World Trade

Organization

 

 

TN/RL/W/158

28 May 2004

 

 

(04-2293)

 

 

Negotiating Group on Rules

Original:    English

 

 

 

PROPOSALS ON FAIR COMPARISON

 

Paper from Chile;  Colombia;  Costa Rica;  Hong Kong, China;  Japan;

Korea; Rep. of;  Norway;  Switzerland;  the Separate Customs Territory

of Taiwan, Penghu, Kinmen and Matsu;  and Thailand

 

 

            The following communication, is being circulated at the request of the Delegations of Chile;  Colombia;  Costa Rica;  Hong Kong, China;  Japan;  Korea, Rep. of;  Norway;  Switzerland;  the Separate Customs Territory of Taiwan, Penghu, Kinmen and Matsu;  and Thailand.

 

_______________

 

 

            This paper addresses four interrelated issues:

 

1.         Comparison at the “same level of trade” – Symmetrical adjustments between export price/constructed export price and normal value

 

2.         Exclusion of certain types of export sales from the calculation of export price and constructed export price

 

3.         The method of comparison under Article 2.4.2

 

4.         Model matching

 

            The Friends of the Anti-Dumping Negotiations (“FAN”) feel that it is useful to address these four issues in a single paper.  All of these issues relate to the fundamental objective expressed in Article 2.4 of achieving a “fair comparison” of export price and normal value.  The FAN note, however, that the issues raised in this paper are not exhaustive.  The concept of “fair comparison” applies to many provisions of the ADA, and some aspects of it will be raised in separate papers. 

 

            The discussions in the Negotiating Group may assist in improving this proposal.  Consequently, we reserve our right to modify or complement the proposal as appropriate.

 

            In preparing and/or analyzing specific provisions, it is clear that amendment of the existing text may have an impact on other Articles of the ADA, which have so far not been explicitly addressed.  These links cannot be fully addressed until we have seen a comprehensive overview of proposed amendments.  Consequently, we also reserve the right to make proposals on provisions which may not have been explicitly addressed so far for clarification or improvement.

 

DESCRIPTION OF THE PROBLEMS

 

In General:

 

            Article 2.4 of the ADA requires that authorities make a “fair comparison” between export price and the normal value of the product under consideration.  Authorities must ensure that a fair comparison is made at the same level of trade, with due allowance being made for differences that affect price comparability.  However, the language of the current ADA is very general, and lacks specific disciplines.  As a result, authorities often fail to make a fair comparison.  This paper proposes changes to Article 2.4 that would address some of the more significant problems. 

 

Issue 1:           Comparison at the “same level of trade” – Symmetrical adjustments between       export             price/constructed export price and normal value[1]

 

            Dumping margin calculations are made on a “net price” basis.  The allowances that are used to adjust the gross price to a net price basis form a fundamental part of the dumping margin calculation.[2]   A fair comparison requires not only that exporters be given fair treatment, but also that the comparison between export price and normal value – taking into account all appropriate allowances -- results in a reasonable, symmetrical calculation.

 

            Article 2.4 further requires that the fair comparison between export price (EP)/constructed export price (CEP) and normal value (NV) must be made at the same level of trade.  Some authorities do not make the symmetrical adjustments that are necessary to place sales at the same level of trade. 

 

            Specifically, some authorities fail to make symmetrical adjustments for selling expenses, typically deducting all selling expenses from EP/CEP, but none (or only a portion of them) from NV.  This results in an NV that is at a higher level of trade than the EP/CEP.  This clearly is inconsistent with the intent of the fair comparison requirement.

 

            In CEP cases, some authorities also make an asymmetrical adjustment for profit.  They deduct profit from the CEP, but not from the NV.  This puts CEP at an even lower level of trade, and thus increases the difference in level of trade from the NV, which is further inconsistent with the intent of the “fair comparison” requirement.

