PUBLIC VERSION OPINION AND ORDER

DONALD POGUE, District Judge

This case involves a challenge to the United States Department of Commerce's ("Commerce" or "the Department") data choices and adjustments in its calculation of an antidumping ("AD") duty on goods produced in a non-market economy ("NME"). Specifically, Plaintiff Zhejiang DunAn Hetian Metal Co., Ltd. ("DunAn") challenges the Department's data selection, use of partial adverse facts available ("AFA"), and scrap offset methodology in Commerce's final affirmative determination AD duty order regarding frontseating service valves ("FSVs") from the People's Republic of China ("China").

The court has jurisdiction pursuant to 28 U.S.C. § 1581(c).

Currently before the court is DunAn's USCIT R. 56.2 Motion for Judgment Upon the Agency Record.

Standard of Review

The statutory provision which supplies the standard for review for Commerce's final determination requires that the court shall "hold unlawful any [agency] determination, finding, or conclusion found . . . to be unsupported by substantial evidence on the record, or otherwise not in accordance with law." Tariff Act of 1930, § 516A (b)(1)(B)(i), 19 U.S.C. § 1516a (b)(1)(B)(i) (2006). See also 19 U.S.C. §§ 1516a (a)(2)(A)(i)(II), 1516a (a)(2)(B)(i); Huaiyin Foreign Trade Corp. (30) v. United States, 322 F.3d 1369, 1374 (Fed. Cir. 2003).

Discussion

By its instant motion, DunAn seeks (1) recalculation of the Indian import statistics used to value brass bar; (2) replacement of the labor wage rate — calculated in accordance with the Department's regulatory regression analysis — with an Indian labor rate of $0.21 per hour; (3) reversal of the Department's application of partial AFA to DunAn's December 2007 U.S. sales data and inventory carrying costs ("ICC"); and (4) revision of the Department's methodology for recognizing the value of DunAn's recycled brass scrap.

The court will, in turn, analyze each of these values.

A. Commerce's Selection of a Value for Brass Bar

Commerce's selection of a value for brass bar is governed by 19 U.S.C. § 1677b(c), which requires that Commerce choose data that is the "best available information" on the record. Here, Commerce selected an average unit value ("AUV") derived from the World Trade Atlas ("WTA") Indian import statistics for the POI. DunAn does not challenge Commerce's use of the WTA data set in general or the particular HTS classification used. Rather, DunAn argues that some aspects of the WTA data set are incorrect and should be eliminated. Specifically, DunAn challenges the sufficiency of the evidence to support Commerce's inclusion, in the WTA data, of brass bar values for Indian imports from France, Japan, and the United Arab Emirates ("UAE").

The relevant WTA data for HTS 7407.21.10 "Bars of Brass" are as follows: Country Quantity Value AUV (Kgs) (Rupees) TOTAL 91,884 27,588,000 300.25

Failing to remove imports from France, Japan, and the UAE constituted error, according to DunAn, because other record evidence, i.e., data from Infodrive India, demonstrated that shipments from the above countries did not consist of brass bar, and, thus, that WTA data as to the UAE, France, and Japan were flawed and unreliable.

The Infodrive India data as to France, Japan, and the UAE are as follows: HTS Code Actual Product Value Description Origin Qty (Kgs) (Rupees) AUV

As a result, the dispute between the parties on this issue turns on Commerce's assessment of these Infodrive data. Explaining its decision not to exclude these imports, Commerce stated:

we find that the Infodrive data contain insufficient product information in the description of the line items to enable the Department to make a definitive determination that these line items are misclassified. Specifically, the product description in the Infodrive data are such that, given the dependency upon the chemical make-up of the underlying products, they could be properly classified within the Indian HTS category where they are, or in the category addressed by DunAn. Thus, the Department cannot determine, due to lack of product detail, i.e., chemical properties, the precise chemical makeup of these line items. Accordingly, without clear evidence to the contrary, the Department will not speculate that these materials have been misclassified. Therefore, . . . the Department has determined to include imports from Japan, France, and the UAE in calculating the surrogate value for brass bar in the final determination because the record evidence does not demonstrate that the imports from these countries were misclassified.

DunAn argues that Commerce should have accorded more weight to the Infodrive India data and that Commerce should have, based on these data, eliminated the three countries' data as unreliable. DunAn in addition contends that it is irrelevant which HTS classification applies; "[t]he issue is not whether these shipments have been properly classified within the HTS [but rather] is whether these shipments are representative of the factor input Commerce is attempting to value." (Pl.'s Br. 13.) DunAn offered record evidence to demonstrate that bronze, beryllium copper, and cupronickel have different chemical properties and have different uses than brass. ( See id. 12-13;) DunAn's Second Surrogate Value Submission at Exs. 3A-3E. Therefore, according to DunAn, because the data from these three countries distorted the resulting brass bar surrogate value and because the Department's usage of that WTA data is not a selection of the "best available" evidence, Commerce's determination is unsupported by substantial evidence.

The government responds that the complete WTA data set under HTS category 7407.21.10 was the best available information on the record. Commerce chooses surrogate values on a case-by-case basis, and prefers to use "public, country-wide data." (Def.'s Opp'n to Pl.'s Mot. for J. Upon the Agency R. ("Def.'s Br.") 27 (quoting Mittal Steel Galati S.A. v. United States, 31 CIT 1121, 1124, 502 F.Supp.2d 1295, 1298 (2007)).) Consequently, the government argues, Commerce used "the full WTA data set, with the exception of imports from non-market economy countries, countries known to have country-wide export subsidies, or unidentified countries. because [the data] are publicly-available, broad-based, product-specific, tax-exclusive and contemporaneous." ( Id. (citing Decision Mem. at 22-23).) In addition, the government points out that Commerce did in fact examine the Infodrive India data, but did not find them definitive, particularly because DunAn failed to "provide any specific evidence that entries containing a certain amount of a chemical were inappropriately considered `brass bar' such as would be classified under the HTS category for brass bar." ( Id. 27, 28.)

