SUMMARY

The Department of Commerce (Department) has prepared these results of redetermination pursuant to the remand order of the U.S. Court of International Trade (Court) in Krupp Thyssen Nirosta GmbH and Krupp Hoesch Steel Products, Inc. v. United States (Krupp Thyssen Nirosta), Court No. 99-08- 00550, Slip Op. 00-89, (Ct. Int'l Trade 2000). This action arises out of the Department's Final Determination of Sales at Less Than Fair Value: Stainless Steel Sheet and Strip in Coils From Germany, 64 Fed. Reg. 30,710 (June 8, 1999); as amended 64 Fed. Reg. 40,557 (July 27, 1999) (collectively, Final Determination) and concerns the margin calculated for Krupp Thyssen Nirosta GmbH and Krupp Hoesch Steel Products, Inc. (collectively, KTN). KTN challenged various aspects of this final determination, resulting in the instant remand order from the Court. What follows below is a brief discussion of the factual and procedural background of the antidumping duty order.

BACKGROUND

On May 19, 1999, the Department released its final determination of sales at less-than-fair-value in the investigation of stainless steel sheet and strip in coils from Germany. Following timely allegations from respondent KTN and petitioners⁽¹⁾ of ministerial errors, and following the International Trade Commission's July 19, 1999 final determination that imports of stainless sheet in coils were causing material injury to the U.S. industry, on July 27, 1999, the Department published in the Federal Register its amended final determination and antidumping duty order. KTN challenged various aspects of the Final Determination before the Court. On July 31, 2000, in Krupp Thyssen Nirosta the Court sustained the Department's Final Determination with respect to the use of facts available for unreported home market resellers and its decision to use an adverse inference in allocating U.S. sales through an affiliated reseller (the U.S. Reseller) where the U.S. Reseller failed to identify the supplier of the merchandise. However, the Court remanded the Final Determination on a number of issues, ordering the Department to i) explain why its choice of adverse facts available for the German resellers was Arationally related to KTN's sales and indicative of its customary selling practices,@ and why these facts available were not unduly harsh or punitive; ii) to explain which data fields in the U.S. Reseller's U.S. cost database were verified or verifiable; iii) explain whether, and to what extent, errors in the U.S. Reseller's cost response tainted its attendant sales database; iv) adduce substantial evidence that KTN had the ability to check the U.S. Reseller's database for errors prior to verification; v) point to additional evidence, aside from computer programming errors, for assigning adverse facts available to the U.S. Reseller; vi) explain why the Department's allocation methodology for the U.S. Reseller's sales of unknown origin was not unduly harsh or punitive; vii) explain its refusal to deduct movement and selling expenses from the U.S. Reseller's gross unit price prior to applying adverse facts available; and viii) exclude the U.S. Reseller's sales of non-subject merchandise (i.e., cut-to-length sheet and strip) from the margin calculation. See Krupp Thyssen Nirosta, passim.

Furthermore, with respect to points (ii) and (iii), the Court ordered the Department to use the U.S. Reseller's data if it found the information was verified or verifiable, Asubject to filling any gaps, as noted in the [C] ourt's opinion, with facts available.@ Krupp Thyssen Nirosta at 19. The Court further held, with respect to points (iv) and (v), that if the Department cannot produce evidence of KTN's ability to check its data prior to verification, and evidence of errors aside from computer programming errors, the Department may not use an adverse inference in selecting among the facts otherwise available. Id.

In accordance with the Court's instructions, we will address the issues remanded for further consideration and treatment. Our discussion of each of the issues arising in this remand order follows.

1. Selection of Facts Available for German Resellers

Background

In the Final Determination the Department applied partial adverse facts available to KTN for certain unreported downstream sales in the home market. The Court found that KTN failed to explain and substantiate its claimed inability to retrieve the requested data and sustained the Department's decision to apply adverse facts available to the German Resellers.

The Court has, however, remanded this issue to the Department to explain why the adverse facts selected for the German Resellers were rationally related to KTN's sales and indicative of its customary selling practices, and why the adverse facts selected for the German Resellers were not unduly harsh or punitive.

Analysis

In accordance with 19 CFR ' 351.403(d), the Department requested that KTN report its affiliates' downstream sales (i.e., sales made by the affiliate) for certain of its home market affiliates, because KTN's sales to its home market affiliates represented more than five percent of its total home market sales. However, KTN failed to report entire databases for two of these home market affiliates, in effect preventing the Department from reviewing the full number of sales required to calculate normal value (NV). KTN's failure to report fully the requested downstream sales data precluded an independent analysis which would allow the Department to analyze for itself the nature of these sales - what products were sold; to whom they were sold; the prices they commanded; and their suitability for comparison to sales of similar products in the United States. These steps are not special requirements tailored for this particular case only; rather they are standard prerequisites to calculating a fair and reliable margin under circumstances involving affiliated parties and downstream sales. In sum, the missing data in this case are of great significance to our analysis for they represent a large volume of KTN's home market sales and would allow us to compare these home market downstream sales with U.S. sales for potential matches, in order to calculate a margin. Therefore, by failing to report such sales, the respondent has limited the information available to the Department for review in calculating a margin or, in the absence of the requested data, in applying adverse facts available.

In choosing the adverse facts available to apply, the Department relied upon KTN's and an affiliate's (NSC's ⁽²⁾) sales data as the source for surrogate information. In other words, the Department did not use third party data or other secondary information as a surrogate for KTN's downstream sales information. Rather, we used sales information that the respondent itself placed on the record since the respondent's data best reflect the company's normal sales practices. Since the Department did not rely on secondary information, but instead relied on the respondent's own data, there is no requirement to corroborate the reported data.⁽³⁾ Further, the KTN data selected for use as adverse facts available consisted of above-cost sales made at arm's length prices in the ordinary course of trade in the home market. As for NSC, plaintiff states that this company specializes in the resale of second-quality stainless products.⁽⁴⁾ Therefore, we can only conclude that NSC sale prices would be lower than if NSC had sold first-quality merchandise. Accordingly, if the Department's application of adverse facts available prices relied on Ainferior@ NSC sales as a benchmark, such an application would have resulted in applying lower prices to a group of unreported downstream sales (unreported sales which quite possibly consisted of only first-quality merchandise). Clearly, such an application of adverse facts available only benefits the plaintiff since the effect is a lower home market facts available price, and thus a lower final dumping margin.

Beyond NSC's merchandise distinctions described above (which are favorable to the respondent), the plaintiff has not demonstrated that the sales used to calculate NV were dissimilar to other home market sales. Plaintiff has not demonstrated how any differences in home market sales resulted in an unfair application of facts available. As stated, the Department applied facts available on a product specific basis. The products sold and used in the Department's application of facts available were fairly common products, that were sold during a contemporaneous time frame (i.e., within the period of investigation) by companies that were similarly situated to the other German resellers (i.e., affiliated resellers in the home market). See Ta Chen Stainless Steel Pipe v. United States, Court No. 97-08-01344, 2000 CIT LEXIS 107 (August 25, 2000). Finally, there was nothing in the record to suggest that the sales selected to calculate the margin were in any way abnormal. See Id. at * 18.

