IX. SCOPE DETERMINATIONS AND “CLASS OR KIND” ISSUES

The countervailing duty statute gives the Department the responsibility as “administering agency” to determine the scope of a CVD order in terms of the “class or kind” of imported merchandise to be included in the order. 19 U.S.C. § 1671(a)(1). The antidumping statute similarly refers to the “class or kind” of imported merchandise in giving the Department the responsibility to determine the scope of an antidumping order. 19 U.S.C. § 1673(1). Although neither statute defines the term “class or kind” and that term is not defined in the Department’s regulations, the general legal framework for the Department’s scope determinations seems straightforward. As was recently stated by the Federal Circuit in Duferco Steel, Inc. v. United States, 296 F.3d 1087, 1096 (Fed. Cir. 2002), which addressed both antidumping and CVD petitions, where an investigation is commenced by petition, the “purpose of the petition is to propose an investigation” whereas one “purpose of the investigation is to determine what merchandise should be included in the final order.” Hence, said the Court, “[i]t is the responsibility of the agency, not those who initiated the proceedings [the “petitioners”], to determine the scope of the final orders.” Id., at 1097. Nevertheless the application of that general legal framework has given rise to numerous issues in the present proceeding as in other antidumping and countervailing duty proceedings.

We address the following issues raised by the Department’s scope determinations: (A) the Complainants’¹³ claim of violation of due process in the Department’s consideration of scope issues; (B) the role of the petition and the Department’s discretion in making scope determinations; (C) the statutory requirement of industry support for each “class or kind” of merchandise included in the petition; (D) the scope of Panel review of the Department’s scope and “class or kind” determinations; (E) whether evidence in the record supports the Department’s rejection of separate “class or kind” status for contested species and contested products, and (F) whether exclusion is warranted for two other products.

A. Due Process in the Department’s Consideration of Scope Issues

In this proceeding the Department chose to take up issues relating to the scope of the investigation jointly with the same issues raised in the concurrent antidumping proceeding. The Department noted in its Preliminary Determination that it had received scope requests of three kinds relating to some 50 products: scope clarification requests, scope exclusion requests, and requests for determinations of separate classes or kinds. See Notice of Preliminary Affirmative Countervailing Duty Determination: Certain Softwood Lumber Products from Canada, 66 Fed. Reg. at 43186, 43187 (August 17, 2001). Substantially the same scope issues were raised in the antidumping proceeding. At that time, August 17, 2001, the Department excluded some products from the scope of the CVD investigation, but made no findings on requests for separate class or kind treatment. Thereafter, on March 12, 2002, the Department issued a preliminary determination on scope issues and called for briefs to be submitted by March 15, with rebuttal briefs due March 18, and held a hearing on scope issues on March 19, 2002.

Complainants charge that the Department’s truncated process for final determination of CVD scope issues, which they term the “nine days in March” amounts to a denial of due process. The Department responds that it had requested comments on scope issues early in the investigation and specifically invited additional comments in the Preliminary Determination issued August 17, 2001. The Department’s argument is as follows:

[By means of the submission of such comments,] all interested parties had notice of the alternatives the Department was considering for the final determination. That is, the record contained requests by some parties to consider certain products as being outside of the scope of the investigation. The record also contained requests by some parties to classify certain products as a different “class or kind” of merchandise. In respect to both types of requests, the record also contained the petitioners’ comments that a number of these requests be denied. Therefore, because these alternatives were on the record, the Department was not required to reach a full preliminary determination on all scope issues. Rather, because the parties had notice of the alternatives being considered, and could address these alternatives in their case briefs, the Department provided the parties with sufficient notice of these issues under the statutory and regulatory regime. [Department’s Response Brief, vol. V, at K-64-65.]

The Department describes the briefing and hearing scheduled for the “nine days in March” as discretionary and extraordinary, as it was not obligated by statute or regulation to request these written arguments or hold a specific hearing specifically addressing only scope arguments. The Department argues that the scope issues were adequately addressed by the opportunity for briefing and the hearing which was held.

The Panel is not persuaded that the Complainants have been denied due process on consideration of scope issues. Although it would certainly have been more reasonable, and perhaps more appropriate, to have allowed for a more leisurely schedule for written comments and for the scope issue hearing, there seem to the Panel to have been no surprises in the content or in the outcome of the scope proceeding. Complainants surely had ample opportunity to submit scope information and arguments in the period following the August 17, 2001 Preliminary Determination, and there is no reason to believe that Complainants, or their experts, were to any significant degree surprised or disadvantaged by the scope arguments put forward either in writing or at the scope issue hearing held March 19, 2002.

B. The Role of the Petition and the Department’s Discretion in Making Scope Determinations: Defining “Class or Kind”

The threshold substantive issue for the Panel is whether the Department properly determined that this investigation involves but a single class or kind of merchandise. The beginning point in this analysis is not in serious dispute: The Department has the statutory responsibility to determine the class or kind of merchandise subject to an investigation. See Duferco Steel, 296 F.3d at 1097 (“[i]t is the responsibility of [the Department], not those who initiated the proceedings [the “petitioners”], to determine the scope of the final orders.” See also Mitsubishi Elec. Corp. v. United States, 700 F. Supp. 538, 555 (CIT 1988), aff’d 898 F.2d 1577 (Fed. Cir. 1990), and Torrington Co. v. United States, 745 F. Supp. 718, 720-21 (CIT 1990), aff’d 938 F.2d 1276 (Fed. Cir. 1991). Indeed, the Complainants “do not challenge Commerce’s definition of scope.” Rather the Complainants frame the issue for Panel review as “whether there are multiple classes or kinds of merchandise within the investigation’s scope, as Commerce defined that scope”. Ibid.

In determining the scope of the investigation, the Department is guided by, but is not bound by the allegations set forth in the petition. See Duferco Steel, 296 F.3d at 1096. As in many other cases, the Court in Mitsubishi Electric, 700 F. Supp. at 555, linked the Department’s authority to the terms of the petition: “The ITA has the authority to define and/or clarify what constitutes the subject merchandise to be investigated as set forth in the petition containing the intent of petitioner expressed in as specific and definite terms, descriptions, and language as reasonably expected of petitioner....” Thus, although the role of the petition is not specified in the statute, the Department’s practice in determining the scope of the final CVD order is to “reflect[] the intent of the petition”. See Minebea Co. v. United States, 782 F. Supp. 117, 120 (CIT 1992). Indeed, reviewing courts have said that the Department may give “ample deference” to the scope as described in the petition by the petitioners. Torrington Co. v. U.S., 786 F.Supp. 1021, 1026 (CIT 1992), cited with approval in Sundstrand Corp. v. U.S., 890 F.Supp. 1100, 1103 (CIT 1995).

Nevertheless the Department’s definition of the scope of the CVD investigation and order presents both the “class or kind” issue and the issue of the extent to which the Department’s determination of “class or kind” is subject to judicial review. As noted above, although the CVD statute refers to the scope of a CVD proceedings in terms of the “class or kind” of imported merchandise to be included in the CVD order, the statutory term “class or kind” is not defined either in the statute or in the Department’s regulations. See 19 U.S.C. § 1671(a)(1). Hence, according to the Department, the matter of “class or kind” is left to the discretion of the Department, and which is guided by the scope as set forth in the petition, and the extent of the harm alleged by the petitioners.

Thus, says the Department:

[T]he scope of an order, as defined by the Department, may appear narrow or broad, depending upon the harm alleged by the Domestic Industry. This has led to separate orders [for separate classes or kinds of merchandise], for example, for roasted and unroasted pistachios from Iran, on one hand, and [to a single] order covering a large array of textiles from Argentina, on the other. See, e.g., Final Affirmative Countervailing Duty Determination and Countervailing Duty Order; Roasted In-Shell Pistachios from Iran, 51 Fed. Reg. 35679 (Oct. 7, 1986); Final Affirmative Countervailing Duty Determination and Countervailing Duty Order; In-Shell Pistachios from Iran; 51 Fed. Reg. 8344 (March 11, 1986); Certain Textiles and Textile Products from Argentina; Final Countervailing Duty Determination, 43 Fed Reg. 53421 (Nov. 16, 1978). [DOC Response Brief, vol. V, p. K-56.]

The Department explains that if, for example, the “domestic fruit-canning industry” were to claim harm by imports of various forms of fruit, the Department could include in the CVD order both apples and oranges. But, where the “domestic fresh apple industry” claims injury resulting from imported apples, the CVD order would include only apples. .

In the present proceeding the Department relied upon the petition in determining the overall scope of the CVD investigation as addressing “various forms of softwood lumber from Canada” as a single “class or kind” of imported merchandise. While the scope of the harm alleged in the petition does not necessarily bind the Department to the concept of “class or kind” embodied in the petition, courts have affirmed that the Department’s discretion extends to determining the scope of the investigation and order based upon the allegations of harm in the petition. Accordingly the fact that the Department found but a single “class or kind” of merchandise across a large spectrum of softwood lumber products is not in itself a basis for overruling the Department’s determination.

