Societe Nouvelle De Roulements (SNR) v. United States

OPINION TSOUCALAS, Judge: Plaintiff Societe Nouvelle De Roulements (“SNR”), a French manufacturer of ball and cylindrical roller bearings, moves pursuant to Rule 56.2 of this Court for judgment on the agency record challenging the final determination of the second administrative review of antifriction bearings from France as it relates to SNR. See Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From, France; et al; Final Results of Antidumping Duty Administrative Reviews (“Final Results ”), 57 Fed.Reg. 28,360 (1992), and amended by Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France, Germany, Italy, Japan, Sweden, and the United Kingdom; Amendment to Final Results of Antidumping Duty Administrative Reviews, 57 Fed.Reg. 59,080 (1992). SNR alleges that the Final Results are unsupported by substantial evidence and not in accordance with law because the United States Department of Commerce, International Trade Administration (“Commerce”), refused (1) to correct an error in the home market model codes reported by SNR; and (2) to change its computer matching routine to capture, for many U.S. models, the home market identical models which were contained in the database. Non-Confidential Rule 56.2 Motion of Plaintiff SNR (Societe Nouvelle De Roulements) for Judgment on the Agency Record (“SNR’s Brief”) at 1-32. SNR requests that the Court remand this action to Commerce for recalculation of SNR’s dumping margin in accordance with the corrected computer and calculation errors identified by SNR. 1 Amended Complaint at 4. Background On May 15, 1989, Commerce published antidumping duty orders on ball bearings, cylindrical roller bearings and spherical plain bearings and parts thereof (“bearings”) from various countries, including France. See Antidumping Duty Orders: Ball Bearings, Cylindrical Roller Bearings, and Spherical Plain Bearings, and Parts Thereof From France, 54 Fed.Reg. 20,902 (1989). 2 On June 28, July 19 and August 14, 1991, Commerce initiated administrative reviews of those orders with respect to sixty-three manufacturers or exporters, including SNR for the period May 1, 1990 through April 30, 1991. See Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From the Federal Republic of Germany, France, Italy, Japan, Romania, Singapore, Sweden, Thailand, and the United Kingdom; Initiation of Antidumping Administrative Reviews, 56 Fed.Reg. 29,618 (1991); Initiation of Antidumping and Countervailing Duty Administrative Reviews, 56 Fed.Reg. 33,251 (1991); Initiation of Antidumping and Countervailing Duty Administrative Reviews, 56 Fed.Reg. 40,305 (1991). On March 31, 1992, Commerce issued the preliminary results of its second administrative reviews, finding margins for SNR of 0.35% for ball bearings and 0.00% for cylindrical roller bearings. See Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France; Preliminary Results of Antidumping Duty Administrative Reviews and Partial Termination of Administrative Reviews (“Preliminary Determination”), 57 Fed.Reg. 10,859 *691 (1992). 3 On June 24, 1992, Commerce published one joint final determination for the nine administrative reviews. Dumping margins for SNR were established at 11.27% for ball bearings and 18.37% for cylindrical bearings. Final Results, 57 Fed.Reg. at 28,360, 28,361. On July 24, 1992, SNR commenced this action, challenging the Final Results with respect to France. On September 21, 1992, the Court granted The Torrington Company’s (“Torrington”) motion to intervene in this action. Federal-Mogul Corporation’s (“Federal-Mogul”) motion to intervene was granted on October 1, 1992. On December 14,1992, Commerce published amended Final Results, correcting certain clerical errors. See Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France, Germany, Italy, Japan, Sweden, and the United Kingdom; Amendment to Final Results of Antidumping Duty Administrative Reviews (“Amended Final Results ”), 57 Fed.Reg. 59,080 (1992). 