Navneet Publications (India) Ltd. v. United States

OPINION AND ORDER GOLDBERG, Senior Judge: In this action, Plaintiffs Navneet Publications (India) Ltd. (“Navneet”), Marisa International, Super Impex, Pioneer Stationary Pvt. Ltd., SGM Paper Products, Lodha Offset Limited, and Magic International Pvt. Ltd. (collectively, “Plaintiffs”) raise various challenges to the all-others rate that the U.S. Department of Commerce (“Commerce”) imposed in the fifth administrative review of the antidumping duty order on certain lined paper products from India. See Certain Lined Paper Products from India, 78 Fed.Reg. 22,232 (Dep’t Commerce Apr. 15, 2013) (final admin. review) (“Final Results ”). Plaintiffs have moved for judgment on the agency record pursuant to USCIT Rule 56.2. See Pis.’ Mot. for J. on Agency R., ECF No. 34 (“Pis.’ Br.”). For reasons discussed below, the court grants Plaintiffs’ motion in part and remands a portion of Commerce’s Final Results. BACKGROUND On October 31, 2011, Commerce initiated an administrative review of the anti- *1356 dumping duty order on certain lined paper products from India. See Initiation of Antidumping and Countervailing Duty Administrative Reviews, 76 Fed.Reg. 67,-133 (Dep’t Commerce Oct. 31, 2011). The review period ran from September 1, 2010 through August 31, 2011 and covered fifty-seven Indian producers and exporters of the subject merchandise. Id. at 67,134-35. As part of its respondent selection process, Commerce issued quantity and value (“Q & V”) questionnaires to thirteen of the firms for which a review had been initiated. Commerce selected the firms on the basis of Customs and Border Protection data documenting companies that imported subject merchandise into the United States during the review period. See Resp’t Selection Mem. 4, PD 61 at bar code 3053175-01 (Jan. 20, 2012), ECF No. 30 (July 23, 2013) (“Resp’t Selection Mem.”). Only eight of the companies responded to the Q & V questionnaires. Id. One company that responded, Plaintiff Navneet, had also requested individual examination as either a mandatory or voluntary respondent. Voluntary Resp’t Request 1-2, PD 14 at bar code 3043588-01 (Nov. 29, 2011), ECF No. 30 (July 23, 2013). Commerce determined that it could not individually examine all fifty-seven companies subject to the review and instead limited its review to the two respondents accounting for the largest known volume of subject merchandise. Resp’t Selection Mem. 8. The two individually investigated respondents were Riddhi Enterprises (“Riddhi”) and SAB International (“SAB”), and Commerce preliminarily assigned those companies weighted average dumping margins of 3.86% and 2.30%, respectively. See Certain Lined Paper Products from India, 77 Fed.Reg. 61,381, 61,382 (Dep’t Commerce Oct. 9, 2012) (prelim, admin, review) (“Preliminary Results ”). In the Preliminary Results, Commerce also applied an adverse facts available (“AFA”) rate of 36.27% to the five companies that failed to respond to Commerce’s Q & V questionnaires. Id. The AFA rate derived from the highest non-aberrational margin calculated for mandatory respondent Riddhi during the review. See Prelim. AFA Mem. 1, PD 140 at bar code 3099879-01 (Oct. 1, 2012), ECF No. 30 (July 23, 2013). For the remaining companies that were neither individually investigated nor subject to an AFA rate (including all Plaintiffs), Commerce preliminarily calculated an all-others rate of 3.36%. Preliminary Results, 77 Fed.Reg. at 61,382. Relying on 19 U.S.C. § 1673d(c)(5)(A) (2006), Commerce arrived at the all-others rate by weight averaging the weighted average dumping margins of Riddhi and SAB. See Preliminary Results at 61,382 n. 1. That statute governs the calculation of all-others rates in investigations, which are usually based on individually investigated respondent rates unless those rates are zero, de minimis, or based entirely on facts available. See 19 U.S.C. § 1673d(c)(5)(A). Navneet subsequently submitted a rebuttal brief, anticipating that both Riddhi’s and SAB’s margins might fall below a de minimis threshold in the Final Results and that Commerce would need to use an alternative all-others rate methodology. See Navneet Rebuttal Br. 1, PD 172 at bar code 3109445-01 (Dec. 7, 2012), ECF No. 30 (July 23, 2013) (“Navneet Rebuttal Br.”). In its brief, Navneet requested that Commerce continue to calculate the all-others rate by averaging Riddhi’s and SAB’s rates, even if those rates later became zero or de minimis. Id. Navneet advocated this method because it believed that it would have received a zero margin if individually reviewed. Id. at 9. In support, Navneet argued that (1) it would have received zero margins in all other reviews if not for Commerce’s prior practice of zeroing negative dumping margins, *1357 and (2) Navneet’s sales and pricing patterns probably closely resembled those of Riddhi and SAB because it self-requested review. Id. at 9-10. Commerce published the Final Results of its review on April 15, 2013. See 78 Fed.Reg. at 22,232. As Navneet anticipated, Commerce revised the margins for