C. The Statutory Requirement of Industry Support for Each “Class or Kind” of Merchandise Included in the Petition

The examples of scope determinations cited in the previous section point to a link between “class or kind” issues and the degree of industry support needed for a CVD petition. The statute requires that a CVD petition be filed by an “interested party” which the statute defines in terms of producers of “a domestic like product”. See 19 U.S.C. § 1671a(b)(1), referring to 19 U.S.C. § 1677(9)(C), (D), (E), (F) & (G). Complainants argue that “domestic like product” must be defined separately for each class or kind of merchandise included in the petition. Hence, according to Complainants, to qualify a petition under 19 U.S.C. § 1677(9)(E), the petitioners (here the Coalition) must show industry support for each class or kind of merchandise. Thus the petitioners must show that “a majority of [its] members manufacture, produce or wholesale a domestic like product” in respect of each class or kind of merchandise included in the petition.

The Department observes that the issue of industry support will arise only if the Panel finds that separate classes or kinds of merchandise are in fact included in the investigation. Since the Department, which found only a single class or kind involved in this proceeding, did not consider the issue of industry support with respect to the asserted separate classes or kinds of merchandise, such a finding would require the Panel to remand to the Department “for further consideration of the legal and factual arguments” involved in the industry support determination.

The Panel agrees with the Department’s analysis. Accordingly the Panel need not address the relationship between “class or kind” and industry support. Should the Panel find that this proceeding involves more than the single class or kind of imported merchandise, the Panel would remand the issue of industry support for reconsideration by the Department.

D. The Panel’s Standard of Review of the Department’s Scope and “Class or Kind” Determinations

According to the Department, where there is no “clear consensus” in the evidence on the record “as to the particular merchandise which should or should not be covered by the scope [of the proceeding], a court must affirm the Department’s ultimate definition of the scope, as long as its determination is reasonably supported by the record.” DOC Response Brief at K-4. In the present investigation, the Department found that “no product subject to the order warranted separate class or kind treatment.”

The Department states that it uses the following procedure in determining whether to include a particular product within the scope of a CVD investigation and order:

[T]he Department often considers whether or not individual merchandise subject to the order differs from other subject merchandise to such a degree, that separate class or kind treatment is warranted. In determining if products are so distinct and unique as to grant such treatment for countervailing duty purposes, the Department compares the merchandise with all other merchandise covered by the scope. [DOC Response Brief at K-5 (emphasis supplied).]

In their challenge to the Department’s denial of separate “class or kind” treatment for particular products, Complainants assert that if a petition identifies a single class of goods that “in reality” includes more than one class or kind, Commerce is required to subdivide the merchandise into multiple classes or kinds. Only in this way can the investigation properly account for each class or kind, including those which are found not to be properly identified within the Petition. The issue here is not what to do when a petition classifies as a single class or kind merchandise that “in reality” constitutes more than one class or kind, it is by what standard the Department is to decide whether a petition includes one or more than one “class or kind” of imported merchandise.

Where the Department must address claims that multiple classes or kinds of merchandise are at issue, Complainants cite Diversified Products Corp. v. U.S., 572 F. Supp. 883 (CIT 1983), and argue that the Department is required to apply the Diversified Products criteria to determine if a contested item is a separate class or kind of merchandise.

Diversified Products involved an administrative review of an antidumping order which the Department construed as applicable to a product that had not yet been developed at the time of the antidumping investigation. In determining that the newly developed product belonged to the same class or kind of merchandise that was included in the final antidumping order, the Department used the following five broad criteria:

“the general physical characteristics of the merchandise, the expectation of the ultimate purchasers, the channels of trade in which the merchandise moves, the ultimate use of the merchandise, and cost.” 572 F.Supp. at 889. Reviewing the Department’s application of these criteria, the Court upheld the Departme nt’s “class or kind” determination, finding that it was supported by substantial evidence on the record.

The Department points out that Diversified Products and its progeny address a “scope ruling” made by the Department in an administrative review of a final CVD or antidumping order – and not an initial scope determination made by the Department in the course of an investigation. Unlike the initial scope determination, the Department’s “scope ruling” is governed by § 351.225 of the Department’s regulations, which provides a procedure by which parties may request what the Department calls a “scope ruling” to determine whether a product is covered by an existing CVD or antidumping order. Subparagraph (k)(1) of § 351.225 directs the Department to base such “scope rulings” on “[t]he descriptions of the merchandise contained in the petition, the initial investigation, and the determinations” of the Department. Where those criteria “are not dispositive,” subparagraph (k)(2) of §351.225 directs the Department to consider the following five criteria:

(i) The physical characteristics of the product;

(ii) The expectations of the ultimate purchasers;

(iii) The ultimate use of the product;

(iv) The channels of trade in which the product is sold; and

(v) The manner in which the product is advertised and displayed.

The regulation’s “scope rulings” criteria are identical to the Diversified Products criteria, except that the “manner in which the product is advertised and displayed” is substituted for the cost of the product.

Since the present proceeding reviews the Department’s initial scope determination and its refusal to exclude certain products from the scope of the investigation and CVD order, neither the Diversified Products caselaw nor the Department’s regulation essentially incorporating that caselaw applies of its own force.

Nevertheless Complainants assert that the Department “must apply” the Diversified Products criteria to determine whether an item is a separate class or kind of merchandise. Without explicitly abandoning the “must apply” proposition of their Case Brief, Complainants’ Reply Brief cites Torrington, 745 F.Supp. at 720-29, for the proposition that “the Diversified Products factors are applicable in original investigations”. Complainants’ Reply Brief at 4 (emphasis supplied). While Torrington certainly endorses the Department’s use of the Diversified Products factors to address class or kind issues raised in an investigation, see Mitsubishi Electric, cited in Torrington, 718 F.Supp. at 723, Torrington does not support the proposition that the Department “must apply the Diversified Products criteria” in making “class or kind” determinations in an original investigation. Indeed, the Panel has found no authority for that proposition. Hence, although in the exercise of its discretion the Department may choose to apply the Diversified Products criteria to “class or kind” determinations in a CVD investigation, the Panel holds that the Department is not bound by law to do so.

In the present case, although the Department’s scope determinations were not “scope rulings” subject to § 351.225 of the Department’s regulations, the Department “recognize[ed] the reasonableness of the Diversified Products factors” and chose to use those factors “to guide its determination of whether a particular product should be separated, for purposes of the final order, as a separate ‘class or kind.’” In making such determinations, the Department states that it compared the contested softwood lumber products “with other merchandise subject to the order, and although it found each product to contain certain unique qualities, none of these products differed to such a degree from other merchandise subject to the order to warrant treatment as a separate class or kind under the order.” [DOC Response Brief at K-8 & 10-11 (emphasis supplied).]

Complainants object to the Department’s analysis on grounds of “law and logic”. Complainants assert, first, that the way in which the Department applied the Diversified Products criteria is inconsistent with Diversified Products jurisprudence, and, second, that “logic” dictates that the Department must either “compare every product to every other product under the [Diversified Products] criteria (which is impractical)”, or must “compare every challenged product to a benchmark that represents the same set of core products covered by the scope” of the investigation. Complainants’ Reply Brief at 5 (emphasis supplied). On the first of these two arguments, Complainants cite a “scope ruling” case, Bohler-Uddeholm Corp. v. U.S., 978 F. Supp. 1176, 1183 (CIT 1997), for the proposition that “a proper Diversified Products analysis would compare the products in question to the entire class or kind of merchandise” covered by the final order (emphasis supplied by Complainants). Complainants’ Reply Brief at 5. The passage quoted from Bohler- Uddeholm may accurately state the Diversified Products “scope ruling” methodology. However, the issue here is whether Complainants have established that the Department is required by law to apply the Diversified Products methodology to the initial scope determination at issue in the present proceeding. Inasmuch as the Panel has held that that the law does not require the Department to apply the Diversified Products criteria to initial “class or kind” determinations in a CVD investigation, the fact that in this proceeding the Department may have departed from the Diversified Products methodology in making the initial class or kind determination does not invalidate that determination.

The Complainants’ second argument is that the “logic” of class or kind determinations requires the Department to “compare every challenged product to a benchmark”. Whether or not the Complainants’ “benchmark” argument is as logical or persuasive as Complainants contend, their argument is one that the Panel cannot accept. As the Panel observed at the outset of this section of the opinion, although the CVD statute refers to the scope of a CVD order in terms of the “class or kind” of imported merchandise, neither the statute nor the Department’s regulations defines the term “class or kind”. Since the Department has no statutory or other legal responsibility to apply a “benchmark” analysis to the determination of “class or kind” issues, the Panel considers that the Department has discretion to determine how to apply the “class or kind” concept to carry out the intent of the statute in the context of the facts of each case.

To summarize the dispute as to the methodology for making initial “class or kind” determinations, Complainants argue that the fit of any product within the scope of the final investigation and order must be tested by the Diversified Products criteria and will be subject to judicial review under the Diversified Products jurisprudence. In contrast, the Department’s focus is on the “degree” to which a contested product “differs from other subject merchandise” and whether the contested product is “so distinct and unique” as to warrant separate “class or kind” treatment. See DOC Response Brief at K-5 (emphasis supplied). And, in determining whether contested products are “distinct and unique”, the Department examines all other merchandise covered by the investigation.

In this investigation, the Department, guided by the petition, gave broad scope to the CVD investigation and order as embracing but a single “class or kind” of imported merchandise, namely “various forms of softwood lumber from Canada”. The Department’s methodology for initial “class or kind” determinations indicates that the Department treated this determination as a matter of degree. As explained by the Department, the Department’s task is to compare the contested product with all other merchandise covered by the CVD investigation to determine whether the contested product is “so distinct and unique” as to differ from other subject merchandise – in which case the contested product would be given separate “class or kind” treatment.