 

            Finally, while most Members currently require exporters to show that they are entitled to adjustments, the ADA should state clearly that the authorities have the burden to ensure and make “due allowances” for all differences which affect price comparability.[3]  

 

            As such, the application by authorities of the basic principle of a “fair comparison,” which is already embodied in the current ADA, must be disciplined.

 

 

 

Issue 2:                       Exclusion of certain types of export sales from the calculation of export price and constructed export price

 

            The ADA provides for the exclusion of sales “not in the ordinary course of trade” from normal value.  However, the treatment of such sales with respect to export price and constructed export price is not explicitly addressed in the ADA.  Such transactions are not “sales” as such, because those prices do not reflect the value of the merchandise.  Rather, it reflects other considerations, for which a “due allowance” cannot be made.  Therefore, it is impossible for authorities to make a “fair comparison” with respect to such sales.  We are therefore of the view that the ADA should be amended to set forth specific rules with regard to the exclusion of such sales from the calculation of export price and constructed export price that are relevant for the purpose of ensuring symmetry.

 

Issue 3:           The method of comparison under Article 2.4.2

 

            Article 2.4.2 currently provides three alternative methods for comparing export prices with normal value:

 

            (i)         weighted average normal value to weighted average of prices of all comparable export transactions;

 

            (ii)        transaction-to-transaction;  or

 

            (iii)       weighted average normal value to individual export transactions, for so-called “targeted dumping” cases.

 

            We are of the view that only method (i) yields a fair comparison.  Method (ii) provides authorities with unpredictably broad discretion to choose among the transactions to be used as the basis for normal value.  Method (iii) is objectionable because the concept of “targeted dumping” lacks any theoretical or empirical basis, as discussed below.  Moreover, authorities continue to use the distortive practice of “zeroing” in “targeted dumping” cases, even if they no longer practice “zeroing” in other cases following the decision of the Appellate Body in EC - Bed Linen. (p. 2).

 

            Also, some authorities consider that the methods set forth in Article 2.4.2 need not be used in proceedings other than initial investigations (for example, refund and expiry reviews).  As a result, such authorities use different methodologies in such proceedings (such as the “average to transaction” method).  The ADA should make clear that the methodology set forth in Article 2.4.2 must be used in all proceedings under the ADA.

 

            The ADA also should require the authorities to use a consistent and reasonable time period for the computation of weighted-average normal value and export price.  Currently, some authorities use inconsistent methods, which result in distorted dumping margins.  For example, in administrative reviews, some authorities calculate monthly averages for purposes of price-based normal value. 

 

Issue 4:                       Model matching

 

            Article 2.1 of the ADA states that a product is to be considered as “dumped” if the export price of the product is less than the comparable price, in the ordinary course of trade, for the like product when destined for consumption in the exporting country.  Article 2.6 defines the “like product” as “a product which is identical, i.e. alike in all respects to the product under consideration, or in the absence of such a product, another product which, although not alike in all respects, has characteristics closely resembling those of the product under consideration”. 

 

            The ADA thus obliges authorities to determine which models of products used to determine normal value are “like” the models of the exported products.  The methodology to identify models of the like products to be used for comparison with models of the exported products often is crucial to the outcome of the dumping margin calculation, and the determination of dumping.  Yet respondents often have little or no input regarding the methodology. 

 

            The ADA, however, provides no guidance for interpreting the standard “closely resembling”.  In the absence of models of like products that are “identical” to a particular model of the exported products, authorities are free to decide which models for the normal value, if any, “closely resemble” the exported model and thus may be used in the calculation of dumping margins.  This excessive discretion causes uncertainty for exporters.  It also exposes authorities to possible WTO challenges by respondents.  Both respondents and authorities would therefore benefit from clearer rules regarding model matching. 

 

ELEMENTS OF A SOLUTION

 

Issue 1:                       Comparison at the “same level of trade” – Symmetrical adjustments between       export             price/constructed export price and normal value.

