WATSON, Senior Judge:
This consolidated action is before the court for review of the results of a redeter-mination made by the Department of Commerce (“Commerce”) pursuant to this court’s remand in
Daewoo Electronics Co., Ltd. v. United States,
13 CIT -, 712 F.Supp. 931 (CIT 1989). It relates to the question of dumping margins which were determined for color television receivers from Korea.
The court turns first to the question of whether or not Commerce complied with the adjustment for taxes required by § 772(d)(1)(C) of the Tariff Act of 1930, as amended (19 U.S.C. § 1677a(d)(l)(C)). That provision is among those designed to insure that before a comparison is made between the price in the home market of the foreign producer and the price for the U.S. market, the “United States price”, (further categorized as “purchase price” or “exporter’s sales price”) does not come out lower than the home market price of the foreign producer simply because it does not include certain unavoidable amounts which are included in the price of the merchandise when it is sold in the foreign country. In particular, the provision involved here is concerned with the extent to which the foreign market price may be higher due to the fact that it includes a tax imposed when the merchandise is sold in the country of exportation but not imposed when it is exported. To eliminate that possible imbalance or unfairness in the comparison the law provides that the U.S. price should be increased by—
the amount of any taxes imposed in the country of exportation directly upon the exported merchandise or components thereof, which have been rebated, or which have not been collected, by reason of the exportation of the merchandise to the United States, but only to the extent that such taxes are added to or included in the price of such or similar merchandise when sold in the country of exportation.
The Korean taxing authority imposes three taxes on such merchandise when it is not exported. They are, a special excise tax of 28%, a defense tax of 30% of the special excise tax amount, and a value added tax of 10% levied on the sum of the tax base, the special excise tax amount and the defense tax amount.
Commerce ultimately determined that the actual Korean tax base was the net dealer delivered price, that is to say, the price to the first unrelated home market buyer. It therefore concluded that the analogous hypothetical tax base in the export market (for the United States) i.e., the amount which would have been taxed but for the exportation and to which the hypothetical tax had to be added in order to make a fair comparison between prices, was the sale to the first unrelated U.S. buyer.
Zenith Electronics Corporation (“Zenith”) argues that Commerce erred in determining that the taxes forgiven on exportation of this merchandise would have been taxes based on the first sale to an unrelated purchaser in the U.S. market.
Zenith first argues that the Korean tax laws make the tax base the equivalent of the ex-factory price. It points out that for home market sales purposes under the Korean Special Excise Tax Act, television receivers are taxed on “the price at which the goods are carried out from the factory.” [Goldstar January 5, 1990 questionnaire response, Appendix 1-2 at p. VI-288; remand Rec.P-Doc. 37, R. 1, Fr. 428.] It further asserts that the Defense Tax, insofar as it is a percentage of the Special Excise Tax, is also derived from the price at which the
television receivers are carried out of the place of manufacture.
Finally, Zenith asserts that the value added tax has the same focus as the Special Excise Tax because, inter alia, the place of the transaction is “where the moving of the goods starts.”
[Id.
at page VI-268 (Art. 10, para. 1, item 1); Fr. 445.] Zenith also stresses that in response to Commerce’s tax questionnaire in this remand, the Korean respondents all described the Korean home market taxes as percentages either directly or indirectly related to “the ex-factory price.”
Zenith characterizes Commerce’s action as the transformation of taxes forgiven upon exportation in the foreign country into taxes forgiven upon resale in another country, taxes which would hypothetically be levied on the amount of import duty paid to the United States government, and taxes which would not be calculated until months after the taxable event of exportation. Zenith characterizes this as an absurd result without foundation in the evi-dentiary record. Zenith also points out that Commerce’s determination with respect to the tax base was a reversal of its position in its proposed remand results and was reached in response to comments by Daewoo and Goldstar on those results without giving Zenith an opportunity to address those comments. Zenith also notes the absence of a bar in the Korean tax law to the use of prices to related parties as the basis for tax assessments, arguing that even though the sales which it believes form the proper tax base may have been sales to related parties, there is nothing in the record to show that they would not be used as the tax base by the Korean authorities.
