BINATIONAL PANEL REVIEW
PURSUANT TO
THE NORTH AMERICAN FREE TRADE AGREEMENT
Article 1904

Secretariat File No.
CDA-95-1904-04
October 9, 1996

(Continued)

VII. DETERMINATION OF COST OF PRODUCTION

A. Statutory Background and Revenue Canada Determination

Sections 16(b) and 19(b) of SIMA provide that, where domestic market sales of subject goods do not provide for the recovery of the costs of producing and selling goods over time, normal value is to be constructed. Revenue Canada required that constructed value computations be calculated in this case. ³⁸ Savannah does not dispute that constructed value computations were necessary, but objects to Revenue Canada's computations for the cost of production element of the constructed value. ³⁹

The major component of refined sugar is the raw sugar or the sugar beet. Thus, the cost of these inputs is of "particular importance" in this case. ⁴⁰ In making its cost of  production findings, Revenue Canada rejected the methodology offered by Savannah for costing sugar inputs and substituted alternative methodologies.

Revenue Canada adopted two alternative methodologies. In the case of sugar processed from U.S. grown sugar beets, the sugar inputs were not commingled and could be tracked. Revenue Canada relied on the cost of the sugar beet processing provided by Michigan Sugar for cost of production calculations for these identifiable inputs.

In the case of sugar refined from raw cane sugar, the raw sugar inputs were fungible and subject to commingling and thus could not be tracked. Revenue Canada sought to determine the weighted average of costs of the #11 and the #14 raw sugar consumed in production in the month in which the sugar was exported to Canada, calculated on the basis of changes in the physical inventories of #11 and #14 raw sugar.

B. Positions of the Participants

Savannah first alleges that the determination of the methodology for calculating cost of production pursuant to SIMA is not a matter within the expertise of Revenue Canada and that the decision should be reviewed as to its correctness. Alternatively, Savannah urges review of reasonableness in fact and law of Revenue Canada's determination.

Savannah argues that it records cost of production on the basis of whether sugar is sold domestically or exported and that it does not track separate physical inventories by source of the raw sugar. Savannah maintains that its method of tracking costs is in accordance with the Generally Accepted Accounting Principles ("GAAP"), a fact certified by respected public accountants. Savannah points to the GATT Antidumping Code to support its view that Revenue Canada should have adopted Savannah's cost assignment methodology for all sales, whether or not commingling of raw sugar inputs was present, because its was Savannah's normal method and an approved GAAP methodology.

Additionally Savannah argues that where commingling is present, it is impossible to determine the physical proportion of #11 raw sugar and #14 raw sugar used in the production of specific physical subject goods exported to Canada. ⁴¹ Savannah challenges the weighted average costing methodology Revenue Canada used to determine the cost of production in establishing normal value. Specifically, Savannah alleges that the method of costing used by Revenue Canada to construct cost of production related to commingled inputs was irrational because it used Savannah Foods' inventory values, which were based on Savannah's method of accounting, a methodology previously rejected by Revenue Canada.

Revenue Canada argues that determining the appropriate costing methodology is a matter within the expertise of Revenue Canada, since it arises in interpreting and applying SIMA. ⁴² Revenue Canada states that SIMA does not require Revenue Canada to adopt a particular method merely because it is in accordance with GAAP. Revenue Canada maintains it has the discretion to adopt another method of determining costs where that other method is reasonable and consistent with SIMA.

Revenue Canada also asserts that SIMA requires that the construction of the normal values of the goods exported to Canada be based on the actual costs of producing and selling the goods. ⁴³ Revenue Canada notes that its analysis of the accounting methodology employed by Savannah revealed that the costs assigned to sales were from management inventory pools of #11 and #14 raw sugar and that Savannah assigned these costs regardless of whether the refined sugar was made from raw cane sugar or from sugar beet. For this and other reasons, Revenue Canada believes that Savannah's method of assigning costs to sales was not related to the production of the goods sold.

C. Decision of the Panel

1. Review under the reasonableness standard

The Panel's review here is conducted to evaluate the "reasonableness" of agency action. We do not adopt the "correctness" standard urged by Savannah. Revenue Canada has been entrusted to administer the antidumping law. ⁴⁴ Cost of production is a key component of SIMA Section 19(b), which itself is at the core of the antidumping law. Thus, the determination of what constitutes cost of production is central to administration of the law. Revenue Canada's findings may be informed by analogous legal interpretations, but its specialized expertise should prevail in defining concepts central to the law unless that interpretation is unreasonable.

