Catfish Farmers of Am. v. United States

Slip Op. 13 - 63

UNITED STATES COURT OF INTERNATIONAL TRADE

: CATFISH FARMERS OF AMERICA, et al., : : Plaintiffs, : : v. : Before: R. Kenton Musgrave, Senior Judge : UNITED STATES, : Court No. 11-00109 : Defendant, : : and : : VINH HOAN CORPORATION, VINH : QUANG FISHERIES CORPORATION, : H&N INTERNATIONAL, and VIETNAM : ASSOCIATION OF SEAFOOD EXPORTERS : AND PRODUCERS, : : Defendant-Intervenors. : :

OPINION AND ORDER

[Remanding sixth antidumping administrative review for reconsideration of certain aspects.]

Dated: May 23, 2013

Valerie A. Slater, Jarrod M. Goldfeder, Natalya D. Dobrowolsky, and Nicole M. D’Avanzo, Akin, Gump, Strauss, Hauer & Feld, LLP, of Washington DC, for the plaintiffs.

Ryan Majerus, Attorney, Commercial Litigation Branch, Civil Division, U.S. Department of Justice, of Washington DC, argued for the defendant. On the brief were Stuart F. Delery, Acting Assistant Attorney General, Jeanne E. Davidson, Director, Franklin E. White, Jr., Assistant Director, and Courtney S. McNamara, Attorney. Of Counsel was David W. Richardson, Office of the Chief Counsel for Import Administration, U.S. Department of Commerce.

Matthew J. McConkey, Dave M. Wharwood, and Jeffrey C. Lowe, Mayer Brown LLP, of Washington DC, for defendant-intervenor Vinh Hoan Corporation.

Court No. 11-00109 Page 2

Robert G. Gosselink and Jonathan M. Freed, Trade Pacific, PLLC, of Washington DC, for defendant-intervenors Vinh Quang Fisheries Corporation and H&N Foods International.

Mark E. Pardo, Andrew Thomas Schutz, and Jeffrey O. Frank, Grunfeld Desiderio Lebowitz Silverman & Klestadt, LLP, of Washington DC, for defendant-intervenor Vietnam Association of Seafood Exporters and Producers.

Musgrave, Senior Judge: This action contests the final results of the sixth administrative review of the antidumping duty order on three species of Pangasius fish1 conducted by the International Trade Administration of the United States Department of Commerce (“Commerce” or “Department”). See Certain Frozen Fish Fillets from the Socialist Republic of Vietnam: Final Results of the Sixth Antidumping Duty Administrative Review and Sixth New Shipper Review, 76 Fed. Reg. 15941 (Mar. 22, 2011), PDoc 246 (“Final Results” or “Sixth Review”) and the issues and decision memorandum (“I&D Memo”) accompanying those results, PDoc 242. The review period is August 1, 2008 through July 31, 2009.

The plaintiffs, domestic industry petitioners,2 move for judgment on the administrative record. In opposition, the defendant-intervenors argue the Final Results should be

1 The antidumping duty order covers Pangasius hypophthalmus (also identified as Pangasius pangasius), Pangasius bocourti, and Pangasius micronemus. See Notice of Antidumping Duty Order: Certain Frozen Fish Fillets from the Socialist Republic of Vietnam, 68 Fed. Reg. 47909 (Aug. 12, 2003) (“Order”). 2 Plaintiffs are Catfish Farmers of America and individual U.S. domestic catfish processors America’s Catch, Consolidated Catfish Companies, LLC d/b/a Country Select Fish, Delta Pride Catfish Inc., Harvest Select Catfish Inc., Heartland Catfish Company, Pride of the Pond, and Simmons Farm Raised Catfish, Inc. The Final Results cover, inter alia, the mandatory respondent Vinh Hoan Corporation (“VC”), voluntary respondent Vinh Quang Fisheries Corporation, as well as the separate rate respondents An Giang Fisheries Import and Export Joint Stock Company (Agifish), East Sea Seafoods Limited Liability Company (ESS LCC), and Southern Fishery Industries Company, Ltd. (South Vina).

Court No. 11-00109 Page 3 sustained as is on matters affecting them. The defendant argues for remand of some of the issues and for sustaining the results in all other respects. The matter is accordingly remanded, as follows.

Jurisdiction and Standard of Review

Jurisdiction is proper pursuant to 19 U.S.C. §1516a(a)(2)(B)(iii) and 28 U.S.C. §1581(c). Commerce’s antidumping duty determinations are to be upheld unless “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. §1516a(b)(1)(B)(i).

Discussion

The margin of dumping of subject merchandise is determined by comparing its export price or constructed export price with its “normal value” (“NV”), a calculation usually based upon home market or third-country sales, depending upon market viability. See 19 U.S.C. §1675(a)(2). For a producer or exporter subject to a non-market economy (“NME”) such as Vietnam, the statute directs that NV shall be based on factors of production calculated by reference to an appropriate surrogate market-economy country or countries. See 19 U.S.C. §1677b(c)(1); see, e.g., Shakeproof Assembly Components, Div. of Illinois Tool Works, Inc. v. United States, 268 F.3d 1376, 1381 (Fed. Cir. 2001). Commerce is required to use the “best available information” in the selection of surrogate data, and the surrogate country should be, to the extent possible, (1) at a level of economic development comparable to the non-market economy country and (2) a significant producer of comparable merchandise. 19 U.S.C. §l677b(c)(1)&(4).

The plaintiffs’ claims mainly concern aspects of Commerce’s surrogate valuation (“SV”) methodology. That system normally relies on publicly available information and values all Court No. 11-00109 Page 4 factors of production in, or from, a single surrogate country. See 19 C.F.R. §351.408(c). First addressed below are matters on which voluntary remand is requested.

I. Voluntary Remand for Reconsideration of Certain Financial Data Included in Surrogate Financial Ratios

Commerce requests remand in order to reconsider including in its surrogate financial ratio calculations for the Final Results certain financial data for Gemini Sea Food, a Bangladeshi company, as Commerce had omitted to address the plaintiffs’ argument that evidence in the record indicates Gemini received a potentially countervailable government subsidy, and such a circumstance is proper for remand. See SKF USA, Inc. v. United States, 254 F.3d 1022, 1029 (Fed. Cir. 2001). The matter will be therefore be remanded (for reconsidering inclusion of Gemini’s financial data).

II. Voluntary Remand for Reconsideration of Surrogate Value for Fish Waste

In the Final Results, Commerce selected surrogate values for fish waste based upon Philippine import statistics for Harmonized Tariff Schedule (“HTS”) category 0304.90 (other fish meat of marine fish) maintained in the World Trade Atlas (“WTA”), and it rejected price quotes on the record the plaintiffs had obtained from Vitarich Corporation, a Philippine fish and seafood processor, consisting of an April 7, 2010 price list with per kilogram pickup prices of Pangasius fish waste (and trimmings, and fish skins) in Philippine pesos. See Sixth Review I&D Memo at 30-32. The plaintiffs contend these price quotes were accompanied by a supporting affidavit providing in substance the same information as that which accompanied two other price quotes, also of record, from Indian seafood processing companies, that Commerce had previously relied upon in prior proceedings, which reliance was upheld in Vinh Quang Fisheries Corp. v. United States, 33 CIT __, Court No. 11-00109 Page 5 637 F. Supp. 2d 1352 (2009). The plaintiffs contend: that it was unreasonable for Commerce to reject the Vitarich price quotes on the basis that they contained “no official company stamp” without explaining why this was relevant or the relevance of the quotes’ provision on Vitarich company letterhead documentation, that Commerce did not elaborate its concerns over public availability or address that the quote had been obtained upon the request of a member of the public as stated in the affidavit, and that Commerce did not adequately explain why a price quote that slightly post-dated the POR precluded its use when Commerce has “frequently” relied on non-contemporaneous data in other antidumping cases. Pls’ Br. at 32-36, referencing, inter alia, Jinan Yipin Corp. v. United States, 35 CIT ___, 800 F. Supp. 2d 1226, 1292 n.76 (2011) (“The ultimate question to be determined is: Do the price data accurately reflect prices throughout the period of review (whether or not those data are ‘contemporaneous’ and ‘representative,’ as Commerce defines those terms)?”); Sichuan Changhong Elec. Co. v. United States, 30 CIT 1481, 1504, 460 F. Supp. 2d 1338, 1359 (2006) (“Because the selected information appears to be more accurate, it cannot be said that Commerce was unreasonable in choosing it over a more contemporaneous, but less accurate alternative.”); see also Petitioners’ SV Submission (Apr. 8, 2010), PDoc 96, at Ex. 16.3

3 The price quote was accompanied by the affidavit of a Philippine lawyer explaining that she was retained to obtain this price data. PDoc 96 at Ex. 16. The affidavit further explained the prices are on an ex-factory and tax-exclusive base. Id. Along with this information, the attorney also identified from whom she obtained the price quote at Vitarich and included a copy of the Vitarich employee’s business card. Id. Commerce emphasized the facts that the quote “contains no official company stamp, was obtained outside the context of an actual business transaction, lists no terms of payment, does not list the person who provided the price, and was obtained after the POR,” Sixth Review I&D Memo at 28, and it also expressed “concerns as to whether this price quote is truly publicly available, to the extent that anyone from the public could duplicate it,” id., but the plaintiffs point out that Commerce did not address the fact that the affidavit identified the delivery terms and the party offering the price and explained precisely the manner in which the price quote was obtained. See PDoc 96 at Ex. 16.