In response to Zenith’s arguments, Commerce contends that it would have been arbitrary and capricious for it to select a point in the continuum between manufacturing and marketing which differed from the point in Korea at which the taxes were actually imposed. It points out that initially it concluded that the Korean taxing authority would have imposed taxes on the exported merchandise on the basis of the f.o.b. Korean port price. However, when Daewoo pointed out that the taxes in question are imposed in the Korean home market on the net price to the dealer, Commerce assertedly realized that the f.o.b. Korean port price did not properly correspond to the actual Korean tax base, which, in practice, was the net dealer delivered price, i.e., the price to the first unrelated home market buyer. The defendant argues that there is nothing in the law which precludes a conclusion that the Korean taxing authorities would impose a tax on exported colored television receivers at the same point in the export market as they do in their own home market. Zenith contends that the mere fact that, in this case, the net unrelated dealer delivered price happened to be the first price at which the goods left the factory, does not mean that in those situations in which the goods left the factory at a price to a related party that selfsame price would not be the proper tax base.
In the opinion of the Court, the Commerce Department acted without the support of substantial evidence in the record when it sought to find an analogous point for fixing the tax base in the United States and when it concluded that, if the Korean taxing authority were to impose taxes on exports to the United States, which taxes were to be rebated, the amount of those taxes would be determined by multiplying the tax rate by the amount of the first sale to an unrelated party in the United States. This is obviously a response to a hypothetical situation which is unavoidably raised by § 772(d)(1)(C) of the Act but that does not justify departing from the evidence of record without the support of other substantial evidence. The record supports the view that the tax base in the home market is the price at which the goods are carried out of the place of manufacture. This basic fact would indicate that, in the absence of any other evidence in the record that the exportations involved here would be treated differently, the ex-factory price is the only properly supported point for determining the tax base. In this context the plain mandate of the Korean tax law is a matter of fact. The attempt to hypothesize about how the
Korean government would depart from that point is not a matter of fact, but a matter of unsupported speculation.
It does appear that in the Korean market the taxable events were sales to unrelated dealers, but that just happened to be the type of price at which the ex-factory transaction occurred. There has been no showing that the price at which the goods left the factory would have been ignored if it had not been a price to unrelated dealers. There has been no showing, and there can be no presumption, that a price to related parties would be rejected if it happened to be the price at which the goods left the factory. In short, there is no evidence in the record to contradict the plain conclusion based on Korean law, that the price for exportation at which the merchandise left the place of manufacture would have been the tax base, if indeed a tax was to be imposed on exportations. There is no evidence that the Korean tax authorities would reject the price for exportation, or would prefer a U.S. resale price, either in its entirety or somehow modified to remove from it various elements which would bring it closer to an ex-factory price for exportation between unrelated parties. This indicates that the only conclusion supported by evidence in the record is that the tax amount forgiven on exportation is an amount which would have been assessed on the ex-factory price for exportation. Commerce will be required to change its calculation of forgiven taxes accordingly.
The Commerce Department also failed to cap the adjustment at the amount of tax in Korea. The Commerce Department now agrees with Zenith that it inadvertently failed to follow the standards set out in
Zenith Electronics Corp. v. United States,
10 CIT 268, 633 F.Supp. 1382 (1986),
appeal dismissed,
875 F.2d 291 (Fed.Cir.1989) [“Zenith 1”]. In that opinion the Court held that, after determining the amount of taxes applicable to the home market and the U.S. market, the Commerce Department should compare the two and in each instance increase U.S. price by the lesser of those two amounts. Accordingly, the case will be remanded to Commerce to make this particular correction as well.