2. Revenue Canada's rejection of Savannah's methodology

The term "cost of production" is not defined in SIMA, but Revenue Canada Regulation 11(a) defines it as the aggregate of all costs that are "attributable to, or in any manner related to, the production of the goods." ⁴⁵ The regulation does not define the phrase "attributable to."

In determining cost of production and whether costs are "attributable to, or in any manner related to" the cost of production, Revenue Canada's interpretation of Canadian antidumping law can be informed by the Antidumping Code. The Code provides that costs should "normally" be calculated on the basis of the exporter's records, if they are kept according to GAAP and "reasonably reflect the costs associated with the production and sale of the product under consideration." ⁴⁶ In fact, acceptance of a responding company's GAAP methodology is apparently common for Revenue Canada. ⁴⁷

Nevertheless, Revenue Canada rejected Savannah's GAAP methodology for calculating cost of production. In the Statement of Reasons supporting its determination, Revenue Canada stated that "the cost allocation proposed by the exporters does not reflect the actual cost of the goods. Revenue Canada believes that the cost of the refined sugar must reflect the actual cost of the inputs." ⁴⁸ Similarly, Revenue Canada stated that "in constructing the cost of the goods under SIMA, we are required to take into consideration the actual cost of the raw material inputs consumed in producing the goods in question." ⁴⁹

Revenue Canada's employment of the term "actual" to reject Savannah's methodology was not related to the existing regulatory formulation for cost findings. Moreover, Revenue Canada's use of the term "actual" was inconsistent within the context of this case. ⁵⁰ The confusion over use of the term "actual" notwithstanding, this Panel finds that Revenue Canada engaged in more than a futile effort to calculate "actual" costs. Revenue Canada found that Savannah's assignment of #11 costs to all export sales did not reflect the [ sic! ]. ⁵¹ In more detail, Revenue Canada found:

U.S. exporters realize a loss on many export sales by selling sugar refined from high priced #14 raw sugar or sugar beets at low prices dictated by competitive factors on the open, world market. These losses on their export sales [ sic! ]. ⁵²

Revenue Canada noted that "the fact that the exporters' costing methodology is acceptable under GAAP for U.S. tax purposes and for reporting to the Securities and Exchange Commission does not mean that costs reasonably reflect the costs associated with particular sales, which is the requirement under SIMA." ⁵³

Revenue Canada also rejected Savannah's cost allocations because they did not reflect Revenue Canada's observations of the physical production and delivery of the goods to Canada. Revenue Canada believed that "specific identification" most accurately measured the cost of goods produced. Revenue Canada sought to link specific input materials to each export transaction. ⁵⁴ Thus, Revenue Canada rejected assignment of #11 costs to sugar which it could specifically identify as produced and directly exported by Michigan Sugar. ⁵⁵ As discussed below, Revenue Canada used Michigan Sugar's cost records which related to these identified shipments for a portion of its cost of production determination.

With respect to commingled merchandise, Revenue Canada recognized that its preferred approach of "specific identification of the raw material input" would not be possible. ⁵⁶ Nevertheless, Revenue Canada again rejected Savannah's methodology finding that costs assigned from a management inventory pool position were not a proper basis for costing the production of the goods because they did not reflect the variation in pricing of the raw sugar inputs from all sources over the period investigated. Indeed, Savannah admits ⁵⁷ that "[t]here is no direct relationship between the accounting inventory of #11 raw sugar and #14 raw sugar at any given point in time and the physical raw sugar consumed in production of specific subject goods because the inventories of #11 raw sugar and #14 raw sugar are fungible and commingled." ⁵⁸

As discussed below, to cost "commingled" raw sugar used in the production of subject goods for export, Revenue Canada used the weighted average cost of #11 raw sugar and the #14 raw sugar consumed in the production process during the specific month of shipment. ⁵⁹ Revenue Canada acknowledged in its brief that "while this did not reflect the actual cost of production of the goods, it results in a reasonable cost associated with the production of the goods for the purposes of SIMA." ⁶⁰

In sum, Revenue Canada did not reject Savannah's costing methodology based merely upon reflexive invocation of the term "actual." Rather, Revenue Canada compared Savannah's cost assignments with its operations to determine if the assignments made reasonably reflected costs for particular sales during the period of investigation. Revenue Canada determined that an alternative methodology was available which better reflected input costs.

