanted them (re)classified, but it did so under protest, noting at the time that it intended to raise this issue in its administrative case brief. In that brief, La Molisana argued Commerce had failed to follow the established methodology for shape classification in this order and further argued against the methodology imposed by Commerce in this review, i.e., Appendix III governs a shape’s classification regardless of any information on the record as to production line speed for that shape. See LM Case Brief (Oct. 6, 2015), CDoc 240, at 2-4.

In the IDM, Commerce refused to allow the reclassification of the shapes but did not directly address whether the methodology had shape-shifted to sales-based from production-based; rather, Commerce indicated that the shape methodology was essentially immutable, that the list had been “refined”, and that Commerce need not consider certain proffered facts. Commerce stated:

Although a 75 percent throughput rate has been used to distinguish pasta shape for specialty long and short pasta cuts, we have never allowed respondents to reclassify pasta shape classifications based on its own reported line speeds without providing the requisite evidence to support such a reclassification. Furthermore, we have not allowed respondents to reclassify pasta shapes that are already included in the pasta shape classification table.

IDM at 6.

B. Argument

La Molisana argues generally that the above statement in the IDM is factually inaccurate, both with respect to the review in question and also with respect to categorical statements Court No. 16-00047 Page 6 about what has and has not been allowed in prior reviews of this order. First, it notes that the statement is directly contradicted by Commerce’s instructions in Appendix III, in which is stated in relevant part: “Please note that any revisions in shape classification” et cetera. See IQ at 151, supra. Thus, La Molisana argues that the questionnaire instructions themselves clearly contemplated revisions to the existing list of shapes, and that if Commerce’s subsequent statement that the list of shapes is essentially immutable and that respondents are not allowed to reclassify shapes were true, then any provision providing for the “revision” of shapes would be rendered a nullity.

Second, La Molisana notes that Commerce has, in fact, allowed respondents to reclassify shapes that were included in the pasta shape classification table. In the I&D Memo for the fourth (1999-2000) administrative review,2 for example, Commerce stated: “For those cuts which PAM believed were specialty cuts yet the Department considered a regular cut (or vice versa), it provided the line speed data. We reviewed this information and accepted the revised shape classification as provided by PAM.” AR4 I&D Memo at issue 19. Thus, Commerce has expressly allowed parties in other segments to re-classify shapes on the shape list. Further, La Molisana criticizes the defendant’s attempt to dismiss Commerce’s own decision in AR15 through claiming that such decision stood for the proposition that Commerce did not reclassify shapes, but La Molisana points out that stance is undercut by the language of AR15 itself:

In the investigation the Department used a 75 percent throughput rate to distinguish pasta shape for specialty long and short pasta cuts. We do not use the throughput rates to further differentiate between specialty short and super-specialty short or specialty long and super-specialty long. Based on the throughput rates of the pasta at issue, we can classify them as specialty short and specialty long because the line

2 See I&D Memo accompanying Certain Pasta From Italy, 67 Fed. Reg. 300 (Jan. 3, 2002) (inter alia, AR4 final).

Court No. 16-00047 Page 7

speed or the throughput rates are 75 percent or less than the corresponding line speed or throughput rates of the short and long pasta.

AR15 I&D Memo at cmt 1. As this language makes clear, and as previously noted, Commerce since the investigation has used speed as the dividing point between special and regular cuts, and it is production speed, not sales characteristics, which apparently controls that distinction. And in considering the record of AR5, Certain Pasta From Italy, 68 Fed. Reg. 6882 (Feb. 11, 2003), the court previously acknowledged “other evidence on the record that appears to show that line speed is a shorthand for shape.” New World Pasta Co. v. United States, 28 CIT 290, 309 n.22, 316 F. Supp. 2d 1338, 1355 n.22 (2004). La Molisana contends that if Commerce’s standard is that the production speed drives whether or not a cut is standard or special but for the administrative review at bar refused to consider any evidence showing the relative production speeds of certain cuts, then Commerce has not actually abided by its own standard but made an arbitrary decision not based on an actual review of production data, which is not in accordance with Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Insurance Co., 463 U.S. 29, 43 (1983) (an agency’s determination is arbitrary and capricious if the agency “entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency”, etc.).