The Department's decision to rely upon the respondent's own data rather than upon potentially more adverse data from petitioners or an unaffiliated producer, stems from the Department's strong commitment to using a facts available methodology which relies upon the most reliable information available that is indicative of the respondent's customary practices. In this case, such information is the respondent's own data. There is no information more relevant to and indicative of a respondent's customary business practices than the respondent's own sales data.

From the respondent's reported data we selected as the facts available the highest normal value per control number (i.e. per specific grade and size of coil) located in either the KTN or NSC databases, and applied these model-specific normal values to the appropriate sales of the two affiliated resellers in question.⁽⁵⁾ Not only did the Department rely upon the respondent's own data, but the Department applied adverse facts available on a model-specific basis (as opposed to applying adverse facts available rates which would not be linked to the reported price of a similar model). Thus, we ensured that our application of facts available was indicative of customary selling practices for each specific product. Accordingly, the adverse facts selected for the German Resellers were rationally related to KTN's sales and indicative of its customary selling practices.

Based on the facts of the investigation, the Department does not believe that utilizing the respondent's own data as the basis of adverse facts available is unduly harsh or punitive. The Department did not apply a facts available price higher than the highest rate per model indicated by the respondent's own data. For example, the Department did not use a more adverse facts available approach which would have assigned the absolute highest home market sale price of any product to all other varying product types. Neither did the Department rely upon third party data, such as the petition, as a source of facts available. Instead, the Department has applied to the unreported sales only prices which the respondent itself has applied to its sales under customary selling practices.

The Department reviewed its application of adverse facts available to the respondent's unreported downstream sales, and has confirmed that the adverse facts available had a limited effect upon the final calculated margin. Our review of the data indicates that of the Aobservations@ (i.e. individual sales) to which we applied adverse facts available, [ * * * ] percent never matched U.S. sales, and thus had no effect on NV or on the resulting margin. Only [ * * * ] percent of the adverse facts available sales were included in the larger group of home market sales which formed the basis of NV. Of the total number of home market sales reported to the Department, the total number of adverse facts available sales actually used in the Department's calculations constituted a maximum of only [ * * * ] percent, a relatively small amount.⁽⁶⁾

At the Court's request, the Department attempted to confirm the extent to which adverse facts available increased KTN's home market selling prices. Of the total number of sales to which facts available were applied which could be used to calculate NV, not more than [ * * * ] percent had an adverse facts available price more than double the actual price reported by KTN for the sale(s). In other words, the vast majority ( [ * * * ] ) of the adverse facts available sales used to calculate NV did not have a price more than double the actual price reported by KTN.⁽⁷⁾ Furthermore, only one adverse facts available price is found to have been more than 250 percent of the reported KTN price.⁽⁸⁾ While there may be one such adverse facts available price at this level (i.e. above 250 percent), the fact remains that the price adjustment ⁽⁹⁾ is irrelevant. The price adjustment simply ensures that the downstream sale price is equal to a price KTN reported for a sale of an identical model. Additional analysis reveals that no adverse facts available price which was a potential match to a U.S. sale was more than 500 percent the price reported by KTN. Such increases to NV stand in sharp contrast to KTN's analysis of the application of adverse facts available with respect to the German resellers' sales. See KTN's March 1, 2000 56.2 Brief at 24-25. The Department's analysis (of only the relatively few potential adverse facts available sales actually used in calculating NV) reveals that the extent to which it increased NV is considerably less than what the plaintiff would lead the Court to believe.⁽¹⁰⁾ After verifying that the vast majority of adverse facts available sales were removed from our calculations during the model match process (i.e. they were not matched to U.S. sales, and accordingly had no effect whatsoever upon NV), the Department's analysis of the data discloses that the overall increase to the weighted-average price of the remaining matching home market sales ( [ * * * ] adverse and [ * * * ] non-adverse matching sales) is only [ * * * ] percent.⁽¹¹⁾ This collective increase to NV for these sales is based upon the facts available to the Department (the respondent's own Abenchmark@ data) as described above.

While we could assume that the German reseller's data are comparable to KTN's data (and that prices range in accordance with the various high and low KTN prices), it is equally valid to assume that their data were not comparable, and hence KTN's failure to submit the data. In the absence of any evidence to the contrary, the Department chooses to make the latter assumption. This presumption reflects a common sense inference that the highest KTN or NSC price per model is the most probative evidence of current sales because, if it were not so, the reseller, knowing the alternative penalties, would have produced current information showing these sales to have lower prices. See Rhone-Poulenc, Inc., v. United States, 899 F. 2d 1185, 1190 (Fed. Cir. 1990).

The Department has concluded that its application of adverse facts available not only had a relatively limited effect on KTN's home market NV, but similarly had a relatively limited impact on the resulting margin. To cross check our findings, we have re-calculated KTN's dumping margin while not applying any facts available to KTN's unreported downstream sales. The resulting margin after leaving KTN's reported home market prices intact was [ * * * ] percent. Therefore, the Department's application of adverse facts available resulted in an increase to the margin of a total of [ * * * ] percent percentage points.⁽¹²⁾ Additionally, the Department has recalculated KTN's margin after excluding all of KTN's [ * * * ] sales for which downstream sales were not reported. The resulting margin after excluding all such sales was [ * * * ] percent (a drop of [ * * * ] percentage points). From these results, it is clear that adverse facts available does not have a significant impact on the margin and is not unduly harsh or punitive.

While an argument could be made that including high margins in the application of adverse facts available produces aberrant results, such an argument is premised upon the false notion that, despite the major void of downstream sale information, there exists an undisputed downstream benchmark price for each product which the Department should use as the starting point for adverse facts available. On the contrary, KTN's failure to report certain downstream sales precludes the Department from testing the missing downstream sales prices and, possibly, identifying such other data. Further, market realities for resellers provide support for the adverse facts chosen. Typically, the prices for downstream sales from the affiliates to the first unaffiliated customer are higher because resellers can only make profits by charging more for their merchandise than what they paid to acquire them. It would be impractical for resellers to sell merchandise at the prices paid for the merchandise, since under such circumstances, resellers would not earn any profit, but would rather incur losses due to expenses exceeding revenue. This would mean that the reported transfer prices from KTN or NSC to the affiliated parties would have been lower than the downstream sales from the German reseller to the first unaffiliated customer. Thus, any arguments that seek to use less adverse facts available are premised upon the understanding that the German resellers do not maximize their profits on resales. However, there is no record evidence to support such an illogical hypothesis. If KTN or NSC is able to sell a product to a reseller at a certain price, the reseller would gain little from re-selling that same product for less or equal that same price. This scenario makes absolutely no sense under normal business practice in a market economy such as in Germany.⁽¹³⁾ Thus, any arguments that our application of adverse facts available produced aberrant results would be based on conjecture, given the absence of the requested and relevant downstream sales data. The Department did not apply to German resellers any price higher than a price KTN (or NSC) has reported being able to establish for a sale of a particular model. The Department's practice when selecting an adverse rate from among the possible sources of information is to ensure that the margin is sufficiently adverse Aas to effectuate the purpose of the facts available role to induce respondents to provide the Department with complete and accurate information in a timely manner.@ Static Random Access Memory Semiconductors from Taiwan; Final Determination of Sales at Less than Fair Value, 63 Fed. Reg. 8,909, 8,932 (February 23, 1998) (SRAMs from Taiwan). In accordance with all applicable statutes and Department practice, the Department's application of adverse facts available is the best possible alternative under the above mentioned circumstances.