Court No. 11-00109 Page 6

The plaintiffs argue Commerce’s rejection of the Vitarich quote (as well as the two Indian price quotes) is contrary to Commerce’s previous position and not adequately explained. Without admitting error, Commerce requests remand in order to reconsider its surrogate fish waste valuation. The matter will be remanded therefor, but upon remand Commerce will also address the plaintiffs’ concerns as articulated in their briefs. If on remand Commerce continues to be inclined toward reliance upon HTS data, it will clearly explain why neither the Vitarich price quote nor the previously-relied-upon Indian price quotes for fish waste were not the best available information to value fish waste as compared with the HTS data. See, e.g., PDoc 96 at Ex. 24; Petitioners’ Case Brief (Jan. 7, 2011), PDoc 222, CDoc 61, at 25 & n.74.

III. Surrogate Values for Broken Meat, and Fish Skins

In the Final Results, Commerce also had to select surrogate values for broken fish meat and fish skin. See Sixth Review I&D Memo at 32-33. To value respondents’ broken meat by-product, Commerce used import price statistics from the WTA for Philippine HTS category 0304.90, (other meat of marine fish). To value the fish skin by-product, Commerce selected WTA import price statistics for Bangladesh HTS category 2301.20 (flours, meals, and pellets, of fish or of crustaceans). Sixth Review I&D Memo, at 32-33. The plaintiffs here repeat that it was erroneous to rely upon such broad “basket” import statistics without considering the relative importance of product specificity in the process of surrogate valuation of the broken meat and fish skin by-products after they pointed out that HTS 0304.90 is a basket HTS category that by definition encompasses many types of meat from many species of fish and includes import data from countries that have no known production of Pangasius, PDoc 222 at 25 & n.75, and also that HTS 2301.20 includes a Court No. 11-00109 Page 7 variety of fish and crustacean products and does not accurately reflect the value of the respondents’ fish skin input, Petitioners’ Rebuttal Brief (Jan. 18, 2010), PDoc 227, at 193. Rather, the plaintiffs contend, the price quote they obtained from Vitarich, supra, is reliable and “highly product-specific” to Pangasius “trimmings” and “skin.” Pls’ Br. at 36, referencing PDoc 96 at Ex. 16.

In determining what constitutes “best available information,” Commerce must evaluate record evidence according to its surrogate value selection criteria. Commerce recognized that the Vitarich price quote “may be more specific,” but in “considering the other criteria” Commerce found the Philippines import data superior because the “Vitarich price quote is not contemporaneous, does not represent a broad market average, and is not publicly available.” Sixth Review I&D Memo at 29. Commerce therefore determined not to use that price quote for purposes of valuing the broken meat and fish skin. The defendant asks that this determination be sustained, but because Commerce’s reasoning here appears intertwined with its rejection of the Vitarich price quote in the context of valuing the fish waste, it is appropriate that Commerce reconsider the broken meat and fish skin valuations from a clean slate, alongside its reconsideration of the proper valuation of fish waste, supra.

IV. Surrogate Country Selection

The plaintiffs’ main challenge is to Commerce’s consideration of the data leading to its selection of Bangladesh as the primary market surrogate. See Sixth Review I&D Memo at 7-14.

A. Background

The selection of a surrogate country involves four steps. See Import Administration Policy Bulletin 04.1 (Mar. 1, 2004) (Non-Market Economy Surrogate Country Selection Process), Court No. 11-00109 Page 8 Commerce will (1) compile a list of countries that are at a level of economic development comparable to the country being investigated, (2) ascertain which of those countries produce comparable merchandise, (3) determine which of those are significant producers of comparable merchandise, and (4) if the selection process retains more than one country at this point, determine which country has the “best factors data” based upon the data’s quality (i.e., their reliability, accessability and public availability). See id. at 3. Commerce generally chooses for the administrative proceeding the most appropriate surrogate country by reviewing these criteria, but on occasion economic comparability cannot be determined until after the significant producer requirement is met. See id. The plaintiffs do not complain of the general adherence to this process during the administrative review at bar.

During the Sixth Review, Commerce was faced with having to determine whether the record with respect to the Philippines or Bangladesh contained the best available information for valuing factors of production. At the preliminary stage, as it had in prior reviews, Commerce determined that whole live fish accounted for the largest percentage of subject merchandise NV and were therefore its most significant input. Certain Frozen Fish Fillets from the Socialist Republic of Vietnam: Notice of Preliminary Results . . . of the Sixth Antidumping Duty Administrative Review and Sixth New Shipper Review, 75 Fed. Reg. 56062, 56066 (Sep. 15, 2010) (“Preliminary Results”) (concluding that the primary consideration must be “the availability and reliability of the surrogate values for whole live fish on the record”), PDoc 164. This is uncontested.

At this point, it is appropriate to summarize Commerce’s surrogate selection process during the prior administrative review. See Certain Frozen Fish Fillets from the Socialist Republic Court No. 11-00109 Page 9 of Vietnam, 75 Fed. Reg. 12726 (Mar. 17, 2010) (inter alia final review results) (“Fifth Review”) and accompanying I&D Memo.4 In that precedent Fifth Review, the plaintiffs argued for valuing the factors of production based on data they had submitted from the “Fish Pond Report” maintained by the Philippine Department of Agriculture, Bureau Agriculture of Statistics (“BAS”) for the country’s pangas fish production, as supported by the affidavit of an official of BAS. The respondents argued in favor of using the “FAO Report data” from United Nations Food and Agriculture Report: Economics of Aquaculture Feeding Practices in Selected Asian Countries for Bangladesh.

Commerce considered both contentions. Regarding the Fish Pond Report, although the supporting affidavit from the BAS official attested that the data in the report were finalized, Commerce voiced “concern with the quality and reliability of the chart and the data contained within it” due to the presence of “#DIV/0!” symbols in some of the data fields, the fact that it bore a handwritten title, and the fact that the affiant “affirms” that the data had yet to be finalized. Commerce found that the data had “yet to be presented in its normal publication” and thus found the data not yet publicly available, therefore not reliable, and therefore not the best available information with which to value the main input (fish). See Fifth Review I&D Memo at 8-10. Regarding the FAO Report, Commerce found that although it was not contemporaneous with the POR, it satisfied the other surrogate value selection criteria as to public availability, specificity to the input, and tax and duty exclusivity. Therefore, “taken as a whole,” Commerce concluded the FAO Report remained the best information available to value the main input. Id. at 10.

For the review now at bar, the plaintiffs again submitted the Philippines Fish Pond Report data. See PDoc 96 at Ex. 5-A and Ex. B-9; see also Petitioners’ Rebuttal Factor Value Data

4 Available at http://ia.ita.doc.gov/frn/summary/VIETNAM/2010-5853-1.pdf

Court No. 11-00109 Page 10 (Apr. 29, 2010 ), PDoc 110, and Petitioners’ SV Submission (July 9, 2010), PDoc 132, at Att. 1. The submission passed preliminary muster in relevant part, and Commerce was again faced with the choice of either the Philippines or Bangladesh as surrogate after certain other data and possible surrogates had been rejected. See Preliminary Results, 75 Fed. Reg. at 56066-67. This time, Commerce found each sets of data from their respective countries to be publicly available, tax and duty free, representative of broad market averages, and indicative of country-wide Pangasius production. See id. at 56067.

The preliminary determination notes that the Bangladesh FAO Report data specified coverage of “Pangasianodon hypothalmus”5 whereas the Philippines data identified the broader genus Pangasius, but Commerce concluded (and the parties do not appear to contest) there was nothing in the record from which to “determine that any difference between the two sources would necessarily generate a difference in price.” Id. Commerce also looked at the contemporaneity of the data sets and found the Philippine data contemporaneous with the period of review because they were from 2008, whereas the Bangladesh FAO Report data were from 2005. Id. Largely on this distinction, Commerce selected the Philippines as the primary surrogate country for the preliminary Sixth Review results. Id. at 56066-67. This selection resulted, ceteris paribus, in antidumping duty rates of $4.22 per kg for Vinh Hoan, $2.44 per kg for Vinh Quang, and $0.93 per kg for CL-Fish. The rate for Vinh Hoan was also assigned to the separate rate respondents Agifish, ESS LCC, and South Vina. See generally id. at 56065-69.

The parties thereafter submitted second and final surrogate country data. E.g., PDocs 194-197, 199, 215, and rebuttal commentary thereon, e.g., PDocs 210-212, 236. The plaintiffs’

5 I.e., Pangasius hypothalamus. Cf. note 1.

Court No. 11-00109 Page 11 submission provided updated Philippine BAS data showing approximately 34 tons of Pangasius produced in 2009 in addition to approximately 12 tons produced in 2008. Petitioners’ Factor Data, PDoc 196, at Ex. 1 (Table 51) (“FS 07-09”). The Vietnam Association of Seafood Exporters and Producers (“VASEP”), as an interested party before Commerce, also provided updating data that purportedly covered weekly wholesale prices gathered via “structured collection” methodology from all 70 regions of Bangladesh during the POR. These data consisted of spreadsheets VASEP had obtained from the Bangladesh Ministry of Agriculture’s Department of Agricultural Marketing. See VASEP’s SV Submission (Nov. 12, 2010), PDoc 195, at Ex. 7 (“DAM 08/09 data”).