4 SNR’s amended margin for ball bearings was 15.96%. The percentage for cylindrical bearings remained at 18.37%. Discussion The Court’s jurisdiction in this action is derived from 19 U.S.C. § 1516a(a)(2) (1988) and 28 U.S.C. § 1581(c) (1988). The Court must uphold Commerce’s final determination unless it is “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(l)(B) (1988). Substantial evidence is “more than a mere scintilla. It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 477, 71 S.Ct. 456, 459, 95 L.Ed. 456 (1951) (quoting Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 217, 83 L.Ed. 126 (1938)). “It is not within the Court’s domain either to weigh the adequate quality or quantity of the evidence for sufficiency or to reject a finding on grounds of a differing interpretation of the record.” Timken Co. v. United States, 12 CIT 955, 962, 699 F.Supp. 300, 306 (1988), aff'd, 894 F.2d 385 (Fed.Cir.1990). 1. Home Market Model Codes 5 In its home market database, SNR reported model codes, ie., product identification numbers or “IDNUMs,” which contained periods or blanks separating, within an alphanumeric string, the alphabetic from the numeric portion. In the U.S. market database, SNR’s reported model codes contained no such periods or separators. For example, SNR listed a particular model code in the home market as “N.307” but as “N307” in the U.S. market. SNR’s Brief at 13. The model codes in the two markets, therefore, were incompatible and prevented matching between comparable bearing models. In a letter to Commerce, dated April 3, 1992, SNR stated that the Preliminary Results had failed to match numerous U.S. and home market models. 6 See France P.R.Doc. No. 250, Reel 4, Frs. 868-871. SNR contends that the error relating to its home market IDNUMs falls under the “egregious and obvious” standard espoused in Tehnoimportexport v. United States, 15 CIT 250, 766 F.Supp. 1169 (1991) and thus Commerce abused its discretion in faffing to correct the *692 error. SNR’s Brief at 14, 22. In support, SNR argues that its inclusion of dots and blank separators in the home market IDNUMs caused thousands of identical price-to-price comparisons to register as “non-identical” and, consequently, artificially inflated SNR’s final dumping margin by approximately seven percentage points. Id. at 24-25. In addition, SNR argues that the disparity between its home market and U.S. IDNUMs is obvious from the printout of its Section C dataset of record. Id. at 19-24. After the Final Results were issued, SNR suggested the insertion by Commerce of a two-line computer command into the computer programming to compress SNR’s home market IDNUMs and eliminate the effect of dots and spaces contained therein. See June 24, 1992 letter to Commerce, France P.R.Doc. No. 308, Reel 4, Fr. 1383. In its response dated July 21, 1992, Commerce declined to make changes in SNR’s databases. See France P.R.Doc. No. 319, Reel 4, Fr. 1421. SNR protests Commerce’s refusal to correct this error and faults Commerce for not generally using the “compress” command which SNR claims is standard for “Statistical Analysis Software” (“SAS”) routines. SNR’s Brief at 14 n. 8; 22 n. 21. According to SNR, correcting the error requires no new factual information. Id. at 24. In the alternative, SNR argues that, even if the observation that its home market IDNUMs contained dots or separators is considered new factual information, this “fact” is self-verifying. Id. at 24 n. 23. SNR also contends that the negligible burden of correcting this error is outweighed by the preference for accuracy in dumping determinations. Id. at 26. Commerce responds that it examined SNR’s April 3, 1992, allegation that certain sales of U.S. and home market models which were identical had not been matched and found no error in its computer program. Defendant’s Memorandum in Opposition to Plaintiff’s Motion for Judgment Upon the Agency Record (“Commerce’s Brief”) at 7. In addition, Commerce argues that SNR made no attempt to correct the error and only made the actual nature and location of this error known after issuance of the Final Results. Commerce’s Brief at 6-7. Commerce also contends that its standard SAS program does not normally contain a “compress” command precisely because some respondents include periods, spaces, and other figures to signify some significant characteristic of particular models. Id. at 10, 11 & n. 7. According to Torrington, action by this Court to correct the error in SNR’s reporting would encourage respondents to avoid participation in administrative reviews. Response of The Torrington Company in Opposition to Plaintiff’s 56.2 Motion for Judgment on the Agency Record at 12. Federal-Mogul also opposes SNR, pointing out that, by compressing U.S. codes but reporting home market identifiers with separators, SNR disregarded specific questionnaire requirements that IDNUMs be reported consistently. Opposition of Federal-Mogul Corporation, Defendanh-Intervenor, to Plaintiff’s Motion for Judgment on the Agency Record at 11. It is a well-settled principle of administrative law that an agency has broad discretion to fashion its own rules of administrative procedure, including the authority to establish and enforce time limits concerning the submission of written information and data. Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 544-45, 98 S.Ct. 1197, 1212, 55 L.Ed.2d 460 (1978). In conducting administrative reviews pursuant to section 751(a) of the Tariff Act (19 U.S.C. § 1675(a) (1988)), Commerce requests pertinent information from those who will be affected by the results of the review. In order to comply with statutory requirements, Commerce promulgated a regulation which provides participants a specified period within which to submit the requested information. Commerce’s regulation which sets forth the time limits governing the submission of factual information provides: § 353.31 Submission of factual information. (a) Time limits in general. (1) Except as provided in paragraphs (a)(2) and (b) of this section, submissions of factual infor- *693 mation for the Secretary’s consideration shall be submitted not later than: (ii) For the Secretary’s final results of an administrative review under § 353.22(e) or (f), the earlier of the date of publication of notice of preliminary results of review or 180 days after the date of publication of notice of initiation of the review; or (2) Any interested party ... may submit factual information to rebut, clarify, or correct factual information submitted by an interested party ... at any time prior to the deadline provided in this section for submission of such factual information or, if later, 10 days after the date such factual information is served on the interested party---- 19 C.F.R. § 353.31(a) (1992). The regulation establishing the deadline for the submission of factual information warns that: (3) [Commerce] will not consider in the final determination or the final results, or retain in the record of the proceeding, any factual information submitted after the applicable time limit. [Commerce] will return such information to the submitter with written notice stating the reasons for return of the information. 19 C.F.R. § 353.31(a)(3). Hence, after the time period expires, Commerce “must refuse to consider untimely submitted information and instead rely on the best information available.” Ansaldo Componenti, S.p.A. v. United States, 10 CIT 28, 36, 628 F.Supp. 198, 204 (1986). The administrative record contains evidence that Commerce’s questionnaire required that product code information for U.S. and home market sales of identical bearings be submitted in identical form. Commerce’s questionnaire instructed that IDNUMs “must be reported in a consistent manner across all computer files.” See Antidumping Questionnaire, Gen. Issues P.R.Doc. No. 26, Reel 1, Frs. 1705, 1722. Therefore, SNR provided Commerce with data in an unreasonable format when Commerce had specified consistency in reporting. SNR submitted the information which Commerce requested but its home market data contained dots and separators. SNR’s response did not inform Commerce that the product codes presented a potential problem or suggest computer instructions to avoid any possible difficulties. Also, SNR had ample time, within the regulatory time limits, to discover any possible errors in its original submission and to propose a remedy or attempt to submit a corrected computer tape. SNR did none of this. Instead, SNR waited until April 3, 1992, to inform Commerce that certain sales of U.S. and home market models which it claimed were identical had not been matched. SNR’s letter did not specify that the failure to find matches was caused by inconsistencies in the model codes or suggest any remedy. See France P.R.Doc. No. 250, Reel 4, Fr. 868. Therefore, SNR’s initial response and its letter were inadequate in highlighting the exact problem and in not presenting a way to adjust the codes to allow matching of U.S. and home market sales. The facts of this case are similar to those in Neuweg Fertigung GmbH v. United States, 16 CIT 724, 797 F.Supp. 1020 (1992). After publication of the final determination in that case, Neuweg suggested programming instructions to rectify product code problems in its