In sum, the adverse facts available selected for the German Resellers were rationally related to KTN's sales and indicative of its customary selling practices since the Department used KTN's own reported sales information, from a contemporaneous time period, that were identical in physical characteristics to the missing home market sales, on a model specific basis, as the benchmark for applying adverse facts available. On this basis the Department concludes that the adverse facts selected were not unduly harsh or punitive. Moreover, it is clear that the majority of sales to which adverse facts available were applied were eventually eliminated from the margin calculation process, and the few remaining adverse facts available sales had a relatively limited effect on both NV and KTN's resulting dumping margin.

2. U.S. Reseller's Data Errors

Background

In the Final Determination the Department applied facts available with respect to the U.S. Reseller's sales and cost data because the Department found that in many instances the further manufacturing data submitted to the Department could not be verified. See Final Determination, 65 Fed. Reg. at 30,714.

The Department also found that the U.S. Reseller's response was filled with errors, and determined that the U.S. Reseller did not cooperate with the Department by acting to the best of its ability in responding to our requests for information. Therefore, the Department determined that the use of adverse facts available is appropriate for the U.S. Resellers's data. Id.

The Court found that the record evidence warranted application of facts available with respect to the erroneous further manufacturing data; however, it disagreed that the errors the Department cited justified the application of facts available to the U.S. Reseller's cost data. See Krupp Thyssen Nirosta, Slip Op. at 10. The Court remanded this issue to the Department for further explanation. Id. Further, the Court held that the Department's decision to apply adverse facts available to all of the U.S. Reseller's cost data was not supported by substantial evidence. Id. at 11. For the data in question, the court ordered the Department to:

1) explain whether the fields, besides the further manufacturing fields, in the U.S. Reseller's cost database were verifiable; and

2) use the remaining fields in the U.S. Reseller's cost database if the Department finds that these fields were verified or verifiable.

The Court further determined that the U.S. Reseller's sales data appeared to be unaffected by the computer programming errors, thereby calling into question the Department's decision to apply total facts available to the U.S. Reseller's entire database. On remand the Court directed the Department to point to substantial evidence that the U.S. Reseller's sales database was affected by the errors in the U.S. Reseller's cost database and, in the absence of such explanation, to use U.S. Reseller's actual data, subject to filling any gaps, as noted in the court's opinion, with facts available. Id. at 11 and 12.

Further, the Court ruled that the Department's decision to apply an adverse inference was not supported by substantial evidence, finding the Department's conclusion that KTN could have checked its further manufacturing data for accuracy amounted to an unwarranted inference from the record. Id. at 14. On remand the Court directed the Department to point to substantial evidence in the record demonstrating that KTN had the ability to run checks on the data to discover and correct the errors in the computer programing prior to verification and, in the absence of such evidence, to refrain from drawing an adverse inference in choosing from the facts available for the further manufacturing data. Id. at 14 and 15. The Court also directed the Department to point to additional evidence beyond the failure to perform checks before drawing adverse inferences for any remaining use of facts available. Id.

Analysis

Because section 772 of the Tariff Act, 19 U.S.C. ' 1677a, directs the Department to Aconstruct@ an export price (EP) from a later resale of the merchandise in the United States, the statute requires us to make a number of adjustments to the Astarting@ price before it can be compared to NV. These adjustments include additions for packing, import duties and countervailing duties for export subsidies. See section 772 of the Tariff Act; 19 U.S.C. ' 1677a (1999). These adjustments also include deductions for movement charges, export taxes and reimbursed antidumping duties. Id. Pursuant to section 772(d) of the Tariff Act, we also must reduce the constructed export price (CEP) starting price by the amount, if any, of certain other expenses to construct a price which is equivalent to an export price from the foreign country. These additional adjustments are as follows:

a. Expenses generally incurred by or for the account of the producer or exporter in the United States in selling the merchandise under investigation or review are deducted from the starting price. See Section 772(d)(1) of the Tariff Act; 19 U.S.C. ' 1677a(d)(1) (1999). These expenses are referred to as ACEP deductions.@

b. Any increased value, including additional material and labor, resulting from a process of manufacture or assembly performed on the imported merchandise after importation and before its sale to the first unaffiliated customer. This adjustment falls under the title of Afurther manufacturing.@ See Section 772(d)(2) of the Tariff Act; 19 U.S.C. ' 1677a(d)(2) (1999) (emphasis added).

c. The profit allocated to the expenses described above (i.e., CEP deductions and further manufacturing costs) also is deducted from the starting price. Section 772(f) of the Tariff Act; 19 U.S.C. ' 1677a(f) (1999); and section 351.402(d) of the regulations provide the special rule for determining the amount of CEP profit to be deducted under section 772(d)(3).

As the statute makes clear, these expenses are treated as costs and are deducted from the U.S. price. It is these last two adjustments which are at issue in this remand. KTN made sales to its affiliated U.S. reseller which underwent further manufacturing. The U.S. Reseller then made the first sale to an unaffiliated customer. Because we classify these sales made by the U.S. Reseller as CEP sales, we must reduce the starting price for the increased value, including additional material and labor, resulting from the further processing of subject stainless sheet prior to the first sale to an unaffiliated customer. As demonstrated in the discussion above, the further manufacturing adjustment is a direct reduction to the starting U.S. price which is used in the calculation of the dumping margin. Therefore, the further manufacturing adjustment is of extreme importance to the dumping analysis. We address each of the court's specific points below.

A. Whether the Fields, Beside the Further Manufacturing Fields, in the U.S. Reseller's Cost Database Were Verified or Verifiable

The only costs reported by the U.S. Reseller in its cost database were the further manufacturing costs. See November 16, 1998 Section E response, Conf. Doc. 36. While only one field, total further manufacturing (reported in U.S. Reseller's databases as TOTFMG), appeared on the sales data file, there are numerous cost fields that accumulate, or sum up to TOTFMG. The detail of the costs accumulating to the TOTFMG field is included in the section E cost file for further manufacturing. See Id. at 24. TOTFMG represents the sum of the various fields for each type of cost incurred by the U.S. Reseller to further manufacture the subject merchandise before it was sold to the first unaffiliated customer. These costs included materials (reported under the field name FURMAT in U.S. Reseller's databases), labor (FURLAB), variable overhead (FURVOH), fixed overhead (FURFOH), general and administrative expenses (FURGNA), and interest (FURINT). Id. at 19-23. The errors in the U.S. Reseller's cost allocation program noted at verification⁽¹⁴⁾ do not affect only one or two of these cost fields; rather, they impact the costs assigned to each of these fields. These errors were the result of logic or programming errors in the U.S. Reseller's computer program. As noted in the U.S. Reseller Cost Verification Report, fully 40 percent of the tested data were found to be in error. This indicated to the verifiers that the computer generated response was unreliable. As a result, our finding that the U.S. Reseller's cost allocation program was inaccurate and unreliable rendered all of the fields in the U.S. Reseller's cost database unverified or unverifiable.