As discussed in the Standard of Review portion of this Opinion, Canadian administrative law consistently holds that a determination of a tribunal may be reasonable even when another reasonable alternative has been rejected. Indeed, the tribunal determination may be reasonable even when the rejected alternative might be more reasonable than the tribunal's approach in the circumstances presented. ⁶¹

Specifically as to GAAP methodologies, "Revenue Canada is not required by the SIMA to adopt a particular method just because it is in accordance with GAAP." ⁶² The offered GAAP may provide "solid guidelines for allocations" but Revenue Canada "is not bound" to those methods. ⁶³ Thus, "SIMA leaves Revenue Canada the discretion to adopt other methods where . . . the facts of the case make the other method reasonable and that method is consistent with SIMA." ⁶⁴ We examine below whether the alternative costing methodology adopted by Revenue Canada was reasonable.

3. Revenue Canada's cost of production determination for inputs from sugar beets was reasonable

The Panel has sustained Revenue Canada's action to investigate Michigan Sugar as an exporter. Accordingly, refined sugar exports to Canada from beet processing plants operated by Michigan Sugar were properly subject to investigation.

As indicated earlier, Revenue Canada believed that "specific identification . . . that is, if the input materials could be specifically identified to each export transaction," most accurately measured the cost of goods produced. ⁶⁵ Savannah's records permitted Revenue Canada to specifically identify exports to Canada which were produced at Michigan Sugar's facility and exported directly to Canada, i.e., the subject goods exported were not processed from raw sugar inputs which had been commingled.⁶⁶ It was undisputed that the input material used at beet processing plants to produce refined sugar was raw sugar beets.⁶⁷ Revenue Canada used the cost of sugar beets and sugar beet processing, as presented by Michigan Sugar, to construct the cost of the refined sugar made from sugar beets at its plants.

The Panel holds that Revenue Canada's effort to match input materials to export transactions is a reasonable interpretation of SIMA, supported by SIMA Section 19(b) and Department Regulation 11(a). Moreover, Revenue Canada's approach is also consistent with Section 15(e) which provides for the comparison of values to be made at the place of shipment. ⁶⁸ Revenue Canada's established policy is to reflect the "differences in costs, efficiencies, economies of scale, technology, plant equipment and production methodology between plants, all of which can impact on the costs and profitability of the goods produced and sold by the various plants." ⁶⁹ Thus, Revenue Canada's determination to attribute the costs which it traced from the specific shipment of those subject goods to the production of the goods from identifiable inputs (not commingled raw sugar) was reasonable in law and fact.

4. Remand of Revenue Canada's cost of production determination for inputs from commingled inventory

The more difficult question involves costing inputs which cannot be identified by production site. As noted earlier in this opinion, Savannah commingles the raw sugar inputs produced in the United States and inputs purchased on the world market. Subject goods, produced from commingled sugar, were exported to Canada. Because of Savannah's objection that Revenue Canada's methodology relied upon the same assumptions Revenue Canada had rejected, the Panel sought to reconstruct the specific mechanism for calculating costs attributed to production from commingled inputs. By examination of the data base upon which the findings were founded the Panel attempted to evaluate fully the reasonableness of Revenue Canada's actions.

As noted earlier, the "Sugar to be Re-exported in Refined Form" program ("Re-export Program") permits licensed U.S. refiners to import raw sugar at the prevailing low world market price, for the production of refined sugar or sugar containing products destined for export. The Re-export Program, through its substitution provisions, "permits the exportation of refined sugar prior to the importation of the corresponding quantity of #11 raw sugar." ⁷⁰ Thus, export credits can be earned under the Re-export Program which permit sugar companies, including Savannah, to later import equivalent quantities of #11 raw sugar.⁷¹

The Re-export Program emphasizes maintenance of records to account for the quantity of imports and exports. ⁷² Under the Re-export Program, "the refiner must maintain records and ensure that a quantity of refined sugar is exported (or transferred to a manufacturer of sugar containing products for export) within a specified time period to correspond with the quantity of #11 raw sugar imported." ⁷³ Revenue Canada could accurately identify and quantify total monthly export sales and #11 purchases from records used by Savannah for eventual reconciliation and apportionment of costs.