Riddhi and SAB down to zero. See id. at 22,234. Commerce also calculated a new AFA rate of 22.02% (again, based on Riddhi data) and reduced the number of uncooperative respondents subject to that AFA rate to four. Id. However, Commerce did not adopt Navneet’s proffered method for calculating the all-others rate. Instead of assigning the remaining fifty-one companies a margin of zero percent, Commerce calculated a margin of 11.01% — the simple average of the zero percent rates assigned to the two mandatory respondents and the 22.02% AFA rates assigned to two of the uncooperative respondents. Id. at 22,233. The instant case ensued. SUBJECT MATTER JURISDICTION AND STANDARD OF REVIEW This Court has jurisdiction pursuant to 28 U.S.C. § 1581(c) and must uphold Commerce’s determination unless it is “unsupported by substantial evidence on the record, or otherwise not in accordance with the law.” 19 U.S.C. § 1516a(b)(l)(B)(i). Record evidence is substantial if a reasonable mind would accept it as adequate to support a conclusion. Nippon Steel Corp. v. United States, 337 F.3d 1373, 1379 (Fed.Cir.2003). The Court reviews the substantiality of the evidence “by considering the record as a whole, including evidence that supports as well as evidence that ‘fairly detracts from the substantiality of the evidence.’ ” Huaiyin Foreign Trade Corp. v. United States, 322 F.3d 1369, 1374 (Fed.Cir.2003) (quoting Atl. Sugar, Ltd. v. United States, 744 F.2d 1556, 1562 (Fed.Cir.1984)). The Court applies the rubric established in Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), to assess whether Commerce’s statutory construction accords with law. Specifically, the Court determines whether Congress has directly spoken to the question at issue. Id. If Congress’s intent is clear, the Court must give effect to that unambiguously expressed intent. Id. However, if the statute is silent or ambiguous, the Court assesses whether Commerce’s interpretation “is based on a permissible construction of the statute.” Id. at 843, 104 S.Ct. 2778. DISCUSSION Plaintiffs raise two challenges to Commerce’s calculation of the all-others rate in this review. Plaintiffs first contend that Commerce unlawfully incorporated an AFA rate assigned to uncooperative, uninvestigated respondents into the all-others rate calculation. Plaintiffs alternatively assert that the all-others rate did not reflect economic reality for uninvestigated respondents and that Commerce’s methodology was, thus, unreasonable. As set forth below, the court denies Plaintiffs’ motion as it pertains to the first issue, but agrees that Commerce did not support its all-others rate with substantial evidence and remands for further consideration. I. Legal framework for the calculation of “all-others” rates in anti-dumping duty administrative reviews In administrative reviews, Commerce “review[s] ... and determine^] ... the amount of any antidumping duty” and assesses final duties for companies for which a review has been requested. 19 U.S.C. § 1675(a)(1)(B). However, Commerce need not investigate every company sub *1358 ject to a review if Commerce reasonably determines that calculation of individual dumping margins is not practicable due to the large number of respondents. Id. § 1677f-l(c)(2). If Commerce reaches that conclusion, it may limit its review to a sample of mandatory respondents (often accounting for the largest export volumes of subject merchandise). See id. To arrive at margins for uninvestigated, cooperative respondents, Commerce calculates an all-others rate using the methodology found at 19 U.S.C. § 1673d(c)(5). Although § 1673d(c)(5) expressly applies only to investigations, Commerce also uses that statute to inform its analysis in administrative reviews. See I & D Mem. 13, PD 188 at bar code 3129602-01 (Apr. 9, 2013), ECF No. 30 (July 23, 2013) (“7 & D Mem.”). Section 1673d(c)(5)(A) instructs Commerce as a “[gjeneral rule” to calculate all-others rates using the weighted average of the weighted average dumping margins established for individually investigated respondents, excluding any zero or de minimis rates and rates based entirely on facts available. If no rates remain after making these exclusions, the statute directs Commerce to use “any reasonable method.” Id. § 1673d(c)(5)(B). The Statement of Administrative Action (“SAA”) accompanying the Uruguay Round Agreements Act 1 clarifies that the “expected method” under § 1673d(c)(5)(B) “will be to weight-average the zero and de minimis margins and margins determined pursuant to the facts available [calculated for individually investigated companies], provided that volume data is available.” See H.R.Rep. No. 103-316 (1994), reprinted in 1994 U.S.C.C.A.N. 4040, 4201. However, “if this method is not feasible, or if it results in an average that would not be reasonably reflective of potential dumping margins for non-investigated exporters or producers, Commerce may use other reasonable methods.” Id. To summarize, then, § 1673d(e)(5) and the SAA collectively establish the following hierarchy when calculating all-others