original submissions. The Court upheld Commerce’s resort to best information available (“BIA”) for sales which could not be matched because “[ultimately it is the respondent’s responsibility to make sure that [Commerce] understands, and correctly uses, any information provided by the respondent.” Id. at 728, 797 F.Supp. at 1024. In this case, Commerce examined each of the U.S. sales transactions which SNR claimed did not register a match and ruled out computer error. See France P.R.Doc. No. 290, Reel 4, Frs. 1249, 1251. It is not unreasonable to contemplate that Commerce did not have the time or resources to figure out how to adjust SNR’s product codes at that time. SNR waited until June 24, 1992, the day that Commerce issued the Final Results, to tell Commerce the exact nature of the error and to suggest a “compress” computer command to rectify the problem. See France P.R.Doc. No. 308, Reel 4, Fr. 1383. It is clear from the administrative record *694 that SNR’s new factual information was untimely under 19 C.F.R. § 353.31(a) as it was presented after publication of the Preliminary Results and over 180 days from the publication of the initiation of this review. The respondent in an antidumping proceeding bears the burden of preparing and providing Commerce with an accurate submission within the prescribed statutory deadline. See NSK Ltd. v. United States, 17 CIT 590, 593, 825 F.Supp. 315, 319 (1993); see also Chinsung Indus. Co. v. United States, 13 CIT 103, 106, 705 F.Supp. 598, 601 (1989). Respondents “must submit accurate data” and “cannot expect Commerce, with its limited resources, to serve as a surrogate to guarantee the correctness of submissions.” Murata Mfg. Co. v. United States, 17 CIT 259, 265, 820 F.Supp. 603, 607 (1993) (citation omitted). Thus, in general, Commerce is not required to correct a respondent’s errors when the respondent has reported erroneous data and has failed to timely correct it. See NSK Ltd., 17 CIT at 592, 825 F.Supp. at 318; NSK Ltd. v. United States, 16 CIT 745, 749, 798 F.Supp. 721, 725 (1992), aff'd, 996 F.2d 1236, 1993 WL 165133 (Fed.Cir.1993). See also Sugiyama Chain Co. v. United States, 16 CIT 526, 530-33, 797 F.Supp. 989, 994-96 (1992). Notwithstanding Commerce’s regulatory deadlines, where corrected information has been untimely filed and rejected by Commerce, the court may remand the case to Commerce for recalculation if the error complained of “was so egregious and so obvious that the failure to correct it was an abuse of discretion and undermined the interests of justice.” Tehnoimportexport, 15 CIT at 259, 766 F.Supp. at 1178. The error must be obvious from an examination of the administrative record in existence at the time the error is pointed out and the newly submitted information must be obviously correct. RHP Bearings v. United States, 19 CIT -, -, 875 F.Supp. 854, 857 (1995); NSK Ltd., 16 CIT at 749, 798 F.Supp. at 725. The question, therefore, is whether SNR’s original submission was obviously in error and whether SNR’s newly submitted information was obviously correct. The Court finds that the error alleged in SNR’s April 20, 1992, letter was clearly not obvious from the administrative record existing at the time that SNR highlighted the error, otherwise it would not have taken SNR almost eight months from the time it submitted the computer tape containing the erroneous information on October 21, 1991, to discover the error. See France P.R. Doc. No. 121, Reel 3, Fr. 1556. The fact that SNR did not specify the exact nature of the error in its April 3, 1992 letter to Commerce further supports the Court’s belief that the error was not obvious. In addition, if SNR considers this error egregious in effect, it has not demonstrated it. The record merely documents that SNR’s concern regarding unmatched sales was limited to 89 transactions according to its April 3, 1992 letter to Commerce, and to [ ] transactions according to its June 24, 1992 letter. See, respectively, France P.R. Doc. No. 250, Reel 4, Fr. 868 and No. 308, Reel 4, Fr. 1383. Hence, the Court finds that Commerce’s decision not to make changes in SNR’s database to eliminate the effect of dots and spaces contained in SNR’s home market IDNUMS is supported by substantial evidence on the record and is in accordance with law. 2. Commerce’s Family Model Matching Methodology In making identical comparisons between bearing models in the home and the U.S. markets