In their administrative case and rebuttal briefs, the plaintiffs asserted that their Philippine FS 07-09 data were relevant and representative of farm-gate prices for whole live Pangasius, and they argued: that Commerce should not rely on either the Bangladesh FAO Report data or the DAM 08/09 data, that the spreadsheets therefor refer only to “pangas” prices and not the specific species covered by the Order, that the DAM 08/09 data were not part of an official published government report and should be rejected for the same reasons Commerce had rejected the plaintiffs’ submission of Philippine BAS data during the Fifth Review, that the DAM 08/09 data spreadsheets also contained the same missing-data symbols (i.e., “#DIV/0!”) that had informed Commerce’s rejection of the BAS data in the Fifth Review, that the DAM 08/09 data were wholesale and not farm-gate prices and therefore at a different level in the chain of distribution, that even though the DAM data were supposedly collected based upon interviews with farmers there is no quantity associated with the prices, and that the data were generally unreliable based on an affidavit obtained from a Bangladeshi lawyer who had interviewed DAM officials responsible for their collection. See generally PDoc 227 at 39-150. That affidavit asserts that DAM officials purport to Court No. 11-00109 Page 12 collect estimates of price averages from interviews with businessmen and customers but do not attempt to validate the data, and that the affiant was not provided, despite request, with a copy of the questionnaire used to collect the data. See id. at 56.

For the Final Results Commerce changed its preliminary determination and decided that the primary surrogate country should be Bangladesh on the basis of the DAM 08/09 data. Regarding the argument that the DAM 08/09 data should be rejected upon the same rationale Commerce had employed to reject the Fish Pond Report in the Fifth Review (to wit, that the DAM 08/09 data set is not an official, published government source and cannot be considered publicly available), Commerce disagreed “that the attributes of the DAM 08/09 data are so similar to those of the Fish Pond Report” that they would warrant rejection as a viable source to value the whole fish. After acknowledging that the Fish Pond Report’s lack of publication had been cause for concern in the Fifth Review, Commerce explained that this was “rooted in the fact that Petitioners claimed the data were to be published” or “source data to be used in a yet-to-be determined manner for official publication in the Fisheries Situationer,” i.e., the data may not have been finalized or were in draft form prior to publication. In contrast, Commerce noted the DAM 08/09 data were accompanied by a letter from a deputy director of the agency in the Bangladesh Ministry of Agriculture in charge of “collecting and disseminating the wholesale market price of various agricultural commodities, livestock, and fisheries, including Pangas.”6 Sixth Review I&D Memo at 12, quoting PDoc 195 at Ex. 7. The official in that letter declares that the DAM 08/09 “data can be provided to any member

6 In the process, Commerce minimized the affidavit that had accompanied the Fish Pond Report as not an “official statement from the Government of the Philippines, but rather solely a personal affidavit by the statistician in charge of compiling the data.” In this Sixth Review, Commerce does not directly fault the affidavit of this same person, averring in her (also apparently same) capacity as the “incumbent Chief of the Fisheries Statistics Division (“FSD”)” of BAS.

Court No. 11-00109 Page 13 of the public upon request, free of cost” and that the price data contained within is “country-wide data” representing “all months of years 2008 and 2009, covering all districts of Bangladesh.” Due to this “official certif[ication] as to the nature and breadth of the DAM 08/09 data, the completeness of the data, and the availability of the data to the public upon request,” Commerce concluded that “the DAM 08/09 data does [sic] not appear to be incomplete or not finalized” and thus found that they constitute publicly available information.

After rejecting the domestic petitioners’ argument that the use of “pangas” among the DAM 08/09 data spreadsheets is non-specific to the Order and therefore renders them unuseable, Commerce turned to address their argument that the spreadsheets contain the same “#DIV/0!” symbols that were a large cause for rejection of the Fish Pond Report data in the Fifth Review. In this instance, having found, a priori, the data implicitly finalized, Commerce reasoned

there is a clear distinction in this case, as the term appears in the DAM 08/09 data when there is no data for any given district of any given week for that month. In other words, if there were no Pangas pricing data available from [a particular] district for any of the weeks in a month, the monthly average column will show the term “#DIV/0!.” In fact, in every instance where the term “#DIV/0!” appears there is no weekly price data for that district, thereby causing the monthly average column to generate the formulaic term “#DIV/0!.” Therefore, it is reasonable to conclude that this term is generated simply as a function of the mathematical formula trying to perform a calculation on cells with no data in them. As such, we do not find the appearance of this term of any significance such that it would question the DAM data’s quality or reliability as Petitioners have argued.

Sixth Review I&D Memo at 11.

Commerce then addressed the domestic petitioners’ argument that the DAM 08/09 data are not be the best available information with which to value the whole fish input because those data are wholesale prices, not farm-gate level prices, and may thus include delivery costs, taxes and duties, and/or mark ups for wholesaler’s profit. Commerce stated it is “unclear whether the DAM Court No. 11-00109 Page 14 08/09 data wholesale prices necessarily include other costs[.]” Commerce then declared that the prices in FS 07-09 “do not contain only farm[-]gate prices” based upon its “plain reading” of the affidavit of the aforementioned chief7 of FSD, to wit, that the prices “quoted by the aqua farm farmers/operators (or other Respondents, as the case may be)” are “farm-gate or first-point-of-sale” prices. Commerce noted that the domestic petitioners at the public hearing “attempted to explain that what was meant . . . with respect to ‘first point of sale’ prices was a reference to the place of sale, not the format of the sale” but Commerce found “no record evidence to further clarify or corroborate” the domestic petitioners’ explanation. Rather, Commerce found that the statements in the affidavit “suggest[ ] that the prices in the FS 07-09 include prices other than strictly farm-gate, i.e., prices for different channels of distribution.” Commerce therefore found “the issue of whether the DAM 08/09 data or whether the FS 07-09 data represents solely farm[-]gate prices sheds little, if any, light in our analysis because both sources can be considered equally to contain information which suggests the prices are not solely farm-gate prices.”

Lastly, Commerce addressed the domestic petitioners’ argument that the DAM 08/09 data are not better than the prices in the FS 07-09 because they do not contain information as to the quantities sold, and therefore it cannot be determined whether the prices are based on commercial sales volumes of whole Pangasius or based on estimates or isolated spot prices. Commerce pointed to the DAM official’s affidavit, wherein is stated that “all the price information therein are in Bangladeshi Taka on a per Quintal basis, i.e., per 100 kg”, that the price data was “collected using a scientific method” and a “structured questionnaire”, involving interaction “with a network of all leading aqua farmers and wholesale traders as well as through direct market enquiry by visiting

7 Commerce repeatedly deflates to “statistician”. E.g., Sixth Review I&D Memo at 11.

Court No. 11-00109 Page 15 mandi (marketplace)”, that the weekly data are collected and forwarded to DAM, and that the monthly average price is based on such weekly price data points. Commerce characterized the domestic petitioners as “ultimately concerned with the overall reliability of the DAM 08/09 data due to the absence of volumes sold” and then it found that the DAM official’s explanation on the data collection methods in part “addresses any concerns with respect to reliability.” Commerce further explained that although it prefers to rely on data that contain volume and value information,

we have also used sources for major inputs in other cases that do not contain specific volume or value data used to generate the prices. For example, in two recent antidumping duty investigations where wire rod is the main input used to produce the subject merchandise (steel wire garment hangers and steel nails), the Department relied on a source that did not contain volume data.[ ] Both of these cases cited to others involving similar fact patterns, e.g., one relied on a publication the Department uses in cases involving chemical inputs and another involved frozen shrimp where the SV for the main input, raw shrimp, is derived from a source without quantity data.[ ] In other words, all other factors being equal, we found these data sources to be the best available information with which to value the major inputs, even in the absence of volume information.

Id. at 12 (footnote omitted). Thus Commerce found these “facts” rendered the absence of volumes with respect to the DAM 08/09 data “of lesser concern.” In the end, Commerce found

that both sources are publicly available, from a potential surrogate country, contemporaneous with the POR, broad market averages, are equally specific to the main input. Simultaneously, both can be considered equally to contain information which suggests the prices are not solely farm-gate prices. Given this degree of equivalence with respect to these factors, we examined the information upon which the Bangladeshi and Philippine potential surrogate whole live fish values were based, concluding that the Bangladeshi data represent a fuller set of data more appropriate for use as an SV. Therefore, as a result of the totality of the information considered above, we conclude that the DAM 08/09 data represent the best available data on the record with which to value the whole live fish input. Given the significance of the whole live fish input in the calculation of NV, we therefore conclude that the choice of Bangladesh offers more reliable SV information and thus select Bangladesh as the primary surrogate country for purposes of these final results.

Court No. 11-00109 Page 16 Id. at 13-14. This decision resulted in antidumping duty rates of $0.00 for Vinh Hoan, Vinh Quang, and CL-Fish, and $0.02 per kilogram for Agifish, ESS LCC and South Vina. 76 Fed. Reg. at 15944.

B. Analysis

The plaintiffs argue the defendant and defendant-intervenors have failed to demonstrate the legal viability of the DAM 08/09 data as the “best available information” because the determination was based on conclusions or assumptions contradicted by substantial record evidence and because Commerce did not, contrary to its stated policy and practice, consider the totality of available data for the full range of reported factors of production in its analysis.