Revenue Canada found that the production process for refining raw cane sugar uses a rapid turnover of inventory and a short production time. ⁷⁴ Therefore, Revenue Canada sought to establish the weighted average cost of the #11 and the #14 raw sugar consumed in the production process during the specific month of shipment in its calculations of cost for export sales. ⁷⁵

For commingled sugar, Revenue Canada calculated the weighted average cost of production in any month by using the opening inventory balances of the #11 and #14 raw sugar physical inventory, adding the receipts of raw sugar during that month, and subtracting the ending inventory balances. There appears to be no objection to the mechanics of Revenue Canada's methodology. In performing calculations Revenue Canada was "just doing the arithmetic." ⁷⁶

Revenue Canada's calculations were based on documentation (VE11 (Revised)) provided by Savannah which reflected Savannah's utilization of the benefits of the Re-Export Program. The column of data describing the cumulative status of export sales volume and the #11 purchases [sic!]. The data indicated that [sic!] ⁷⁷

When the #11 raw sugar was in a negative position, Revenue Canada assumed a zero inventory balance for #11 raw sugar. This reading of VE11 is not disputed. Counsel for Savannah summarized: "when the inventory position is negative, yes, there is nothing there. There is nothing there. Why is there nothing there? Because we already made the export sales in question. . . . It is only zero or it is only negative because we made the sale and the goods haven't yet arrived." ⁷⁸

Rather, underlying the reliability issue here is that Savannah has apparently always commingled raw sugar inputs and cannot physically differentiate the goods. Even though the data may accurately reflect the cumulative movement of raw sugar during the period for which such records were maintained, determination of #11 inventory levels is dependent upon the composition of the inventory from which the movements were first measured. Thus, the movement data apparently reflect the cumulative status of export sales and purchases, but incorporate to some degree the opening inventory composition when the current system of record keeping was established. Nothing presented to the Panel indicates that Revenue Canada addressed this infirmity in the submitted data. To the contrary, Revenue Canada consistently demanded that physical inventory volumes be reported and accepted VE11 Revised as establishing accurate physical inventory levels. ⁷⁹ Thus, the Panel finds that Revenue Canada has not addressed the issue of the composition of opening inventory and remands this matter for consideration.

Revenue Canada's reasoning must be transparent. The determination issued by Revenue Canada and the briefs filed in this proceeding did not specifically detail the computations performed or the data base assumptions made for commingled sugar inputs.

The Panel explored Revenue Canada's methodology through questioning at oral argument and by direct examination of the record, but the Panel should not be required to reconstruct the decision-making process. Given the high level of deference accorded Revenue Canada's decisions, it should provide a clear statement of the agency's path of reasoning and computational framework.

With respect to Revenue Canada's use of VE11 (Revised) and transparency of decision-making, the Panel also notes that it is unclear whether the sugar attributable to Michigan Sugar shipments was segregated from the #14 inventory data reported on VE11 (Revised) and whether segregation would affect inventory computations.

Finally, the Panel notes that the time period for data evaluation selected here, i.e. monthly computations, significantly affected the margin calculations. If data had been considered on an annual basis, quite different dumping results might be determined. Nevertheless, Revenue Canada justified this methodological decision in a perfunctory manner. ⁸⁰ For example, Revenue Canada's explanation did not refer to past administrative practice in cases involving goods with production and inventory turn-over times consistent with those experienced here. The selection of the time period for computation of cost of production was an important methodological decision, one which warrants full explication.

At the end of this opinion, the Panel provides remand instructions which direct Revenue Canada to re-evaluate its methodology for costing commingled inputs in light of the problem of establishing historical physical baseline inventories for #11 and #14 raw sugar. The remand instructions also direct Revenue Canada to justify its methodological approach in the other areas discussed in this opinion. Nothing in the Panel opinion or remand instructions should be construed as expressing a Panel preference for a particular cost of production calculation methodology for commingled inputs.

VIII. REVENUE CANADA'S DATE OF SALE DETERMINATION WAS REASONABLE

A. Factual Background and Revenue Canada's Determination

To determine which transactions are subject to investigation Revenue Canada must determine whether the "date of sale" falls within the period of investigation. Savannah and ED&F Man entered into a contractual relationship on December 8, 1992 ("ED&F contract") for the purchase of subject goods. The period of investigation covered January 1, 1994 to February 28, 1995. Some shipments pursuant to the ED&F contract were made before the period of investigation, while other shipments were made during the period of investigation.