for this review, Commerce’s computer program first grouped models into their respective families and then searched for identical matches within each family. SNR’s submitted data contained errors in listing the proper bearing family for a number of its bearing models. These errors became material because of Commerce’s family model matching methodology. SNR explains that it timely notified Commerce that “several” of the U.S. sales which had errors in the U.S. model code (“FAMILYE”) had identical matches in the home market database. SNR’s Brief at 26, 29. SNR contends that the identical models were not captured because Commerce used an impermissible “shortcut” model match sorting routine which sorted the home market and *695 U.S. databases initially by family, rather than by identical model, i.e., product code/model or IDNUM. Id. at 27-30. According to SNR, if the “product code/model” had been Commerce’s first sorting criterion, the computer program would have selected the correct model matches. Essentially, SNR argues that if Commerce had searched for identical, rather than similar models, in accordance with statutory preference, the typographical errors in its family coding would have been irrelevant. Id. at 27-31. SNR complains that the U.S. sales which had identical home market model matches but were not matched, were subject to a BIA rate of 56.50%. Id. at 30. In a letter to Commerce, dated April 3, 1992, SNR notified Commerce that its U.S. family descriptors contained typographical errors. See France P.R. Doc. No. 250, Reel 4, Frs. 868-871. On June 24, 1992, after receiving full disclosure of Commerce’s Final Results methodology, SNR filed comments with Commerce concerning the typographical errors in the FAMILYE and advocating the “product code/model-first” approach. See France P.R. Doc. No. 308, Reel 4, Frs. 1383, 1384. As noted above, Commerce declined to make changes in SNR’s databases. See France P.R. Doc. No. 319, Reel 4, Fr. 1421. SNR now requests that the Court instruct Commerce to adjust its model match sorting methodology to allow for identical matches in the entire home market dataset by the variable “IDNUM,” without regard to family designator. SNR’s Brief at 32. Commerce responds that SNR’s new information submitted to correct its list of certain bearing families was untimely and that the errors in SNR’s family descriptors were not obvious from the record in existence at the time SNR first highlighted the errors. Commerce’s Brief at 12 n. 8; 13. Commerce also defends its searching of identical matches only within the same families as a prudent programming instruction that makes efficient use of computer time. Id. at 14. SNR’s arguments fail to invalidate Commerce’s decision. First, SNR errs in claiming that its new information to correct its listing of certain bearing families was timely. SNR’s information was untimely pursuant to 19 C.F.R. § 353.31(a)(l)(ii) as it was submitted after the publication of the Preliminary Results and not within 180 days of the publication of the notice of initiation of this administrative review. As noted, a respondent “must submit accurate data” and “cannot expect Commerce, with its limited resources, to serve as a surrogate to guarantee the correctness of submissions.” Murata Mfg., 17 CIT at 265, 820 F.Supp. at 607 (citation omitted). The errors pointed out by SNR are typographical errors which should have been picked up before SNR’s data was submitted to Commerce. SNR cannot shift the burden of proofreading onto Commerce. This Court has commented that “if the burden of compiling, checking, rechecking, and finding mistakes in the submission of Plaintiffs were placed upon Commerce, it would transform the administrative process into a futility.” Sugiyama, 16 CIT at 531, 797 F.Supp. at 994. The Court has consistently held that Commerce is not required to correct a respondent’s errors when the respondent has reported erroneous data and has failed to timely correct it. See NSK Ltd., 17 CIT at 592, 825 F.Supp. at 318; NSK Ltd., 16 CIT at 749, 798 F.Supp. at 725. See also Sugiyama, 16 CIT at 533, 797 F.Supp. at 995. As noted above, in order for Commerce to correct an error after respondent submits untimely corrected data, the error in respondent’s original submission must be obvious from the administrative record in existence at the time respondent highlights the error. RHP Bearings, 19 CIT at -, 875 F.Supp. at 