The government is correct. The problem with DunAn's claim is that it does not focus on the specific information in the administrative record. In scrutinizing the Infodrive India data, Commerce confronted two inconsistent descriptions of the imports from the three countries at issue. Whereas the Infodrive India spreadsheet listed an "actual product description" for France, Japan, and the UAE, specified as copper and bronze, beryllium copper, and cupronickel, respectively, the spreadsheet also categorized the imported metals as "brass" under the HTS Code. Bronze, brass, beryllium copper, and cupronickel are all copper alloys and, although having distinct properties and uses, the types of metals at issue here are relatively similar in chemical makeup. See supra notes 13, 15-17. The accuracy of labels assigned to these imports, therefore, depends very much upon the types and percentages of metals contained therein. Commerce did not have this information. Thus, even had Commerce solely relied upon the Infodrive data, Commerce would still have had to choose between the "brass bar" classification listed and the perhaps conflicting product description. It was therefore reasonable for Commerce to rely upon the Infodrive India HTS classifications rather than the product descriptions, the former being consistent with the WTA data. As the court frequently emphasizes, "the possibility of drawing two inconsistent conclusions from the evidence does not prevent [Commerce's] finding from being supported by substantial evidence." Matsushita Elec. Indus. Co. v. United States, 750 F.2d 927, 933 (Fed. Cir. 1984) (quoting Consolo v. Fed. Mar. Comm'n, 383 U.S. 607, 619-20, 86 S.Ct. 1018, 16 L.Ed.2d 131 (1966)). It is not the court's place to choose between the classification and the description based on this record. See U.S. Steel Group v. United States, 96 F.3d 1352, 1357 (Fed. Cir. 1996).

DunAn cites to opinions of the Court which, according to DunAn, "admonished the Department . . . for failing to accord sufficient weight to Infodrive India data or other trade intelligence data that indicated the Indian import statistics utilized for the surrogate value were not representative of the input being valued." (Pl.'s Br. 9-12 (discussing Taian Ziyang Food Co. v. United States, ___ CIT ___, ___, 637 F.Supp.2d 1093, 1149 (2009); Zhengzhou Harmoni Spice Co. v. United States, ___ CIT ___, ___, 617 F.Supp.2d 1281, 1325, 1327 (2009); Globe Metallurgical, Inc. v. United States, No. 07-00386, 2008 WL 4417187, at *5-7 (CIT Oct. 1, 2008); Longkou Haimeng Mach. Co. v. United States, ___ CIT ___, ___, 581 F.Supp.2d 1344, 1363 (2008)).) See also Dorbest I, 30 CIT at 1698, 1700, 462 F.Supp.2d at 1286, 1288.

The case law DunAn cites, however, focuses on Commerce's failure to provide a sufficient reasoned explanation of its chosen data set. In these cases, the Department wholly failed to address "whether or not Infodrive India casts light on potential inaccuracies" in the WTA data. Dorbest I, 30 CIT at 1695, 462 F.Supp.2d at 1284. See also Zhengzhou, ___ CIT at ___, 617 F.Supp.2d at 1322-23 (rejecting Commerce's cursory presumption that Infodrive data were unreliable without "even a scintilla of evidence of manipulation or distortion or affiliation"), 1326-27; Taian Ziyang, ___ CIT at ___, 637 F.Supp.2d at 1156 ("Commerce simply dismissed [respondents'] concerns" in a "terse three sentences" and "failed to directly confront" respondent's claims), 1159-60. Instead, Commerce "conclude[d] that Infodrive data [are] unreliable or contain [] misclassifications, while simultaneously claiming that [WTA data are] both reliable and contain [] no inaccuracies." Dorbest I, 30 CIT at 1697, 462 F.Supp.2d at 1286.

In the determination at issue in this case, however, the Department directly addressed the relevance of the Infodrive data to the WTA data. Commerce did not merely dismiss the Infodrive India data out of hand, nor did it make a general finding that Infodrive data were unreliable. Rather, Commerce, assuming the Infodrive data were in fact reliable, directly discussed Infodrive India's relevance to the WTA data and found the Infodrive data to be inconclusive.

Further, even if the overall WTA data were being challenged, DunAn did not present Commerce with evidence that HTS 7407.21.10 includes imports that are other than the specific input at issue here. Moreover, DunAn did not provide evidence, other than Infodrive India, to attempt to demonstrate WTA misclassifications. Compare Dorbest I, 30 CIT at 1694, 462 F.Supp.2d at 1283 (noting that respondents provided evidence that Taiwanese export data did not match the WTA-listed Indian imports from Taiwan). See also Allied Pac. I, 30 CIT at 754-55, 435 F.Supp.2d at 1311-12; Zhengzhou, ___ CIT at ___, 617 F.Supp.2d at 1317 (rejecting Commerce's determination on grounds that plaintiffs provided specific proof that the relevant HTS classification was too broad and "the great majority of entries" covered goods other than the subject merchandise (quotation marks citation omitted)). In other words, here, both the WTA data and the Infodrive India data are sufficiently product-specific or representative to be considered the "best available" information. As the finder of fact, Commerce therefore had the discretion to choose between these data sets.

The court's duty, in reviewing Commerce's determination as to whether a certain set of data is the "best available information," is "not to evaluate whether the information Commerce used was the best available, but rather whether a reasonable mind could conclude that Commerce chose the best available information." Goldlink Indus. Co. v. United States, 30 CIT 616, 619, 431 F.Supp.2d 1323, 1327 (2006). Thus, "[i]f Commerce's determination of what constitutes the best available information is reasonable, then the [c]ourt must defer to Commerce." Id. Accord Fujian Lianfu Forestry Co. v. United States, ___ CIT ___, ___, 638 F.Supp.2d 1325, 1333 (2009); Zhejiang I, 2008 WL 2410210, *6-7, 2008 Ct. Intl. Trade LEXIS 69, at *20; Dorbest I, 30 CIT at 1676-77, 462 F.Supp.2d at 1269.