rates — (1) the “[g]eneral rule” set forth in § 1673d(c)(5)(A), (2) the alternative “expected method” under § 1673d(c)(5)(B), and (3) any other reasonable method when the “expected method” is not feasible or does not reasonably reflect potential dumping margins. See 19 U.S.C. § 1673d(c)(5); SAA, 1994 U.S.C.C.A.N. at 4201. Notably, while a particular method may be reasonable as a legal matter, the method may still be unreasonable as applied in a particular case. See Yangzhou Bestpak Gifts & Crafts Co. v. United States, 716 F.3d 1370, 1378 (Fed.Cir.2013). As a result, remand may be necessary when Commerce’s all-others rate is unsupported by substantial evidence demonstrating that the rate reflects economic reality for uninvestigated respondents. Id. (noting same regarding separate rates, which Commerce calculates like all-others rates). II. Commerce’s method for calculating the all-others rate in this case was not unreasonable as a matter of law Because the rates calculated for the mandatory respondents in this ease were zero, Commerce proceeded under the “reasonable method” standard of § 1673d(c)(5)(B). However, instead of calculating a rate based exclusively on individually investigated respondents’ (Riddhi’s and SAB’s) zero margins, Commerce calculated the all-others rate by taking the *1359 simple average of Riddhi’s and SAB’s zero percent rates and the 22.02% AFA rates assigned to two of the uncooperative respondents. See Final Results, 78 Fed.Reg. at 22,233. A threshold issue in this case is whether that method was unreasonable as a matter of law. Plaintiffs claim that Commerce contravened law (1) by incorporating an AFA rate into the all-others rate, or (2) at a minimum, by incorporating into the all-others rate an AFA rate assigned to uncooperative, non-mandatory respondents. The court addresses each argument in turn. A. 19 U.S.C. § 1677e does not render Commerce’s methodology unlawful Plaintiffs first contend that the method Commerce selected to calculate the all-others rate in this case was unlawful because it violated another statute, 19 U.S.C. § 1677e. Pis.’ Br. 12-16. Section 1677e governs the use of facts otherwise available to complete a deficient record and, in certain circumstances, authorizes an adverse inference when selecting among facts otherwise available. Specifically, the use of an adverse inference against an interested party under § 1677e(b) necessitates a threshold finding that the “interested party has failed to cooperate by not acting to the best of its ability to comply with a request for information.” According to Plaintiffs, Commerce unlawfully partially based its all-others rate on an adverse inference even though Plaintiffs fully cooperated with Commerce’s requests in the underlying review. The court disagrees. As previously noted, the all-others rate statute expressly permits the inclusion of facts available rates. See 19 U.S.C. § 1673d(c)(5)(B); SAA, 1994 U.S.C.C.A.N. at 4201. Specifically, § 1673d(c)(5)(B) accords Commerce discretion to select among a variety of reasonable methods, including “weight-averag[ing] the zero and de minimis margins and margins determined pursuant to the facts available.” See SAA, 1994 U.S.C.C.A.N. at 4201 (emphasis added). For that reason, the Federal Circuit has already rejected the argument that AFA rates may not be incorporated into the all-others rate. See Yangzhou, 716 F.3d at 1378 (rejecting a similar argument because “§ 1673d(c)(5)(B) and the SAA explicitly allow Commerce to factor both de minimis and AFA rates into the calculation methodology”). In sum, had Congress intended to disallow AFA rates in this context, it would not have specifically authorized the use of such rates. Plaintiffs maintain that a different result is warranted because this case involves an administrative review and 19 U.S.C. § 1673d(c)(5) speaks only to investigations. According to Plaintiffs, § 1673d(c)(5) operates as a “carve-out or exception to the general rules concerning the application of adverse inferences set forth in 19 U.S.C. § 1677e(b).” Pis.’ Reply in Supp. of Mot. for J. on Agency R., ECF No. 46 (“Pis.’ Reply Br.”), at 6. Plaintiffs aver that when § 1673d(c)(5) is not directly applicable (i.e., in an administrative review), the “carve-out or exception” disappears and § 1677e(b) applies with full force to preclude the use of any adverse inferences in the all-others rate. Id. Otherwise stated, Plaintiffs submit that inclusion of an AFA rate may be lawful when calculating an all-others rate in an investigation, but it is necessarily unlawful when calculating an all-others rate in a review. The court does not read the relevant statutes to require this result. Though § 1673d(c)(5) explicitly references investigations, nothing in that statute or in any other statute expressly or impliedly precludes application to administrative reviews. As a result, Commerce has considerable discretion when selecting among *1360 possible methodologies. Although Plaintiffs acknowledge Commerce’s • discretion, they essentially maintain that Commerce could never reasonably exercise that discretion to follow the methodology from the only statute that addresses the calculation of all-others rates in antidumping duty proceedings — § 1673d(c)(5). See Pis.’ Reply Br. 5-6 (noting that such a methodology would be “manifestly impermissible under the second step of the Chevron analysis”). Plaintiffs offer no persuasive justification for this incongruous result, which conflicts with Commerce’s established practice' and Federal Circuit case law. See I & D Mem. 13 (noting Commerce’s practice of using § 1673d(c)(5) as “guidance” in administrative reviews); Yangzhou, 716 F.3d at 1378. Indeed, as noted, the Federal Circuit in Yangzhou summarily rejected the argument that Commerce may never use an AFA rate when deriving a “separate rate” 2 for cooperative, uninvestigated respondents in non-market economy proceedings. See 716 F.3d at 1378. The Federal Circuit reached this conclusion even though — similar to here— § 1673d(c)(5) is silent on the calculation of separate rates. See Baroque Timber Indus. (Zhongshan) Co. v. United States, 38 CIT -, -, 971 F.Supp.2d 1333, 1340 n. 19 (2014) (noting, in non-market economy investigation, that no statute directly speaks to the calculation of dumping margins for separate rate companies). This court is similarly unpersuaded that it is per se unreasonable to partially incorporate an AFA rate into an all-others rate in an administrative review. 3 *1361 B. 19 U.S.C. § 1673d(c)(5)(B)’s “reasonable method” standard does not, as a matter of law, preclude the incorporation of AFA rates assigned to uninvestigated, uncooperative respondents Plaintiffs alternatively argue that § 1673d(c)(5)(B)’s reasonable method standard does not permit the inclusion of an AFA rate assigned to uncooperative, uninvestigated respondents. See Pis.’ Br. 17-18. Thus, while the statute may in certain circumstances permit the inclusion of an AFA rate, Plaintiffs submit that the only permissible AFA rate would be one calculated for an investigated respondent. In support of this argument, Plaintiffs cite Changzhou Wujin Fine Chemical Factory Co. v. United States, 701 F.3d 1367 (Fed.Cir.2012). But the court disagrees that Changzhou establishes Plaintiffs’ proposition. In that case, Commerce was tasked with calculating a separate rate for cooperative, uninvestigated respondents in a non-market economy investigation. See id. at 1370. Applying the expected alternative method under 19 U.S.C. § 1673d(c)(5)(B), Commerce originally based the separate rate on a simple average of the de minimis rate assigned to one mandatory respondent (Wujin Water) and a total AFA rate assigned to the remaining mandatory respondent (Kewei). Id. at 1372. However, Commerce found on voluntary remand that it could no longer corroborate the AFA rate assigned to Kewei. Id. at 1372-73. In its place, Commerce calculated a “new, hypothetical AFA rate” based on Wujin Water’s verified normal value data and unverified U.S. price data taken from a non-cooperating exporter of subject merchandise. Id. at 1373. Commerce calculated this rate solely for incorporation into its separate rate calculations and the separate rate respondents were the only parties affected by the hypothetical rate. Id. Commerce also intentionally selected its data points to avoid a de minimis result, finding that result “would not be sufficiently adverse as to effectuate the purpose of the facts available rule to induce respondents” to comply with Commerce’s requests. Id. at 1378 (emphasis omitted). When employing this methodology, Commerce apparently felt that its “hands [were] tied” by § 1673d(c)(5)(B)’s expected method of averaging the individually investigated respondents’ de minimis and AFA rates. Id. The Federal Circuit disagreed. The court found that the only AFA rates contemplated under that paragraph “are those determined for ‘individually investigated’ parties,” and Commerce’s hypothetical AFA rate was not assigned to any individually investigated party (though it was a proxy for Kewei’s AFA rate). See id. at 1379. Ultimately, the Federal Circuit found that Commerce had a “duty ... to select a method appropriate for the circumstances” and that the method Commerce selected was inappropriate. Id. Specifically, the Federal Circuit found that Commerce could not reasonably “cherry-pick[ ]” data points with the sole purpose of increasing the margin for cooperative separate rate respondents. Id. The Federal Circuit never found that Commerce was legally barred from using an AFA rate calculated for and assigned to an uninvestigated respondent in its-separate rate calculations. Rather, the court found that Commerce could not elevate the *1362 averaging methodology of § 1673d(c)(5)(B) above other, more reasonable methods when the AFA rate at issue was only applied to adversely increase the margin for cooperative respondents and was not even calculated for an “individually investigated” company. See id. By contrast, the AFA rate here was not hypothetical and Commerce did not purport to proceed under § 1673d(c)(5)(B)’s averaging methodology. The rate was actually applied to uncooperative respondents and derived from data that Riddhi submitted during the proceeding. Final Results, 78 Fed Reg. at 22,233. “Although ... questionable in terms of economic reality, this court detects no legal error in” Commerce’s method viewed against the “lenient” requirement that the method simply be reasonable. See Yangzhou, 716 F.3d at 1378. Indeed, the court has previously found lawful a similar methodology. Baroque, 38 CIT at -, 971 F.Supp.2d at 1339, 1341 (finding, in a case where all mandatory respondents received zero margins, that it was “not per se unreasonable” for Commerce to calculate separate rates by simple averaging mandatory respondent rates and an AFA rate applied to the country-wide entity). 