1. Commerce’s Preference For “Farm-Gate” Over Wholesale Prices

Attention drawn in the Sixth Review to distinguishing between farm-gate and wholesale prices reflects the relative importance of level of trade in the administrative analysis. Commerce’s previously-stated preference is for farm-gate prices. See, e.g., Fifth Review I&D Memo, supra, at 15 (stating that farm-gate prices are preferable to downstream “market” prices because market prices may reflect additional other expenses). The Sixth Review record shows that the respondents purchased their whole fish directly from fish farms at farm-gate prices. See, e.g., Vinh Hoan’s Section D response (Jan. 6, 2010), PDoc 70, at 5-6; QVD’s Section A Response (Dec. 8, 2009), PDoc 53, at 20. The defendant argues (or admits) that this “record does not support a clear analytical distinction between farm gate and wholesale prices” but that “[b]oth are prices which a producer of fish fillets could pay for the whole live fish input.” Def’s Resp at 21 n.3. The contention veers into post hoc rationalization, given that Commerce stated as follows:

it is uncertain the extent to which prices clearly identified as being farm-gate prices or wholesale prices are relevant in the surrogate valuation analysis most importantly

Court No. 11-00109 Page 17

because surrogate valuation seeks to determine the price a respondent would pay for an input if it were to be producing in the surrogate country, not necessarily what producers of that input in the surrogate country receive.

Sixth Review I&D Memo at 11 & n.37.

This does not adequately address deviating from Commerce’s previously-expressed preference, ceteris paribus, for farm-gate prices. In the Fifth Review, for example, Commerce found farm-gate prices quite relevant when “determin[ing] the price a respondent would pay for an input if it were to be producing in the surrogate country[.]” In that review, Commerce rejected data pertaining to the “Pangas Thesis” precisely because it is “unclear whether [its] methodology relies on farm gate prices or market prices, and if market prices, what movement or other expenses are included in those prices.” Fifth Review I&D Memo at 15 (italics added). Deviation from practice may be upheld if the agency’s reasons therefor are valid, Allegheny Ludlum Corp. v. United States, 346 F.3d 1368, 1373 (Fed. Cir. 2003), but Commerce’s determination does not adequately explain why the distinction between farm-gate pricing and wholesale pricing is now irrelevant to determining the price a respondent would pay for an input if it were producing in the surrogate country.

2. Contradictory Record Evidence

a. Commerce’s Finding That The Philippines’ Data and Bangladeshi Price Data “Equally” Contain “Not Solely Farm-Gate Prices”

Commerce concluded “the issue of whether the DAM 08/09 data or whether the FS 07-09 data represent[ ] solely farm[-]gate prices sheds little, if any, light in our analysis because both sources can be considered equally to contain information which suggests the prices are not solely farm-gate prices.” Sixth Review I&D Memo at 11. This muddles the record. The defendant contends “Commerce meant that the DAM data are not farm-gate prices and [that] the BAS data Court No. 11-00109 Page 18 contain some prices other than farm-gate prices.” Def’s Br. at 23. That is also post hoc rationalization. If it accurately portrays what Commerce meant, Commerce should have so stated. Even then, the explanation is artificial, as the uncontested record shows that the Bangladesh DAM 08/09 data solely reflect wholesale-level prices, which circumstance does not “suggest” the inclusion (or rather straw-man disinclusion) of farm-gate prices whatsoever. Commerce’s articulation thus disingenuously put the data sets on “equal” footing and detracted from the analysis.

b. Commerce’s Interpretation of the Philippines Price Data

The conclusion that the Philippine FS 07-09 data “contain information which suggests that the prices are not solely farm-gate prices” depended upon interpreting the two affidavits pertaining to the FS 07-09 data. Considering the affidavit from the chief of the Philippines’ BAS’s Fisheries Statistics Division (“FSD”), Commerce found nothing of record to corroborate the petitioners’ interpretation of it and stated that a “plain reading . . . suggests that the prices in the FS 07-09 include prices other than strictly farm-gate, i.e., prices for different channels of distribution.” Sixth Review I&D Memo at 11. The affidavit is indeed plain, but it leads to the opposite conclusion. Relevant statements therein are as follows:

. . . Respondents must be aquafarm farmers, operators or caretakers. Other possible respondents are aquafarm traders and persons knowledgeable in the production of aquaculture in the locality. *** . . . Among the information included in the data collection and gathering are the price/value of the product per kilogram, volume of production in metric ton[,] and harvest area in terms of hectarage. The prices quoted by aquafarrm farmers/operators (or other respondents, as the case may be) are also referred to as first-point-of-sale- price or farm-gate price.

. . . The price stated in the Fisheries Statistics of the Philippines and in the Fisheries Situationer is referred to as the farm-gate price or the price quoted by the

Court No. 11-00109 Page 19

aquafarm farmers/operators at their first point of sale. The price stated is also tax exclusive. *** . . . The volume and value data for Pangasius in the attached schedule, entitled “Freshwater Fishpond, 2008,” is the complete and final compiled information collected for pangasius produced in the in the Philippines for the year 2008. This report forms part of the Bureau’s working papers and contains the statistical data used to prepare the official Fisheries Statistics of the Philippines and the Fisheries Situationer publications. . . .

. . . The Freshwater Fishpond, 2008 Report includes the quantity, value and weighted-average unit price data for whole live Pangasius produced and sold in the Philippines in 2008. . . .

PDoc 132 at Att. 1, ¶¶ 9, 13-14, 18-19 (italics added).

Taking into account whatever in the record supports the agency’s finding as well as fairly detracts, it cannot be concluded that interpreting this affidavit as referring to two different “channels of distribution” and implicit selling price points was reasonable. Contrary to the finding of no information of record to support the plaintiffs’ interpretation, the BAS information of record provides that agency’s clear definition of “farm-gate prices”: they are equivalent to “first point of sale” prices, to wit, “prices received by farmers and livestock raisers for the sale of their produce at the first point of sale, net of freight costs.” See Petitioners’ SV Submission (Dec. 13, 2010), PDoc 210, at Ex. 8 (italics added). That describes not “two different channels of distribution” but the same price point -- for production -- that does not include the cost of a different (or further) channel of distribution. “The” price stated in the Fisheries Statistics of the Philippines and in the Fisheries Situationer is thus plain, and the “or” employed in the above affidavit is conjunctive, not disjunctive. To imply or conclude that “the” price stated therein and published is one of commingled, different price points from different channels of distribution is to ignore, unreasonably, BAS’s stated Court No. 11-00109 Page 20 statistical focus on “the” price of live, whole fish. Substantial record evidence thus does not support the contrary administrative interpretation of this affidavit.8

A second affidavit of record influenced the agency’s opinion that “there is some evidence that the data for one of the Philippine Regions may include a small volume of further processed whole live fish which is meaningful to the analysis because prices of cleaned and cut fish are substantially higher than those for whole fish.” This affidavit was procured from an official of the Philippines Bureau of Fisheries and Aquatic Resources (“BFAR”), which is wholly separate and distinct from the BAS that prepares and publishes the Fisheries Statistics that would include the FS 07-09, who attests that he was the project leader of Pangasius development in “Region 2” and that

BFAR Region 2 demonstration farms produced 3200 kilograms of [Pangasius] fish in 2008 and 2009. This production was included in official government surveys and is accounted for in the Fisheries Statistics of the Philippines, 2007-2009.

Pangasius prices in BFAR Region 2 were significantly higher than those in other parts of the Philippines in 2008 and 2009. The prices were higher because the region is land-locked and is otherwise isolated from sources of marine[ ] and brackish water fish, making fish relatively scarce and thus more expensive than other parts of the Philippines.

8 BAS had also defined “traders” as “those who buy and sell goods or commodities” and “wholesalers” as “those who buy in bulk from farmers/raisers/fishermen and fellow traders.” The only reasonable interpretation of “aquafarm trader” in the affidavit is for obtaining his or her knowledge of “production of aquaculture in the locality” (including “price/value of the product per kilogram, volume of production, in metric ton and harvest area in terms of hectarage”) and quotation of first-point-of-sale price, i.e., the farm-gate price. Even construing, arguendo, “other respondent” in the affidavit to encompass a fellow-trader “wholesaler,” it is still plain that for purposes of FSD’s statistical information gathering any “prices quoted by” such wholesaler would still have to be first- point-of-sale prices (i.e., the farm-gate price) based upon such wholesaler’s knowledge of what he or she paid or would have paid to purchase product from an aquafarm farmer. That would be the only meaningful construction, because a price “quoted by” a wholesaler for a sale from that wholesaler, as defined by BAS, to another purchaser would not be a “first-point-of-sale” price, as defined by BAS.

Court No. 11-00109 Page 21

Some 10-15 percent of Pangasius fish harvested in Region 2 in 2008 and 2009 were sold cleaned, cut or otherwise not live or in whole form. Prices of the cleaned cut fish are substantially higher than those for whole fish.

VASEP Second SV Rebuttal Submission, PDoc 211, at Att. I,¶¶ 2-6.

The plaintiffs admit the possibility that some Pangasius were sold in further processed forms in Region 2, but they argued to Commerce that the BFAR official did not claim he had knowledge of the data collection procedures of BAS or indicate that the prices included in the FS 07-09 data reflected any sales of cleaned, cut, or non-whole Pangasius for Region 2, as he stated only that “production” is accounted for in BAS’s publication. The plaintiffs here contend Commerce simply connected two disparate statements and speculated that BAS included among the “substantially higher” prices of the Pangasius sold in Region 2 those that were “sold cleaned, cut or otherwise not in live or whole form” into its price reports.