Accordingly, because Commerce's reading of the evidence is reasonable, the court rejects DunAn's challenge to Commerce's selection of WTA data, including data on imports from France, Japan, and the UAE.

B. Commerce's Calculation of Labor Costs

Next, DunAn challenges Commerce's derivation of a wage rate for DunAn's labor using a "regression-based" calculation that included wage rate and income data from countries (a) not economically comparable to China and (b) not significant producers of merchandise comparable to the subject FSV. DunAn instead advocates the use of an Indian wage rate as the surrogate wage rate for DunAn's labor. For the following reasons, the court sustains the Department's use of the labor regression model in this determination.

In its wage rate calculation, Commerce performs a simple regression to estimate the linear relationship between yearly per capita gross national income ("GNI/capita") and hourly wage rate ("wage"). To describe this relationship, Commerce uses publicly available GNI/capita and wage data from 61 market economy countries. GNI/capita serves as the independent variable (x) and wage as the dependent variable (y); predictably, these variables have a positive linear relationship. Commerce then estimates the wage for an NME country by using the NME GNI/capita.

Accordingly, although Commerce used Indian data for other FOPs, in calculating a 2005 surrogate wage for merchandise from China, Commerce regressed 2005 market economy wages on 2005 GNI/capitas, as follows:

Exhibit

Expected Wages of Selected Non-Market Economy Countries, supra note 27. Thus, Commerce created a 2005 NME regression line equation to describe the relationship between the 61 countries' wages and GNI/capita:

Wage[predicted] = 0.257585 + 0.000448 * GNI/capita[entered]

NME wage[predicted] = 0.257585 + 0.000448 * NME GNI/capita [entered]

See id. That is to say, by entering in an NME's GNI/capita, Commerce could predict the NME's wages.

Using this equation, Commerce then entered China's 2005 GNI/capita — $1,740 — into the above equation and derived a wage of, approximately, $1.04. Id. Commerce selected this surrogate wage rate as the labor FOP for the calculation of DunAn's COP. DunAn notes that, had Commerce limited its surrogate values to that of India as it did for other FOPs, DunAn's surrogate wage would have instead amounted to $0.21. (Pl.'s Br. 39;) Expected Wages of Selected Non-Market Economy Countries, supra note 27.

Like the plaintiff in Dorbest I and in the subsequent decision post-remand, Dorbest Ltd. v. United States ("Dorbest II"), ___ CIT ___, 547 F.Supp.2d 1321 (2008), DunAn attacks 19 C.F.R. § 351.408(c)(3) as invalid both facially and as-applied in this case. The court will address each of DunAn's arguments in turn.

1. Facial Challenge

First, DunAn argues that Commerce's regulation is contrary to the statute and, therefore invalid on its face. Following Dorbest I, the court rejects DunAn's facial challenge.

Section 1677b does, of course, require Commerce to utilize, to the extent possible, the prices or costs of factors of production in one or more market economy countries that are —

(A) at a level of economic development comparable to that of the nonmarket economy country, and

(B) significant producers of comparable merchandise.

DunAn attempts to distinguish Dorbest I on the grounds that Dorbest I did not address the substantial producer prong of section 1677b (c). However, the court's reasoning for economic comparability applies equally to the significant producer requirement, as to choice of an underlying data set, should such a requirement indeed exist. Therefore, Dorbest I is persuasive here.

DunAn also attacks Dorbest I by highlighting an excerpt from the legislative history of section 1677b(c). DunAn cites the Conference Agreement discussion of this provision contained in the Omnibus Trade and Competitiveness Act of 1988: "[t]he [FOPs] would be valued from the best available evidence in a market economy country (or countries) that is at a comparable level of economic development as the country subject to investigation and is a significant producer of the comparable merchandise." H.R. Rep. No. 100-576, at 590 (1988) (Conf.Rep.), reprinted in 1988 U.S.C.C.A.N. 1547, 1623. But the language from the Conference Report does not provide any more guidance than the wording of section 1677b(c), nor has DunAn demonstrated that the Report is in any way inconsistent with the court's reading of the statute.

Finally, DunAn points the court to other decisions from this Court that have invalidated 19 C.F.R. § 351.408(c)(3) as an impermissible and unreasonable implementation of section 1677b(c). See, e.g., Taian Ziyang, ___ CIT at ___, 637 F.Supp.2d at 1136-38; Allied Pac. Food (Dalian) Co. v. United States ("Allied Pac. II"), ___ CIT ___, 587 F.Supp.2d 1330 (2008). These opinions take the position that the regulation is invalid because it precludes Commerce from considering investigation- and product-specific wage data. See Allied Pac. II, ___ CIT at ___, 587 F.Supp.2d at 1356 ("The regulation requires a single calculated wage rate to be determined annually. . . . [and therefore] [t]he regulation does not permit a surrogate labor rate to be determined for an individual proceeding"), at 1358 ("Commerce's response in the preamble [to the regulation] falls short of a plausible explanation of why Commerce considered it acceptable to foreclose consideration of data specific to the type of labor required to produce comparable merchandise. . . . [Commerce's] rationale [is] insufficient to justify a regulation that disallows the use of data on the cost of a specific type of labor"). But the statute imposes a data selection requirement of country comparability, not merchandise specificity; nor does the statute require the use of data generated for an individual or specific investigation. See 19 U.S.C. § 1677b(c)(4) (Commerce "shall utilize . . . the prices or costs of factors of production in one or more market economy countries that are . . . significant producers of comparable merchandise"). Consequently, the Department's labor rate regulation is not, on its face, inconsistent with the statute.