In response to the Department's Section E questionnaire, KTN developed its cost allocation computer program to quantify the further manufacturing cost incurred for each type of further processing performed on each sale. KTN used its cost allocation computer program to develop the costs reported in each field in the section E response. KTN then merged the cost data with the sales-specific data drawn from its invoice records to compile a single integrated sales response comprising both further manufactured and non-further manufactured sales. The Department attempted to verify the U.S. Reseller's section E further manufacturing cost data and found that approximately 40 percent of the items tested at verification had serious errors, the nature of which are fully described in the cost verification report.⁽¹⁵⁾

The section E questionnaire also instructs respondents to report variable(s) that will link each product that was further manufactured in the U.S. to the product(s) as imported. See November 16, 2000 section E response, Conf. Doc. 36 at 18. KTN's response did not provide this variable as requested, stating that since the cost information is already merged with the sales data, there is no need to provide a linking variable. Id. Apparently, KTN based its decision to omit linking variables on the false premise that an integrated cost and sales file would provide the keys necessary to link imported products to further manufactured sales. In this regard, as demonstrated by the U.S. Reseller's Cost Verification Report, KTN's computer program overwhelmingly failed.⁽¹⁶⁾ KTN's omission of the linking variable is further evidence that KTN did not act to the best of its ability in providing the Department with reliable information. The Department believes that KTN should be held accountable for withholding the requested information (linking variable), and providing error-filled surrogate information (the integrated cost and sales file). By not providing a separate linking field as requested by the Department, KTN relied upon the accuracy of its computer program to generate accurate data. That KTN did not provide a reliable integrated computer file, despite the company's heavy reliance on these data, magnifies the severe limitations of the entire database and of the limits placed on the Department's ability to use such data.

Finally, during verification the volume of information that must be verified in a very limited amount of time is large. Therefore, the Department does not look at every transaction but, rather, samples and tests the information provided by respondents, in order to determine if the data supplied are reliable and accurate.⁽¹⁷⁾ The Department's long standing practice is that if no errors are identified in the sampled transactions, the untested data are deemed reliable. Conversely, if errors are identified in the sample transactions, the Department increases the sample size to determine if the errors were limited or more pervasive. In this case the Department verifiers found errors, increased the sample size, and found that the errors were not limited to the initial sampled transactions but also similarly affected the additional sample transactions, leading to the conclusion that the errors were much more pervasive and systemic. When the Department finds that the errors are pervasive, the untested data are presumed to be similarly tainted absent satisfactory explanation and quantification by the respondent.⁽¹⁸⁾

Based on the results of the cost verification (e.g., a 40 percent error rate in sampled transactions) and the fact that the U.S. Reseller's data indicate that approximately [ * * * ] percent of its sales underwent further processing⁽¹⁹⁾ the Department correctly concluded that all fields in the U.S. Reseller's further manufacturing cost database were unverified or unverifiable and therefore unreliable. Because the only fields in the U.S. Resellers's cost database were further manufacturing fields, and because the Department could not verify these further manufacturing fields, there are no U.S. Reseller cost data that are verified.

B. If the Remaining Fields in the U.S. Reseller's Cost Database Were Verified or Verifiable the Department must Use the Actual Numbers

As shown above, there are no fields in the U.S. Reseller's cost database that were verified or verifiable. Hence, the Department cannot use the numbers provided by the U.S. Reseller.

C. The Department must Explain Why the U.S. Reseller's Sales Database Was Affected by the Errors in the U.S. Reseller's Cost Database

While the further manufacturing cost may only be a single field in the U.S. Reseller's sales data file (TOTFMG), its impact on the calculated CEP and, therefore, the dumping margin, is significant. This is true because the TOTFMG field is a direct reduction from the U.S. Reseller's sales price to the first unaffiliated customer in the United States in order to arrive at the CEP. As a result, an error in this field will corrupt the resulting CEP that is compared to NV to calculate the dumping margin.

More importantly, the cost allocation program developed by the U.S. Reseller served as the linkage between the further manufactured sale and the product- specific classification (i.e. the control number, or CONNUM) of the product used to produce the further manufactured sale.⁽²⁰⁾ Absent a reliable link, we have no assurance that the U.S. Reseller's CEP sales are being compared to the correct home market sales. Without assurance that we are comparing sales prices for identical or similar merchandise, the resulting dumping margin is meaningless.

The errors found at the U.S. Reseller cost verification pertaining to the U.S. Reseller's cost allocation program directly impact both the accuracy of the total further manufacturing costs allocated to each sale, and the reliability of the linkage between the further manufactured sale and the input CONNUM. The extent of the errors found in the cost database, particularly classifying a product as further processed when, in fact, it received no further processing,⁽²¹⁾ or assigning no further processing costs to a further manufactured product,⁽²²⁾ undermines the credibility of the data in their entirety, both sales and cost. Consequently, neither the further manufacturing data nor the non-further manufactured data can be deemed reliable for calculating the margin, since both groups of sales are tainted by the programming errors. While KTN has argued that these errors were minimal and isolated,⁽²³⁾ the simple fact remains that 40 percent of the tested items were unverified. Further, the further manufacturing errors clearly carry through to the reported sales database. As a result, the U.S. Reseller's sales database is unusable for the dumping calculation.⁽²⁴⁾

The unverified further manufacturing cost data in this case are analogous with the unverified cost data in several other cases where the Department found that without verified cost data, the associated sales data were rendered useless.⁽²⁵⁾ If the Department were to accept sales information when the respondent's cost information (a substantial part of the response) does not verify, respondents would be in a position to manipulate margin calculations by permitting the Department to verify only that information which the respondent wishes the Department to use in its margin calculation (i.e. the information which would yield a favorable outcome for the respondent), thereby evading statutorily mandated reductions to those sales prices.

The extent of the errors found in the further manufactured data, e.g., identifying a product as further processed when in fact it received no further processing,⁽²⁶⁾ or failing to identify the nature of the sales,⁽²⁷⁾ necessarily impacts the credibility of the data as a whole. Furthermore, as discussed immediately below, the errors in the computer program, along with other attendant errors in compiling the U.S. Reseller's response, led to failings in the sales database unrelated to the U.S. Reseller's reported further manufacturing costs. These failings impeach the reliability of the reported sales data independent of the errors limited strictly to further manufacturing costs. Consequently, neither set of data can be deemed reliable for calculating the margin. See Final Determination, 64 Fed. Reg. at 30,739. Among the serious flaws in the U.S. Reseller's sales database we found:

1. Failed sales trace

The U.S. Reseller was unable to provide any documentation whatsoever to substantiate its invoice number [ * * * ] , which was selected as a Asurprise@ sales trace at verification. This represents a fundamental failure at verification, as sales traces are a procedure required of all respondents at all verifications. A routine part of any sales verification, whether abroad or in the United States, is to present respondents with a number of pre-selected sales for which respondents are expected to present complete documentary traces (customer inquiries, purchase orders, production orders, shipping documents, invoices, records of payment, etc.). Another universal practice, used to confirm the validity of the pre-selected sales traces, is for the Department also to present a number of Asurprise@ sales transactions at the start of verification and repeat this exercise. The Department conducts surprise sales traces in order to prevent any attempt by a respondent to create a spurious paper trail simply for purposes of verification. Sales traces, by providing a complete step-by-step reconstruction of a transaction from first inquiry to final payment, provide conclusive documentary evidence that a reported transaction represents a bona fide sale, and that all prices and expenses reported for that sale are accurate. The use of surprise sales, in addition to the pre-selected sales, is effective and especially critical precisely because the respondent does not know in advance which transactions will be selected. Therefore, unlike the pre-selected sales traces, it is nearly impossible for a respondent to fabricate computer or paper records to support its response. At the U.S. Reseller sales verification, the Department chose an invoice from the sales file as a surprise trace. The U.S. Reseller could not locate the invoice and underlying documents from its archived records, a process that required twenty-four hour notice for retrieval.⁽²⁸⁾ The U.S. Reseller did not inform the Department that it was unable to document the sale until the last day of verification, at which point it was too late to identify an additional surprise sale. That the U.S. Reseller was unable to comply when asked to perform this routine and universal verification procedure leaves the Department with little confidence that the pre-selected sales documentation the U.S. Reseller did provide was complete and genuine.⁽²⁹⁾