The Panel recognizes that, owing to the limited scope of judicial review of expert agency decisions, the Panel is not authorized to second-guess the Department’s expert judgment on such matters of degree. The Panel must therefore affirm the Department’s reasonable interpretations of the governing statute, and, in this instance, the methodology chosen by the Department to give effect to the “class or kind” provision of the statute. The Panel must also affirm factual determinations made by the Department pursuant to that methodology unless such determinations are “unsupported by substantial evidence on the record.” 19 U.S.C. § 1516a(b)(1)(B). The Panel may not reweigh the evidence or substitute its own evaluation of the evidence for that made by the Department. [See the section on Standard of Review in this opinion.] Thus, although we briefly review the record evidence as to the two contested species and three contested products in the next section, our review is limited to assessing whether in those instances the Department’s 48 denial of separate “class or kind” status is reasonably supported by evidence on the record.

E. Whether Evidence on the Record Supports the Department’s Rejection of Separate “Class or Kind” Status for Contested Species and Contested Products

As indicated above, review has been sought in respect of two species, namely Western Red Cedar and Eastern White Pine, and three products that the Department held were covered by the CVD investigation and order: Finger-Jointed Flangestock, Square End Bed Frame Components, and Box-Spring “Kits” that are not exported to the United States in individual packages. With respect to the two contested species and the first two of the four contested products, that is, Western Red Cedar, Eastern White Pine, Finger- Jointed Flangestock, and Square End Bed Frame Components, Complainants argue that each species or product is a separate class or kind of merchandise and hence falls outside the scope of the “softwood lumber products” investigation.

In view of the Panel’s findings, the Panel’s review of record support for the contested scope determinations is based on the Department’s approach to scope issues. Accordingly our inquiry in each case is whether substantial evidence on the record supports the Department’s determination that, based on comparisons “with all other merchandise covered by the scope” of the CVD investigation, the contested product is not “so distinct and unique” as to warrant separate “class or kind” treatment. With respect to the contested species and the products at issue here, the Department either provided a detailed analysis under each Diversified Products factor or addressed only the “factors which arguably distinguished that product from other merchandise covered by the scope of the investigation.” And, in each case, the Department found that although each species or product had properties unique to its species or product type, each was still softwood lumber, and was similar enough under the Diversified Products factors with the other scope merchandise to be included in one class or kind of merchandise.

1. Western Red Cedar and Eastern White Pine

With respect to Western Red Cedar and Eastern White Pine, the Department analyzed all five Diversified Products factors. Although extensive counter arguments have been made by Complainants, those arguments cannot overcome the deference owed to the Department on findings of fact and the application of its methodology for determining whether a contested product is “so distinct and unique” as to differ from other subject merchandise. For example, the Department found that many physical characteristics and end-uses invoked to distinguish these two species from other species of softwood lumber were in fact shared in varying degrees with other softwood lumber species. Such characteristics include attractive appearance, light weight, and natural durability for Western Red Cedar, and structural strength and ability to take stains, paints, etc, for Eastern White Pine; and end uses include high-end use for shingles, paneling, and siding for Western Red Cedar, and millwork, paneling, and siding for Eastern White Pine. The Department made similar analyses of characteristics shared between the two contested species and other species of softwood lumber across the remaining Diversified Products factors, that is, expectations of the purchaser, channels of trade, and manner of advertising.

Ultimately the question for the Panel is whether substantial evidence on the record supports the Department’s determination to decline to accord separate class or kind status to Western Red Cedar and Eastern White Pine. The Department’s determination need be supported only by “substantial evidence”, which does not refer either to the weight of the evidence or to the preponderance of the evidence; and the Panel is not entitled to reweigh the evidence in reaching its conclusion. By this standard and in light of the Department’s methodology for making class or kind determinations, the Panel considers that substantial evidence on the record supports the Department’s determination that neither Western Red Cedar nor Eastern White Pine is “so distinct and unique” as to warrant separate “class or kind” treatment. Hence the Panel affirms the Department’s determination with respect to Western Red Cedar and Eastern White Pine.

2. Finger-Jointed Flangestock

Turning to Finger-Jointed Flangestock (hereinafter “FJF”), it is common ground between the parties that FJF “is a softwood lumber product” from Canada. DOC Response Brief, at K-30 (emphasis in the original). Although FJF is an “engineered wood product” with “physical characteristics which make the product unique”, the Department determined that FJF “was not so unique as to treat it as a different class or kind of merchandise”. DOC Response Brief, at K-30. The Department states that although it had reached this judgment after analysis under all five Diversified Products factors, it had explained the judgment in terms of some, but not all, of the Diversified Products factors. The Department states that it “determined that the merchandise shared common features with other softwood lumber products and therefore only addressed those Diversified Products factors which arguably differentiated this product as unique for purposes of class or kind treatment.” Id. at K-33 (emp hasis supplied).

Complainants chiefly rely on two arguments in urging that the Department’s decision to deny separate “class or kind” status to FJF is contrary to law. First, Complainants argue that the law requires the Department to apply the complete Diversified Products analysis in making class or kind determinations in an original investigation. As shown above, the Panel has rejected this argument on the ground that neither the statute nor the Department’s regulations establish a legal responsibility for the Department to use the Diversified Products analysis in an original investigation. Hence the Department has discretion to invoke Diversified Products analysis as a guide to its determination, as it did in the present case.

Second, pointing to the rather brief explanation of the Department’s FJF “class or kind” determination, i.e., the two paragraphs in the Decision Memo at 165-66. Complainants argue that the determination is not supported by substantial evidence on the record and that the Department failed to provide a clear “path of reasoning which led to the final outcome.” See Wheatland Tube Co. v. U.S., 161 F.3d 1365, 1369-70 (Fed Cir. 1998).

The Panel notes, however, that there is record evidence of the uniqueness of FJF, and that the Department’s “path of reasoning” shows very little more than that FJF is a “softwood lumber product from Canada” having some Diversified Product features in common with other softwood lumber products. Neither element is contested here. What is at issue is whether FJF is “so unique” as to warrant separate “class or kind” treatment in this proceeding.

While the Department acknowledges that the “physical characteristics” of FJF “make the product unique”, it asserts that these physical characteristics are not “in of themselves (sic) sufficient to warrant separate class or kind treatment”. DOC Response Brief at K-30-31. In the Decision Memo at 165, PR Doc. 863, the Department characterized FJF as “another lumber product in the broad field of lumber products with distinct characteristics and end uses” (emphasis supplied). The Department also said: “The particular construction, strength rating, dimension and end-use of flanges cannot be the sole basis for their treatment as a separate class or kind.” Ibid. These statements seem to mean that while the Department accepts that FJF has “unique” physical characteristics and end-uses, a showing of a product’s unique physical characteristics and/or end-uses may not make the product sufficiently “unique” for separate “class or kind” treatment.

The Department further explained that “FJF’s unique length, channel of trade and manner of advertising potentially differentiated this softwood lumber product from other products under the scope.” DOC Response Brief at K-33. As to the latter two factors, the Department “recognized” that it had found that the FJF producer Tembec was able to identify “a distinct channel of trade (sales to I-beam producers)” and “a [distinct] manner of advertising (none since it sells all of its flange stock to I-beam producers)”, but it “found that this channel and manner of advertising were not so distinct from ‘other lumber specialty products’.” Id. at K-31, quoting from Comment 52, PR Doc. 936. This refers to the Decision Memo where the Department said: “[T]his channel of trade and this manner of advertising, distinct as they are from most dimension lumber, are not so distinct from those of other lumber specialty products.” Decision Memo at 165. This seems to mean that a showing of a product’s merely “distinct” channel of trade and/or advertising will not make the product sufficiently “unique” for separate “class or kind” treatment where other subject merchandise shares similar channels of trade and/or advertising.

Putting these statements together as indicating the Department’s “path of reasoning”, it seems that a showing of any combination of a product’s “distinct” and/or “unique” physical characteristics, end-uses, channels of trade, and/or advertising may not make the product sufficiently “unique” for separate “class or kind” treatment where any other subject merchandise shares one or more similar characteristics. Nothing is said about the expectations of purchasers, which is the remaining Diversified Products factor, but, as the Department has said, while it looks to Diversified Products for guidance, it is not obligated to address every Diversified Products factor in an original scope determination.

On the present record the Panel is able to discern the following “path of reasoning”: First, FJF is undeniably “softwood lumber from Canada”. Second, whereas the physical characteristics and end-uses of FJF are “unique”, FJF channels of trade and advertising are merely “distinct” in that these channels also exist for other forms of softwood lumber. We are not clear as to when an “unique” characteristic of a softwood lumber product will be considered by the Department to be sufficiently unique to warrant separate “class or kind” treatment. However, on the final step of evaluating the uniqueness of FJF, the Department appears to consider that evidence showing that channels of trade and advertising similar to those used for FJF are used for other forms of softwood lumber is sufficient to establish that FJF is not “so unique” as to warrant separate “class or kind” treatment.