2. As-Applied Challenge

Second, DunAn argues that the Department should have used the Indian wage rate in lieu of the number for China's wage rate derived from the regression. In this regard, DunAn does not attack the methodology Commerce used in implementing the regulation and does not, other than challenging the legality of the regression model, explain how the Indian wage rate instead constitutes the "best available information." Rather, DunAn argues that Commerce did not explain how use of global regression-based calculation has produced a wage rate that is more accurate for the valuation of labor in the FSVs industry. ( See Pl.'s Br. 39-40; Pl.'s Reply Br. 15.)

But Commerce did in fact explain its reasoning on this issue:

While surrogate values for other FOPs are selected from a single surrogate country, due to the gross variability between wage rates and GNI, we do not find reliance on wage data from a single surrogate country reliable for purposes of valuing the labor input. While there is a strong positive correlation between wage rates and GNI, there is also variation in the wage rates of comparable market economies. For example, even for countries that are relatively comparable in terms of GNI for purposes of factor valuation ( e.g., where GNI is below US$2500), the wage rate spans from US$0.21 to US$2.06. To further illustrate, DunAn advocates that instead of relying on the regression methodology, the Department should value labor using India's single wage rate. Petitioner contends that should the Department consider valuing labor from a single surrogate, other comparable countries should also be considered, such as [Colombia]. Although both India and [Colombia] have GNIs of under U.S. $2500, India's wage rate is approximately U.S. $0.21, as compared to [Colombia's] observed wage rate of U.S. $1.13. The large variance in these two countries' wages — not to mention the variances which occur when wage rates are considered for other market economy countries of economic comparability — illustrate the arbitrariness of relying on a wage rate from a single country. For these reasons, DunAn's suggestion of using a single surrogate country to value labor does not constitute the best available information.

Moreover, earlier in its Decision Memo, Commerce explained the benefits of its global regression model: "[t]he Department . . . considers that the regression methodology constitutes the best available information for purposes of valuing labor" as "[t]he Department's methodology avoids extreme variances in labor wage rates that exist across market economies, and instead, accounts for the global relationship between GNI and wages." Id. at 17. Further:

relying only on data from countries that are economically comparable to each NME would undermine, rather than enhance, the accuracy of the Department's regression analysis. The number of "economically comparable" countries would be extremely small. For example, when examining countries with GNIs that range between U.S. $700 and U.S. $2500 ( e.g., countries that might be considered economically comparable to [China]), there are just nine countries out of a full data set of 61 countries used in the revised wage calculation in May 2008. A regression based on such a small subset of countries would be highly dependent on each and every data point, and thus, the inclusion or exclusion of any one country could have an extreme effect on the regression results from case-to-case, and from year-toyear. Relying on a broad data set, as opposed to data from just the economically comparable countries, maximizes the accuracy of the regression results, minimizes the effects of the potential year-to-year variability in the basket, and provides predictability and fairness. . . .

[T]he purpose in using a regression methodology . . . is to provide a more accurate labor value that is stable and predictable across all cases. The regression methodology accomplishes this by providing a variable average that "smoothes out" the variations in the data and permits, in a predictable manner, the estimation of a market economy wage rate relative to a level of GNI that is as accurate as practicable, with the least amount of volatility across cases.

Considered in light of Commerce's analysis, DunAn did not create a record establishing that the particular Indian surrogate value, $0.21/hour, would somehow be more accurate, and thus better information, than the regression surrogate value derived for China. Compare Dorbest II, ___ CIT at ___, 547 F.Supp.2d at 1328-30 (analyzing plaintiffs' argument that the regression was heteroscedastic); Dorbest I, 30 CIT at 1710-12, 462 F.Supp.2d at 1296-98 (addressing plaintiffs' arguments that the regression model is distorted given the existence of a non-zero y-intercept). Contrary to DunAn's arguments, Commerce's explanation reasonably supports Commerce's use of the broader global regression methodology and preference for the wage rate selected here when compared to India's surrogate wage rate. See Dorbest I, 30 CIT at 1677, 462 F.Supp.2d at 1269 (Commerce must "justify its selection of data with a reasoned explanation."). There is nothing on the record here that would preclude a reasonable mind from preferring the regression-generated wage rate to the specific India-based $0.21 rate.

Accordingly, the court affirms Commerce's use of the regression model in calculating the FOP for labor in this case.

C. Partial AFA Applied as to DunAn U.S. Sales and ICC

DunAn next challenges, as unsupported by substantial evidence on the record, Commerce's decision to apply partial AFA — to DunAn's reported U.S. December 2007 sales and to the ICC for the months of October through December 2007 — because the Department was unable to verify the data DunAn submitted. As the court will explain, the court also affirms this aspect of the Department's final determination.

1. Commerce's Verification of DunAn's U.S. Sales and ICC

The court begins with a summary of the relevant facts at issue during Commerce's verification of DunAn's POI U.S. sales of FSVs.

During the POI and beyond, Dunan's U.S. subsidiary, DunAn Precision, Inc., and a U.S. customer had a purchasing agreement whereby the customer maintained quantities of imported DunAn FSV inventory in the customer's U.S. warehouse. Each month, the customer withdrew FSVs out of the warehouse as needed, and reported the number of used inventory (or "usage") to DunAn Precision. According to DunAn, DunAn Precision would review these "consumption reports" for accuracy, and, if correct, would invoice the customer for the values on the consumption report each month. (Pl.'s Br. 15-16;) Application of Partial Adverse Facts Available for Zhejiang DunAn Precision Industries Co., Ltd., Zhejiang DunAn Hetian Metal Co., Ltd. and their U.S. Subsidiary DunAn Precision Inc. in the Antidumping Investigation of Frontseating Service Valves ("FSVs") from the People's Republic of China, A-570-933, POI 7/1/07-12/31/07 (Mar. 6, 2009), Admin. R. Conf. Doc. 228 (" Application of AFA Mem."), at 2-3. If the consumption reports were inaccurate, DunAn Precision would invoice the customer at the correct value, and "would keep a record of the discrepancy and invoice [the customer] on the corrected quantity." (Pl.'s Br. 16.) Accord Application of AFA Mem. at 2, 3. Thereafter, based on the customer's recent consumption and projected upcoming needs, DunAn Precision would issue FSV orders to DunAn. Application of AFA Mem. at 3.