2. Non-prime merchandise

The U.S. Reseller was unable to substantiate that the majority of the material shipped as non-prime merchandise was actually non-prime. See March 15, 1999 U.S. Reseller Sales Verification Report, Conf. Doc. 95 at 2. At verification we found that two-thirds of the material it sells as non-prime actually has no physical defects.⁽³⁰⁾ In many cases, the U.S. Reseller's computer generated response classified merchandise as non-prime due only to the period of time the product stayed in inventory prior to sale, and not because of any damage or defects. Id. at 6. Assigning a non- prime label to merchandise without regard to product quality distorts the Department's matching of the U.S. sales with sales of identical merchandise in the home market, since our practice is to match sales of prime quality merchandise to prime and secondary merchandise only to seconds. Furthermore, this failing would benefit KTN because prime U.S. merchandise would be matched to low priced home market seconds (i.e., low quality or damaged goods), resulting in an artificially lower margin for those sales.

3. Unidentified supplier sales and missing model match characteristics

The U.S. Reseller failed to identify the supplier of a significant portion ( [ * * * ] pounds) of its sales claiming it did not have the means to track this information.⁽³¹⁾ This lapse in linking sales data meant the Department had no way of identifying who supplied a substantial portion of the U.S. Reseller's material, and to which of the three concurrent investigations, if any, such material should be apportioned. Yet, with no prior experience in either the steel industry or with the U.S. Reseller's specific records, the Department's verifiers discredited this claim. Of seven Aunidentified supplier@ transactions sampled at verification, we were able to trace immediately the outside supplier for three of these using nothing more than a personal computer in the U.S. Reseller's offices (see the discussion of AAllocation of U.S. Reseller's Sales of Unknown Origin,@ infra).⁽³²⁾

The inability of the U.S. Reseller's chosen computer language to assign accurately merchandise by supplier was particularly egregious, affecting sales of both further manufactured and non-further manufactured goods. As the record now stands, the Department has no way of determining if, in fact, we have accounted accurately for all of KTN's sales of each type of merchandise, and whether or not we have accurately captured KTN's total further manufacturing costs (as opposed to costs properly allocated to, alternatively, the two concurrent investigations or to other foreign or domestic suppliers not subject to the investigations).

The Department also found that model match characteristics were missing for some of the unidentified supplier sales due to bad data which the U.S. Reseller's computer programs could not reconcile. As a result, some sales were assigned a quantity of [ * * * ] , some sales were assigned no supplier, and some sales were assigned no value for some of the model match characteristics, such as grade of steel, width, thickness or finish. The Department is incapable of calculating margins for these sales because required data related to such sales are unreliable or unavailable, and we therefore have no means of comparing these sales to home market sales of similar merchandise. Relying upon such corrupted data would compromise the reliability of the Department's model match and margin calculations.⁽³³⁾

While at first glance it appears the missing model match characteristics apply only to further manufactured goods, because of the intrinsic errors in the computer program, we cannot say with certainty that we have isolated with any precision further manufactured versus non-further manufactured sales; hence we cannot determine from the record the extent to which the latter were included in these sales.

4. Early payment discounts

The U.S. Reseller reported that no early-payment discounts were granted during the POI. However, while performing routine verification sales traces, the Department discovered that the company had, in fact, granted early-payment discounts on some sales of subject merchandise. The U.S. Reseller did not identify the sales that were afforded these discounts. Because discounts are a reduction to the U.S. price (regardless of whether the transaction is EP or CEP), they have a direct and consequential impact on the dumping margin (omitting U.S. discounts directly benefitted KTN by falsely raising U.S. prices). That the U.S. Reseller's financial records included nearly $ [ * * * ] in discounts,⁽³⁴⁾ yet the U.S. Reseller reported none, further attests to its failure to use readily-available resources to check its responses for errors;⁽³⁵⁾ that its computer program failed to pick up the discounts and link them to the appropriate sales is additional evidence that errors in the computer program affected the reliability of the price and expense information in the sales data files.

These additional errors noted with the U.S. Reseller's sales data serve only to further impeach the reliability of the sales response, and provide further justification for the Department drawing an adverse inference with respect to the U.S. Reseller's data. The severe and systemic errors in both the further manufacturing data and the sales data mean that the Department cannot obtain an accurate U.S. price as mandated by the statute, thereby rendering both the cost and sales data virtually useless. These facts provide further justification for the Department drawing an adverse inference with respect to the U.S. Reseller's data.

D. If the Department Is Unable to Point to Substantial Evidence Demonstrating That the Sales Database Was Effected by Errors in the Cost Database, it must Use the U.S. Reseller's Actual Data Subject to Filling Any Gaps with Facts Available

Because the Department has demonstrated that the U.S. Reseller's sales database was fatally effected by the errors in the sales and further manufacturing cost database, it is not possible for the Department to fill gaps and use any of the U.S. Reseller's submitted data.

E. The Department must Point to Substantial Evidence in the Record Demonstrating That KTN Had the Ability to Run Checks to Discover and Correct the Errors in the Computer Programming Prior to Verification

The Department found that KTN had the ability to run checks on the data to discover and correct the errors found at verification. This ability is demonstrated repeatedly throughout the record of this case.

KTN submitted no fewer than three section E further manufacturing cost databases during the course of this proceeding. Each cost database was submitted with the required certification as to the accuracy of the information. Because the U.S. Reseller submitted various versions of the cost information, it is evident that the U.S. Reseller was conducting checks on the submitted data and finding areas that needed to be corrected. On November 16, 1998, KTN submitted the section E response for the U.S. Reseller. Page 16 of the response states:

[t]o provide the Department with the most accurate further processing costs possible, the Company has calculated the further processing costs on a transaction specific basis. By providing the Department with the requested further processing cost information in the format, the Company will provide all required levels of detail regarding the cost of further manufacturing as laid out in the field descriptions below. In addition, the Company has already linked the cost incurred for each combination of input and output with the sales file. This represents the most accurate indication of the further processing expenses that can be obtained from the Company's records.

November 16, 1998 Section E Response, Conf. Doc. 97 at 16.

On January 15, 1999 KTN submitted the supplemental section E response for the

U.S. Reseller. Page 4 of the response states: A [i]n practice, Reseller has found its internal schedule of process costs to be a highly reliable system for determining actual processing costs.@ January 15, 1999 Supplemental Section E Response, Conf. Doc. 72 at 4. To draw this conclusion was an open admission to the Department that it could rely on the cost allocation computer program because the company had run checks on the results of the program and had found the results to be reliable.