“Substantial evidence” is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Consolidated Edison Co. of New York v. NLRB, 305 U.S. 197, 229 (1938). Commerce concluded BAS included the prices of sales of “cleaned, cut or otherwise not in live or whole form” Pangasius from BFAR Region 2 into its price reports, since it noted that the inclusion of such (obviously dead) fish “may explain” the “observation of price volatility in the FS 07-09 data for Region 2.” Sixth Review I&D Memo at 13. That conclusion runs counter to the BFAR official’s expressed reason for why prices for Pangasius for BFAR Region 2 as a whole were “significantly higher” as well as BAS’s focus on the production and price of whole, live fish.

The court is very much aware that Commerce’s mandate requires it to assess often conflicting or unclear evidence, that judicial review refrains from re-weighing the evidence leading Court No. 11-00109 Page 22 to an administrative determination, e.g., Matsushita Elec. Indus. Corp. v. United States, 750 F.2d 927, 936 (Fed. Cir. 1984), and that “the possibility of drawing two inconsistent conclusions from the evidence does not prevent an administrative agency’s finding from being supported by substantial evidence,” Consolo v. Federal Maritime Comm’n, 383 U.S. 607, 620 (1966) (citation omitted) (see, e.g., American Silicon Technologies v. United States, 261 F.3d 1371, 1376 (Fed. Cir. 2001), but speculation does not amount to reasonable inference, as it provides no factually-grounded basis for sustaining an agency’s determination. See, e.g., Fifth Review I&D Memo at 9 (wherein Commerce reasoned it would not be appropriate to “draw conclusions about the range of prices” from the Philippines Fish Pond Report “given the nature of the pool of respondents and the location and time period of the data collected”). While it is unclear from the record whether the BFAR official’s and FSD’s official’s statements are even inconsistent, it is, however, clear that Commerce’s conclusion impugns the stated focus of BAS’s statistical reports in addition to the veracity of the FSD official’s statements. If that was intentional, the issue, of whether the price-volatility finding is mere speculation or reasonable inference, fortunately need not be addressed at this time, because remand of the surrogate country determination as a whole is otherwise required. See infra.

On remand, the “observed price differential” for Region 2 may also need proper context. In addition to the foregoing, Commerce noted that the Philippines whole fish prices in 2008 ranged from 77.14 to 128.82 Php/kg, i.e., by 67%, and it also noted that “similar volatility is not seen in the Bangladeshi data[.]” Sixth Review I&D Memo at 13. The plaintiffs point out that this latter observation is inaccurate, as the DAM worksheets show price variations that are even wider among different Bangladeshi districts than those of the Philippine BAS data: for example, in January 2008, Court No. 11-00109 Page 23 the monthly average prices for “pangas--small” ranged from 2938 to 7100 Tk/quintal, i.e., a range of 142%. See PDoc 195 at Ex. 7. If Commerce again reaches the issue of price volatility on remand, it should reasonably address the plaintiffs’ concerns regarding such directly contradictory evidence. See, e.g., Huvis Corp. v. United States, 570 F.3d 1347, 1351 (Fed. Cir. 2009). The court here again emphasizes it is not substituting judgment for that of Commerce on these issues, it is merely observing;9 Commerce’s expressed preference for farm-gate prices may give way to a reasonable determination that they are not the “best” data for purposes of surrogate country selection if it provides a reasonable explanation for the choice, but thus far that explanation is lacking.

c. Commerce’s Interpretation of Bangladeshi Price Data

After determining that the issue of whether the DAM 08/09 data or the FS 07-09 data represent “solely” farm-gate prices “sheds little . . . light . . . because both sources can be considered equally to contain information which suggests the prices are not solely farm-gate prices[,]” Commerce then concluded the price data reflected in the Bangladeshi DAM 08/09 data, reflecting a per-100 kilogram basis, represent “a fuller set” than the Philippines data and are reliable because they were “collected using a scientific method”. This determination requires reconsideration and a fuller explanation.

9 The plaintiffs also ask the court to take judicial notice of the fact that in the seventh administrative review Commerce rejected the worksheets as not publicly available, see Certain Frozen Fish Fillets From the Socialist Republic of Vietnam: Final Results and Partial Rescission of the Seventh Antidumping Duty Administrative Review, 77 Fed. Reg. 15039 (Mar. 12, 2012) and accompanying I&D Memo. A court may do so (e.g., Borlem S.A.-Empreedimentos Industrais v. United States, 913 F.2d 933, 940 (Fed. Cir. 1990)), and the plaintiffs have also submitted a “notice of supplemental authority” concerning the eighth administrative review covering the 2010-2011 period, see 78 Fed. Reg. 17350 (Mar. 21, 2013), but there is no need to refer to those determinations at this point.

Court No. 11-00109 Page 24

The plaintiffs had argued, as described above, that the DAM 08/09 data do not indicate specific quantities associated with the prices indicated, or, for that matter, what type of prices are indicated (whether “actual”, mere estimates, or isolated spot prices), and, further, that there is no basis in the record for ascertaining that the prices would reflect the type of normal commercial quantities producers or exporters of subject merchandise would require. Commerce agreed that its preference is to rely on data that contain volume and value information, but it determined that the instant review presented a “similar fact pattern[ ]” to certain of its precedents wherein Commerce had used sources for major inputs without such data.10 The defendant adds that if “the factual circumstances warrant using prices that have no associated quantities, Commerce will do so.” Def’s Br. at 28-29. But it does not elaborate on what those “factual circumstances” are, and neither does Commerce.

The plaintiffs, however, argue that in both of the referenced determinations, Commerce had valued steel inputs using a publicly available industry data bank that represented “national-level steel monitoring by a joint government/industry board” in India and which was used throughout the Indian steel industry as a market index for steel prices, whereas for this review no such “market” representation exists with respect to the DAM data, as is evident in the affidavits they submitted from Bangladeshi Pangasius farmers and from an attorney that had conducted interviews with those farmers and with the DAM official that supplied the wholesale price data to the

10 Sixth Review I&D Memo at 12, referencing Certain Steel Nails from the People's Republic of China, 73 Fed. Reg. 33977 (June 16, 2008) (inter alia, final LTFV determination) and accompanying I&D Memo at cmt. 10, and Steel Wire Garment Hangers from the People's Republic of China, 73 Fed. Reg. 47587 (Aug. 14, 2008) (final LTFV determination) and accompanying I&D Memo at cmt 4.

Court No. 11-00109 Page 25 respondents. See PDoc 210 at Ex. 13. Countering, but without addressing the purported farmer affidavits of record, the defendant contends the plaintiffs’ Bangladeshi attorney’s affidavit is “self-serving”11 hearsay, whereas the “official letter” concerning the DAM data “is a direct representation from the Government of Bangladesh[, and a]s such, the data attains the status of being inherently objective data”. Def’s Br. at 31-32.

If an affidavit is made from personal knowledge and sets forth specific facts, then whether it is “self-serving” is beside the point. See, e.g., Caterpillar Inc. v. Sturman Industries, Inc., 387 F.3d 1358, 1374-75 (Fed. Cir. 2004); Payne v. Pauley, 337 F.3d 767 (7th Cir. 2003). On the charge of hearsay, Commerce’s hearings are subject to neither the Administrative Procedure Act, see 19 C.F.R.§351.310(d)(2), nor, e.g., the Federal Rules of Evidence.12 The defendant’s argument rather concerns credibility, and its latter point, for that matter, would be just as apt with respect to the affidavit of the Chief of FSD of BAS, concerning her government’s “inherently objective data.”

In any event, the explanation is not part of Commerce’s determination. Commerce did not, in fact, address this affidavit at all. But if matter is in the record and relevant, it must be addressed, and the affidavit submitted by the plaintiffs, concerning DAM’s price data collection methodology, appears of record and relevant. It should, thus, have been addressed, including, at a

11 The plaintiffs especially take issue with Commerce’s disregard for the contents of the affidavit of the attorney, whom they had dispatched to interview the same DAM official that had supplied the wholesale price data to the respondents. The affidavit purports the DAM official’s description of the “scientific method” as involving interviews of local wholesale businessmen who provide estimates, and then the DAM officials “just record the average price of pangas based on these estimates provided during these interviews” and without validation. 12 Even then, and if pursuant thereto, one exception to Fed. R. Evid. 802’s proscription against the admissibility of hearsay, of course, is “a statement describing or explaining an event or condition, made while or immediately after the declarant perceived it”. Fed. R. Evid. 803(1).

Court No. 11-00109 Page 26 minimum, notice of any concerns regarding its veracity. See 19 C.F.R. §351.301(c)(3)(iv) & (5)(i); cf., e.g., Certain Color Television Receivers from the People’s Republic of China, 69 Fed. Reg. 20594 (Apr. 16, 2004) (final LTFV determination), I&D Memo at cmt. 9 (contacting Infodrive India, the company from which parties had obtained proposed surrogate value data, in order to better understand the company’s data collection methods). A lack of attempted corroboration cannot, ceteris paribus, reasonably result in construal against the submitter.