During verification, DunAn submitted to Commerce DunAn Precision's invoices and financial statements together with a "Sales Reconciliation" worksheet harmonizing the two. Verification of the U.S. sales questionnaire responses of Zhejiang DunAn Precision Industries Co., Ltd., Zhejiang DunAn Hetian Metal Co., Ltd., and their U.S. subsidiary DunAn Precision Inc. in the Antidumping Investigation of Frontseating Service Valves ("FSVs") from the People's Republic of China, A-570-933, POI 7/1/07-12/31/07 (Jan. 15, 2009), Admin. R. Conf. Doc. 254 (" Verification Mem."), at 7. DunAn noted that payment received from its customer in December 2007 was thirty cents less than the value on its December 2007 invoice, but, at the time, provided no reason for this seemingly minor discrepancy. Application of AFA Mem. at 3.

Subsequently, in its investigation into other DunAn Precision sales records, Commerce discovered that, for the month of December 2007, the consumption report and the invoice did not match. Although the total value noted in the invoice deviated by only thirty cents from the consumption report — consistent with the December 2007 discrepancy between payment and invoice noted above — the quantities in the December 2007 consumption report differed significantly from those in the invoice. Verification Mem. at 8; Application of AFA Mem. at 3. Specifically, the consumption report quantity for one FSV model vastly exceeded that in the invoice, and, for another FSV model, the invoice quantity vastly exceeded that in the consumption report. Application of AFA Mem. at 5; Decision Mem. at 49-50.

Commerce asked for an explanation of these discrepancies. Mr. Qi responded that he did not have one. Verification Mem. at 8. Later, he told Commerce that he "remembered" that "when he first received the December 2007 [] monthly consumption report, he noticed an abnormally large withdrawal" of a certain FSV model; after verifying that the amount in the consumption report was inaccurate, he corrected the error with the customer. Id. Accord Application of AFA Mem. at 3. However, DunAn did not have any record of the discrepancy and correction, despite DunAn Precision and Mr. Qi's policy to maintain such documentation. Verification Mem. at 8; Application of AFA Mem. at 2, 3. Moreover, Mr. Qi did not explain this deviation from policy and claimed not to know the location of any of Mr. Han's reports. Application of AFA Mem. at 3.

The only document DunAn voluntarily produced as to the discrepancy was an email exchange between Mr. Han and an employee of the U.S. customer, in which the latter sent the consumption report, asking Mr. Han to "[p]lease review this and confirm this payment is accurate." U.S. Sales Verification Ex. 7. Mr. Han responded that "[t]he numbers showing on the report of the month of Dec.[are] right [and] [p]lease go ahead [and] arrange the payment." Id. DunAn explained that the e-mail shows an assent to value rather than quantity. Verification Mem. at 9; Application of AFA Mem. at 4. DunAn also claimed that the U.S. customer likely misreported numbers on its consumption reports for financial reasons. Further, DunAn argued that the consumption report numbers were abnormally large in comparison to other 2007 orders, and, in any event, it would have been impossible for the customer to withdraw so many of the relevant model of FSVs from inventory, as DunAn Precision did not have enough of the model in stock. Verification Mem. at 10; Application of AFA Mem. at 5.

Confusing matters further, however, Mr. Qi maintained separate "monthly inventory reports" or "MIRs," which Commerce obtained for the months October 2007 through March 2008. Conspicuously, the January 2008 MIR was singularly structured. Verification Mem. at 11; Application of AFA Mem. at 5-6. Unlike the other MIRs, the January 2008 MIR did not account for the final inventory from the previous month, that is, the report did not account for December 2007 inventory remaining after subtracting out the U.S. customer's purchases that month; the January MIR instead used the U.S. customer's consumption numbers. Verification Mem. at 11-12; Application of AFA Mem. at 5-6. The remaining 2008 MIRs carried over the sales from December 2007, but the amount carried over still came from the December consumption report, not the December invoice. Verification Mem. at 12; Application of AFA Mem. at 6. Commerce was concerned that this MIR structure, by eliminating December 2007 net inventory, served to side-step, rather than to reconcile, the very quantity discrepancies at issue. Verification Mem. at 12; Application of AFA Mem. at 6. At first, Mr. Qi could not provide any explanation for the difference in the January 2008 MIR or how the MIRs resolved the quantity/inventory December 2007 conundrum and could not even remember making the report. Verification Mem. at 9, 12. Mr. Qi finally answered that DunAn Precision reported its numbers in that way for tax reasons. Verification Mem. at 12.

In essence, Mr. Qi admitted that the inventory numbers contained in the October, November, and December 2007 MIRs were incorrect. But while Commerce pointed out that this recordation of inventory indicated that the consumption reports were accurate, Mr. Qi again, without explanation, maintained that the invoice quantities were correct. Verification Mem. at 12. Mr. Qi still refused to answer why consumption report numbers, rather than invoice numbers, were used in the MIRs, except to indicate that the U.S. customer [[]]. Application of AFA Mem. at 6.

As a result of the Commerce's inability to verify DunAn's conflicting sales data for December 2007, and because of DunAn's lack of clarity regarding these data, Commerce applied AFA to DunAn's December 2007 entries at an AD margin of 55.62 — the margin from the initiation and the highest margin calculated for the proceeding. Decision Mem. at 52. Moreover, because Mr. Qi provided evidence that inventory numbers were incorrectly reported in the MIRs for the latter three months of 2007, Commerce applied AFA as to the ICC for these months and used the highest ICC expense calculated for any sale during the POI. Id.