In mid-February 1999 the U.S. Reseller submitted a revised U.S. sales listing. At verification we asked the U.S. Reseller about certain differences which existed between the February 1999 and January 1999 sales listings. In response, the U.S. Reseller cited three reasons for the identified differences:

First, it attributed some of the difference to its programming. It explained that the more recent listing was produced after it had changed some of the programming to eliminate sales that were not real sales, but that had been created as a result of inaccuracies accompanying the merging of the detail file and the inventory file. Second, it stated that there had been some changes in the reporting of suppliers. It made these changes as a result of having manually reviewed many of the invoices. Third, in reviewing its list of suppliers Reseller noticed that certain suppliers had been assigned several redundant listings in its vendor file, not all of which had been reported.⁽³⁶⁾

Furthermore, each questionnaire, supplemental questionnaire and verification agenda that was sent to KTN and the U.S. Reseller contained statements for the respondent to contact the Department were it to encounter any difficulties in complying with the requests for information. At no point did KTN or the U.S. Reseller inform the Department that it could not comply with the requests for the U.S. Reseller's information.

On February 23, 1999, one week in advance of the scheduled cost verification of the

U.S. Reseller, the Department issued the cost verification agenda identifying what products we were going to test and what steps we were going to perform, in order to allow company officials to prepare for the verification. See February 23, 1999 U.S. Reseller Verification Agenda, Conf. Doc. 88. Upon receipt of the agenda, the U.S. Reseller never contacted the Department indicating that it was not going to be able to prepare for the verification.

In checking the data and in preparing for the verification the U.S. Reseller found several items that were incorrectly calculated and presented these corrections at the beginning of verification. See March 18, 1999 U.S. Reseller Cost Verification Report, Conf. Doc. 97 at 2. These corrections included the identification of an error in the computer program where processing charges were being allocated to sales upon which no processing was performed. The errors that the U.S. Reseller presented at the beginning of verification are prima facie evidence that the U.S. Reseller had the resources to check its program because it obviously did run some form of checks. Even more telling is the fact that the U.S. Reseller claimed to have run a complete check on the accuracy of its data in less than a day subsequent to the Department verifiers completing their testing.⁽³⁷⁾ The U.S. Reseller's ability to isolate certain errors in the cost allocation program indicates that KTN had the ability to identify and correct errors prior to verification.

In sum, the record contains ample evidence that the U.S. Reseller had the incentive and the ability to determine whether its computer program provided accurate data and to correct any errors, as demonstrated by: (i) the repeated runs of the computer program to generate a new or updated Section E response; (ii) the certifications made to the Department as to the accuracy of the data; and (iii) the statutory requirements that data reported must be accurate. Notwithstanding these facts, the U.S. Reseller failed to address the substantial errors discussed herein.

F. If the Department Is Unable to Point to Substantial Evidence in the Record of KTN's Ability to Discover and Correct the Errors in the Computer Programming Prior to Verification, the Department must Not Draw an Adverse Inference in Choosing from the Facts Available for the Further Manufacturing Data

The Department demonstrated that KTN had the ability to detect and correct the errors in the cost allocation program prior to verification and failed to do so. This failure on the part of respondent precluded the Department from being able to perform the dumping analysis on the

U.S. Reseller's further manufactured sales. Hence, the Department has no choice but to draw an adverse inference in choosing the facts available for the further manufacturing data.

G. The Department must Point to Additional Evidence in the Record, Besides the Errors in the Computer Program, Before Drawing Adverse Inferences for Any Remaining Use of Facts Available

In addition to the fact that the further manufacturing cost data and linking variables were so fatally flawed that they rendered the U.S. Reseller sales database useless, there were additional errors in these data that undermined their overall reliability. These errors are discussed in detail under 2.C, supra.

3. Allocation of U.S. Reseller's Sales of Unknown Origin

Background

Prior to verification, KTN reported a large number of the U.S. Reseller sales transactions for which the manufacturer was not identified. The U.S. Reseller stated at verification that if material from its warehouse is sold to another location, the warehouse subsequently will enter the merchandise into its own inventory by recording itself as the supplier. Thus, the U.S. Reseller asserted that it has no method of confirming the origin of these sales without conducting a manual audit of all related documentation.

During verification, however, the Department found that in at least some instances it was possible, through a rudimentary search of the U.S. Reseller's existing computerized records, to identify the supplier. Of seven Aunidentified supplier@ transactions sampled at verification, we were able to identify immediately the outside supplier for three of these using nothing more than a personal computer at the U.S. Reseller's offices. See March 15, 1999 U.S. Reseller Sales Verification Report at 10. Based on our findings at verification, and because KTN failed to report the requested information prior to verification (despite ample time to do so), for the Final Determination we determined that the U.S. Reseller failed to cooperate by acting to the best of its ability in compiling information essential to our analysis. We thus made an adverse inference in apportioning the unidentified transactions. As an adverse inference, we treated all of the unidentified merchandise as having originated with one of the three respondent firms in the concurrent investigations (i.e. Germany, Italy, and Mexico). To apportion the unidentified sales among the three investigations, we adjusted the quantity for each of the unidentified sales on a pro rated basis, using the verified percentages of the U.S. Reseller's merchandise supplied by each respondent mill.

The Court has sustained the Department's decision to draw an adverse inference based on substantial evidence. However, the Court questioned whether our decision is unduly harsh or punitive, since we allocated unidentified origin sales to KTN at a percentage more than double the amount of the Averified@ percentage of KTN's sales by volume to the U.S. Reseller during the period of investigation. Krupp Thyssen Nirosta at 17. Therefore, the Court remanded this issue to the Department to explain why the allocation methodology for the sales of unidentified origin is not unduly harsh or punitive, and meets the purpose of the facts available statute.

Analysis

The Department's application of adverse facts available to sales of unidentified origin is not unduly harsh or punitive, but rather is entirely justified and reasonable under the statute governing the application of facts available. As stated above, from our findings at verification the Department has determined that the U.S. Reseller had the means to ascertain the origin of a significant portion, if not all, of its sales of Aunidentifiable origin.@ Moreover, the Department's inability to accurately allocate these sales to KTN and the Mexican and Italian respondents is due to the U.S. Reseller's failure to sort through its files and computerized record system and report the identity of the supplier for these sales. The Department has no means of conducting an independent evaluation of this large quantity of sales to determine whether the patterns found for the identified universe of transactions hold equally true for merchandise which was transferred between or among the U.S. Reseller's locations.⁽³⁸⁾ The Department cannot reasonably assume that the source of the group of merchandise of unidentified origin is the same as for the group of merchandise of identified origin. Such an inference would have been extremely suspect considering that the U.S. Reseller had the ability to provide relevant information on the origin of many of these sales, but chose not to do so. Had the Department apportioned sales of unidentified origin to KTN in accordance with the percentage of verified KTN sales to the U.S. Reseller ( [ * * * ] percent), the Department would not have been applying adverse facts available to such sales but, rather, would have been applying non-adverse, or neutral, facts available to these sales. (We note that the Court has sustained the Department's decision to draw an adverse inference with respect to the allocation of sales of unidentified merchandise).