In a broad scope investigation (such as softwood lumber), the Department’s reasoning seems to make it virtually impossible for any specialty product (like FJF) to receive separate “class or kind” treatment. Although (like FJF) the specialty product may be considered “unique” by the Department in respect of physical characteristics and enduses, and (like FJF) the product is supplied directly to the end-user and therefore not advertised, the product will not qualify for separate “class or kind” treatment. This will be true whenever, as here, the investigation includes a broad array of specialty products some of which (like FJF) will naturally be supplied directly to the end-user and not advertised. This was essentially acknowledged by the Department in the Decision Memo, at 165, in that the Department characterized FJF as “another lumber product in the broad field of lumber products with distinct characteristics and end uses”, but rejected separate “class or kind” treatment since other softwood lumber products use the same channels of trade and advertising.

Nevertheless the Panel considers that the Department has discretion to adopt such a methodology in its interpretation of the “class or kind” provision of the CVD statute. The Department is entitled to include in “the broad field of [softwood] lumber products” such specialty products as Finger-Jointed Flangestock, and to choose to be guided by Diversified Products factors in “class or kind” determinations. Hence the only question 52 for Panel review is whether substantial evidence on the record supports the Department’s determination that Finger-Jointed Flangestock is not “so distinct and unique” as to warrant separate “class or kind” treatment. Since sufficient record evidence supports the Department’s finding that Finger-Jointed Flangestock shares common features (channels of trade and manner of advertising) with other softwood lumber products, the Panel affirms the Department’s determination that Finger-Jointed Flangestock does not warrant treatment as a separate class or kind of merchandise.

3. Square End Bed Frame Components and Box-Spring Bed Frame Kits

Separate class or kind treatment for Square End Bed Frame Components (hereinafter “SEBF Components”) presents much the same issue as Finger-Jointed Flangestock. SEBF Components are undeniably “softwood lumber products from Canada”, and, whereas the physical characteristics and end-uses of SEBF Components may be “unique”, SEBF Component channels of trade and advertising are merely “distinct” in that these channels also exist for other forms of softwood lumber. For the reasons stated by the Panel in the preceding paragraph with respect to the “class or kind” treatment of Finger-Jointed Flangestock, the Panel considers that substantial evidence on the record supports the Department’s determination that SEBF Components are not entitled to separate “class or kind” treatment.

There is, however, one complicating factor: SEBF Components are “extremely similar” to Radius End Bed Frame Components (hereinafter “REBF Components”), a softwood lumber product that was excluded from the scope of the investigation with the support of the petitioners.

Petitioners specifically identified SEBF Components as within the scope of the petition, but agreed to the exclusion of REBF Components because, whereas SEBF Components “cannot easily be distinguished from other softwood lumber products”, REBF Components “can visibly be identified by the unique ‘radius cuts’ on both ends”. DOC Response Brief at K-42, citing Comment 52, PR Doc. 936. Complainants’ object to the disparate treatment of SEBF Components (included in the investigation) and REBF Components (excluded from the investigation):

Square End Bed Frame Components necessarily must belong to the same “class or kind” of merchandise as excluded Radius End Bed Frame Components. Examination of the Diversified Products criteria reveals that Square End and Radius End Bed Frame Components are exactly the same under each of the Diversified Products factors, except for the obvious fact that the ends of the Radius End components are rounded. The radius cut corners on the ends of some bed frame components but not others hardly suffice to distinguish the two sets of components. … Square End and Radius End Bed Frame Components are not sold or marketed as different articles of commerce, or for different uses, or to different users, or through different channels of trade. [Complainants’ Case Brief at 71.]

The Department agreed that “SEBF Components and REBF Components are the same class or kind of merchandise”, but argues that this “does not mean that SEBF Components should be included in the exclusion from the order” since:

It is within the discretion of the Department to exclude products which would otherwise be covered by the scope of an investigation. … There is no statutory, regulatory or other legal authority which has ever required the Department to broaden an exclusion, because it determined that other products were within the same “class or kind” as that exclusion. [DOC Response Brief at K-43.]

The Panel does not consider that the Department’s exclusion of REBF Components would necessarily require a similar exclusion for SEBF Components.

Complainants also objected to the Department’s treatment of box-spring bed frame kits. As proposed by the petitioners, the Department determined that box-spring frame kits should be excluded from the order, but only if the kits were “individually packaged”. Complainants objected, arguing that “the exclusion should be expanded to include ‘kits imported in bulk form’ because ‘individual bundling’ is neither common in the industry, nor economical.” DOC Response Brief at K-38 (citing record documents). The Customs Service confirmed that the industry-wide practice is “to ship identical bedframe components bundled together, not in kits.” Id. at K-39. But Customs also said “that it would be ‘problematic to administer and enforce’ any exclusion involving kits, whether or not individually bundled, because no HTS classification number recognizes these collections of box-spring frame components.” Ibid. With this evidence on the record, the Department considered that “from an administrative/enforcement perspective any exclusion of box-spring frame components would be difficult, if not impossible, to enforce.” Id. at K-40.

Nevertheless, since “petitioners offered on the record their belief that the scope of the order should narrowly exclude box-spring frame components shipped in individual packages”, the Department excluded individually packaged box-spring bed frame kits from the countervailing duty order. Ibid. In supporting the exclusion of box-spring frame kits, petitioners argued that individual packaging was needed “to prevent manipulation or evasion [of the CVD order] as a result of this exception”. Hence petitioners would only support exclusion of “kits which were bundled in such a way that an ‘individual package or bundle’ would make one bed frame.” Id. at K-38 (citing record documents).

The Department supports the exclusion order by reference to the broad discretion enjoyed by the Department in determining the scope of an investigation. Since the petitioners supported an exclusion limited to kits shipped in individual packages, and gave reasons on the record in support of that limitation, the Panel considers that the Department’s determination to exclude only individually packaged box-spring bed frame kits is supported by substantial evidence on the record.

F. Product Exclusions

In addition to claims for separate “class or kind” status, two types of claims for exclusion of particular products have been submitted to the Panel for review in these proceedings. One is a request by the Canadian Parties to exclude products using lumber that originates in the Maritime Provinces but is reprocessed in other Canadian provinces. The second is a claim by Anderson Wholesale, Inc., to exclude used railroad ties from the scope of the investigation. The Department rejected both claims.

1. Reprocessed Maritimes-Origin Lumber

The Department gave the following reason for rejecting the request to exclude reprocessed Maritimes-origin lumber:

Once the lumber is processed in a non-Maritimes province, it becomes indistinguishable from lumber of other origin and may benefit from subsidies bestowed on the mill. To the extent that a company processing the Maritimes-origin lumber can demonstrate that it does not receive any other subsidies, the company can be excluded from the CVD investigation, and the Maritimes-origin lumber processed in an unsubsidized mill outside the Maritimes provinces is not subject to CVD duties.¹⁴ [Comment 57 at Exclusion 1, PR Doc. 936; DOC Response Brief at K-69.]

Since Petitioners did not allege that softwood lumber originating in the Maritime Provinces is subsidized and should be included in the investigation, softwood lumber produced in the Maritime Provinces and exported directly to the United States is not included in the CVD order. In addition, Petitioners agreed to a Government of Québec request to exclude U.S.-origin lumber that underwent “minor processing” in Québec and was then reexported to the United States. The U.S.-origin lumber exclusion provides:

U.S. Lumber shipped to Canada for minor processing and imported into the United States is excluded from the scope of the CVD investigation if the following conditions are met: 1) the processing occurring in Canada is limited to kiln-drying, planing to create smooth-to-size board, and sanding, and 2) if the importer establishes to Custom’s satisfaction that the lumber is of U.S. origin. [Quoted from DOC Response Brief at K-68 note 15.]

Nevertheless, Petitioners opposed any exclusion for lumber originating in the Maritime Provinces and reprocessed in Québec, stating that it is “virtually impossible to determine whether a shipment of softwood lumber from a particular company consists only of lumber produced from Maritimes-origin wood, unless that company processes Maritimes-origin wood exclusively.” See DOC Response Brief at K-68. The Panel observes, however, that the exclusion of US-origin lumber reprocessed in Canada does not require that the exporting company process U.S.-origin wood “exclusively.”

With apparent reference to the inconsistent treatment of reprocessed Maritimes-origin lumber as compared to reprocessed U.S.-origin lumber, the Department asserts that the law does not require it to “treat merchandise within the scope [of the investigation] in a uniform manner with merchandise outside of the scope.” The Department relies upon its “broad discretion” both to define the scope of the investigation and to “exclude any merchandise from the scope of an order, as long as the exclusion does not interfere with the relief requested in the petition.” DOC Response Brief at K-68-69.

The Panel does not agree that the Department’s rejection of an exclusion for reprocessed Maritimes-origin lumber lies within the Department’s discretion. In a CVD proceeding, as in any statutory proceeding, it is the responsibility of the Department to give effect to the will of Congress. While the intent of Congress is often difficult to discern, the CVD statute is designed to impose countervailing duties only on imported merchandise that benefits from subsidies conferred by the government of the exporting country. See 19 U.S.C. § 1671(a). In determining the scope of the proceeding, the Department may choose to be guided by the intent of the petition, but neither the petition nor the concerns of the Petitioners can give the Department a legal basis to impose countervailing duties where they are not warranted under the statute.