In its first decision in this case, the Court forbade Commerce to assume that home market taxes passed through fully to consumers in Korea. The Court required Commerce to measure the amount of pass-through. On remand, Commerce performed an econometric measurement of pass-through and determined that all of the commodity tax was passed through to consumers. That econometric measurement is now under attack by Zenith and by the International Union of Electronic, Electrical, Technical, Salaried and Machine Workers, AFL-CIO; International Brotherhood of Electrical Workers; Independent Ra-dióme Workers of America; and Industrial Union Department, AFL-CIO; [the Unions].
What Commerce did was to set up a model of the economic factors involved in the Korean home market. The intention was to see whether the commodity tax would be fully included in the price as a result of factors controlling the production side of the market, and then see whether the demand side of the market would allow the full commodity tax to be included in the price. The econometric method makes quantitative estimates of the various economic factors which affect supply and demand and indicates to what extent those two elements of the market will permit a tax to be included in the price. The technique by which these analyses are made is a mathematical one using mathematical and statistical methods to analyze data regarding costs, prices, sales volume and similar factors to arrive at the most likely description of the behavior of those involved in the market.
The form taken by the mathematical and statistical analyses is that of graphs with one line or curve representing the supply and the other line or curve representing the demand. The relationship between these lines can be used to reveal the effect which an increase in price has on consumption. The price indicated by the intersection of these lines is the “equilibrium price,” the price that consumers are willing to pay and
manufacturers are willing to accept. Depending on the position and shape of the lines, the equilibrium price can change. Depending on the shape of the graphic lines such graphs can indicate whether an increase in price will be fully accepted by consumers, or will be met with a response which causes a lower quantity of the merchandise to be sold.
The responsiveness of these lines to the economic factors which affect them is termed their “elasticity.” For example, the demand curves for such things as are considered the necessities of life tend to have relatively
inelastic
demand curves, that is to say, the demand will not fall off significantly in response to price increases or rise much in response to price decreases. On the other hand, the demand curves for so-called luxury items tend to have relatively elastic curves which means that demand is very responsive to changes in the price of these commodities.
On the supply side, the elasticity of the supply curve relates to the economic factors which arise in producing the commodity in various quantities at various scales of production. When a supply curve is inelastic that indicates that increased costs of making and selling the commodity do not affect the producer’s willingness to supply the commodity at a higher price. On the other hand, an elastic supply curve is one which shows that the increased costs of producing or selling the commodity reduces the amount the producers are willing to supply.
As applied in this case, the supply curve discussed in general terms above, was replaced by cost curves for the producers. These cost curves were found to be such that increases in production would not increase the cost of producing the additional television products. This can be stated as a finding that in the Korean color television market the cost elasticity is one, or, that the production of the product shows constant returns to scale. The fact that additional production does not raise unit costs leaves a manufacturer free to increase its production in accordance with its motivation of maximizing its profits. The production side of the Korean television market was found to exhibit imperfect competition of an oligopolistic type which means that only a few suppliers exist, with each one having a meaningful brand identity, consumer loyalty and some power to set its own prices. It is clear from the record that on the production side of a market such as this, when an excise tax is imposed on the producers, they will have the will and the capacity to raise their prices by the amount of the tax, if we consider only the factors operating on their side of the market equation.
There has been no significant dispute with this aspect of the determination by Commerce on remand. It can therefore be concluded that it would be the tendency of the Korean firms to raise their prices in the full amount of the excise tax unless such price increases would be influenced by factors coming into play from the demand side of the market. This gives central importance to the conclusions reached by Commerce with respect to the demand curve.
The demand curve will indicate to what extent higher prices might mean lost sales and might require the firm to balance the increased price per unit with the decreased quantity of units sold.
Zenith argues that when it came to the formulation of the demand curve, Commerce did not proceed in accordance with the law. Zenith claims that Commerce’s choice of a demand curve for color television receivers in Korea is an assumption which is not based on substantial evidence, which cannot be reconciled with known characteristics of that market, and which was not adequately tested against the data base relevant to choosing a demand curve.
The demand curve in issue has a form which indicates that demand elasticity in the Korean market is constant. This means that when the factors on the supply side, namely, the cost elasticity of one, show that the tax will be fully included in price by the producers, demand for the television receivers will not significantly decrease when the prices rise. This would indicate that all of the tax imposed on the
producers can and will be passed through to the consumers.