For adverse facts available, we apportioned all of the unidentified sales among the three investigations by adjusting the quantity for each of the unidentified sales on a pro-rated basis, using the verified percentages of U.S. Reseller's merchandise supplied by each respondent mill. See Final Analysis Memorandum, May 19, 1999, at 11. Such a methodology is neither unduly harsh nor punitive, for several reasons. As we have stated previously, adverse facts available must be sufficiently adverse as to Ainduce respondents to provide the Department with complete and accurate information in a timely manner.'' SRAMS from Taiwan, 63 Fed. Reg. at 8,932. The respondent has not afforded the Department any reasonable benchmark for ascertaining what allocation methodology would be more appropriate, i.e., would be sufficiently adverse to induce cooperation while not being unduly harsh or punitive. It is entirely possible that KTN benefitted from withholding information on the U.S. Reseller's suppliers that would result in higher margins than those actually calculated absent respondent data. Because KTN did not provide the information on the U.S. Reseller's purchases necessary to complete our analysis with respect to the group of unidentified origin sales, it is impossible for us to determine the actual percentage of those sales that originated from KTN in relation to the sales of the accompanying respondents. It is certainly possible that in allocating all of the sales of unidentified origin to the three respondents in this investigation in accordance with each of their verified percentages of sales to the U.S. Reseller, we may have underestimated the actual volume of unidentified origin merchandise that originated from KTN, as opposed to the two accompanying respondents. This is so because the U.S. Reseller did not provide concrete evidence to refute the possibility that the vast majority of the unidentified sales originated from KTN. Despite this possibility, the Department in applying an adverse inference did not apply all of the sales of unidentified origin to KTN. It is conceivable that the Department could have imposed an even more adverse inference to these sales of unidentified origin than the inference used for the Final Determination. Put simply, the Department used the only information KTN made available in order to select the most appropriate adverse inference, without imposing the most severe adverse inference. It is insupportable for a respondent with the means to provide required information to the Department to choose not to give such information, but yet attempt to impose limitations on the Department's use of adverse inferences.

We further note that the application of adverse facts available had a limited impact on the dumping margin. In response to the Court's order, the Department has reviewed the resulting impact on the margin from our adverse facts available methodology. Our adverse inference resulted in an increase to the dumping margin of [ * * * ] percent. In other words, had the Department allocated sales of unidentified origin to KTN in line with KTN's verified relative

percentages (a non-adverse inference), the resulting margin would have been [ * * * ] percent,

instead of [ * * * ] percent. In sum, the effect of an adverse inference in this matter was limited,

and comparing results fully supports the Department's position that our facts available inference did not have unduly harsh or punitive effects on the dumping margin. Rather, our inference should encourage cooperation in the future.

4. The Department's Treatment of U.S. Reseller's Movement and Selling Expenses

Background

In applying facts available to the U.S. Reseller's sales, the Department did not deduct

movement and selling expenses from its reported U.S. price. Rather, it assigned the adverse facts available margin percentage rate to the weighted-average gross unit value for the U.S. Reseller sales.

Both the adverse facts available margin percentage rate applied to the U.S. Reseller sales and the calculated weighted-average unit value for the U.S. Reseller sales affect the magnitude of the overall weighted-average margin for all U.S. sales, as the product of the two reflects an estimated dollar margin for each U.S. Reseller sale. These estimated dollar margins for the U.S. Reseller sales, in turn, are added to the margins calculated for KTN's properly reported U.S. sales to enable the calculation of a single weighted-average margin percentage rate for all of KTN's U.S. sales. The Department has argued that the calculated adverse facts available margin percentage rate already reflected deductions for movement and selling expenses. KTN, in turn, noted that this reasoning actually supports KTN's case for deducting movement and selling expenses from the U.S. Reseller gross unit prices as well.

The Court has ordered the Department to explain its decision not to deduct verified

movement and selling expenses from the U.S. Reseller gross unit prices.

Analysis

As explained in section 2.C, above, systemic errors in the cost and sales portions of the U.S. Reseller's responses rendered the attendant sales data unusable for purposes of calculating KTN's overall weighted-average dumping margin. As we explain at page 18, in some cases, such as the instant case, errors in portions of a response render the response unusable in its entirety. Accordingly, the Department's longstanding practice is to disregard the response entirely, and not attempt to "cherry-pick,@ or choose putatively useable data from the surrounding dross in a rejected response.⁽³⁹⁾ In keeping with that practice, when the Department determined that the

U.S. Reseller's sales data were entirely useless for purposes of calculating KTN's dumping margin, the response was, so to speak, set aside, and the Department turned elsewhere for surrogate information to serve as the facts otherwise available, in accordance with section 776 of the Tariff Act; 19 U.S. C. ' 1677e.

The adverse facts available margin rate applied by the Department was the highest

non-aberrational margin calculated on KTN's properly reported U.S. sales ( e.g., KTN's U.S. sales through Krupp-Hoesch Steel Products). Because the margin used as adverse facts available was calculated on KTN's properly reported U.S. sales, this margin necessarily reflected various adjustments to U.S. gross unit price, including those for movement and selling expenses incurred in the United States, since all of these variables were verified. However, because the response of the U.S. Reseller was deemed unusable in its entirety, as discussed earlier, deduction of various reported expenses drawn from a thoroughly discredited response would have been inappropriate. Therefore, in applying facts available the Department multiplied the adjusted calculated margin by the U.S. Reseller gross unit price. In short, the Department deducted movement and selling expenses when they originated from a properly reported and verified response (i.e., the calculated non- aberrational margin was based on a verified gross unit price adjusted to account for verified movement and selling expenses) but did not do so where the veracity of the entire response was suspect (i.e., the sales portion of the U.S. Reseller response). To make any adjustments to an unverified gross unit price based upon unverified data would render the calculation meaningless.

5. Disregarding Sales of Non-subject Merchandise

Background

On the first day of verification the U.S. Reseller presented to the Department a list of three corrections to be made to its databases. Included in this list are sales of what the

U.S. Reseller claimed to be non-subject merchandise (e.g., cut-to- length sheet). See March 15, 1999 U.S. Reseller sales verification report at 2 and 4. In the Final Determination the Department demurred, suggesting that KTN's claim could not be substantiated given that the U.S. Reseller's overall response had been deemed unreliable. Upon further consideration, the Department acknowledged that it had the ability to remove these sales from the U.S. Reseller database and requested that this issue be remanded to the Department. The Court ordered that the Department exclude the non-subject merchandise from its calculations.

Analysis

The Department has excluded the non-subject sales from the calculation of the dumping

margin. In order to exclude sales of non-subject merchandise the Department first identified

these sales as suggested by the plaintiff. See March 2, 2000 Plaintiff's brief at 28. We then

deducted the total quantity of these sales from our calculations. The removal of non-subject sales

from the U.S. Reseller's database increased the weighted average sale price from [ * * * ] to

[ * * * ] U.S. dollars per pound. We accounted for this change in our recalculation of KTN's dumping margin. See October 30, 2000 Analysis Memorandum.

RESULTS OF REDETERMINATION

The amended weighted-average margin for KTN published in the Department's antidumping duty order of July 27, 1999 was 25.37 percent. 64 Fed. Reg. at 40,557. As a result of the changes in the Department's exclusion of non-subject merchandise sold through the

U.S. Reseller, as detailed above, the Department has determined that the weighted-average margin for KTN is 25.33 percent. Upon a final and conclusive court decision, the Department will issue an amended final determination. We will issue cash deposit instructions directly to the Customs Service.