 

Proposals: 

 

1.         Symmetrical adjustments

 

(1)        Basic requirement of “symmetrical adjustments”

 

  • Clarify the first sentence of Article 2.4 by qualifying the fair comparison requirement: “A fair comparison shall be made between the export price or constructed export price and the normal value, with symmetrical adjustments thereof.” 
  •  

(2)        Definition of “symmetrical adjustments”

 

  • Amend Article 2.4 to add a definition for the phrase “symmetrical adjustments”:  “Where a category of adjustment (e.g., due allowance for differences in selling or other expenses, and profits) is applicable both to the export or constructed export price and to the prices used for normal value, authorities must make comparable adjustments to both prices.  Moreover, the adjustment to both prices must fully reflect the actual amount of the expense incurred.”

 

(3)        Adjustments for selling expenses and profits

 

  • Amend Article 2.4 to clarify that all selling expenses related to sales activities for the product under consideration and like products must be fully deducted from their EP, CEP, and NV.

 

 

(4)        Standard elements

 

  • Add an Annex to the ADA, showing the standard elements (starting price and adjustments to starting price) for the calculation of normal value (NV), export price (EP) and constructed export price (CEP).

 

2.         Burden of ensuring a fair comparison

 

  • Amend Article 2.4 to state that “the obligation to ensure a fair comparison lies on the authorities”.[5]  Further, delete from the third sentence of Article 2.4 the phrase “are also demonstrated to”.

 

  • Amend the beginning of the fifth sentence of Article 2.4 to delete the conditional clause at the beginning of the sentence “If in these cases price comparability has been affected…”  Instead, state an absolute obligation on the part of the authorities:  “The authorities shall establish the normal value …” 

 

Explanations:

 

1.         Symmetrical adjustments

 

            Pursuant to Article 2.4, a “fair comparison” between CEP/EP and NV must be made at the same level of trade.  Some authorities fail to comply with this requirement, and compare EP/CEP and NV at different levels of trade.  This is because such authorities make asymmetrical adjustments.

 

            Based on current practice as explained below, it is clear that the current provisions of Article 2.4 are not specific enough to ensure a fair comparison.  The best way to do so is to amend the ADA to provide clear and unambiguous requirements for symmetrical selling expense and profit adjustments[6] as set forth in the proposal above.

 

(1)        Adjustments for selling expenses

 

            As noted above, some authorities make asymmetrical adjustments for selling expenses in EP and CEP cases.  Specifically, they deduct all of the respondent’s direct and indirect selling expenses from the EP/CEP, but none of (or only a portion of) the respondent’s indirect selling expenses from NV.[7]  As a result, the net EP/CEP is at the lowest level of trade (net of all expenses related to selling functions), but the net NV is at a higher level of trade (because it includes some or all expenses related to selling functions).  This is inconsistent with the intent of the fair comparison requirement.

 

(2)        Adjustments for profits

 

            Besides this, CEP cases present additional problems with respect to profits.  In CEP cases, the respondent sells to the importing country via an affiliated reseller.  The affiliated reseller typically is the respondent’s national distributor, and sells to local distributors or end-users. 

 

            To construct the export price, authorities start from the affiliated reseller’s gross invoice prices to its unaffiliated customers.  They deduct all selling expenses and profit incurred or realized by the affiliated national distributor. Thus, the export price is constructed at the level of the producer’s sales to a national distributor.  The price includes only those selling expenses and profit that the respondent incurs and realizes as the producer and the shipper. 

 

            In the domestic market, respondents do not normally sell to national distributors.  Instead, they sell directly to local distributors or end users.  The respondent therefore acts as a national distributor itself, and incurs and realizes selling expenses and profits associated with the national distribution functions.  Some authorities do not deduct any portion of profit from NV, and (as noted above) deduct none of (or only a portion of) the indirect selling expenses from NV.  As a result, the net NV includes selling expenses and profits relating to the respondent’s functions as a national distributor.  However, the CEP includes expenses and profits only with regard to the respondent’s functions as a producer/shipper.  The normal value thus is at a higher level of trade (“LOT”) than the CEP.  These methodologies result in comparisons between CEP and NV at substantially different levels of trade. 