Revenue Canada determined that the material terms of the transaction were not established by the ED&F contract, and thus the date of entry into the contract was not the date of sale. Revenue Canada found that the sales invoice was the first document to firmly establish the material terms of the sale and used the invoice date as the date of sale. When the sales invoice was not available, the date of shipment was deemed to be the date of sale.

B. Positions of the Participants

Savannah maintains that the determination of what constitutes a sale or an agreement to sell is governed by law of contract, sale of goods and commercial transactions. Savannah argues that these areas are outside the expertise of Revenue Canada and Revenue Canada's determination of date of sale should be reviewed as to its legal correctness, rather than for reasonableness of Revenue Canada's action.⁸¹

Savannah characterizes the ED&F contract as an "agreement to sell" which qualifies as a "sale" within the plain meaning of SIMA. Savannah argues that, because the term "agreement to sell" was included in the definition of "sale" in SIMA, Revenue Canada need not inquire further as to whether the material terms of the sale are ascertained. ⁸² Savannah notes that Revenue Canada had the statutory authority to expand the period of investigation under subsection 15(d) of SIMA, but failed to do so. Accordingly, because the agreement to sell occurred outside the period of investigation, Savannah believes that the transactions made pursuant to that contract should not have been included in the calculation of export price. Savannah also argues that, in any event, the ED&F contract was sufficiently definite at the date of signing for Revenue Canada to find that the material terms of the transaction were established at that time.

Revenue Canada argues that, because "date of sale" is not defined in SIMA, Parliament clearly intended that Revenue Canada exercise its expertise in making that determination. Revenue Canada refers to the ED&F contract as an "umbrella agreement" ⁸³ which did not provide for all terms and conditions of sale. Instead, the date of sale was properly determined on the basis of either the date of the sales invoice issued by Savannah Foods before the merchandise was shipped, or, if the date of the sales invoice was unavailable, the date of shipment. ⁸⁴ According to Revenue Canada, these dates marked the first point in time when the material terms of the ED&F Contract could be ascertained.

C. Decision of the Panel

The determination of date of sale is a crucial aspect of an antidumping investigation because the date of sale defines the sales that will be included in the comparison of normal value and export price in order to determine whether the merchandise in question has been dumped. Because Revenue Canada is responsible for determining the methodology for calculating export price, it must have the authority to determine what constitutes the date of sale. This function is reasonably related or ancillary to calculating export price. Revenue Canada properly states that "the objective of selecting a date of sale in anti-dumping investigations is specific to the objectives of SIMA and, in this respect, differs from any other legislation governing commercial transactions." The interpretation of SIMA by Revenue ⁸⁵ Canada with respect to what constitutes the date of sale is well within Revenue Canada's statutory function and that interpretation should be reviewed under the reasonableness standard.

SIMA provides that "a sale includes leasing and renting, an agreement to sell, lease or rent and an irrevocable tender." ⁸⁶ For purposes of the Panel analysis, we accept that the ED&F contract was an agreement to sell and a "sale" under the SIMA. ⁸⁷ Nevertheless, whether or not an agreement to sell may constitute a sale under the SIMA, such does not establish that the date of the agreement is the date of sale.

Although SIMA itself is silent with respect to the methodology for determining the "date of sale," Revenue Canada properly derives guidance from the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade which provides that the appropriate date of sale may be determined from the date of the contract, the purchase order, the order confirmation, or the invoice, whichever establishes the material terms of sale. ⁸⁸ Moreover, investigation of when the material terms of a transaction were established to find the "date of sale" is apparently in accord with an established practice of the tribunal. ⁸⁹

The record is sufficient to support as reasonable Revenue Canada's conclusion that the material terms of the sale were not contained in the ED&F Contract of December 8, 1992. Many of the essential terms of the contract were still subject to some modification by both the buyer and seller until that point in time at which the sales invoice was issued by Savannah or the goods were actually shipped to the buyer.