The government defends its determination of the nature of the demand curve as preferable over other forms on theoretical grounds and as supported by the economic literature. The government also asserts that its demand curve, termed an “iso-elastic demand function” was to be preferred because the alternatives were not supported by the data contained in the study of the Korean market. However, the alternatives do not appear to have been tested against the actual data.
This brings us to another issue connected to the dispute over the measurement of pass-through. The data base to which Commerce applied its measurement of the market characteristics was one which was aggregated. This means that Commerce combined the data for all of the Korean respondents and for the different types of television receivers. Commerce also transformed the quarterly data of two respondents into semi-annual data in order to accommodate the fact that the third respondent had only supplied semi-annual data. Zenith argues that this aggregation of the data base precluded a determination of whether the amount of pass-through differed between the various respondents and between various types of television receivers. Commerce argues that such an approach was preferable to excluding the respondent which supplied semi-annual data from the data base and relying on the best information available to make a calculation for that respondent. Commerce also argues that it is not uncommon for empirical research on economic phenomena to accomplish a sound result based on limited data.
After close study of this aspect of the dispute, and even after allowing for the fact that econometric techniques cannot be expected to have perfectly accurate methodology, the Court finds unacceptable defects on the demand side of the Commerce determination. Although Commerce makes theoretical objections to the various alternative shapes of demand curves proposed by Zenith, it does not appear to have actually made a full fledged attempt to apply those curves to the data base and to determine whether they might not conform to the data and explain its behavior better than the iso-elastic form used to reach the determination challenged here. What this amounts to, in the opinion of the Court, is a lack of adequate connection between a crucial determination and the evidence in the administrative record. This is simply another way of saying that the demand function and the results which arise from its use are not supported by substantial evidence on the record. It may have a theoretical basis and it may have support in the literature but it has not been shown to be sufficiently supported by facts in the record. In the absence of reliance on evidence derived from the data, which justifies the choice of one demand curve over another, the Court cannot affirm the results of this remand. Unless that is done, the results have the appearance in the end of being ordained by selection of the demand curve rather than arising from, and being based on, the data in the record. In other words, if one demand curve is to be selected as the one which most accurately reflects the realities of the market, Commerce is obliged to explain what data in the record qualifies that curve for a preferred role, and disqualifies the other possible demand curves. It is not enough to simply disqualify alternatives on theoretical or academic grounds.
The Court is also of the opinion that the aggregation of the data for purposes of measuring the pass-through cannot be reconciled with the obligation to base such findings on substantial evidence. It appears to the Court that the amalgamation of all the data from different respondents and different products into one composite is equivalent to the creation of a non-existent single firm, producing an undifferentiated product, and therefore does not have sufficient connection to the reality of the marketplace. Conceivably, such an aggregation might be justifiable if there was no other way to treat the data. However, in the past, Commerce has successfully disaggregated such data and analyzed it by producer and by product category. In fact, this was done in an antidumping case in
volving Japanese television receivers in which less data was available than in this case. There is a contradiction between the way Commerce used two levels of disaggregated measurement in a remand determination involving Japanese television receivers, in which even less data was available, and the way it treated data in this remand involving Korean television receivers. See Determination on Remand (dated April 14, 1987) filed in consolidated C.I.T. No. 85-06-00788, Perloff Report (dated April 8, 1987) at 6.
It is the opinion of the Court that in order to test the extent of pass-through of taxes to consumers, at a minimum, the relevant data must be utilized, as it has in the past, in a way which recognizes the separate identity of different producers and the existence of major commercial distinctions or types of products within the market for television receivers. For the reasons given above, Commerce’s measurement of pass-through cannot be affirmed. This aspect of the case must be remanded to Commerce to perform the analysis on a disaggregated basis, to consider alternative forms for demand curves and explain the evidence from the data which justifies its choice of one form over others.