 

            The ADA also should specify the methodology to calculate the profit that may be deducted from CEP and NV. Authorities consider that they cannot use the affiliated reseller’s actual profit for CEP profit.  This is because, when an authority resorts to CEP, it has deemed that the affiliation renders the prices from the producer to the reseller unreliable.  If these prices are unreliable, the reseller’s profits on its resales also are unreliable.[8]

 

            However, the total profits of the respondent (i.e., the profits of the producer and those of the reseller together) are reliable.  This aggregate profit reflects the total amount realized by the respondent on sales to unaffiliated parties. The authorities shall allocate a portion of the respondent’s total profits to each CEP sale. 

 

            It is reasonable to allocate profits based on selling expenses.  When a company must undertake additional selling functions in order to make a sale, it incurs additional expenses, and presumably would not incur these efforts and expenses unless it expected a greater profit.  In contrast, where the company engages in a lower level of selling functions, it has lower expenses and probably would expect a lower profit. 

 

            Thus, authorities shall allocate a portion of the respondent’s total profits to each CEP sale, based on the selling expenses that the respondent incurred on that sale.

 

            Our proposed methodology determines profits for CEP and NV in a consistent manner.[9]  Further, it uses the respondent’s own data, and provides transparency and predictability for both authorities and responding parties.

 

2.         Burden of ensuring a fair comparison

 

            Most Members place the burden on the exporter to demonstrate that it is entitled to adjustments, particularly those that lower the dumping margin.  This practice places an undue burden on exporters, and is inconsistent with the finding of the Appellate Body  in US – Hot-Rolled Steel case, that Article 2.4 places the burden on the authorities to ensure a fair comparison.  The ADA should clearly state that authorities bear the burden of ensuring a fair comparison.  The proposed changes would clarify without any ambiguity that authorities have the burden to ensure and make due allowance for all differences “which affect price comparability”.

 

Issue 2:                       Exclusion of certain types of export sales from the calculation of export price and constructed export price

 

Proposal:

 

  • Amend Article 2.1 to provide that the following types of sales shall be excluded from the calculation of export price and constructed export price: sales of samples, sales to employees, barter sales, sales of the like products to a toller or subcontractor for further-manufacture, upon the condition that the further-manufactured products will be returned to the responding party, and certain sales to affiliated parties.[10]

 

Explanation:

 

            The exclusion of sales “not in the ordinary course of trade” often has a significant effect on the calculation of the weighted-average normal value.  Authorities often exclude low-priced sales (samples, employee sales, etc.) as “not in the ordinary course of trade” from the normal value. These transactions are not “sales” as such.  In particular, those prices do not reflect the value of the merchandise alone; rather, they reflect other considerations, for which a “due allowance” cannot be made.  Therefore, it is impossible for authorities to make a “fair comparison” with respect to such sales.  For example, the purpose of sample transactions is not to realize revenue on those specific transactions, but rather to stimulate customers’ demand for the product.  Similarly, employee sales often are made at below-market prices as a part of the employee’s fringe benefits.  The price is more closely related to the employee’s benefit package than to the value of the product.  In case of barter sales or sales to a toller or subcontractor, the parties agree to the “sales” price at any level, so long as the “sales” price is fully reflected to the price of counter-transactions.

 

            Excluding from normal value sales that are “not in the ordinary course of trade” results in a higher normal value.  The FANs, therefore, have submitted in their “Proposal on Determination of Normal Value” (TN/RL/W/150) that the ADA defines the types of sales of the like product that should be considered as not in the ordinary course of trade, and thus should be excluded from normal value. 

 

            The ADA does not, however, explicitly require authorities to exclude such sales from export price and constructed export price.  As a result, unusual (and often low-priced) sales are included, which results in a lower export price. As noted above, the price of such sales reflects considerations other than the value of the merchandise, and due allowance cannot be made with respect to these considerations.  Thus, in many cases authorities compare an artificially high normal value (excluding low-priced sales) with an export price that includes all sales.  This can significantly raise the dumping margin, particularly for relatively new exporters who have low shipment volumes and some “sample” shipments. 