2. Analysis

a. Application of "Facts Otherwise Available"

The administrative record contains substantial evidence supporting Commerce's application of AFA as to the December 2007 sales. In accordance with 19 U.S.C. § 1677e(a)(2)(D), Commerce noted significant irreconcilable differences, in December sales, between the consumption reports and the invoices, and thus could not verify these sales numbers.

When a respondent has not provided Commerce with accurate, verifiable record evidence, the statutory provision for application of facts otherwise available is intended to permit Commerce to fill the gap. Ningbo Dafa Chem. Co. v. United States, 580 F.3d 1247, 1255 (Fed. Cir. 2009); Statement of Administrative Action, H.R. Rep. No. 103-316, at 869 (1994) ("SAA"), reprinted in 1994 U.S.C.C.A.N. 4040, 4198. See also Nippon Steel Corp. v. United States, 337 F.3d 1373, 1381 (Fed. Cir. 2003); Dorbest I, 30 CIT at 1737, 462 F.Supp.2d at 1318 ("Section 1677e(a) requires that there be a gap in the record of verifiable information due to a party's failure to supply necessary or reliable information in response to an information request from Commerce" (citing NTN Bearing Corp., 368 F.3d at 1377 ("All that is required is that the necessary information be unavailable on the record.")) (other citations omitted)). Accordingly, Commerce's use of facts otherwise available is warranted when Commerce cannot verify the accuracy of respondent's data or cannot reconcile the information produced. 19 U.S.C. § 1677e (a); Heveafil Sdn. Bhd. v. United States, 58 Fed.Appx. 843, 847-48 (Fed. Cir. 2003). As to its procedures for verifying information provided in respondent's questionnaire responses, Commerce is accorded broad discretion. Heveafil, 58 Fed.Appx. at 847; Micron Tech., Inc. v. United States, 117 F.3d 1386, 1396 (Fed. Cir. 1997).

DunAn points to other documents, separately verified by Commerce, such as the invoices and consumption reports through November, that purportedly demonstrate the accuracy of its December invoices. Because of these other invoices and documents, DunAn claims that no "gaps" existed on the record to justify the use of AFA. DunAn contends that Shandong Huarong Mach. Co. v. United States, 30 CIT 1269, 435 F.Supp.2d 1261 (2006) supports the conclusion that, because it provided Commerce with "necessary information," Commerce could not apply facts otherwise available. Shandong, 30 CIT at 1301-02, 435 F.Supp.2d at 1289.

The government responds, however, that, when attempting to verify the accuracy of sales and ICC, Commerce "discovered documents that contradicted those previously provided to it." (Def.'s Br. 17.) Because of these inconsistencies, Commerce was unable to verify the information on the record and could not rely on the accuracy of DunAn's documentation. The government also argues that DunAn's consumption reports prior to December 2007 are not relevant to the question of whether records for December 2007 were accurate, and shipment records only display import numbers and do not account for the debated consumption numbers. ( Id. 21.) As to DunAn's complaints that Commerce did not utilize the consumption reports instead of the invoices, the government argues that Commerce did not accept the accuracy of the consumption reports, but, rather, determined that the consumption reports indicated contradictions in DunAn Precision's records that prevented Commerce from verifying the December sales numbers. ( Id. 21-22.)

The government is once again correct. Nothing in Shandong indicates that Commerce may not apply facts otherwise available when information is unverifiable; unverifiable "necessary information" creates a "gap" in the administrative record to the same degree as a complete absence of "necessary information." See Heveafil, 58 Fed.Appx. at 847-48. See also SAA at 869, reprinted in 1994 U.S.C.C.A.N. at 4198. Such information is still "missing" from the record that would serve to verify the contradictory sales numbers. A respondent's submission of unverifiable evidence, rather than no evidence at all, does not save the respondent from Commerce's reasonable use of facts otherwise available. In any event, the court would direct DunAn to read the clear and unambiguous language of section 1677e(a) that instructs that facts otherwise available are appropriate either when "necessary information is not available on the record" or when "an interested party or any other person . . . provides such information but the information cannot be verified. . . ." 19 U.S.C. § 1677e(a).

For the same reason, as to ICC, the court cannot mandate that Commerce ignore purportedly unreliable statements made by Mr. Qi in favor of what Commerce has reasonably determined to be unverifiable information.

Finally on this issue, DunAn argues that the inventory reports show that it was impossible to fulfill the consumption reports' requirements, and claims that the December 2007 withdrawal was clearly so large as to be aberrational. Commerce replied to DunAn's concerns that "DunAn attempts to demonstrate the accuracy of its records by pointing to the very records that could not be substantiated at verification." Decision Mem. at 51.

The court again agrees with the government's position. Given that, for the relevant valves, Commerce could verify neither the exact number of December 2007 sales nor the total inventory in stock between October and December 2007, it is reasonable for Commerce to decline to square alleged December inventory amounts and sales, even in light of DunAn's proffered import data and even assuming, arguendo, that this import information is correct. Moreover, because its December 2007 sales were unverifiable and because of Mr. Qi's admission, it was also within Commerce's discretion to refuse to find that anomalies in December sales numbers definitively either verified or discredited the accuracy of various other information on the record.

As such, Commerce's use of facts otherwise available, as to the December 2007 sales and ICC, is sustained.

b. Application of Adverse Inferences

The administrative record also reflects that Commerce supported, with substantial evidence, its decision that DunAn did not act "to the best of its ability" to aid Commerce in resolving record discrepancies. 19 U.S.C. § 1677e(b).

In making its determination whether or not to utilize an adverse inference, Commerce need only make a "factual assessment of the extent to which a respondent keeps and maintains reasonable records and the degree to which the respondent cooperates in investigating those records and in providing Commerce with the requested information." Nippon, 337 F.3d at 1383. Moreover, there is no mens rea requirement to warrant an adverse inference, and Commerce may use adverse facts regardless of a respondent's motivation or intent. Id.