4 III. Commerce’s calculation method was arbitrary and unsupported by substantial evidence and was, accordingly, unreasonable as applied “Nevertheless, ‘[w]hile various methodologies are permitted by the statute, it is possible for the application of a particular methodology to be unreasonable in a given case.’ ” Yangzhou, 716 F.3d at 1378 (quoting Thai Pineapple Canning Indus. Corp. v. United States, 273 F.3d 1077, 1085 (Fed.Cir.2001)). Thus, the court must consider Plaintiffs’ secondary argument that substantial evidence does not support the reasonableness of Commerce’s methodological choice in this case. Specifically, the court must determine whether Commerce “articulate[d] a satisfactory explanation for its action” that is not based on “mere conjecture or supposition.” Id. (quoting, in part, 19 U.S.C. § 1677(7)(F)(ii)). In undertaking this assessment, “ ‘form should be disregarded for substance and the emphasis should be on economic reality.’ ” United States v. Eurodif S.A., 555 U.S. 305, 317-18, 129 S.Ct. 878, 172 L.Ed.2d 679 (2009) (quoting Tcherepnin v. Knight, 389 U.S. 332, 336, 88 S.Ct. 548, 19 L.Ed.2d 564 (1967)). This is because Commerce’s overriding purpose in administering antidumping law is to accurately calculate dumping margins. Yangzhou, 716 F.3d at 1379. In the same vein, to be supported by substantial evidence, “rate determinations for nonmandatory, cooperating separate rate, respondents must ... bear some relationship to their actual dumping margins.” See id. at 1380. As set forth below, the court finds that the record does not support Commerce’s action. At the administrative level, Commerce identified several facts purportedly sup *1363 porting its methodological choice in this case. First, Commerce noted that the general rule identified in § 1673d(c)(5)(A) was unavailable because both Riddhi and SAB received zero margins in the Final Results. See I & D Mem. 13. Commerce further found that it could not apply its preferred, alternative “reasonable method” of using margins previously calculated for respondents because those margins were the product of “zeroing,” 5 and Commerce no longer zeroes in administrative reviews. Id. at 14. Thus, Commerce opted for the simple average of Riddhi’s and SAB’s zero rates and two AFA rates assigned to the respondents that did not respond to Commerce’s Q & V questionnaire. Id. Commerce justified this choice by first concluding that it could not conduct a full respondent selection analysis without possessing complete Q & V responses. Id. Without conducting a full analysis, Commerce did not know whether it would have selected two uncooperative companies instead of Riddhi and SAB for individual review. Id. And because Commerce might have selected other respondents if it had a complete universe of Q & V data, Commerce could not conclude that Riddhi’s and SAB’s zero rates approximated the pricing behavior of the uninvestigated, cooperative respondents. Id. That explanation, without more, did not rise to the level of substantial evidence supporting Commerce’s methodological choice in this case. See Baroque, 38 CIT at -, 971 F.Supp.2d at 1343 (“The mere presence of non-cooperating parties ‘fails to justify [Commerce’s] choice of dumping margin for the cooperative uninvestigated respondents.’ ” (quoting Amanda I, 33 CIT at 1420, 647 F.Supp.2d at 1381)). Initially, Commerce’s rationale relied exclusively on the fact that limited data prevented Commerce from confirming the representativeness of Riddhi’s and SAB’s zero rates. However, that the record was so limited stems in no small part from Commerce’s decision to individually investigate only two companies. Commerce may not “explain the absence of evidence by invoking procedural difficulties that were at least in part a creature of its own making.” Yangzhou, 716 F.3d at 1378; accord Albemarle Corp. v. United States, 37 CIT -, -, 931 F.Supp.2d 1280, 1293 (2013) (“[T]he state of the record is not the fault of the separate rate respondents. The available data ... were limited by the Department’s decision to individually examine only two mandatory respondents.”). Furthermore, even if Commerce’s concerns regarding the representativeness of Riddhi’s and SAB’s zero rates might justify using a methodology other than the expected methodology under § 1673d(c)(5)(B), those concerns