While an agency’s “decision of less than ideal clarity” may be sustained if its “path may reasonably be discerned”, Bowman Transportation, Inc. v. Arkansas-Best Freight System, Inc., 419 U.S. 281, 286 (1974), the Sixth Review determination remains unclear as to what circumstances would permit reasonable reliance upon price data without associated volume or value information, or why the circumstances here so warrant. In essence, Commerce effectively states it has “done so in the past” in “similar” circumstances, and the reader is left unclear as to the parameters of what those similar factual circumstances are.13 Since remand is otherwise required, if Commerce again reaches this issue on remand, it will need to explain with precision what those circumstances are, and state why the present ones are similar.

Commerce and the defendant also call attention to the fact that the DAM worksheets provided 2828 data points whereas the FS 07-09 data had only 12 price points, and that country-wide pangas production in Bangladesh during 2008-2009 totaled 59474 metric tons (“MT”), which

13 See Sixth Review I&D Memo at 12 & n.41, referencing Certain Frozen Warmwater Shrimp From the Socialist Republic of Vietnam: Final Results of the First Antidumping Duty Administrative Review and First New Shipper Review, 72 Fed. Reg. 52052 (Sep. 12, 2007) and accompanying I&D Memo at cmt. 1. As summarized, supra, Commerce simply states with respect to that determination: “the SV for the main input, raw shrimp, is derived from a source without quantity data.”

Court No. 11-00109 Page 27 compared favorably versus the total volume of pangas production covered by the Philippine FS 07-09 during the same period: 47.14 MT as surveyed, or 2264 MT if sourced from Status of the Pangasius Industry in the Philippines. See, e.g., Def’s Br. at 28-29. The plaintiffs take issue with the fact that the figures for “country-wide pangas production for Bangladesh” stated in the Sixth Review came not from the DAM 08/09 data but from the Statistical Yearbook of Bangladesh published by the Bangladesh Department of Fisheries. The defendant admits there is “some uncertainty” as to the total production figures, but it argues that even taking the highest of the alleged Philippine production figures on the record, 2264 MT, and comparing it with the lowest production figure of record for Bangladesh, 58474 MT, “it is clear that Bangladesh has a much larger pangas producing industry than the Philippines.” Def’s Br. at 16, referencing Sixth Review I&D Memo at 13-14. That may be true, but once again that is not quite what Commerce stated. And cf. Sixth Review I&D Memo at 6, quoting, in part, Policy Bulletin 04.1 (“The statute does not require that the Department use a surrogate country . . . that is the most significant producer of comparable merchandise. . . . The extent to which a country is a significant producer should not be judged against . . . the comparative production of the five or six countries on [the Office of Policy]’s surrogate country list.”) (italics in original). Commerce only remarked that it “believe[d] these distinctions should be considered in the context of comparing these two competing data sources.” Sixth Review I&D Memo at 13. On the other hand, if Commerce is only inclined to tailor a finding equivalent to the defendant’s on remand, then remand would be futile. Cf. Bowman, supra, 419 U.S. at 286.

Be that as it may, Commerce’s determination that the Bangladeshi DAM 08/09 are “fuller” relies on a comparison of the 2828 data points in that set that were obtained from 64 of the 70 reporting districts, versus, for the FS 07-09 set, 12 data points obtained from the 5 of the largest Court No. 11-00109 Page 28 pangas-producing provinces out of 81 provinces in the Philippines. The plaintiffs argue that because the DAM data are weekly and the FS 07-09 are annual and at a higher level of aggregation, this is like comparing apples and oranges and therefore not meaningful. The court cannot agree that the comparison is not without some quantum of probative value on data set “fullness”, at least of country-wide wholesale prices of Pangasius in Bangladesh; theoretical conversion of the FS 07-09 data to weekly data points would still result, at least in absolute terms, of a set approximately onequarter the size of the DAM 08/09 set, assuming the court’s back-of-the-envelope calculation is correct. The court can agree, however, that it would be illogical to infer from such observation that the DAM 08/09 are therefore “better” than the FS 07-09 data (e.g., that the DAM 08/09 data are therefore more “comprehensive”), because the comparison does not prove that the Philippine sampling methodology does not provide statistically equivalent representation, in comparison with the DAM 08/09 data, of country-wide farm-gate prices for Pangasius. That does not address Commerce’s implicit and additional consideration of Pangasius production as being greater in Bangladesh than in the Philippines, supra; it merely leads back to whether farm-gate or wholesale pricing are the “better” data in this instance -- concerning which Commerce found both data sets equivalent as far as being “publicly available, from a potential surrogate country, contemporaneous with the POR, broad market averages, [and] equally specific to the main input” in any event.

With respect to the reliability of the data sets, for the DAM 08/09 price data Commerce gave credence to the representation of the Bangladeshi official that they were “collected using a scientific method . . . using a structured questionnaire”. As indicated above, that would also appear true of the FS 08-09 data, but in any event, Commerce stated “this further explanation on the data collection methods provides additional information which, in part, addresses any concerns with Court No. 11-00109 Page 29 respect to reliability,” and that it has also used sources for major inputs in other cases that do not contain specific volume or value data. See Sixth Review I&D Memo at 12-14; see also supra, note 10. That may sound reasonable up to a point, but it does not address the plaintiffs’ central contention that there is “no record evidence that links these two different Bangladeshi sources or any other basis for assuming that the Bangladesh worksheets cover more ‘sales’ or quantities than the Philippine national statistics.” Pls’ Reply at 11.

Statistics require proper context. To take the plaintiffs’ example, “if each data point in the DAM internal worksheets represented 5 kgs -- and nothing on this record indicates any quantities associated with each data point, or if there are any -- then the 2,828 data points would represent pricing for only 14.14 MT, which is less than the FS[ ]07-09 total quantity of sales.” Id. at n.13 (italics omitted). This seems a primary reason the plaintiffs have been arguing that, ceteris paribus, the DAM 08/09 data are not necessarily “better” than the FS 07-09 data, and the court agrees that these aspects of the Sixth Review determination, concerning the meaning of the absence of volume information among the DAM 08/09, and whether that set’s prices would be representative of commercial quantities of whole fish sales, also require re-examination or clarification. In that regard, while Commerce did not address the reliability of the FS 07-09 data similarly, it apparently deemed them, as above indicated, reliable. See Sixth Review I&D Memo at 12 (“we disagree . . . that the record evidence . . . is unreliable”).

The plaintiffs also contend Commerce acted inconsistently with respect to determining that the DAM spreadsheets were “finalized” despite the numerous instance of Court No. 11-00109 Page 30 “#DIV/0!” symbols in certain fields,14 when in the precedent Fifth Review Commerce had precisely pointed to the presence of those symbols among the data proffered by the plaintiffs, as obtained from the Government of the Philippines, as reason for not finding their data finalized. See Fifth Review, I&D Memo at 9-10. That was not, however, Commerce’s sole reason for rejection in the Fifth Review. The defendant points out that in contrast to the Fifth Review, for the Final Results Commerce had before it an “official endorsement” from the Bangladeshi government, i.e, PDoc 195 at Ex. 7. The Sixth Review I&D Memo, beginning on page 9, restates that Commerce had “legitimate concerns” during the Fifth Review “that the data may not have been finalized or was in draft form prior to publication” and offers explanation of why, for this Sixth Review,

the DAM 08/09 data do[ ] not appear to be incomplete or not finalized. Here, we have a Bangladeshi Government official certifying as to the nature and breadth of the DAM 08/09 data, the completeness of the data, and the availability of the data to the public upon request.

Sixth Review I&D Memo at 10. Commerce went on to interpret the presence of the #DIV/0! symbol in this review as “the function of the mathematical formula trying to perform a calculation on cells with no data in them” because “in every instance where the term ‘#DIV0!’ [sic] appears there is no weekly price data for that district.” Id. at 11. The court considers that the presence of #DIV/0! symbols on a spreadsheet may reasonably be interpreted to indicate a lack of finality if other

14 Specifically, in the review at bar, Commerce “d[id] not find the appearance of this term of any significance such that it would question the DAM data’s quality or reliability,” as “the term appears in the DAM 08/09 data when there is no data for any given district of any given week for that month.” Sixth Review I&D Memo at 11.

Court No. 11-00109 Page 31 circumstances are present and likewise indicative. For that reason, the court cannot find unreasonableness in Commerce’s consideration of the issue for this Sixth Review.15

3. Surrogate Country Selection in Light of Entire Record

The plaintiffs’ broader argument is that because Commerce found a “degree of equivalence” between the DAM 08/09 and FS 07-09 data with respect to each of the key surrogate value selection criteria, Commerce’s failure to consider the totality of the surrogate value record was an abuse of discretion.16 The defendant argues that the silence in 19 U.S.C. §1677b(c)(l) as to what

15 I.e., the plaintiffs may be justified, as they argue, in complaining of goalpost-shifting at the hands of Commerce in a number of respects including this one, as, for example, in the Fifth Review the Philippines data they submitted were also accompanied by an “official endorsement,” and the record may have involved inappropriate denigration (cf., e.g., notes 6 & 7, supra; cf. also Public Hearing Transcript (Jan 26, 2011), PDoc 234, at 51-52 (regarding unprofessional data set acronym creativity)), but the court cannot conclude that Commerce’s conclusion on the issue for this administrative proceeding was unreasonable, given the standard of judicial review. 16 In particular, the plaintiffs aver that the record contains the financial statements of Philippine companies who produce only frozen fish products and the financial statements for two Bangladeshi companies producing solely frozen shrimp products and one producing both frozen fish and frozen fish products, and that Commerce’s “clear preference” for purposes of surrogate valuation is to select sources that are producers of identical merchandise. Pls’ Br. at 24, quoting Polyethylene Retail Carrier Bags from the People’s Republic of China, 74 Fed. Reg. 6857 (Feb. 11, 2009), I&D Memo at cmt. 2.10. They further aver that the record of Philippine data on secondary material inputs, energy factors, and packing materials is more contemporaneous to the POR than the Bangladesh data for those inputs. To the extent Commerce considered these arguments, it found the inclusion of frozen shrimp production among the available Bangladeshi data inconsequential to valuing the main factors involved in frozen fish production because “the production processes (capital structure) of which we believe to be similar [are] in terms of: cold processing area, freezing machines, and cold storage.” Sixth Review I&D Memo at 22. This observation may not seem unreasonable in isolation, but it was made in the context of considering the suitability of those Bangladesh companies’ financial statements, i.e., after Commerce had determined that the DAM 08/09 data were the “better” data. The observation was not in the context of considering the record as a whole, including all relevant facts, when determining whether the Bangladesh data or the Philippines data were the best information of record available. See Policy Bulletin 04.1 (“the country with the best factors data is selected as the primary surrogate country”) (italics added). Commerce (continued...)