857; NSK Ltd., 16 CIT at 749, 798 F.Supp. at 725. In this case, SNR’s family descriptors contained various typographical errors. For example, as SNR points out in its April 3, 1992 letter to Commerce, in observations 3 and 4, FAMILYE should have been 310 instead of NG31. In observations 11-15, the FAMILYE should have been 309 not NU30. These errors in SNR’s original submission were clearly not obvious errors. It is not necessary to decide whether the effect of the error was egregious. See, e.g., Tehnoimportexport, 15 CIT at 259, 766 F.Supp. at 1178. *696 In addition, the Court believes that Commerce’s computer program was a reasonable and lawful means of attaining efficiency. Commerce’s methodology appears to be designed to match identical merchandise, or to match most similar models based upon bearing families when identical models are not available. This Court has previously upheld Commerce’s use of the family model matching methodology in such a context. See Torrington Co. v. United States, 19 CIT -, -, 881 F.Supp. 622, 635 (1995) (family model matching methodology found to be within Commerce’s broad discretion to devise a methodology for determining what constitutes “similar” merchandise). See also Koyo Seiko Co. v. United States, 66 F.3d 1204 (Fed.Cir.1995) (Commerce has implicit delegated authority to develop model-match methodology for determining what is “such or similar” merchandise under 19 U.S.C. § 1677b). SNR itself concedes that, absent its typographical mistakes in reporting FAMILYE, Commerce’s family-first “matching methodology would not cause otherwise identical bearings to become ‘non-matches’ because bearings with the same model number should always be in the same family.” SNR’s Brief at 29 (emphasis added). Thus, the Court cannot say that Commerce used an impermissible approach under the statute. See 19 U.S.C. § 1677b(a)(l)(A) (1988); 19 U.S.C. § 1677(16) (1988). Accordingly, Commerce’s refusal to change its computer program to minimize the effects of SNR’s flawed bearing FAMILYE data is supported by substantial evidence on the record and is in accordance with law. Conclusion The Court finds that Commerce properly refused to: (1) correct alleged errors in the home market model codes reported by SNR; and (2) change its computer program in order to alleviate the effects of alleged errors made by SNR in reporting families for certain bearing models. Therefore, Commerce’s reliance on SNR’s uncorreeted information to calculate SNR’s dumping margin is sustained. SNR’s motion for judgment on the agency record is denied and this case is dismissed. This case having been duly submitted for decision and the Court, after due deliberation, having rendered a decision herein; now, in accordance with said decision, it is hereby ORDERED that plaintiff’s motion for judgment upon the agency record is denied in all respects and Commerce’s determination is affirmed; and it is further ORDERED that this case is hereby dis- • missed. Footnotes [1] On January 12, 1993, SNR filed an amended complaint. [2] See also Federal Republic of Germany, 54 Fed. Reg. 20,900 (1989); Italy, 54 Fed.Reg. 20,903 (1989); Japan, 54 Fed.Reg. 20,904 (1989); Romania, 54 Fed.Reg. 20,906 (1989); Singapore, 54 Fed.Reg. 20,907 (1989); Sweden, 54 Fed.Reg. 20,907 (1989); Thailand, 54 Fed.Reg. 20,909 (1989); United Kingdom, 54 Fed.Reg. 20,910 (1989). • [3] See also Federal Republic of Germany, 57 Fed. Reg. 10,862; Italy, 57 Fed.Reg. 10,865 (1992); Japan, 57 Fed.Reg. 10,868 (1992); Romania, 57 Fed.Reg. 10,871 (1992); Singapore, 57 Fed.Reg. 10,873 (1992); Sweden, 57 Fed.Reg. 10,875 (1992); Thailand, 57 Fed.Reg. 10,877 (1992); United Kingdom, 57 Fed.Reg. 10,878 (1992). [4] SNR's amended complaint of January 12, 1993, dropped Counts IV, V, and VI of its original complaint in light of the Amended Final Results and otherwise eliminated Counts III and VII. See Amended Complaint at 1-4. [5] A "model code” is a string of characters, each of which represents a separate characteristic and which, together, represent a particular bearing model. Commerce’s computer seeking to identify identical models, will search for an identical string of characters among the model codes in the U.S. and home market databases reported by respondent. [6] The France administrative record is cited as "P.R.” and “C.R.” referring to the public and confidential records, respectively. “Gen. Issues” relates to the multi-country record.