do not absolve Commerce of its duty to verify that the resulting rate reflects economic reality. See Baroque, 38 CIT at -, 971 F.Supp.2d at 1343-44. Otherwise stated, the incomplete Q & V data may provide Commerce with a reason to avoid using the expected methodology — a weighted average of Riddhi’s and SAB’s rates — but it would not justify assigning cooperative, uninvestigated respondents an all-others rate that is completely untethered to their pricing behavior. The 11.01% all-others rate that Commerce selected here appears untethered to respondents’ pricing behavior because (1) it is unsupported by eorro *1364 borative record evidence, and (2) is actually undermined by evidence suggesting that it is not an accurate depiction of pricing during the review period. Regarding the first point, the court notes that Commerce cited no evidence below suggesting that a rate of 11.01% reflects the economic reality of all-others rate respondents. In briefing before this court, the Government attempts to belatedly supplement the record with additional support. Specifically, the Government claims that “[t]he AFA rate of 22.02% was the highest, non-aberrational transaction-specific margin calculated for one of the mandatory respondents in the review, and as such, reflects the economic reality of the non-selected respondents in the review.” Def.’s Resp. to Pis.’ Mot. for J. on Agency R„ ECF No. 41 (“Def.’s Br.”), at 31-32. Quoting Yangzhou, the Government and DefendanUntervenor also claim that the selected all-others rate is reasonable because it is not “exceptionally larger” or “far in excess” of Riddhi’s and SAB’s zero rates. See id. at 32; Def.-Intervenor’s Resp. to Pis.’ Mot. for J. on Agency R., ECF No. 40, at 20 (citing Yangzhou, 716 F.3d at 1376, 1379). Aside from the fact that the court’s review is limited to the agency record, this reasoning is unpersuasive standing alone. While the 22.02% figure derived from actual sales data reported by Riddhi during the review, it was also purposely selected with adversity in mind and constituted but one sale out of many other non-dumped sales. Indeed, “if the presence of [a 22.02% margin] failed to justify assigning an overall above-de minimis rate [to Riddhi], then [that margin] certainly cannot serve to do so for the remaining cooperative companies.” Amanda Foods (Vietnam) Ltd. v. United States, 34 CIT -, -, 714 F.Supp.2d 1282, 1295 (2010). Furthermore, the bare assertion that a 11.01% all-others rate is not “far in- excess” of Riddhi’s and SAB’s rates is not substantial evidence that a rate of 11.01% “reasonably refleet[ed] ... potential dumping margins” for uninvestigated, cooperative respondents. See SAA, 1994 U.S.C.C.A.N. at 4201. Commerce’s sparse reasoning in this case was particularly questionable because there is evidence supporting a lower all-others rate. The all-others rate of 11.01% “represents a historic high” for cooperative respondents in proceedings under this order. 6 See Pis.’ Br. 9. For example, Commerce had previously calculated all-others rates of 1.22% in the first review; 1.34% in the second and third reviews; and 3.05% in the fourth review. Certain Lined Paper Products from India, 74 Fed.Reg. 17,-149, 17,152 (Dep’t Commerce Apr. 14, *1365 2009) (1st admin. review); Certain Lined Paper Products from India, 75 Fed.Reg. 7563, 7565 (Dep’t Commerce Feb. 22, 2010) (2d admin, review); Certain Lined Paper Products from India, 76 Fed.Reg. 10,876, 10,878 (Dep’t Commerce Feb. 28, 2011) (3d admin. review); Certain Lined Paper Products from India, 77 Fed.Reg. 14,729, 14,731 (Dep’t Commerce Mar. 13, 2012) (4th admin. review) (hereinafter, “FR Notices from Prior Reviews ”). Further, cooperative mandatory respondents received margins of 1.22% in the first review (Kejriwal Exports and Kejriwal Paper Limited); 1.34% in the second review (Navneet); 0.43% and 0.28% in the third review (Navneet and Super Impex, respectively); and 2.70% and 3.58% in the fourth review (Navneet and Riddhi, respectively). See FR Notices from Prior Reviews. When placed in context, the 11.01% figure appears aberrational because it is significantly higher than all prior margins calculated for cooperative respondents, and it represents a nearly four-fold increase from the preceding review during a time when mandatory respondent margins dropped to zero. 7 The 11.01% rate appears equally aberrational when placed among other data from this review. Although Commerce questions the reliability of the rates assigned to Riddhi and SAB, those zero rates nonetheless constitute the only contemporaneous evidence of pricing practices among large exporters of subject merchandise and are presumed to represent respondents as a whole. 8 See Amanda Foods (Vietnam) Ltd. v. United States, 36 CIT --, -, 837 F.Supp.2d 1338, 1345-46 (2012). Additionally, the Q & V data on the record also appear to detract from the reasonableness of an 11.01% all-others rate. According to Plaintiffs, the average unit values (“AUV”) of Riddhi’s and SAB’s subject exports during this review were [[Confidential Data Deleted]], respectively. See Pis.’ Reply Br. 23. The other six companies that responded to Commerce’s Q & V questionnaires reported AUVs of between [[Confidential Data Deleted ]]. Id. AUVs provide a “rough, estimated snapshot of a respondent’s pricing practices.” Yangzhou, 716 F.3d at 1376 (quoting Commerce’s remand results). A low AUV may be associated with a higher dumping margin, while a high AUV suggests a comparatively lower margin (if any). See id. Though of limited independent usefulness, 9 the fact that [[Confidential Data Deleted]] received a *1366 zero margin and its reported AUV was apparently the [[Confidential Data Deleted ]] is evidence suggesting that other respondents were also not dumping. Based on the foregoing, the court cannot find that substantial evidence supported Commerce’s all-others rate. Accordingly, remand is necessary so that Commerce can reconsider its methodology as applied in this case. CONCLUSION AND ORDER For the foregoing reasons, the court concludes that Commerce must reconsider the methodology that it used to calculate the all-others rate in the Final Results. Upon consideration of all papers in proceedings in this ease and upon due deliberation, it is hereby ORDERED that the Final Results be, and hereby are, REMANDED to Commerce for reconsideration and redetermination in accordance with this Opinion and Order; it is further ORDERED that Plaintiffs’ Rule 56.2 Motion for Judgment on the Agency Record be, and hereby is, GRANTED IN PART as provided in this Opinion and Order; it is further ORDERED that Commerce shall reconsider its method of calculating the all-others rate imposed against Plaintiffs, and redetermine those margins in accordance with this Opinion and Order; and it is further ORDERED that Commerce shall have ninety (90) days from the date of this Opinion and Order in which to file its Remand Redetermination, that Plaintiffs and Defendant-Intervenor shall have thirty (30) days from the filing of the Remand Redetermination in which to file comments thereon; and that the Government shall have thirty (30) days from the date of filing of Plaintiffs’ and Defendant-Intervenor’s comments to file a response to those comments. Footnotes [1] Congress recognizes the SAA as “an authoritative expression by the United States concerning the interpretation and application” of the Uruguay Round Agreements Act. 19 U.S.C. § 3512(d). [2] In non-market economy proceedings, an uninvestigated company that establishes a certain degree of independence from government control may receive a separate rate from the country-wide rate otherwise assessed against uninvestigated companies operating within the non-market economy. See, e.g., Yangzhou, 716 F.3d at 1373-74. Commerce calculates that "separate rate” using the framework from 19 U.S.C. § 1673d(c)(5). Id. at 1374. [3] Plaintiffs rely on a series of distinguishable cases to support the argument that Commerce may not lawfully incorporate AFA rates under § 1673d(c)(5) absent a finding of non-cooperation on the part of the uninvestigated parties. See Pis.’ Br. 13-16; Pis.’ Reply Br. 10-12. Initially, SKF USA Inc. v. United States, 33 CIT 1866, 675 F.Supp.2d 1264 (2009), did not even involve the calculation of all-others rates or separate rates. The cases that did address such rates also do not directly support Plaintiffs’ proposition. For example, Plaintiffs cite Yantai Oriental Juice Co. v. United States, 27 CIT 477, 487 (2003), for the proposition that a cooperating respondent's "failure to be selected [as a mandatory respondent] is an insufficient basis for the application of AFA” and that a "finding of non-cooperation is required.” Pis.’ Reply Br. 10. However, Yantai cannot reasonably be read to establish that proposition. Though the Yantai court remanded an all-others rate that was partially derived from a countrywide AFA rate, the court found only that Commerce did not rationally connect the resulting margins to separate rate respondents’ likely dumping margins. See 27 CIT at 487-88. The court never found that Commerce was legally barred from using AFA rates at all. Amanda Foods (Vietnam) Ltd. v. United States, 33 CIT 1407, 647 F.Supp.2d 1368 (2009) ("Amanda I”), is distinguishable for similar reasons. In Amanda I, Commerce assigned separate rate companies the rates that those companies had received in prior proceedings under the order. Id. at 1411, 647 F.Supp.2d at 1374-75 Although both mandatory respondents received de minimis rates, Commerce declined to weight average those margins because thirty-five companies received AFA rates and the record was devoid of data regarding those companies’ market shares or pricing practices. See id. at 1420, 647 F.Supp.2d at 1381. Though the court expressed concern with using data from prior proceedings due solely to the presence of uncooperative respondents, the court never found that Commerce could not lawfully incorporate an AFA rate into its calculations (and because Commerce relied on previously-calculated rates, Amanda I did not even concern that precise issue). See id. *1361 Ultimately, nothing in 19 U.S.C. § 1673d(c)(5)(B) limits Commerce's use of AFA rates to situations in which the all-others rate respondents have been uncooperative. While there may be legitimate concerns about whether including such rates accurately reflects uninvestigated respondents' pricing practices, that issue is addressed elsewhere in this opinion. [4] Plaintiffs alternatively contend that even if it were permissible to include AFA rates assigned to uninvestigated respondents in an investigation, it would not be permissible to do so in an administrative review because of the comparatively greater emphasis on precision in administrative reviews. Pis.’ Reply Br. 16-17. While the court agrees that investigations and administrative reviews differ in certain ways, those differences are unimportant in this context because Commerce must always ensure that its "reasonable method” under § 1673d(c)(5)(B) reflects economic reality and results in margins that are as accurate as possible. See Yangzhou, 716 F.3d at 1379-80 (citing, in part, Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1191 (Fed.Cir.1990)). Thus, there are other safeguards that protect against arbitrary, excessively imprecise results. [5] Zeroing is the practice whereby "negative dumping margins (i.e., margins of sales of merchandise sold at nondumped prices) are given a value of zero and only positive dumping margins (i.e., margins for sales of merchandise sold at dumped prices) are aggregated.” Union Steel v. United States, 713 F.3d 1101, 1104 (Fed.Cir.2013). [6] The Government justifies ignoring rates calculated in prior reviews on the basis that those rates are distorted due to zeroing and because agency decisions in other reviews are not binding in the instant review. Def.’s Br. 37-38. As a preliminary matter, Plaintiffs never argue that the previously calculated margins are binding in this review. Instead, Plaintiffs offer that evidence only to show— along with other evidence — that the 11.01% all-others rate might be unreasonably inflated. See Pis.’ Reply Br. 20. In any event, Commerce itself stated that it occasionally considered margins from other reviews when calculating all-others rates. See I & D Mem. 14. The Government’s concerns regarding zeroing are also not a persuasive reason for disregarding all prior-calculated rates. If anything, the fact that the previous margins were significantly lower than the 11.01% all-others despite the presence of zeroing detracts from the reasonableness of Commerce’s determination. While dumping margins based on zeroing might lead some to question whether the resulting margin is artificially inflated, zeroing can never lower a dumping margin. [7] The court has previously expressed concern with reliance on prior rates where those rates were calculated "in another review for other respondents” and might not rationally relate to pricing in subsequent reviews. See Albemarle, 37 CIT at -, 931 F.Supp.2d at 1292; see also Amanda I, 33 CIT at -, 647 F.Supp.2d at 1382. Notably, most of the Plaintiffs here (with the exception of Navneet and Super Impex) have never been individually investigated. Thus, the court's focus on prior rates should not be construed as a suggestion that Commerce use older margins without further explanation. Rather, the court references prior rates along with other data points to illustrate that Commerce did not support with substantial evidence the economic reality of an 11.01% all-others rate. [8] Indeed, Commerce justified its decision to investigate two respondents on its belief "that by selecting the largest exporters as mandatory respondents ... we will examine companies that account for a significant volume of total exports.” Resp’t Selection Mem. 8. If— contrary to what Commerce found in its Respondent Selection Memorandum — Commerce were concerned that Riddhi’s and SAB's data would not provide a representative sample, Commerce should have selected additional respondents for review. [9] In Yangzhou, the Federal Circuit accepted plaintiff Yangzhou Bestpak's challenge to a separate rate that Commerce corroborated using AUV data. 716 F.3d at 1380. In that case, Commerce had investigated two respondents: one respondent that cooperated (and *1366 received a de minimis margin) and another that did not (and received the country-wide AFA rate of 247.65%). Id. at 1375. Commerce calculated the separate rate as the simple average of those two margins and justified the economic reality of the separate rate by reference to AUVs. Id. at 1376. Commerce reasoned that the separate rate reflected economic reality because Yangzhou Bestpak’s AUV fell in between the AUVs for the two mandatory respondents. Id. The Federal Circuit rejected this reasoning as speculative. Id. at 1379. This court does not highlight the AUVs in this case to suggest that Commerce use that data exclusively to corroborate its all-others rate. Rather, as with margins calculated in prior reviews of this order, the court highlights the AUVs because those figures are some evidence detracting from the reasonableness of an 11.01% all-others rate.