 

            In order to implement the requirement for a “fair comparison”, the ADA must explicitly require that such sales be excluded from export price and constructed export price.

 

Issue 3:                       The method of comparison under Article 2.4.2

 

Proposals:

 

1.         Abolish the last two methodologies in Article 2.4.2 

 

  • Abolish the last two methodologies set out in Article 2.4.2 and allow only the comparison of weighted-average normal value with weighted-average export price or constructed export price.

 

2.         Applicability of Article 2.4.2 to all anti-dumping proceedings

 

  • The ADA should make clear that the calculation methodology set forth in Article 2.4.2 (amended as set forth above) shall apply to all investigations and reviews under this Agreement.  Delete the phrase “in the investigation phase,” or provide a definition stating that this term refers to all anti-dumping proceedings.

 

3.         Time periods

 

  • Require that authorities calculate weighted-average normal value and weighted-average export price/constructed export price using all sales during the period of data collection, which the FANs have proposed separately in the proposals on determination of normal value.

 

Explanations:

 

            The FANS have previously proposed, in TN/RL/W/113, that the zeroing practice must be prohibited in all cases.  The present proposal elaborates and supplements that proposal.

 

1.         Abolish the last two methodologies in Article 2.4.2 

 

            Authorities should make dumping calculations only on a “weighted average to weighted average” basis.  That is, they should compare the weighted-average export price to the weighted-average normal value.[11]

 

            We propose to abolish the “transaction to transaction” methodology.  This method permits authorities broad and unpredictable discretion to choose the individual home-market transactions that are to be compared to individual export sales.  Nothing in the ADA stipulates which individual transactions the authorities must use as the basis for normal value under this methodology.  Thus, for example, authorities can compare export transactions to individual home market sales with anomalously high prices, thereby inflating the dumping margin.

 

            We also propose to abolish the “targeted dumping” provision, and the “average to transaction” methodology, which many authorities use in cases of “targeted dumping.” (Some authorities also use this methodology in administrative reviews.)  The second sentence of Article 2.4.2 provides that an individual import transaction may be compared with the weighted-average normal value, where the exporter or producer allegedly targeted dumping to certain regions, purchasers, or periods. 

 

            There is no empirical or theoretical basis for “targeted dumping”.  Prices differ among purchasers (due to different bargaining power, different quantities, etc.), different periods (fluctuations in market price) and even sometimes among different regions.  It would be surprising if such differences did not exist in most cases.  Yet the ADA allows authorities to use the average-to-transaction method whenever such differences exist, provided only that the authority explains “why such differences cannot be taken into account appropriately by the use of a weighted average‑to‑weighted average or transaction‑to‑transaction comparison”.

 

            Moreover, in some cases authorities that have abolished “zeroing” in average-to-average comparisons still use “zeroing” in “targeted dumping” cases.  There is no reason why zeroing should be allowed in the case of alleged targeted dumping when zeroing is prohibited in normal circumstances.

 

2.         Applicability of Article 2.4.2 to All Anti-Dumping Proceedings

 

            Some authorities take the position that Article 2.4.2 applies only to initial investigations, not to other anti-dumping proceedings (such as refund and expiry reviews).  This apparently is based on the phrase “during the investigation phase” which is set forth in Article 2.4.2.  The FANs propose to amend Article 2.4.2 to clarify that the methodology set forth therein applies to all proceedings under the Anti-Dumping Agreement.[12] 

 

3.         Time Periods 

 

            To ensure a fair comparison, authorities should use the data during the period of data collection to calculate weighted-average normal values and weighted-average export prices.  We have proposed the period of data collection in our proposals on determination of normal value.  We have proposed that the period of data collection shall be one year and shall be based on the respondent’s fiscal periods.  We believe this practice yields adequate data in most cases.  Also as discussed in our prior proposal, the ADA should take into account the “cyclical markets” in which some products are sold in order to ensure that one entire cycle is captured.