Third, La Molisana argues that Commerce “did not even try” (LM Br. at 13) to explain the departure from its prior practices and rather appears to have misconstrued both prior practice from other reviews and the positions that it had taken in this review. Commerce claimed that it never allowed “reclassifications” of products on Commerce’s shape list and that a “formal request” (whatever that means) to modify the shape methodology was not made and that insufficient information was placed on the record. Yet, as discussed above, Commerce not only has considered Court No. 16-00047 Page 8 the reclassification of shapes, it has expressly allowed this and expressly contemplated this in the questionnaire.

Fourth, La Molisana contends it was not asking that Commerce “modify” its shape methodology but rather was simply asking that Commerce apply the existing methodology to the facts of record including the information provided by La Molisana, and La Molisana submits that the information placed on the record in this review would appear to be similar to that placed on the record in AR4 (the 4th review of the antidumping duty order on pasta from Italy); and that if the record were not sufficient without supplementation then Commerce should have asked for more information. In AR4, Commerce did not believe the information to be sufficient to sustain the claim and issued a supplemental questionnaire requesting more information. AR4 I&D Memo at issue 17. La Molisana thus argues that to the extent that Commerce believed the information La Molisana submitted was insufficient, then Commerce should have followed extant precedent and asked for further detail from La Molisana; but rather than requesting this information, Commerce essentially enumerated a “new standard” based on sales (i.e., an “immutable” list of shapes) and not on production; and when confronted with the fact that this was a different standard it wound up further mischaracterizing its own practice.

C. Analysis

The court cannot agree that Commerce enumerated a “new” standard. Rather, for this proceeding Commerce appears to have ossified in its approach to “considering” the facts before it. The court, thus, can agree with La Molisana that Commerce’s practice, as established by both administrative and judicial precedent, is quite clear, and that Commerce must follow its prior decisions unless it is able to provide a reasoned explanation for the departure from such decisions. Court No. 16-00047 Page 9 See Citrosuco Paulista, S.A. v. United States, 12 CIT 1196, 1209, 704 F. Supp. 1075, 1088 ( 1988) (“it is . . . a general rule that an agency must either conform itself to its prior decisions or explain the reasons for its departure”); Cinsa S.A. de C.V. v. United States, 21 CIT 341, 349, 966 F. Supp. 1230, 1238 (1997) (“Commerce . . . cannot arbitrarily abandon a relied[-]upon methodology”, it “can reach different determinations in separate administrative reviews but it must employ the same methodology or give reasons for changing its practice”); RHP Bearings Ltd. v. United States, 24 CIT 1218, 1226, 120 F. Supp. 2d 1116, 1124 (2000) (application of a “special rule” in only two prior reviews does not form a “long-established practice” but Commerce is still obliged “to explain the apparent inconsistency of its approach in this review and the two preceding reviews”). See also American Silicon Technologies v. United States, 22 CIT 776, 777, 19 F. Supp. 2d 1121, 1123 (1998).

La Molisana is correct that an immutable list cannot take into account any changes in technology in the production of a cut nor the specific “shape” of the cut produced by each producer nor the speeds of each producer’s machines. However, La Molisana undercuts its own argument to a certain degree in pointing out that each pasta producer uses its own unique dies, and that a shape produced by one producer will not necessarily be identical to that of another producer. The defendant thus correctly implies that the shapes list is not company-specific. See, e.g., Def’s Resp. at 17 (“taken to its logical conclusion, permitting classifications based on company-specific line speeds would result in a system in which (1) physically identical pasta could have different classifications, and (2) the pasta would be classified by something other than its physical characteristics”) (internal quotes omitted; referencing 19 U.S.C. §1677(16)(A)) On the other hand, the defendant goes too far in attempting to characterize La Molisana’s arguments as an “attempt to replace the well-established shape classification in the model match methodology with a system Court No. 16-00047 Page 10 based on company-specific line speeds.” Id. at 9. The court does not interpret La Molisana’s arguments to that extent.