Although the December 8, 1992 contract contained a pricing formula by which price could be determined at some fixed future date, the exact pricing could be affected by elections made by the parties. Moreover, quantity was not fixed, but was established with reference to an "upper and lower range with the final tonnage to be determined by buyer's advice within 90 days of signing of the contract." ⁹⁰ The grade or quality of the merchandise was also subject to availability at the time the product was to be delivered. ⁹¹ Accordingly, Revenue Canada could reasonably conclude that the date of sale for each of these transactions was not the date of the ED&F contract, but a subsequent date which fell within the period of investigation.

The Panel holds that there was sufficient objective evidence on the administrative record to support Revenue Canada's determination that the date of sale for shipments pursuant to the ED&F contract fell within the period of investigation and should thus be included in Revenue Canada's calculation of export price.

IX. CONCLUSION

For the reasons stated above, Revenue Canada's determination is hereby affirmed in part and remanded in part.

REMAND INSTRUCTIONS:

1. Revenue Canada must analyze and explain the reasonableness of its methodology for costing raw sugar inputs from commingled inventories. If Revenue Canada finds it can determine the physical composition of the baseline or opening inventory, it should detail the analysis which permits it to reach this conclusion. If Revenue Canada cannot determine the physical composition of the baseline or opening inventory, Revenue Canada must explain why its methodology is nevertheless reasonable notwithstanding its adoption of opening inventory data provided by Savannah.

2. If Revenue Canada finds that the inability to establish physical baseline inventory levels fatally compromises its methodology, Revenue Canada may determine cost of production for commingled merchandise by such other method it believes is reasonable.

3. Revenue Canada should describe whether segregation of any quantity reflected in Savannah's #14 raw sugar inventory data attributable to Michigan Sugar is appropriate in making its cost calculations for commingled inputs. If segregation is appropriate and was not performed in prior calculations, such an analysis should be conducted and Revenue Canada should report whether the change materially affected its cost computations.

4. Revenue Canada should specifically justify the decision to analyze inventory consumption data for cost calculations on a monthly basis. In its discussion, Revenue Canada should also state whether the period selected materially affected the quantum of dumping found to be present.

The results of this remand shall be provided by Revenue Canada to the Panel within 45 days of this decision.

SIGNED IN THE ORIGINAL BY:

Issued on the 9th day of October, 1996 

ARTICLE 1904 BINATIONAL PANEL REVIEW UNDER THE NORTH AMERICAN FREE TRADE AGREEMENT

ORDER

For the reasons stated in the memorandum opinion, the Panel affirms in part Revenue Canada's final determination. The panel also remands in part Revenue Canada's final determination for further proceedings consistent with the memorandum opinion and remand instructions.

The results of the remand shall be provided by Revenue Canada to the Panel within 45 days of this decision.

SIGNED IN THE ORIGINAL BY:

Issued on the 9th day of October, 1996 

³⁸ Statement of Reasons, No. 4237-80 & 4218-2, October 5, 1995 (hereinafter Statement of Reasons), at 4, 9; see also Revenue Canada's Brief at 12.

³⁹ The constructed value is the "aggregate of (i) the cost of production of the goods, (ii) a reasonable amount for administrative, selling and other costs, and (iii) a reasonable amount for profit." SIMA Section 19(b).

⁴⁰ Statement of Reasons at 4.

⁴¹ Savannah's Brief at 5.

⁴² Revenue Canada's Brief at 47.

⁴³ Revenue Canada's Brief at 48.

⁴⁴ See Beer Panel at 17 ("The interpretation and application of the SIMA falls squarely within Revenue Canada's area of expertise.")

⁴⁵ Canada Gazette Part II, Vol. 118, No. 25 (Dec. 12, 1984).

⁴⁶ Agreement on Implementation of Article VI of GATT, Article 2.2.1.1.

⁴⁷ Certain Cold-Rolled Steel Sheet, CDA-93-1904-08, at 29.

⁴⁸ Statement of Reasons at 6 (emphasis added).

⁴⁹ Memorandum for Mr. Brian Brimble, September 25, 1995, (A.R. Index p. 107, tab 15) (hereinafter Brimble Memo) at 3 (emphasis added).

⁵⁰ The panel notes that Revenue Canada stated that its methodology for commingled inputs (which involved assignment of a blended or weighted-average cost) reflected "the actual cost of the domestic sales of refined sugar made using raw sugar from all sources." Statement of Reasons at 6. This use of the term "actual" is confusing, particularly in light of Revenue Canada's later statement that its approach "did not reflect the actual cost of production of the goods." Revenue Canada's Brief at 12.