The parties agree that neither U.S.-origin nor Maritimes-origin lumber receives the subsidy at issue in this proceeding. When such lumber is reprocessed by companies that also process subsidized lumber, there may be legitimate concern that the resulting product may be “indistinguishable” from lumber that the reprocessing company produces from subsidized logs. In such circumstances, the burden will fall upon the producer/exporter to show that, in fact, a particular shipment is derived from unsubsidized U.S.-origin or Maritimes-origin lumber. The U.S.-origin lumber exclusion does exactly that. The Canadian Parties stress that Maritime-origin lumber must be accompanied by a certificate from the Maritime Lumber Bureau confirming its origin, which may provide a basis for distinguishing reprocessed Maritimes-origin lumber from subsidized lumber. Nothing in the record shows that U.S. imports of reprocessed Maritimes-origin lumber cannot be handled in a fashion similar to U.S. reimports of reprocessed U.S.-origin lumber. Such a procedure would focus the countervailing duties on the subsidized merchandise and ensure that countervailing duties are not imposed on Maritimes-origin lumber when it does not in fact benefit from the subsidies at issue here. The Panel considers that the Department’s blanket refusal to exclude reprocessed Maritimes-origin lumber from the scope of the order is inconsistent with the intent of the statute and therefore not authorized by law.

Accordingly the Panel remands the matter of excluding reprocessed Maritimes-origin lumber to the Department for reconsideration in light of this opinion.

2. Used Railroad Ties

New softwood lumber railroad ties are excluded, but used railroad ties are included within the scope of the CVD order. The exclusion of new railroad ties results from the Customs classification of this product under a Harmonized Tariff System heading (HTSUS 4406) that was not included in the petition, whereas the subheading that covers used railroad ties (HTSUS 4407.1000) is included in the petition. Furthermore, although the petition did not specifically identify used railroad ties, Petitioners affirmed that their intention was to include all products covered by HTSUS 4407.1000, including used railroad ties. Complainant Anderson Wholesale, Inc. (hereinafter “Anderson”), an importer of used railroad ties, requests that the product be excluded from the scope of the order.

Anderson asserts that the used railroad ties currently imported from Canada were manufactured either in the 1950’s and 1960’s, or even earlier, in the 1930’s and 1940’s, and that it is not known whether such ties consist of softwood or of hardwood lumber, and whether they were manufactured in the provinces subject to the CVD order in this proceeding, or elsewhere, such as in the Maritime Provinces or in the United States. What is known, however, is that imported used railroad ties are no longer fit to be used as railroad ties; they are used as landscape timbers.

Anderson argues that used railroad ties must be excluded because new railroad ties are excluded. However, since Anderson accepts that new and used railroad ties have totally different uses, and are classified by the Customs Service under separate HTSUS headings, the Department is well within its discretion in treating the two as distinct products, excluding one and including the other within the scope of the investigation.¹⁵

In rejecting Anderson’s exclusion request, the Department relied on the Customs classification and the fact that the petition covers used railroad ties, and asserted that “even damaged, used and old softwood lumber products continue to fall under the scope of our investigations.” Comment 52, PR Doc. 936, quoted from DOC Response Brief at K-51. While the Department’s comments may be accurate, they do not address two threshold issues presented by the exclusion request: First, are the used railroad ties at issue softwood lumber products? And, second, do currently imported used railroad ties benefit from the subsidies at issue in this proceeding?

Since Complainant Anderson accepts that its imports of used railroad ties may be either hardwood or softwood lumber, or both, the Department has discretion to treat the imports as softwood lumber unless Complainant can establish otherwise. And, since Complainant offered no record evidence that its used railroad ties are hardwood lumber, Complainant cannot obtain an exclusion on that basis. However, as Petitioners state in their brief, should a specific shipment of used railroad ties be shown to be of hardwood, that shipment would fall outside the scope of the CVD order since the order extends only to imports of softwood lumber products.

On the second question, Anderson presents two distinct claims that current imports of used railroad ties do not benefit from “stumpage” subsidies. Anderson’s first claim is that currently imported railroad ties may have been produced in any one of three jurisdictions: in the United States, in the Maritime Provinces, or in the provinces whose lumber exports are subject to countervailing duties. Since Anderson accepts that its railway tie imports may have been produced in the affected provinces, the Department has discretion to include such imports in the CVD order unless Anderson can establish that a particular shipment had been produced elsewhere. As the Panel held with respect to reprocessed Maritime-origin lumber, where an importer is able to show that its products have not benefited from the subsidies at issue in this proceeding, the Department must exclude those imports from the scope of the CVD order.

Anderson’s second claim relevant to receipt of subsidies is that currently imported railroad ties were produced in the 1950’s or 1960’s, or earlier, and hence should not be subject to the current CVD order even where they were produced in the affected provinces. In response Petitioners argue that Canadian softwood lumber subsidies “date from the beginning of the century.” Since nothing further was said on this issue in the Final Determination, the Department apparently relied on the Petitioners’ response in rejecting Anderson’s request for exclusion. See Final Determination at 166-67 note 484, PR Doc. 936.

Even assuming that Petitioners’ “beginning of the century” reference is historically accurate, there is no basis in the record to attribute to imports of used railroad ties the Department’s findings that currently produced softwood lumber products benefit from the current subsidies at issue here. The railroad ties at issue here were produced forty and fifty years ago under different subsidies practices that have not been identified and have not been investigated by the Department in this proceeding. Accordingly the Panel vacates the Department’s rejection of an exclusion for used railroad ties and remands the matter to the Department for reconsideration in light of this opinion.¹⁶

X. UPSTREAM SUBSIDIES

An upstream subsidy is a subsidy on an input the benefit of which is passed through to a downstream product. The provision regarding upstream subsidies permits the assessment of countervailing duties on the downstream product notwithstanding that there is no subsidy provided directly to that product. Conversely, no countervailing duty is to be assessed on the downstream product if the input is not subsidized.

In its Decision Memo, Commerce states that the statutory provision on upstream subsidies does not apply because “no reasonable grounds” exist to believe or suspect that provincial stumpage confers a competitive benefit on the subject merchandise in the form of an upstream subsidy as defined in the law. The Department’s rationale for this conclusion is, briefly stated, that both timber (the input) and the downstream products (lumber and remanufactured lumber products) are subsidized because the recipient of the subsidy is the producer of the subject goods. Restated, the manufacturers of the subject lumber and the remanufactured products and the recipients of the stumpage subsidies, are in each case one and the same entity. Further, to the extent that there may be producers of subject goods to whom the subsidy is not passed through, because the investigation is being conducted on an aggregate basis, no upstream analysis is required. Rather, any such company is entitled to a subsequent review to determine a company specific rate. In the case of a party to whom no benefit is passed through, the company specific rate would presumably be zero.

According to section 1671(e) of Title 19:

[w]henever the administering authority has reasonable grounds to believe or suspect that an upstream subsidy, as defined in section 1677-1(a)(1) of this title, is being paid or bestowed, the administering authority shall investigate whether an upstream subsidy has in fact been paid or bestowed, and if so, shall include the amount of the upstream subsidy as provided in section 1677-1(a)(3) of this title.

Section 258 of the URAA defines an upstream subsidy as:

… any countervailable subsidy, other than an export subsidy, that (1) is paid or bestowed by an authority . . . with respect to a product (hereafter in this section referred to as an `input product’) that is used in the same country as the authority in the manufacture or production of merchandise which is the subject of a countervailing duty proceeding; (2) in the judgment of the administering authority bestows a competitive benefit on the merchandise; and (3) has a significant effect on the cost of manufacturing or producing the merchandise.

19 U.S.C. § 1677-1(a)(1).

In determining whether a competitive benefit has been bestowed:

the administering authority shall decide that a competitive benefit has been bestowed when the price for the input product referred to in subsection (a)(1) of this section for such use is lower than the price that the manufacturer or producer of merchandise which is the subject of a countervailing duty proceeding would otherwise pay for the product in obtaining it from another seller in an arms -length transaction.

19 U.S.C. § 1677-1(b)(1).

Part 351.523 of the Regulations provides the specific requirements for the initiation of an upstream subsidy investigation. Before investigating the existence of an upstream subsidy, the Department must have a reasonable basis to believe or suspect that the following elements exist: (1) a countervailable subsidy, other than an export subsidy, is provided with respect to an input product; (2) inter alia, the government sets the price of the input product so as to guarantee that the benefit provided with respect to the input product is passed through to producers of the subject merchandise; and, (3) the ad valorem countervailable subsidy rate on the input product is equal to, or greater than, one percent. 19 C.F.R. § 351.523(a)(1). An “input product” is defined in the Regulations as “any product used in the production of the subject merchandise.” 19 C.F.R. § 351.523(b).

The Canadian Parties argue that the provincial stumpage programs involve the harvesting of timber, and result in the production of logs. They assert that logs themselves are not within the scope of the investigation, but instead are inputs to the production of the subject merchandise, i.e., softwood lumber.

The Canadian Parties assert that Section 701 (e) of the Tariff Act is unambiguous. Congress’ intent in passing the upstream subsidy provision was to address issues such as those found in prior softwood lumber cases and to require the Department to make specific factual findings showing that purchasers of inputs receive a countervailable benefit. Canada asserts that where a subsidy is alleged with respect to an input product, the allegation must be analyzed under the separate upstream subsidy provisions of the statute. Where the purported subsidy is granted to upstream input products -- products that are used in the manufacture or production of the merchandise being investigated -- the link between the purported subsidy and the subject merchandise is, at most, indirect. Therefore, the statutory rules require a determination that the benefit to the original recipient of the subsidy has been “passed through” to the downstream producer of the subject merchandise.