Troy H. Cribb
Assistant Secretary
    for Import Administration

Date: October 30, 2000

1. ¹ Petitioners are: Allegheney Ludlum Corp., Armco, Inc., J&L Specialty Steel, Inc., Washington Steel Division of Bethlehem Steel Corp., United Steelworkers of America, AFL- CIO/CLC, Butler Armco Independent Union and Zanesville Armco Independent Organization.

2. ² Nirosta Service Center.

3. ³ See section 776(c) of the Tariff Act of 1930, as amended (the Tariff Act); 19 U.S.C. ' 1677e(c) and the Statement of Administrative Action, H.R. Doc. No. 103-316, vol. 1, at 870.

4. ⁴ KTN's March 1, 2000 56.2 Brief at 4.

5. ⁵ In other words, in each application, we applied NV from one model to an identical model.

6. ⁶ KTN and NSC reported [******] home market observations for the period of investigation. The Department applied adverse facts available to [*****] sales. From this group, only [***] sales, or [*****] percent of the [*****] sales passed the criteria established by our model match program and were potential matches for U.S. sales.

7. ⁷ A maximum of [***] adverse facts available sales could possibly be matched to U.S. sales. Of this number, [***] sales were not more than double the KTN reported prices.

8. ⁸ In other words, this price constitutes only [*****] percent of the total number of reported home market sales, or [******] of all adverse facts available sales.

9. ⁹ In other words, the difference between the reported price (to the affiliate) and the facts available price assigned to this sale (since the downstream price was not reported).

10. ¹⁰ Plaintiff's analysis of the data (as referenced on page 7 and 8 of the Court's Slip Op.) seems to indicate that the Department vastly increased NV. However, the plaintiff's analysis does not confirm whether the comparison of prices is made across many distinct product groups or, rather, within groups of similar products. Given the facts of this case, the Department's analysis compares only the prices of identical merchandise, and furthermore, only for those observations which passed the model match test and which could have a potential impact on the margin.

11. ¹¹ Of the total number of reported home market sales ([******]), a total of [*****] sales passed the Department's model match test (this means the remaining ****** were therefore not used in our calculations). These [*****] Amatching home market sales@ comprise the [***] matching sales to which facts available were applied and the remaining [*****] matching sales which did not receive any facts available. The Department's application of adverse facts available to the sub-group of [***] sales raised the normal value of the collective [*****] matching sales by only [****] percent.

12. ¹² The final amended margin was 25.37 percent. See Final Determination, 64 Fed Reg. at 30,749.

13. ¹³ If KTN can establish a Afavorable price@ for its sales, it would make sound business sense for a reseller to, at minimum, attempt to match such favorable selling prices. For a reseller to set its optimum resale price lower than the primary seller's actual favorable selling price would not make practical business sense. As stated, the Department applied to affiliated resellers' sales the highest price per model that KTN was able to establish under its normal business operations. Considering market economy business realities, it is conceivable that in applying only the highest KTN price per model to the German Reseller's sales (and not a price higher than the highest price per model), that the Department may have underestimated the actual prices for the downstream sales which the respondent failed to report.

14. ¹⁴ The U.S. Reseller's Cost Verification Report identified several errors in the cost allocation program: (1) incorrectly reported quantity extra charges; (2) incorrectly reported output weights; (3) incorrectly allocated processing costs to a sampled product that received no processing; (4) incorrectly reported bottom-finishing costs for certain products; (5) improperly reported re-spinning processing.

See March 18, 1999 U.S. Reseller Cost Verification Report, Conf. Doc. 97 at 14.

15. ¹⁵ See March 18, 1999 U.S. Reseller Cost Verification Report, Conf. Doc. 97 at 14 -17.

16. ¹⁶ See March 18, 1999 U.S. Reseller Cost Verification Report, Conf. Doc. 97 at 14.

17. ¹⁷ See, e.g., Bomont Industries v. United States, 733 F. Supp. 1507, 1508 (CIT 1990) (A[v]erification is like an audit, the purpose of which is to test information provided by a party for accuracy and completeness@) and Monsanto Company v. United States, 698 F. Supp. 275, 281 (CIT 1988) (A[v]erification is a spot check and is not intended to be an exhaustive examination of a respondent's business''). Note, that during this investigation of one respondent (KTN) the Department was required to visit five cites to verify the sales and cost data of the home market and U.S. sellers.

18. ¹⁸ See, e.g., Tatung Company v. United States, 18 CIT 1137 (December 14, 1994).

19. ¹⁹ The U.S. Reseller's sales database indicates that of the [*****] reported Aobservations@ (i.e., sales), [*****] sales were accompanied by positive amounts reported for FURMANU, indicating that the merchandise underwent further manufacturing.

20. ²⁰ In its Section E questionnaire response, KTN reported that A[t]he Company has assigned a unique control number to each sale of a further manufactured product based on the characteristics of the stainless steel coil used as the input in the production of the product being sold@. See November 16, 1998 Section E response, Conf. Doc. 97 at 17.

21. ²¹ See March 15, 1999 U.S. Reseller Cost Verification Report, Conf. Doc. 97 at 15.

22. ²² Id. at 16.

23. ²³ See KTN's March 1, 2000 56.2 Brief at 7-14.

24. ²⁴ In addition to the cost data problems which affected the usability of the sales data, at the

U.S. Reseller sales verification the Department found that there were additional errors in the sales data which undermined their overall reliability. See March 15, 1999 U.S. Reseller Sales Verification, Conf. Doc. 95 at 2, and the discussion, infra.

25. ²⁵See e.g., Notice of Final Determination of Sales at Less Than Fair Value: Steel Wire Rod from Germany, 63 FR 8953, 8955 (Feb. 23, 1998).

26. ²⁶ See March 15, 1999 U.S. Reseller Cost Verification Report, Conf. Doc. 97 at 15.

27. ²⁷ See the discussion of sales of unidentified origins, infra.

28. ²⁸See March 23, 1999 KTN Case Brief, Conf. Doc. 98 at 65-66.

29. ²⁹See March 15, 1999 U.S. Reseller Sales Verification Report, Conf. Doc. 95 at 2.

30. ³⁰At the U.S. Reseller sales verification the Department found that four of the six sales of

Anon-prime@ merchandise we reviewed were, in fact, prime merchandise. The prime versus non-prime misclassifications were independent of the further manufacturing errors previously discussed.

31. ³¹ See May 19, 1999 Analysis Memorandum, Conf. Doc. 104 at 11.

32. ³²See also March 15, 1999 U.S. Reseller Sales Verification Report, Conf. Doc. 95 at 10.

33. ³³ Without sufficient product information (i.e. model match criteria), the Department is unable to identify identical or similar type merchandise. If the Department cannot match identical or similar type merchandise between the home market and U.S. market sales, the Department is unable to calculate a dumping margin between the two groups of sales.

34. ³⁴ See March 15, 1999 U.S. Reseller Sales Verification Report, Conf. Doc. 95 at 18.

35. ³⁵ The early payment discount inconsistency is independent of the further manufacturing errors previously discussed.

36. ³⁶ See March 15, 2000 U.S. Reseller Sales Verification Report, Conf. Doc. 95 at 11 (emphasis added).

37. ³⁷ See KTN's March 1, 2000 56.2 Brief at 15.

38. ³⁸ The U.S. Reseller explained during verification that once it transfers inventory between its locations, its computerized inventory system issues a new stock number, thereby erasing the original link with the supplying mill. See Final Determination, 64 Fed Reg. at 30,742.