⁵¹ Brimble Memo at 5.

⁵² Id.; see Revenue Canada's Brief at 49.

⁵³ Id.

⁵⁴ Brimble Memo at 3.

⁵⁵ Savannah's counsel stated that the Michigan Sugar shipment was treated "notionally" as a domestic sale by Complainant. "In other words, we will take the Michigan Sugar, ship it across the border, but really treat the Michigan Sugar purchase as a domestic purchase for our sale in the United States." Public Hearing Tr. at 95 (Pearson), see also id. at 338 (Pearson).

⁵⁶ Statement of Reasons at 6.

⁵⁷ Brimble Memo at 5; see Revenue Canada's Brief at 49, 55; Officer Memo to File (March 10, 1995), A.R., Index p. 113, tab 8, pp. 19 - 21 (no public version).

⁵⁸ Public Hearing Tr. at 268 (Pearson).

⁵⁹ Revenue Canada's Brief at 12.

⁶⁰ Id.

⁶¹ Certain Cold-Rolled Steel Sheet, CDA-93-1904-08, at 29.

⁶² Id.; see Savannah's Brief at 38.

⁶³ Certain Cold-Rolled Steel Sheet, CDA-93-1904-08 at 31.

⁶⁴ Id. at 29.

⁶⁵ Brimble Memo at 3; see Revenue Canada's Brief at 21.

⁶⁶ See Public Hearing Tr. at 270 and Exporter Decision Memorandum (August 14, 1995), A.R. Index p. 113, tab 3 (no public version).

⁶⁷ Confidential Portion of the Hearing Transcript, July 10, 1996, (hereinafter Confidential Hearing Tr.) at 15 (Pearson) (non-confidential statement) "The goods that were shipped from Michigan plant locations was all sugar beet-based sugar."

⁶⁸ Savannah's counsel agreed that "Section 15(e) of the statute basically says that normal values should be established at the plants from where the goods are sent to Canada. We don't have a problem with that construction." Public Hearing Tr. at 141 (Pearson).

⁶⁹ Brimble Memo at 3.

⁷⁰ Statement of Reasons at 5.

⁷¹ Revenue Canada's Brief at 14; Statement of Reasons at 5.

⁷² Id.; see also Confidential Hearing Tr. 93 (Woyiwada) (no public version).

⁷³ Statement of Reasons at 5.

⁷⁴ Revenue Canada's Brief at 13.

⁷⁵ Revenue Canada's Brief at 12; Statement of Reasons at 6.

⁷⁶ Confidential Hearing Tr. at 8 (Pearson) (non-confidential statement). Revenue Canada "just did the normal arithmetic of adding inventory and subtracting deliveries and coming out with consumption. This is just straight plus, minus, equals." Id. at 9 (Pearson) (non-confidential statement).

⁷⁷ Confidential Hearing Tr. at 11-12 (Pearson).

⁷⁸ Public Hearing Tr. at 319 (Pearson).

⁷⁹ "VE11, according to the documents, the correspondence that I quoted earlier, shows physical inventories." Hearing Tr. at 268 (Woyiwada).

⁸⁰ Panel members Trakman and McGill believe Revenue Canada's rationale for monthly calculations to be reasonable in light of the short production cycle for subject goods and rapid inventory turnover of raw sugar inputs.

⁸¹ Savannah's Brief at 53.

⁸² Public Hearing Tr. at 47 (Pearson).

⁸³ Revenue Canada's Brief at 66.

⁸⁴ Id. at 27.

⁸⁵ Id. at 65-66. 

⁸⁶ SIMA section 2(1).

⁸⁷ Public Hearing Tr. at 292-294 (Woyiwada).

⁸⁸ Revenue Canada's Brief at 65, citing Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade, 1994, Article 2, footnote 8, sub-paragraph 2.4.1.

⁸⁹ See, e.g., Acceptance of an Undertaking and Suspension of the Antidumping Investigation Respecting Certain Cast Steel Grinding Balls Originating In or Exported From the United States of America and Produced By or on Behalf of Capitol Castings Inc., its successors and assigns, of the United States of America, File No. 4258-79 (APW/824), December 15, 1989, at 6-7.

⁹⁰ Savannah's Brief at 21.

⁹¹ Id.