 

Issue 4:                       Model matching

 

Proposals:

 

1.         Impose disciplines on authorities’ selection of the characteristics to be used in identifying the “identical” and “most closely resembling” models; impose limits on products that may be deemed “Closely Resembling”  

 

  • Amend the ADA to require that, for purposes of model matching, the authorities must use all characteristics that have a significant effect on the commercial value or the end use of the product including its technical specifications.  Moreover, the authorities must set a hierarchy among these characteristics.  The hierarchy must be based on the effect of each characteristic on the product’s commercial value or its end use (including technical specifications), relative to the effect of each of the other characteristics.   Model matching characteristics must, where possible, be based on product coding used by the respondents in the course of business.  Where this is not possible, model-matching characteristics must take into account industry product standards used in the country of exportation.   

 

  • The FANs will further discuss possible ways of imposing outer boundaries on the selection of models for normal value that may be deemed to closely resemble the exported models.  One useful way to accomplish this could be to examine the difference in the variable costs of manufacture of the models.

 

2.         Calculation of allowances for differences in physical characteristics

 

  • Amend the ADA to require that, where there are no models for determining normal value that are identical to an export model, allowances for differences in physical characteristics for purposes of Article 2.4 shall be made based on the difference in the per-unit variable cost of manufacture between the exported model and the model used as the basis for normal value.    

 

3.         Require authorities to permit responding parties to comment on model matching

 

  • Amend the ADA to require that, where a product under consideration consists of more than one model, responding parties be given adequate opportunities to propose model matching characteristics that determine identical and most resembling models, and to respond to those proposed by the petitioners and the authorities. 

 

Explanations:

 

            Model matching methodologies determine the universe of models that authorities will use to calculate the “normal value” for a particular exported model.  This can significantly affect the result of an anti-dumping investigation. 

 

            As discussed above, some authorities define “models” based on “product codes”.  The codes are used to identify products that are “identical” or, if there are no “identicals”, to establish which model among various models sold in home market (or third country) is most “closely resembling” the exported model under consideration.  We believe, as discussed above, that the ADA should set forth rules both to guide authorities in the creation of the criteria for determining “identical” and most “closely resembling” products, but also to impose some objective limit on products that can be considered “closely resembling”.

 

            First, it would be useful to specify that the product coding must be based on all characteristics that have a significant effect on the commercial value or the end use, including the technical specifications of the product, and that the product coding must be based on that used by the exporters in the regular course of business or, alternatively, industry product standards of the exporting country.  The ASTM and DIN standards for steel are examples of such industry product standards.  This would provide an objective and verifiable basis for constructing product codes.  It also would make product matching more predictable for exporters.

 

            Second, it is important to ensure that models are similar enough to provide a meaningful comparison.  Again, model matching is highly case-specific, and it is not possible to devise a comprehensive definition of “closely resembling” that could apply to all cases.  However, it is important to define the “outer boundaries” of models that could be eligible for comparison.  For example, because cost is an important factor in determining price, it does not seem appropriate for authorities to compare models that differ significantly in cost.  A large-screen projection TV should not be compared with a portable small-screen model.  The cost structures of these models would differ so much that no adjustment could adequately account for the resulting price differences. As discussed above, clearer rules regarding model matching would benefit responding parties as well as authorities.

 

            We also are of the view that respondents should have opportunities to comment on model matching.  First, as discussed, model matching is an issue of fundamental importance which can profoundly affect the dumping margin calculation.  Second, respondents have expert knowledge of the products and product standards.  Thus, respondents’ input could be invaluable to authorities. Yet authorities often create product codes without significant input from respondents.  Typically, the petitioner will propose the codes.  Authorities often use these codes in the questionnaire, without giving responding parties an adequate opportunity to comment.  Respondents usually find it impossible to convince authorities to change product codes after the questionnaire has been issued. 