In short, it is the opinion of the Court that the evidence of record with respect to actual transactions in the market must be given a significant role in the choice of the form to be applied in the course of econometric measurement. The form cannot be allowed to become a Procrustean bed into which the evidence is fitted. Without a strong demonstration of linkage between the data and the chosen form, the threat exists that the administrative process can become a matter of choice between theoretical techniques which are equally defensible in the abstract, but which do not have a proper grounding in substantial evidence.
The Unions have made an additional argument against the correctness of the Commerce Department’s procedure in reaching a determination regarding the extent to which the commodity tax was passed through to the consumer. The Unions argue that Commerce should only have adopted a methodology to measure pass-through, and should have left it as the burden of the respondents to prove what the pass-through was in their transactions.
The government responds that its procedure was mandated by the terms of the Court’s remand and furthermore, is the most efficient way to accomplish the process. With respect to this disagreement, the Court notes only that its instructions to Commerce to find a methodology for measuring pass-through should be understood as obligating the agency only to establish the methodology. From that point it is free to proceed as it did here or to impose an appropriate burden of proof on the respondents. In this case, the establishment of the methodology was required but the procedure by which it was to be applied to the facts was left to the authority of Commerce.
The Unions correctly point out the importance of the adjustment for tax pass-through in the administration of the anti-dumping law and its great potential for affecting dumping margins, but those factors indicate only the essentiality of adherence to legal standards and do not dictate that the burden of proof for such adjustments must be wholly placed on the respondents.
Both Zenith and the Unions have requested the Court to certify its ruling for immediate appeal so that the recurring question of pass-through measurement may be decided with finality. However, the Court does not find that the issue satisfies the requirements for certification.
The Korean interests continue to argue for- a circumstances of sale adjustment to account for differences in the amount of commodity tax included in foreign market value and in United States price. This recurring argument is based on the fact that dumping margins, to a certain extent, become increased by a so-called “multiplier effect” when the adjustment for foreign commodity taxes required by 19 U.S.C. § 1677a(d)(l)(C) (1988) is accomplished by adding to U.S. price the amount of home market taxes that would have been imposed on the sale had it been made in the
home market. Commerce performs this adjustment by multiplying the U.S. sale price by the home market tax rate and adding the resulting amount to U.S. price.
Repeating the example given in Footnote 9 of
Zenith Elec. Corp. v. United States,
10 CIT 268, 273, 633 F.Supp. 1382, 1386 (1986), if the home market price, excluding the tax, is $100, and the U.S. price is $90, the dumping margin would be $10. If a 50% foreign commodity tax was to be added to both the home market and the U.S. price, the foreign price would become $150, the U.S. price would become $135, and the margin would now become $15, $5 more than the theoretical neutral dumping margin. With respect to this occurrence, the Court adheres to its view that a circumstance of sale adjustment, to fully offset the absolute difference in tax amounts arising from a pre-existing dumping price, is not in accordance with the law. However, the Court has noted that it is possible for the multiplier effect to result from differences in the home market and U.S. tax bases which are not related to pre-existing dumping margins. (10 CIT at 281 n. 22, 633 F.Supp. at 1393 n. 22.) The Court now clarifies its views on how the previously expressed opinion with respect to the use of circumstances of sale adjustments should be applied in this case.
It is the opinion of the Court that the consequences of pre-existing dumping margins are not appropriate for correction as circumstances of sale. This would go beyond the intention of the statute and would represent an unreasonable interpretation of its terms. However, by the same token, when the “multiplier effect” can be attributed in whole or in part to bona fide differences in the circumstances of sale, then those differences should be accounted for and the multiplier effect should be mitigated to the extent that it is generated by those factors. Commerce can do this by multiplying the net adjustment to home market and U.S. price for differences in such things as moving expenses, selling expenses, duty draw-back, and physical characteristics of the merchandise, by the tax rate, and then deducting that additional amount from home market price in arriving at foreign market value. So long as this can be done without eliminating the consequences arising from pre-existing dumping margins, it would be the proper application of the law and of its intent that adjustments be made for those factors which arise from genuine differences in the circumstances of sale of the sales being compared. The Court views Commerce’s expressed reluctance to make such secondary-level adjustments as tantamount to a decision not to do so and therefore considers it appropriate to issue directions on this point. Accordingly, on this remand, Commerce will be required to make circumstances of sales adjustments consistent with the aforementioned distinctions.