Where the producer of the allegedly subsidized input product (the timber harvester) and the producer of the downstream product (the lumber producer or the remanufacturer) are one and the same entity, the Canadian Joint Parties agree that Commerce need not do an upstream analysis as the alleged subsidy is provided directly to a producer of the subject merchandise. However, it is claimed that Commerce, in this case, has unlawfully presumed that the two are the same, contrary to the evidence.

Canada argues that this presumption is in violation of the statute and that the Department has ignored substantial record evidence establishing that many lumber producers purchase their log inputs in arm’s-length transactions and thus receive no benefits directly from provincial stumpage programs. They state that the evidence on the record establishes that independent harvesters harvest timber, and that unrelated sawmills often purchase at arms-length from those independent harvesters. The Canadian Parties point out that nearly one-third of the Crown harvest in several provinces was harvested by harvesters that did not own sawmills. The Canadian Parties claim that Commerce had conceded that some of the harvest goes to loggers not owning sawmills, and that the provinces in their questionnaire responses indicated that these were arms-length transactions. The Canadian Parties state that there was also a request on the record for an upstream subsidies investigation, and that it was Commerce's responsibility to honor that request.

Specifically with respect to remanufacturers, Canada challenges Commerce’s decision not to conduct an upstream analysis to determine whether all remanufacturers received countervailable benefits, because some remanufacturers may be stumpage holders and the Department was conducting its investigation on an aggregate basis. Canada is not satisfied with Commerce’s decision to leave the interests of 61 remanufacturers, who may not have received countervailable benefits, to be examined only as part of a subsequent “review” when the proper relief would be considered.

Finally, the Canadian Parties point out that in Lumber III, which was also an aggregate case, the Department acknowledged that it was appropriate to account for arms-length purchases. The Department’s refusal to do so in Lumber III was based upon the claimed lack of evidence regarding the extent of such purchases, but the principle is undisputed. Because the Department did not do so in this case, the Canadian Parties claim that Commerce’s subsidy calculation is invalid and overstates the amount of benefit received by producers of the subject merchandise.

Canada specifically requests that the Panel remand the case to the Department with instructions to exclude arms-length purchasers from application of the order and to exclude from the numerator of the subsidy calculation the benefit attributed to the volume of logs purchased by sawmills in arms-length transactions (and the volume of lumber purchased by remanufacturers in arms-length transactions). In Canada’s view, a remand to Commerce to conduct an upstream subsidy investigation at this time would not be appropriate nor an effective remedy.

Commerce responds by pointing out that when it investigates a product that is upstream to another, but both products are subject merchandise, both products are investigated without the necessity of a separate upstream subsidy analysis because neither is upstream to subject merchandise. Since the scope of this investigation covers both dimension lumber and remanufactured products, and since both are produced by stumpage holders who are directly receiving and benefiting from the countervailable subsidies, an upstream subsidy investigation is not required.

The Department further states that in an aggregate investigation, an upstream subsidy investigation is not required for the small minority of softwood lumber not directly subsidized. Where it is not “practicable” to determine individual exporter or producer subsidy rates, because of the large number of exporters or producers involved in the investigation, Commerce has the option of limiting the examination to a statistically valid sample of exporters or producers accounting for the largest volume of subject merchandise that can reasonably be examined, or determining a "single country-wide rate." According to Commerce, no upstream subsidy analysis is necessary or appropriate with respect to lumber under investigation, because the “vast majority” of lumber producers receive the stumpage subsidies directly (i.e. through input logs), not indirectly through independent suppliers.

According to the Department, the record demonstrates that the great majority of producers of subject merchandise are tenure holders and thus direct recipients of the benefit provided by the stumpage programs. In British Columbia, for example, 90 percent of the timber harvest originates from Crown lands. More than 83 percent of this timber is harvested by tenure holders that are, by the terms of the tenure agreements, required to own sawmills. Only about nine percent of the Crown timber has been shown to be harvested by entities that do not own sawmills. As to the remaining eight percent, the evidence is inconclusive regarding the requirement of sawmill ownership in the tenure agreements. Thus, somewhere between nine percent and seventeen percent of Crown timber in British Columbia is harvested by entities that do not own sawmills.

In Quebec, the Forest Act explicitly requires that tenures be granted only to persons authorized to operate a wood processing plant. Similar requirements are found in Ontario, Alberta, Saskatchewan, and Manitoba. For these provinces, the Department estimates that about 95% of timber is harvested by entities that own sawmills.

The Department claims that in spite of the Canadian Parties' reliance on the independent logger they fail to identify any such entities or provide information about the portion of stumpage purchases that may be attributed to them. According to the Department, there are very few transactions by independent harvesters. And those transactions that do occur are restricted by requirements that they sell to mills within the province. These restrictions hamper the harvesters' ability to negotiate freely and force them to sell to specified customers. As the British Columbia government noted, independent loggers operate as employees or contractors for holders of private lands or Crown tenures.

Finally, Commerce asserts that when it conducts an investigation on an aggregate basis, no company-specific rates are determined, except to the extent it has found it practicable to consider certain company-specific exclusion requests. For other producers, Commerce states that a review is the appropriate avenue to determine if there are specific companies that do not receive countervailable benefits.

The Coalition supports Commerce’s views, citing Delverde v. United States, 989 F. Supp. 218, 224 (CIT 1997), to argue that, when Commerce investigates a product that is upstream to another but both products are subject merchandise, both products are investigated without the necessity of a separate upstream subsidy allegation, because “neither is upstream to subject merchandise.” Accordingly, since both first-mill lumber and remanufactured lumber are subject merchandise, the Department may properly investigate both without an upstream analysis. Furthermore, for the Coalition, even if the Department had determined that some independent loggers provided some logs to lumber producers that required an upstream analysis, the Canadian Parties had not provided the Department with the data necessary to conduct such an examination.

The questions which the panel must answer, then, are as follows. First, there is the factual issue as to whether all buyers of logs are the same parties who directly benefit from the stumpage programs so that, as a matter of law, the alleged subsidy can be attributed to the logs which are made available to such mills or producers (remanufacturers) who are not direct beneficiaries of the stumpage programs. The second question is whether it is permissible to avoid the analysis on the basis that the investigation is conducted on an aggregate basis.¹⁷

As a preliminary matter, the Panel observes that there is no need for a passthrough analysis when both parties in an arm’s length transaction are producers of subject merchandise. The only situation where a pass-through analysis may be required is where the tenure holder is an entity that does not produce subject merchandise, or, in other words, does not own a sawmill.

There is a factual dispute as to the prevalence of independent loggers in this investigation. While the Canadian Parties assert that independent loggers represent a significant portion of the market, the Department has found that overwhelmingly, tenure holders own lumber processing operations. In spite of these assertions, the Department concedes in its brief that loggers that do not own sawmills harvest at least 9 percent of Crown timber in British Columbia. That number is said by the Department to be 5 percent for Saskatchewan and Manitoba. These numbers represent a significant portion of the Crown timber harvest as to which there is uncertainty as to whether the benefit attributable to stumpage pricing may fairly be included in the benefit calculation.

While the Department thus concedes that some independent loggers exist, it asserts that sales by such loggers are de facto not arms length sales as the independent loggers are required to sell logs to sawmills pursuant to certain regulatory and contractual provisions. The effect of these restrictions is to force independent loggers to pass the benefit they received by way of their tenures to the purchasing sawmills. The Department describes several mechanisms that achieve this result. In British Columbia, for example, the Department acknowledges that between 9 and 17 percent of Crown timber is harvested by entities that do not own sawmills. Yet the Department found that the British Columbia Forest Act requires that all timber from B.C. lands be processed in British Columbia. In addition, the independent loggers in question are small business operators. These circumstances, taken as a whole, indicate that sales of logs from independent loggers to sawmills that produce subject merchandise are not arms length transactions between independent parties. The Department provided a similar analysis for Quebec, Ontario, and Alberta. On the other hand, the Department conceded that in Saskatchewan and Manitoba there are no similar legal requirements to force harvesters to sell to mills within the province. The Department's decision not to conduct a pass-through analysis for these two provinces was based solely on the fact that about 95% of Crown timber in both provinces was harvested by entities that also hold provincial licenses to operate sawmills.

Factual determinations by the Department must be affirmed if they are supported by substantial evidence on the record. The Panel does not find that, with respect to British Columbia, Quebec, Ontario, and Alberta, there is no substantial evidence to support the Department's determination that sales to sawmills by independent loggers in these four provinces are de facto not arms length transactions. The record evidence sustains the Department's finding that the legal requirements imposed upon the sale of logs by entities that do not own sawmills operate to limit their ability to set the price. The Canadian Parties have not challenged this finding. For these reasons, the Department’s conclusion that producers of subject merchandise in the four provinces who buy logs from independent loggers have in fact received the stumpage benefit appears not to be inconsistent with the law or to be unreasonable. The Panel affirms the Department's decision not to conduct a pass-through analysis with respect to such transactions in these four provinces.

With respect to Saskatchewan and Manitoba, the evidence in the record is quite different with respect to the Department’s decision not to conduct a pass-through analysis. The Department has not addressed the fact, which it accepts, that in both Saskatchewan and Manitoba about 5% of Crown timber is harvested by independent loggers. The Department also notes in its brief that independent loggers in these two provinces are not hampered by requirements regarding domestic processing.