The statutory standard does not require perfection; nonetheless, though mistakes may sometimes be discounted, a respondent cannot be "inattentive [], careless [], or inadequate [in] record keeping." Id. at 1382. Commerce may presume that a respondent is familiar with its own records. Id. Commerce may also assume that respondents are familiar with rules and regulations that apply to the import activities undertaken. Id. As a consequence, in order to avoid an adverse inference, a respondent must:

(a) take reasonable steps to keep and maintain full and complete records documenting the information that a reasonable importer should anticipate being called upon to produce; (b) have familiarity with all of the records it maintains in its possession, custody, or control; and (c) conduct prompt, careful, and comprehensive investigations of all relevant records that refer or relate to the imports in question to the full extent of the importers' ability to do so.

Errors resulting in failure to provide information ( e.g., computer errors or mistaken advice from an attorney) will not absolve a respondent from a Commerce determination that the respondent has failed to cooperate. See PAM, S.p.A. v. United States, 582 F.3d 1336, 1339 (Fed. Cir. 2009). A respondent may not rely on excuses that its employee designated to prepare the response either does not know about the needed records, Nippon, 337 F.3d at 1383, or has a lack of familiarity with the respondent's accounting records. Heveafil, 58 Fed.Appx. at 849. Nor can the respondent avoid adverse inferences if it finds that records no longer exist or cannot be located. See id.

DunAn argues that DunAn Precision merely failed to maintain sufficient records to account for discrepancies between consumption reports and invoices. These failures, however, even if inadvertent, support Commerce's decision to apply an adverse inference. See Nippon, 337 F.3d at 1383. Further, Commerce met with Mr. Qi several times and, in some instances, Mr. Qi could not explain the inconsistencies in DunAn Precision's documents. Of those inconsistencies for which he could account, Mr. Qi informed Commerce that DunAn Precision made adjustments in its accounting records — also provided to Commerce — but, at the same time, kept MIRs that conflicted with the invoices. Yet Mr. Qi continued to insist that the invoices were correct. This lack of clarity indicates, at best, negligent record keeping, inadequate knowledge of existing records, insufficient inquiry into these records, and failure to adequately prepare Mr. Qi for Commerce's investigation. See id.; Heveafil, 58 Fed.Appx. at 849; PAM, 582 F.3d at 1339.

DunAn once again relies on Shandong, 30 CIT at 1301-02, 435 F.Supp.2d at 1288-89, as support for its position. But, consistent with Shandong, it is within Commerce's discretion to make factual conclusions based upon the administrative record so long as a reasonable fact finder could make such conclusions. DunAn has not demonstrated to the court that, on this administrative record, a reasonable fact finder could not come to the conclusion that Commerce reached here. See U.S. Steel Group, 96 F.3d at 1357-58; In re Alonso, 545 F.3d 1015, 1019 (Fed. Cir. 2008).

On the record before the court, it was reasonable for Commerce to find that either DunAn was not completely forthcoming to Commerce or that DunAn Precision was at least negligent with its record keeping. Therefore, the court sustains Commerce's application of an adverse inference against DunAn.

c. Use of AFA Margin of 55.62 percent

DunAn also contests the resulting AFA rate. Commerce applied a 55.62 percent dumping margin to DunAn's December 2007 sales of the two FSV models at issue, as 55.62 percent was the initiation rate and the highest rate in the proceeding. Application of AFA Mem. at 9; Decision Mem. at 52.

The statute explicitly authorizes Commerce, in determining an appropriate AFA rate, to rely on any information placed on the record, including information derived from the petition. 19 U.S.C. § 1677e(b); 19 C.F.R. § 351.308; F.lli De Cecco Di Filippo Fara S. Martino S.p.A. v. United States, 216 F.3d 1027, 1032 (Fed. Cir. 2000); SAA at 870, reprinted in 1994 U.S.C.C.A.N. at 4199.

That said, Commerce may not "overreach reality" in resorting to an adverse margin. De Cecco, 216 F.3d at 1032. See also PAM, 582 F.3d at 1340; 19 U.S.C. § 1677e(c). When using "secondary information," such as that from the petition, to create a proxy margin, Commerce must "to the extent practicable, corroborate that information from independent sources that are reasonably at [its] disposal." 19 U.S.C. § 1677e(c); 19 C.F.R. § 351.308(d). "Commerce evaluates whether secondary information has probative value by assessing its reliability and relevance." KYD, Inc. v. United States, ___ CIT ___, ___, 613 F.Supp.2d 1371, 1378 (2009) (citation omitted). See also Mittal Steel Galati S.A. v. United States, 31 CIT 730, 734, 491 F.Supp.2d 1273, 1278 (2007).

Accordingly, the rate chosen must attempt to be a "reasonably accurate estimate" of respondent's actual rate, "albeit with some built-in increase intended as a deterrent," and must be corroborated with information on the record. De Cecco, 216 F.3d at 1032. See also id. ("Congress could not have intended for Commerce's discretion to include the ability to select unreasonably high rates with no relationship to the respondent's actual dumping margin."). Within these constraints, Commerce is entitled to use the highest margin applied to DunAn or any other respondent. See Heveafil, 58 Fed.Appx. at 846, 849-50; De Cecco, 216 F.3d at 1032.

DunAn first argues that Commerce may not apply the China-wide rate because DunAn is independent from the Chinese government. See Qingdao Taifa Group Co. v. United States, ___ CIT ___, ___, 637 F.Supp.2d 1231, 1240-41 (2009), aff'd on other grounds, 581 F.3d 1375 (Fed. Cir. 2009); Gerber Food (Yunnan) Co. v. United States, 29 CIT 753, 771-72, 387 F.Supp.2d 1270, 1287 (2005). But the government responds that Commerce utilized the petition rate, rather than purposefully applying the China-wide rate, and claims that Commerce did not specifically group DunAn within the China-wide entity which received the China-wide rate: "Commerce's use of the petition rate as [AFA] for DunAn's December 2007 sales quantity [] cannot be equated to Commerce treating DunAn as part of the China-wide entity." (Def.'s Br. 22.)