Several of the parties have brought to the attention of the court a number of errors in calculations about the existence of which there appears to be no serious dispute. What resistance there has been to the corrections of those errors appears to arise from procedural objections. Commerce has taken the position that it was barred from addressing errors not pointed out in the Court’s remand decision by this Court’s ruling in
Zenith Electronics Corp. v. United States,
699 F.Supp. 296 (C.I.T. 1988),
aff'd,
884 F.2d 556 (Fed.Cir.1989) that Commerce was prohibited from making changes in the final result without the express authorization of the Court. Although that holding should not have prevented any application to the Court for the purpose of correcting errors, the focus of the Court at this time is to determine whether the errors should be corrected now. The important thing is to correct the errors and not to be impeded by any obstacles other than those required by law. See
Serampore Industries PVT, Ltd. v. U.S. Dept. of Commerce,
696 F.Supp. 665 (CIT 1988).
The Court’s review of the record and consideration of the arguments of the parties lead to the following conclusions:
Zenith demonstrated that the foreign market values for Goldstar fell short of the amount equivalent to the 50.04% of average home market prices net of tax.
Goldstar has revealed that this substantial understatement arose from a ministerial error by Commerce, namely, its failure to define fields in its data base that were sufficiently large to accommodate the tax values reported by Goldstar. This amounts to a clerical error and should be corrected on remand. The government argues that the need for this correction was not raised previously. What is important, however, as Goldstar itself notes, is that the erroneous home market tax values did not play a role in Commerce’s calculations prior to the issuance of the draft remand results. Furthermore, it was only the final administrative remand results that brought the use of erroneous home market tax amounts under scrutiny by counsel. Even if the Court was to set stricter standards for raising such issues, it would be reasonable to consider this as the first appropriate time for raising that error.
Goldstar has suggested alternative means by which that error may be corrected, but because those means involve converting Goldstar’s submitted raw data file into an SAS data set, a technique which has been prone to problems, the Court adopts Zenith’s suggestion that the most reliable means of correction would be for Commerce to calculate the tax amounts in a consistent manner for both markets. Therefore, on this remand, Commerce is instructed to calculate Goldstar’s home market tax amounts by multiplying the tax bases (net home market prices) by 50.04%. The Court notes that this error was clearly inadvertent in nature and does not indicate in any way that Goldstar under-reported home market taxes.
Goldstar points out that it was adversely affected by Commerce’s failure to eliminate the deduction from Goldstar’s exporters sale price [“ESP”] of an amount representing accounts receivable from the parent company. In addition, the dumping margin calculation was affected by a clerical error regarding a reduction of duty drawback by a wastage factor, which error was first brought to Commerce’s attention during the remand proceeding. On this remand, Commerce will be required to correct both of those errors.
With respect to Goldstar, Zenith points out that Commerce failed to analyze one Goldstar purchase price transaction, but nevertheless included full U.S. price for that transaction in the denominator used to ascertain Goldstar’s weighted-average margin. Although Goldstar and Commerce correctly point out that the Court did not specifically address this error in its first remand decision, it is unquestionably an error. Its triviality is not apparent and it should be corrected on this second remand.
Zenith has also pointed out errors in the calculation of average values used to generate foreign market values for Daewoo. This arose from the use of numerous home market “sales” between related Daewoo parties, which Daewoo acknowledges were only paper transactions and not the sort of sales which should be included in these calculations. It is unquestionable that consideration of these paper transactions skewed the average home market prices derived by Commerce and resulted in an understatement of foreign market values for Daewoo. In the Court’s opinion, this error is of a clerical or ministerial type and the Court approves the method suggested by Zenith for correcting it. The proper way to correct this error is for Commerce to correct the home market data base used to calculate average fair market values for Daewoo by excluding consideration of the related party paper transactions altogether, and Commerce is directed to do this by means of the program instructions suggested by Zenith in its initial brief.