With regard to Saskatchewan, the Department found that about 86% of the harvest was attributable to four major Forest Management Areas tenure holders, all of which own sawmills. See Decision Memo at 137. The remaining 14% of the Saskatchewan harvest was conducted pursuant to Forest Products Permits (FFP). Id. However, the Department has not determined whether FFP holders are required to or in fact own sawmill operations. Ibid.

As to Manitoba, the Department analyzed the various types of license agreements required to harvest Manitoba timber. Id. at 128. As discussed in the Department's brief, the license agreements do not require domestic processing. See ibid.

In short, the evidence on the record does not warrant an assumption that independent harvesters are limited in their ability to set the sale price of their logs and thus are not capable of retaining the stumpage benefit in sales to sawmills. Thus the record lacks a reasoned basis for the Department's decision not to conduct a pass-through analysis for independent logger transactions in Saskatchewan and Manitoba. Nevertheless we are mindful that the parties bear the burden of developing the record in a countervailing duty proceeding. See Mannesmannrohren-Werke AG, 120 F. Supp. 2d 1075, 1087 (CIT 2000). As set out in the Department’s regulations, “The Department obtains most of its factual information in antidumping and countervailing duty proceedings from submissions made by interested parties during the course of the proceeding.” 19 C.F.R. § 351.301(a). Since the record lacks evidence of the extent of arms length transactions in sales by independent loggers to sawmills in Saskatchewan and Manitoba, the Panel affirms the determination of the Department not to undertake a pass-through analysis of such transactions.

XI. COMPANY EXCLUSIONS AND COMPANY-SPECIFIC DUTY RATES

A. Company Exclusions

The Department’s regulations allow it to exclude from an aggregate case an exporter or producer for which it determines an “individual net countervailable subsidy rate of zero or de minimis.” 19 C.F.R. § 351.204(e)(1). The Department uses this mechanism to exclude from a countervailing duty order parties that receive no subsidy or receive a subsidy that is de minimis.

In the Notice of Initiation, Commerce stated that “it anticipates creating a system that will allow us to expeditiously process and rule on these exclusion requests without compromising the integrity of the CVD law, while, at the same time, ensuring fairness and transparency in the treatment of the exclusion requests.” 66 Fed. Reg. at 21335. Commerce explicitly sought “the cooperation of the Government of Canada and the provincial governments in implementing any such system.” Id.

In response, the Government of Canada filed a proposal in May 2001 setting forth a process by which Commerce could review company exclusions. The Government of Canada agreed to collect and file exclusion requests and certifications, to categorize the applicants and to provide company-specific data. The Department took no action on the proposal, but in August 2001, it issued a decision memorandum that established a system to process exclusion requests. In the memo, the Department accepted the Canadians’ proposal, with certain clarifications. The Department also stated that it would consider applications for exclusion prior to the issuance of the preliminary determination if filed within eight days. Despite the short deadlines, the Canadian Parties filed 95 applications for exclusion.

In the Preliminary Determination, Commerce stated that, with one exception, it was not granting any of the exclusion requests because certain information had not been provided. The Department also stated that it “intend[s] to continue working closely with both individual requesters and the [Government of Canada] on this issue as the investigation proceeds.” 66 Fed. Reg. 43188.

The Canadian Parties claim they then expended significant resources to revise the company exclusions process to satisfy Commerce. For example, a specific office was established by the Government of Canada to oversee the exclusion request process and an independent consulting firm was retained to ensure quality control. Significant time was spent gathering the required information and categorizing the applicants into groups. Ultimately, more than 350 requests for exclusion were filed.

Despite receiving cooperation from the Canadian and provincial governments, and despite their spending considerable time developing and implementing a system to submit requests for company exclusions, Commerce declined to consider the requests submitted on the grounds of impracticability.

In the Final Determination, the Department stated that it found it practicable to consider only 30 of the more than 350 company requests for exclusion it had received. In its Decision Memo, Commerce explained that its methodology for examining exclusion requests was based on an “input source” criterion, because this was simple, factual and easily verifiable. Using an “input source” criterion, Commerce considered whether the subject logs were sourced from the Maritime provinces, from private lands or sources other than Crown land. Company exclusion applications that relied on other bases for exclusion, such as arm’s length transactions, were not considered or granted. This, Commerce explained, was because such transactional bases for exclusion required extensive investigation and analysis, which would have been too difficult to do for each company within the time limits of the Final Determination. It would also be beyond the scope of the limited exercise envisioned under the regulations for consideration of company exclusions in an aggregate case.

Section 777A(e)(2)(b) of the Act gives the Department discretion to “determine a single country-wide subsidy rate” when “it is not practicable to determine individual countervailable subsidy rates . . . because of the large number of exporters or producers involved.” 19 U.S.C. § 1677f-1(e)(2).

Commerce’s regulations state that it “will exclude from an affirmative final determination under section 705(a) . . . any exporter or producer for which the Secretary determines an . . . individual net countervailable subsidy rate of zero or de minimis.” 19 C.F.R. § 351.204(e)(4).

The Department’s regulation vests the agency with the discretion to “consider and investigate” exclusion requests “to the extent practicable.” 19 C.F.R. § 351.204(e).

The authority to grant exclusions from an aggregate subsidy rate is conditioned by specific requirements aimed at ensuring that the basis upon which the request is granted is reliable or verified. Pursuant to 19 C.F.R. § 351.204(e)(4), an exporter or producer that desires exclusion from an order must submit the following information:

(i) A certification by the exporter or producer that it received zero or de minimis net countervailable subsidies during the period of investigation;

(ii) If the exporter or producer received a countervailable subsidy, calculations demonstrating that the amount of net countervailable subsidies received was de minimis during the period of investigation;

(iii) If the exporter is not the producer of the subject merchandise, certifications from the suppliers and producers of the subject merchandise that those persons received zero or de minimis net countervailable subsidies during the period of investigation; and,

(iv) A certification from the government of the affected country that the government did not provide the exporter (or the exporter’s supplier) or producer with more than de minimis net countervailable subsidies during the period of investigation.

The Canadian Parties argue that Commerce’s decision to consider only 30 applicants is an abuse of discretion. Although the Canadian Parties agree that the regulations give the Department some discretion on exclusions, they assert the discretion is limited. Canada claims that it is an abuse of discretion to commit to consider applications for exclusion and then claim impracticability.

According to the Canadian Parties, the statute required Commerce to conduct an upstream subsidy investigation to determine whether any alleged subsidy was passed through to downstream purchasers. (See the previous section on upstream subsidies for a discussion of the issue.) They claim that if the Department had not refused to do so, two-thirds of the exclusion applicants who applied as arm's length purchasers of logs or lumber would have been found unsubsidized in the upstream subsidies investigation and would have had no reason to apply for exclusions. Canada asserts that Commerce cannot lawfully claim that the number of applicants made its task impracticable, when the number of exclusion applicants was a direct result of its own failure to abide by the statute on upstream subsidies.

According to Canada, Commerce's claim that it was too complex to determine affiliation, or to determine whether transactions were at arm's length, show that its line drawing was not reasonable. There were groups of applicants that had no affiliation with the tenure holder, and the Department’s own “Exclusions Memo” stated that transactions between unaffiliated parties are deemed to be at arm's length. Each of the company applicants -- at least those whose applications were submitted by the Government of Canada and many others -- was advised on the definition of affiliation and certified to its affiliates or their absence. The Canadian Parties assert that Commerce did not adhere to its own bright line rule of excluding only companies basing their applications on source, and had abused its discretion in the exclusion process.

Canada points out that in Lumber III, Commerce had recognized that remanufacturers, who hold no crown stumpage rights, are not affiliated with crown stumpage holders, and purchase their inputs at arm's length, should be excluded from a countervailing duty order, going well beyond the regulatory requirements for exclusion. The Department’s procedures for accomplishing this are either through company exclusion requests or through the investigation and promulgation of company-specific rates.

For its part, Commerce argues that it has ample discretion to decide whether to consider requests for company exclusion. The Department states that the statute and regulation make clear that it may consider requests only to the extent practicable, and that it is not required to grant all requests for exclusion. Commerce argues that it was within its discretion to decide the basis for what is practicable, and that it was reasonable to conclude that considering over 350 company exclusion requests would not be practicable. The Department considers that, in administering the law, it has discretion in the choice of methodology as long as the chosen methodology is reasonable and its conclusions are supported by substantial evidence on the record.

Commerce also states that it had not received all of the information required under 19 C.F.R. § 351.204(e)(4)(i)-(iv) at the time of its decision.

According to Commerce, conclusions reached by the Panel in Lumber III remain valid, specifically where the Panel:

upheld the Department’s decision not to exclude 334 companies because the investigation would be “impracticable”;

accepted that when country-wide aggregate rates are assessed, a CVD is imposed on the merchandise equal to the amount of the country’s net subsidy, not the amount of the subsidy in fact received by the company; and

recognized that Commerce was not legally required to grant company exclusion requests in an aggregate case.

In sum, Commerce asserts that, in investigations conducted on an aggregate basis, the exclusion of companies is totally within its discretion. Its decision was consistent with Lumber III, was supported by substantial evidence on the record and was in accordance with the law.