Consistent with the government's arguments, in Commerce's determinations it referred to the high margin given to DunAn merely as the rate in the petition, and did not, in using this rate, deny DunAn a separate rate. Final Determination at 10,889; Decision Mem. at 52; Frontseating Service Valves from the People's Republic of China, 73 Fed.Reg. at 62,956; Application of AFA Mem. at 9. Compare Qingdao, ___ CIT at ___, 637 F.Supp.2d at 1240, 1242 (reversing Commerce's decision, when applying AFA, to apply the PRC-wide rate instead of a separate rate). As a consequence, whether or not Commerce can apply the China-wide rate to DunAn is irrelevant here.

Be that as it may, Commerce's use of the petition margin, when based upon secondary information such as the petition rate, still must approximate DunAn's actual rate, and Commerce must corroborate the use of the rate with evidence on the record. See PAM, 582 F.3d at 1340; De Cecco, 216 F.3d at 1032.

In this regard, DunAn does not argue that Commerce failed to adequately corroborate, as required by 19 U.S.C. § 1677e(c), Commerce's use of the 55.62 percent petition rate. Rather, DunAn argues that the applied rate was not directly related to the unverified information at issue. DunAn complains that, while Commerce found DunAn's sales quantity data to be unverifiable, Commerce, in its calculations, nonetheless used this same quantity data when calculating DunAn's weighted-average dumping margin. To DunAn, this choice was inconsistent.

Yet DunAn's argument is unpersuasive because it "conflate[s] Commerce's determination to reject as unreliable [certain information DunAn] submitted with Commerce's determination to use as AFA `other' record evidence. . . ." Wash. Int'l Ins. Co. v. United States, No. 08-00156, 2010 WL 446048, *5-6, 2010 Ct. Intl. Trade LEXIS 13, at *16-17 (CIT Feb. 9, 2010) (citation omitted). Commerce may, as here, use the quantity information rejected as unverifiable because it is DunAn's own number. DunAn cannot now complain that Commerce used, when applying AFA, the information DunAn itself submitted during the investigation. See id.

The court thus sustains Commerce's use of 55.62 percent as DunAn's AFA rate.

D. Offset for Recycled Brass Scrap

The court also rejects DunAn's final ground for remand. In using brass bar to manufacture its FSVs, DunAn produces by-product brass scrap. DunAn sells some of this scrap, but much of the brass scrap is recycled and integrated into later production of other FSVs. Responding to DunAn's case brief in the administrative proceeding below, the Department allowed an offset, in part, for the reduced value of brass scrap, and applied this offset to DunAn's normal value. See Decision Mem. at 58-59.

DunAn claims that Commerce's scrap offset methodology is contrary to law. Specifically, DunAn argues that Commerce reduced its calculation of DunAn's COP by the reduced value of brass scrap rather than subtracting the brass bar created from scrap from the total brass bar used. This method of calculating the offset, according to DunAn, leads to significant undervaluation of DunAn's cost savings from the use of recycled brass scrap.

Commerce did not address this argument in its Decision Memorandum. Accordingly, DunAn asks that the court direct Commerce to explain its choice of offset calculation.

The court's analysis of this issue begins with the recognition that it is Commerce's consistent practice to grant, from the COP, an offset of the scrap's sales value. In addition, the court sees no conflict between the offset methodology and the governing statutes and regulations, and thus determines the Department's practice to be reasonable and hence in accordance with law.

DunAn cites Commerce decisions to demonstrate that "the Department's practice is not to value by-products reused in production." (Pl.'s Reply Br. 13 (quoting Coated Free Sheet Paper from the People's Republic of China, 72 Fed.Reg. 60,632 (Dep't Commerce Oct. 25, 2007) (final determination of sales at less than fair value), and accompanying Issues and Decision Memorandum, A-570-906, POI 04/01/06-09/30/06 (Oct. 17, 2007), at 36, available at http://ia.ita.doc.gov/frn/summary/PRC/E7-21041-1.pdf (last visited Apr. 19, 2010)) (quotation marks omitted).) Pursuant to this practice, Commerce will not add the value of the brass scrap as an FOP in calculating the COP, as the scrap came from the already-valued brass FOP.

But this is a different issue. Here, Commerce indeed did not add the sales value of the scrap to DunAn's COP. In fact, Commerce granted DunAn an offset because DunAn used some of the scrap. If DunAn's argument were correct, namely that the values of brass bar and brass bar scrap were equivalent and therefore were completely fungible, then the sales value of the scrap would equal the sales value of the brass bar and the value Commerce applied to the scrap would not be in dispute. Rather, for whatever reason, brass scrap does not equal brass bar, and it was thus reasonable for Commerce to treat these inputs differently based upon the sales value of each.

Moreover, the court does not agree that Commerce's failure to address this one issue constitutes error. Commerce is presumed to have considered all the record evidence, see Thomas v. Office of Pers. Mgmt., No. 2009-3107, 2010 WL 391327, at *2 (Fed. Cir. Feb.4, 2010) (per curiam); Gonzales v. West, 218 F.3d 1378, 1381 (Fed. Cir. 2000); Nucor Corp. v. United States, 28 CIT 188, 233, 318 F.Supp.2d 1207, 1247 (2004), aff'd, 414 F.3d 1331 (Fed. Cir. 2005), and need not address every argument raised by a respondent in its briefing. Timken U.S. Corp. v. United States, 421 F.3d 1350, 1354-56 (Fed. Cir. 2005). As has been explained many times before, the court will remand if Commerce "failed to consider an important aspect of the problem. . . ." Motor Vehicle Mfr. Ass'n of the U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983). By following its reasonable established practice, Commerce has not so failed.

Conclusion

For all of the foregoing reasons, Plaintiff's Motion for Judgment Upon the Agency Record is DENIED. Judgment will be entered for Defendant.

It is SO ORDERED.