Zenith has also pointed out another error in Commerce’s analysis of Daewoo’s ESP transactions which apparently arose from the incorrect designation of a certain home market model as the comparison model for model TCK405P. The Court agrees with Zenith that the appropriate solution is to designate home market model TCK406PW as the comparison for sales of export model TCK405P and Commerce will be required to correct the error in this manner on remand.
Zenith has also raised an objection to Commerce’s treatment of certain expenses incurred by Goldstar in replacing its own brand name plate with private brand name plates on certain shipments of television receivers. In the remand of this case, Commerce had been directed to address the adjustment to Goldstar’s U.S. price for the cost of replacing the brand name plates. (712 F.Supp. at 958.) On remand, Commerce treated that expense as an indirect selling expense and deducted the cost of the replacement name plates in recalculating U.S. price for ESP sales. In addition, Commerce allocated the expense by apportioning it over all ESP sales, not simply over those television sets on which the name plates were changed. As to the second step, Commerce now agrees that it was an error to allocate the expense of changing the name plates over all television sets and it will be required to make an appropriate allocation over the television sets affected by the name plate change on this remand.
The characterization and treatment of these nameplate costs, prior to allocation, still remains in dispute. Zenith argues that such an expense must be accounted for as a deduction from ESP of post-importation U.S. value-added, under 19 U.S.C. § 1677a(e)(3), which provides for adjustment for “any increased value, including additional material and labor, resulting from a process of manufacture or assembly” performed on imported merchandise in the United States. The government argues that the replacement of nameplates is not the type of activity contemplated by § 1677a(e)(3) and suggests that it is not a sufficiently substantive part of the manufacture or assembly of the television set. The Court is of the opinion that the work of replacing name plates falls literally within the meaning of the statutory terms and that it would not be in accordance with the statute to develop additional standards for the substantiality of manufacturing or assembling work. The affixing of labels is part of the process of assembly at the very least and its relative importance in that process should not be the determinant of how it is to be accounted for. It was unlawful to treat such expenses as indirect selling expenses and subject to offset by indirect home market selling expenses. It was also error to consider such costs in the calculation of the ESP offset cap. Commerce will be required to correct its error by attributing such costs directly to the sales of private labeled sets as further processing costs provided for in 19 U.S.C. § 1677a(e)(3).
For the reasons given above, it is hereby ORDERED, ADJUDGED AND DECREED that the ease is remanded to Commerce for the correction of errors found in its first redetermination on remand in accordance with this opinion; that it shall use ex-factory prices to determine the taxes forgiven by reason of exportation of Korean television receivers and shall cap the upward adjustment of U.S. taxes for such forgiven taxes passed through to home market purchasers of comparison merchandise; that with respect to the measurement of tax pass-through, Commerce shall perform the measurements on a disaggregated basis for each respondent and for the different types of television receivers; and that Commerce shall consider alternative forms of demand curves and specify the evidence in the record which makes its chosen curve a superior representation of the facts with respect to demand in the home market.
It is further ORDERED that Commerce shall include in its tax determination a circumstance-of-sale adjustment for those differences arising from bona fide circumstances of sale and not from pre-existing dumping margins; and it is further ORDERED that Commerce shall correct the understatement of foreign market values for Goldstar, shall adjust the U.S. price for Goldstar’s replacement of nameplates, correct its failure to analyze one Goldstar purchase price transaction, shall exclude the paper transactions in the calculation of Daewoo’s average values, and designate the proper model for comparison with export model TCK405P.
It is further ORDERED that Commerce shall issue a redetermination on remand, consistent with this opinion, within 60 days and transmit said redetermination to the Clerk of this Court. In all other respects, the motions pending before the Court are denied.