Further, the defendant argues “La Molisana has not demonstrated that the well-established shape classification list, which was proposed by Italian pasta respondents in the

original investigation,[ ] and further refined in subsequent reviews, is unreasonable” Id. at 14

(footnote omitted). The argument presumes perfection in Appendix III for this review, i.e., that the listed shapes and their corresponding CONNUM categories are indeed immutable, and yet, as also acknowledged, the list has been subject to at least four revisions over the years. While it appears true, as the defendant also contends, that the list reflects a “shape-based approach [to] match[ ] foreign pasta to domestic pasta based on the ‘physical characteristics’ of the pasta, which is precisely what 19 U.S.C. §1677(16)(A) requires”3, it is plain that the list is also divided over those “physical characteristics” by presumed production line speed at the start of each administrative review segment, and it is well-established that each such segment is to be considered a separate proceeding with its own unique facts. E.g., Shandong Huarong Machinery Co. v. United States, 29 CIT 484, 491 (2005). The questionnaire itself contemplates the possibility of revisions and, by implication, the right of parties to seek them. La Molisana placed on the record specific information about the line speeds used for the production of its various cuts of pasta and why the cuts that it sought to re-classify were produced at the same line speeds as exemplar cuts. See CDocs 38-54, supra, at Exs B-1 and C-1. La Molisana argues that this information demonstrated that some of the cuts which Commerce had placed on the list as “special” cuts, were in fact produced at the same speed as that

3 Def’s Resp. at 15, referencing New World Pasta Co. v. United States, 28 CIT 290, 306, 316 F. Supp. 2d 1338, 1352 (2004).

Court No. 16-00047 Page 11 of standard cuts. The record being plain that consideration of this claim is what La Molisana sought, to which the record and argument here reflect no adequate rejoinder, the matter will therefore be remanded therefor.

II

The second issue concerns Commerce’s differential pricing analysis. In the preliminary results Commerce determined that the differential pricing analysis did not produce a meaningful difference between the use of average to average (“A-A”) when compared to average to transaction (“A-T”). Certain Pasta From Italy, 80 Fed. Reg. 47467 (Aug. 7, 2015) (prelim. results) and accompanying preliminary decision memorandum, PDoc 183 (Aug. 3, 2015), at 8. Anticipating that when Commerce corrected any errors it might revisit this issue, La Molisana submitted arguments in its case brief challenging Commerce’s differential pricing analysis and the potential use of the A-T methodology. CDoc 240 at 11-12. These included a discussion of the seasonality of the product and other factors that could explain a difference in price.

In the Final Results, as reflected in the IDM, Commerce did not consider La Molisana’s arguments because they were mooted by the finding in the preliminary results that the difference was not meaningful. However, in the amended final results Commerce determined to apply the differential pricing analysis and to use A-T methodology. Amended Final Calculation Memorandum, PDoc 237, CDoc 264 (Mar. 7, 2016). Commerce did not address any of the arguments La Molisana presented but appears to have simply ruled that since the A-T methodology produced a higher rate than the A-A methodology, it was more accurate and thus should be used. La Molisana complains Commerce neither stated why a “higher” rate is necessarily more accurate nor addressed any of the concerns raised in its case brief. Court No. 16-00047 Page 12

The defendant voluntarily requests remand in order to place its analysis and determination on the record. The defendant-intervenors oppose remand and argue for sustaining Commerce’s determination because when La Molisana alleged clerical error in the Final Results, it must have understood that the consequence of that allegation would result in halving its dumping margin, which in turn would trigger a re-calculation that would invoke use of the A-T methodology. But the court has been provided no basis on which to draw such ultimate conclusion, and it is still a fundamental tenet of these matters that an agency must “articulate a[ ] rational connection between the facts found and the choice made.” Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168 (1962). See also SEC v. Chenery Corp., 332 U.S. 194, 196 (1947) (courts judge the propriety of agency action solely on the ground invoked by the agency and are powerless to substitute “a more adequate or proper basis”). Remand to Commerce to address the record is appropriate. See SKF USA Inc. v. United States, 254 F.3d 1022, 1029 (Fed. Cir. 2001).

Conclusion

The matter is hereby remanded to the Department of Commerce, International Trade Administration with instruction to reconsider La Molisana’s arguments on the application of the speed/shape methodology to the facts of record and also to supplement the record with articulated consideration of La Molisana’s administrative case brief comments regarding the differential pricing issue. The results of remand shall be due November 30, 2017, and within five days of such filing the parties shall confer and file with the court either a proposed briefing schedule for commenting on the remand results or other indication of proceeding to final disposition. So Ordered.

Dated: August 23, 2017 /s/ R. Kenton Musgrave New York, New York R. Kenton Musgrave, Senior Judge