Court No. 11-00109 Page 32 constitutes “best” available information provides Commerce “broad discretion to determine the definition of ‘best available information’ in a reasonable manner on a case-by-case basis”. Def’s Br. at 17, quoting Goldlink, 431 F. Supp. 2d at 1327 (citation omitted). In this instance, the defendant argues, “degree of equivalence” does not mean Commerce found the data “equivalent” but only with respect to certain factors, and that Commerce did, in fact, consider the entire record, ultimately finding the DAM 08/09 data a “fuller set” based on the number of data points as an indication of broader market coverage than the BAS data.

Generally speaking, the “fullness” of a data set does not address its suitability for the purposes sought by Commerce. Cf. Laizou Auto Brake Equipment Co. v. United States, 32 CIT 711, 717 (2008) ( “[i]t is clear that a larger data set, in and of itself, is not necessarily better in valuing factors of production than a smaller one”). The defendant adds, then, that the determination is also based on the fact that Bangladesh produced significantly more Pangasius than the Philippines during the POR. Def’s Br. at 12-13, referencing Sixth Review I&D Memo at 13-14; see also supra. Based upon these and other17 significant differences, the defendant argues it was reasonable for Commerce to find the DAM 08/09 data the best information available without the need to examine other factor

16 (...continued) also considered contemporaneity, but only in the context of the data for the “main input,” i.e., whole live fish. Commerce did not address the plaintiffs’ argument that Philippine secondary material data were more contemporaneous because it “ha[d] selected Bangladesh as the primary surrogate” and its “practice is to rely on upon the surrogate country for all SVs whenever possible.” Sixth Review I&D Memo at 21. That is, once again, putting the cart before the horse. 17 The defendant also points to Commerce’s finding of price volatility in the BAS data that can be explained by the record indication of BAS data including cleaned and cut fish with “substantially higher” prices than whole and live fish. See Sixth Review I&D Memo at 13-14. As indicated above, this finding will be re-examined and/or may be rendered moot upon remand.

Court No. 11-00109 Page 33 data as a “tie breaker” between the two data sets. The defendant-intervenors add that “the law does not oblige Commerce to preempt every possible riposte or to organize its determination according to a party’s particular taste.” Def-Ints’ Br. at 5.

Commerce’s finding of a “degree of equivalence” concerned purported farm-gate price data and wholesale price data. As above mentioned, Commerce determined along the path of equivalence as far as finding both sets “publicly available, from a potential surrogate country, contemporaneous with the POR, broad market averages, [and] equally specific to the main input[.]” Also as above mentioned, Commerce observed there is no specific evidence on the record as to what costs or expenses are included or not included in the DAM wholesale prices, Sixth Review I&D Memo at 11-12. That, however, is aside from Commerce’s stated preference for using farm-gate pricing, as is the fact that the DAM 08/09 data provide, arguendo, broader market coverage than the BAS data, as well as the fact that Bangladesh produced “significantly,” arguendo, more Pangasius than the Philippines during the POR. At the relevant point of “equivalence,” it is unclear to the court why Commerce’s analysis of the data for the main input did not, then, abide its policy of examining the totality of available data, as the plaintiffs argue. Commerce’s Policy Bulletin 04.1states that “if more than one country has survived the selection process to this point, the country with the best factors data is selected as the primary surrogate country” (italics added), and the data for each surrogate factor of production are supposedly accorded equal importance. See, e.g., Folding Metal Tables and Chairs from the People’s Republic of China, 76 Fed. Reg. 2883 (Jan. 18, 2011) (final review results), I&D Memo at cmt. 1C (“in selecting a surrogate country, we do not give more importance to financial ratios than to surrogate values for raw materials, but instead equally consider Court No. 11-00109 Page 34 all surrogate data in selecting a surrogate country”) (italics added). Cf. Camau Frozen Seafood Processing Import Export Corp. v. United States, 36 CIT ___, ___, 880 F. Supp. 2d 1348, 1360-61 (2012) (“Commerce’s conclusion that Bangladesh’s wage rate is the best available information for valuing the wage rate in Vietnam must be based on a reasonable reading of the entire record”) (italics added).

Surrogate valuation is not an exact science, but Commerce must approximate the factors of production as accurately as feasible. Certainly Commerce has discretion as to what information in the record is “best,” and there may be instances where data are clearly “better,” but if information is “available” in the record, the statute does not confer discretion to avoid addressing it. Cf. id. Commerce in this instance avoided considering the factors data for secondary material inputs, energy, and packing materials by determining, a priori, that the DAM 08/09 data represented the “best” data, resulting in the selection of Bangladeshi as the primary surrogate country. That logic precluded a fair selection of “the country with the best factors data . . . as the primary surrogate country” in the context of the record as a whole, including whatever “fairly detracts” as well as supports the determination of what is the “best” available information. The analysis is thus marred.

As it is unclear what impact any particular factor has had on Commerce’s analysis to this point, remand of the entire issue of surrogate country selection as a whole is appropriate, and without precluding reconsideration of the entire record for and against the selection of the primary surrogate country upon which to value the respondents’ factors of production. If Commerce also deems it necessary to gather additional information, it has the discretion to reopen there record. Court No. 11-00109 Page 35

V. Use of Unadjusted Factor Useage Data

In the underlying review, the respondent Vinh Hoan submitted farming factor (input usage) data requested by Commerce, including for its affiliate, Van Duc. Vinh Hoan reported Van Duc’s factor usage data for fish feed, labor, fingerlings, medicines and salts and lime upon the basis of Van Duc electronic books and records. See, e.g., Vinh Hoan Verification Report (Dec. 20, 2010), PDoc 214, CDoc 60, at 11. At verification, Commerce noticed there had been no recording of feed for the first five months of the POR. Id.18 Van Duc explained that the farms were new and the booking procedures had not been followed but averred that all of the activity in those months was accounted for in the warehousing record system in the January 2009 records accounting for the usage for the preceding five months. Id. The plaintiffs requested that Commerce either reject Vinh Hoan’s reported farming factors data outright or, pursuant to 19 U.S.C. §1677e, assume that the highest reported monthly usage of the inputs applied to the missing months. See CDoc 61 at 28-31 and Atts B and C. For the Final Results, Commerce accepted Vinh Hoan’s arguments that it had “fully” reported its farming factors. The Sixth Review I&D Memo quotes from a portion of the verification report as follows:

When asked, Van Duc officials provided us proof for January 2009 electronic warehouse-out slips. . . . Company officials explained that each of the warehouse-out

18 “Company officials explained that activity for the first five months of the POR was not recorded for feed until January 2009, as the farms were in their infancy, they did not know the procedures, and the records had not been delivered to [Van Duc headquarters].” PDoc 214, CDoc 60, at 11. When reviewing labor, it also found that “activity for the first six months of the POR was not recorded . . . until February 2009, as the farms were in their infancy, they did not know the procedures, and the records had not been delivered to [Van Duc headquarters].” Id. at 14. Verification exhibits showed similar problems for fingerlings, medicines, salt and lime. See id. at Exs 20-24.

Court No. 11-00109 Page 36

slips for January 2009 for each pond for each farm has a note detailing how much of the total activity is attributable to prior months.

Sixth Review I&D Memo at 35. Commerce did not note any discrepancies with this explanation and the documents provided at verification, and found that “the company explained how consumption was accounted for in its normal books and records” and “verified the consumption reported by Vinh Hoan by tying the numbers to the general ledger and/or financial statements.” Id.

The plaintiffs argue this shows only that Commerce verified some of Vinh Hoan’s consumption amounts for those select months for which data were reported and does not explain why no adjustment was made to account for their “critical omission,” or articulate how the margin calculation for Vinh Hoan is “as accurate as possible” given the verified evidence that Van Duc did not report full farming factors for the full POR, or explain why Van Duc’s explanation “was an adequate remedy for the fact that key consumption data remained unreported for multiple months of the POR.” Pls’ Br. at 38 (emphasis in original). The plaintiffs contend Commerce failed to provide a reasoned and adequate explanation of why “some adjustment” on the basis of facts available to account for the missing data was unnecessary. See 19 U.S.C. §1677e(a) (providing that Commerce “shall” use facts available if “necessary information is not available on the record” or if submitted information “cannot be verified”).