 

            Because product codes are very case-specific, we are of the view that, although general rules regarding product codes (such as those set forth above) would be useful to provide necessary guidance for authorities in this matter, it is equally important that responding parties be given reasonable opportunities to propose product codes, and to comment on those proposed by the petitioner and the authorities. 

 

__________

 

 


[1] Constructed export price may be used in some cases involving sales between affiliated parties.  The definition of affiliated parties and the treatment of export sales to affiliates as “unreliable” are discussed in our separate paper regarding “affiliated parties.”

[2] As the Appellate Body stated in US – Hot-Rolled Steel, if proper allowances are not made, then the comparison is, “by definition, not ‘fair,’ and not consistent with Article 2.4.”  See United States – Anti-Dumping Measures on Certain Hot-Rolled Steel Products from Japan, WT/DS184/AB/R (24 July 2001), para. 176.  The Appellate Body thus clarified that proper allowances are central to the validity of the determination.

[3] See United States – Anti-Dumping Measures on Certain Hot-Rolled Steel Products from Japan, op. cit., para. 178. 

                [4] The profit shall be calculated as follows:

 

                (a)           Calculate the ratio of the selling expenses incurred on the individual CEP sale to the respondent’s total production and selling costs and expenses for all sales of CEP and NV.

                (b)           Multiply ratio (a) by the total actual profit earned by the respondent from the sales of CEP and NV. 

This calculation yields the portion of the "total profit" that is allocable to the selling expenses for the individual CEP sale.

 

            The authorities shall calculate CEP profit based on the respondent's own accounting records.

[5] See United States – Anti-Dumping Measures on Certain Hot-Rolled Steel Products from Japan, WT/DS184/AB/R (24 July 2001), para. 178 ("[w]e would also emphasize that, under Article 2.4, the obligation to ensure a "fair comparison" lies on the investigating authorities, and not the exporters.  It is those authorities which, as part of their investigation, are charged with comparing normal value and export price and determining whether there is dumping of imports")

[6] The phrase "symmetrical adjustments" does not mean that identical amounts must be deducted from EP and NV.  Rather, it means that where an adjustment is made to the EP or CEP, the parallel and comparable adjustment must be made to NV, provided that the adjustment relates to sales in the comparison market.

[7] For example, in EP cases, some authorities deduct indirect selling expenses from NV only if the respondent pays commissions to unaffiliated parties on export sales, and does not pay commissions on domestic market sales.  In such cases, the authorities deduct indirect selling expenses from NV, but only up to the amount of the commission that was paid on export sales.  In CEP cases, some authorities deduct indirect selling expenses from normal value only up to the amount of indirect selling expenses deducted from the CEP, and only where normal value is determined at a more advanced level of trade ("LOT") than the CEP, but the data do not enable the authority to quantify the amount attributable to the difference in levels on price comparability.  These authorities simply disregard all home-market indirect selling expenses that exceed the cap.

[8] Authorities use CEP where they consider that the export price from the producer to the affiliated reseller is unreliable.  This price is the basis for the costs that the reseller keeps in its normal accounting.  In its accounting, the reseller's profit is the difference between its revenue and its costs.  Because the reseller's costs (prices from the producer) are unreliable, its profit also is unreliable.

[9] To calculate the profit, some Members use a standard profit margin based on information obtained from an independent importer or importers of the product concerned.  This method lacks transparency and predictability.

[10] The treatment of certain sales to affiliated parties is discussed in our paper regarding "affiliated parties" (TN/RL/W/146 (11 March 2004)).  Also, we do not propose that below-cost export sales be excluded from export price/constructed export price because this would be inconsistent with the purpose of anti-dumping measures, and would lead to unreasonable results.  Specifically, if  below-cost export sales were excluded, then exporters could export at less than normal value without being found to dump – so long as the export sales were below cost.

[11] The related issue of "model matching" is discussed in issue 4 below.

[12] We have made a proposal to the same effect in our submission on reviews, TN/RL/W/83.