The two primary objectives of verification are to verify the accuracy of data submitted in a response, and to verify that relevant data were not omitted from the response. 15 Antidumping Manual §II.A. (Dep’t Comm. 2009). When calculating Vinh Hoan’s NV, Commerce relied upon Van Duc’s farming factor usage rates without resorting to facts available because it found that the allegedly “missing” data were not, in fact, missing, and Commerce did not find any discrepancy in Court No. 11-00109 Page 37 Vinh Hoan’s responses. See Sixth Review I&D Memo at 35, and Verification Report at 11, PDoc 214. According the general presumption of administrative regularity to which Commerce is entitled, the court cannot second-guess that Commerce did not properly verify how the usage data were accounted for in Vinh Hoan’s books and records in the absence of specific contrary evidence of record. The determination is therefore supported by substantial record evidence and is otherwise in accordance with law.

VI. Ministerial Error Allegations

After publication of the final results, the plaintiffs alleged two ministerial errors, namely that Commerce should adjust the surrogate financial ratio calculations to account for inventory changes incorrectly excluded from the selling, general and administrative expenses (“SG&A”) and profit calculations the changes in finished goods inventories derived from the financial statements of Apex, Gemini, and Fine Foods, and that Commerce had inadvertently failed to value the electricity and coal used to generate Vinh Hoan’s by-products. PDoc 251. With respect to Fine Foods, Commerce responded that there had been no error.19 With respect to the other alleged ministerial errors, Commerce admitted error, but declined to amend the Final Results, contending that correction would not affect the margins of any of the respondents. Id.

19 See Ministerial Error Allegations Memorandum to Gary Taverman, Acting Deputy Assistant Secretary for AD/CVD Operations, dated April 13, 2011, PDoc 249, at 3. Commerce also stated that the plaintiffs’ challenge was actually methodological, but that reason does not appear valid, as the only apparent avenue to raise the plaintiffs’ claim was ministerial, given Commerce’s departure from its Preliminary Results and its sua sponte decision to use Fine Food’s financial statements for the purpose of calculating surrogate financial ratios. The plaintiffs contend they only submitted those statements for the purpose of valuing whole live fish, not for the purpose of calculating surrogate financial ratios, and Commerce’s decision was without the benefit of briefing or argument from the parties. See Sixth Review I&D Memo at 22.

Court No. 11-00109 Page 38

Regarding the plaintiffs argument on Commerce’s calculation of SG&A and profit derived from Apex’s and Gemini’s financial statements, Commerce’s practice is to account for inventory changes in the calculation of the denominators of surrogate financial ratios. See, e.g., Seamless Refined Copper Pipe and Tube from the People’s Republic of China, 75 Fed. Reg. 60725 (Oct. 1, 2010) (final less than fair value (“LTFV”) determination), I&D Memo at cmt. 2 (explaining that Commerce’s practice is to (1) adjust the materials, labor, and energy expenses for changes in work-in-process inventories, and (2) adjust those expenses as well as factory overhead expenses included in the denominator of the SG&A and profit ratios “for changes in finished goods inventories”); Polyethylene Terephthalate Film, Sheet, and Strip from the People’s Republic of China, 73 Fed. Reg. 55039 (Sep. 24, 2008) (final LTFV determination) and accompanying I&D Memo at cmt. 3.

Normally, de minimis non curat lex, and “ the administering authority may . . . decline to take into account adjustments which are insignificant to the price or value of the merchandise.” 19 U.S.C. §1677f-l(a)(2). Commerce’s regulations define an insignificant adjustment as “any individual adjustment having an ad valorem effect of less than .33 percent or any group of adjustments having an ad valorem effect of less than 1 percent, of the export price, constructed export price or normal value, as the case may be.” 19 C.F.R. §351.413. Here, except for Commerce’s averment, the court has no basis for concluding what effect, if any, correction of the admitted ministerial errors, in addition to correction of any other identified errors, would have on the analysis. Given that even a “trifle” among the calculations in this instance may mean the difference between a finding of dumping and a finding of no dumping, it is appropriate that Commerce correct for error, particularly where remand is otherwise required. Cf., e.g., NTN Bearing Court No. 11-00109 Page 39 Corp. v. United States, 74 F.3d 1204 (Fed. Cir. 1995) (“where a remand is made to correct clerical errors made by the ITA, it would be paradoxical to deny consideration at the same time of similar errors of others”); Brother Industries, Ltd. v. United States, 15 CIT 332, 771 F. Supp. 374 (1991) (“court-ordered amendments of ministerial errors are not destructive of the ITA’s ability to manage its proceedings” and “a party’s ministerial or clerical errors have warranted correction where remand has been necessary on other grounds”) (citations omitted). On remand, Commerce is requested to incorporate correction (depending, of course, upon Gemini’s financial data’s continued inclusion in the analysis; see supra) and also include the value of coal and electricity in the by-product analysis.

Also, in the Sixth Review results, Commerce regarded the plaintiffs’ argument that it had not accounted for the increases in Fine Foods’ inventories of fish and shrimp when calculating the denominators of Fine Food’s surrogate overhead, SG&A and profit ratios, and Commerce denied error, taking the position (in contrast to the above) that “[t]he treatment of changes in inventory is done on a case-by-case basis”. After determining that the company’s statements lacked sufficient detail to account for work-in-progress inventory changes, Commerce intentionally excluded such incorporation into the denominator of its ratio calculations. See PDoc 249 at 3. Here, the plaintiffs repeat their contention that Fine Foods’ financial statements contained specific and discreet line items that would allow accurate capture in the relevant financial ratio denominators Commerce employs to account for inventory changes, e.g., the line items “Add: Opening Stock/Inventories” and “Less: Closing Stock/Inventories” in the cost of goods sold calculation. See PDoc 96 at Ex. 20, pp. 23. Note 1.18 to the financial statements (“Valuation of Inventories”) indicates that the inventories Court No. 11-00109 Page 40 were “of fisheries” as opposed to finished processed fish.20 Id., p. 20. Note 7 thereto (“Inventories”) itemizes the different materials included in inventory at the company’s two separate locations and aggregates the total inventories under the row labeled “Total fish.” Id., p. 23.

The defendant responds Commerce properly determined that the financial statement did not contain sufficiently detailed information to make the adjustment and that the plaintiffs are incorrect because the portions of the financial statement plaintiffs identify in their brief do not contain the necessary information. The defendant argues that in order to make the change in inventory adjustment, the financial statement must identify what type of inventory to which the numbers refer (raw materials, processing material, by-product, packing inventory, trade good inventory, and self-produced finished goods inventory) and that the Fine Foods financial statement does not identify the type of inventory to which the line items refer. The defendant further explains that the type of inventory is critical in determining where to include the inventory changes in the financial ratio calculations, i.e., whether for overhead (“OH”), SG&A, and profit: if there is a change in raw materials inventory, it is included in the denominator of the OH calculation (as the plaintiffs argue should be the case here), but if there is a change in the processing material inventory, it is included in the numerator of the OH calculation, and if the inventory change is for trade goods or self-produced goods it is included in the SG&A and profit ratios. No adjustment is made if the change is in packing or by-product inventory.

20 Specifically, Note 1.18 states that “management has valued inventories as mentioned in the subsequent paragraphs”; that “[a]ll the fishes except those kept and reared for breeding are listed in the inventory as Trading Stock of fisheries”;and that “[a]ll these Trading Stocks of fisheries have been valued at estimated net realized values as per management’s best estimate considering various market factors like volatility, demand and supply and the choices of customers.” PDoc 96 at Ex. 20, p. 20.

Court No. 11-00109 Page 41

The court can agree with Commerce that the Fine Foods financial statement does not have the degree of specificity that Commerce requires, as it is unclear what “trading stock” means, unless that is a term of art (of which the court has not been apprised), and there is no indication of the extent to which work-in-process is included in that term, or in the “break-up” of inventory into “pangas,” “tilapia,” “common carp,” “fingerlings,” etc., on page 23 of the financial statements of Fine Foods, whose principal activities include “processing fish and marketing the same products in local and foreign market,” see id. at p. 17, note 1.3, and as there is no separate “finished goods” inventory itemization, especially of frozen processed finished goods, the summary might or might not be inclusive of a raw materials inventory itemization. See id. at p. 27, n. 22. Therefore, the figures under “Inventories,” in note 7, could be finished product, or work-in-process, or some combination, as Commerce implicitly determined. That does not mean, however, that it was appropriate for Commerce to avoid the obligation to use facts available pursuant to 19 U.S.C. §1677e(a), given that Commerce’s determination is that Fine Foods’ financial statement offers appropriate surrogate values, that Commerce’s indicated practice (to the extent the court has been able to discern it) is to account for inventory changes in the calculation of the surrogate financial ratios, that such practice is indicative of the necessity of that accounting, and that there are, clearly, changes in inventory indicated on Fine Foods financial statements. At a minimum, Commerce must address any viable substitute(s) therefor (or lack thereof) for its financial ratio calculations, and more clearly explain its reasoning. Court No. 11-00109 Page 42

Conclusion

For the above reasons, the matter must be, and hereby is, remanded for reconsideration and further explanation in accordance with the foregoing.

The results of remand shall be filed by September 3, 2013, comments thereon, if any, by October 3, 2013, and rebuttal commentary, if any, by October 18, 2013.

So ordered.

/s/ R. Kenton Musgrave R. Kenton Musgrave, Senior Judge

Dated: May 23, 2013 New York, New York