The Coalition agrees with Commerce, arguing that neither the statute nor the regulations require the Department to consider all exclusions in an aggregate case. At most, the law gives the agency discretion to determine whether it is practicable to consider and investigate, based on, inter alia, the number of exclusion requests, the complexity of the issues that must be considered and the overall complexity of the case. Company exclusion investigations, it is argued, are of secondary importance to the statutorily mandated issues, such as completing the aggregate investigation itself. The Department properly balanced the mandatory tests against the discretionary tests, and made an appropriate decision in light of its available resources.

The Panel considers the arguments raised by the parties in accordance with the standard of review set forth above, namely whether Commerce’s decision is reasonable, supported by substantial evidence and is otherwise in accordance with the law.

The Panel first notes that the language of Section 777A(e)(2) of the Act and the applicable regulations clearly provides the administering authority with considerable discretion to decide whether to consider requests for company exclusion.

The Panel is mindful that the panel in Lumber III concluded that the “bright line” criterion the Department had chosen in that case to limit its investigation was reasonable, and that a Panel may not substitute its own bright line nor remand the decision, unless the decision is an “abuse of discretion” or a decision that “no reasonable agency could have made in the circumstances.” The Lumber III Panel had also concluded that it was proper for Commerce to determine a country-wide aggregate rate and apply it to all companies, even those claiming exclusion based on zero or de minimus subsidies.

Nonetheless, the Panel finds that Commerce failed to properly apply its own “input source” criterion by failing to grant applications submitted by all companies that relied on the source of their lumber as the basis for exclusion. These companies included those that salvage used railroad ties, barnboard or other “old wood” from demolished buildings, riverbeds, etc. These applicants submitted letters indicating that they did not hold Crown stumpage and did not benefit from any subsidy. The timber harvesting happened decades ago and was clearly not done during the POI. These applicants could not have benefited from the alleged subsidy and presented the source of their inputs as the basis of the application. As Commerce failed to provide a rational basis for not considering these exclusion applications, especially in the face of its own policy of considering applications based on “input source,” the Department’s determination in this regard is unsupported by substantial evidence and is contrary to law. The Panel remands this issue to Commerce for consideration of those additional companies whose applications were based on input source.

The Panel is troubled that Commerce gave assurances to the Canadian Parties that their exclusion requests would be seriously considered and sought their cooperation in establishing a system of review, when it should have fully anticipated the number of requests it would receive given the number submitted in past lumber cases. Given that in Lumber III Commerce received 334 requests for exclusion, the agency should have allocated the proper amount of resources to adequately consider a comparable number in this case after it announced its intention to establish a system to process the requests received. Moreover, it was the Department’s own delay that, at least in part, created the impracticability.

The Panel believes that by conducting the case on an aggregate basis, as an exception to the general rule in favor of establishing individual company CVD rates, the bar has been raised for Commerce to consider applications for exclusion fairly and to allocate the resources necessary to do so. This is particularly relevant in instances such as this where the Department has declined to conduct an upstream subsidy analysis. The Panel is disappointed that Commerce failed to do so.

Nonetheless, with the exception of the one category of applicants noted above whose applications were based on input source, the Panel is of the opinion that the statute and the regulatory grant of discretion is sufficiently broad so as to permit Commerce to reasonably conclude that consideration of the more than 350 applications for company exclusion would be impracticable. Thus, the Panel concludes that Commerce’s decision is supported by substantial evidence and is otherwise in accordance with the law.

1. Goodfellow

Goodfellow Inc. (“Goodfellow”), a Canadian softwood manufacturer, submitted a brief contesting Commerce’s failure to consider its request for a company exclusion on the grounds that it had not received countervailable subsidies. The company explains that it purchased lumber for the production of remanufactured products from unaffiliated suppliers in ‘arm’s length transactions.’ Goodfellow also claims that lumber purchased for same-condition re-sales was purchased in ‘arm’s length transactions’ and thus any benefit received from such sales was de minimis. Goodfellow maintains that Commerce’s failure to consider exclusion requests is particularly egregious in light of the agency’s failure to conduct an upstream subsidy analysis.

For the reasons set forth above in relation to arguments raised by the Canadian Parties, the Panel determines that Commerce was not obligated to consider or grant Goodfellow’s request for a company exclusion and that the agency’s decision in this regard was supported by substantial evidence and was otherwise in accordance with the law. Nevertheless, the Panel’s disappointment as expressed above is particularly relevant in this case.

2. MLTC/NorSask

The Meadow Lake Tribal Council (“MLTC”) and NorSask Forest Products Inc. (“NorSask”) submitted a brief in support of Canada's claim that provincial stumpage did not constitute countervailable subsidies. MLTC/NorSask specifically argue that the First Nations of Northern Saskatchewan (“First Nations”) cannot be said to benefit from a countervailable subsidy, because aboriginal rights over their traditional lands and their treaty rights are "sui generis" and constitute a "unique market characteristic," relying on certain Canadian caselaw which they claim supports such principles.

Because the unique characteristics of the First Nations had not been considered by Commerce, MLTC and NorSask ask the Panel to remand to Commerce with instructions to: a) provide a reasoned analysis as to the unique characteristics of the aboriginal rights over their traditional lands; b) determine that exports of softwood lumber into the United States from MLTC/NorSask are exempt from CVDs; and c) determine that exports of softwood lumber from First Nations' forestry operations constitute a special category of exclusion from the imposition of CVDs.

Commerce argues that MLTC and NorSask fail to show how, under U.S. trade law, the existence of aboriginal rights in their lands constitutes a special characteristic that would negate the existence of a benefit. MLTC/NorSask argued in their exclusion request that, because of their aboriginal proprietary rights in the land, benefits received do not amount to a countervailable subsidy. Commerce maintains that this is not a proper use of company exclusion requests, which are a vehicle for exporters or producers to demonstrate that they fulfil the minimum certification requirements for potential exclusion from a CVD order.

The Department further argues that it does not have the authority to consider the status of aboriginal rights on Canadian lands or matters of self-government arrangements of the First Nations, which are matters of Canadian law. These issues are not relevant to its determination or to its consideration of MLTC/NorSask's exclusion request. If MLTC/NorSask's request for exclusion had been rejected, it was because it had neither provided supporting calculations nor the government certification required under regulations at 19 C.F.R. § 351.204(e)(4).

Finally, the Department claims it had not received the information it requires for providing an exemption when it conducts a CVD investigation on an aggregate countrywide basis. In particular, there must be a certification that an exporter or producer received zero or de minimus countervailable subsidies. If de minimis subsidies are alleged, supporting calculations must be provided and there must be a government certification accompanying the exclusion request.

For the reasons set forth above, the Panel finds that Commerce had the discretion not to consider MLTC/NorSask’s exclusion request and that the agency’s decision was supported by substantial evidence and was otherwise in accordance with the law.

B. Company-Specific Duty Rates

Two companies, Canfor Corporation (“Canfor”) and Tembec Inc. (“Tembec”) requested a company-specific countervailing duty rate.

Canfor claims to be the largest producer in Canada and exporter to the United States of softwood lumber covered by this investigation. On May 30, 2001, Canfor filed a letter with Commerce requesting that a company questionnaire be prepared and submitted to Canfor. Commerce did not respond to the request. In the Final Determination, Commerce stated that the statute and regulations generally do not provide for voluntary respondents in aggregate cases.

Tembec filed a request for a company-specific duty rate with Commerce on December 21, 2001 and re-submitted its request on February 21, 2002. The December 21 submission contained what Tembec claimed to be the information required for the calculation of an individual countervailable subsidy rate. The Department stated in its Decision Memo that the company had not filed its request for a company-specific determination within the deadline set in section 782 (a)(1)(B) of the Tariff Act. Commerce also stated that Tembec had not provided the necessary certifications required under section 351.204 (e)(4) of the Act.

Section 777A(e)(2) of the Tariff Act (19 U.S.C. § 1677f-1(e)(2)) provides that, where it is not practicable to determine individual countervailable subsidy rates under section 777A(e)(1) of the Act because of the large number of exporters or producers involved in the investigation or review, the administering authority may:

(A) determine individual countervailable subsidy rates for a reasonable number of exporters or producers by limiting its examination to (i) a sample of exporters or producers . . ., or (ii) exporters and producers accounting for the largest volume of the subject merchandise from the exporting country . . .; or,

(B) determine a single country-wide subsidy rate to be applied to all exporters and producers.

This is an exception to the general rule established in section 777A(e)(1) of the Tariff Act (19 U.S.C. § 1677f-1(e)(1)), which provides that "the administering authority shall determine an individual countervailable subsidy rate for each known exporter or producer of the subject merchandise.”

These provisions were introduced by the URAA of 1994. The SAA of the URAA explains that:

[s]ection 265 of the implementing bill repeals section 706 (a)(2). It eliminates the presumption in favor of a single country-wide CVD rate and amends section 777A of the Act to establish a general rule in favor of individual CVD rates for each exporter individually investigated. In addition, instead of examining a limited number of individual exporters and producers, section 777 A (e)(2)(B) would permit Commerce to calculate, on the basis of aggregate data, a single country-wide subsidy rate to be applied to all exporters and producers of the subject merchandise.

Pursuant to 19 U.S.C. § 1677m(a), in any investigation or a review in which the administering authority has limited the number of exporters or producers examined, or determined a single-country-wide rate, the administering authority is to establish an individual countervailable subsidy rate for the exporters or producers selected for examination, if:

such information is so submitted by the date specified: (A) for exporters and producers that were initially selected for examination; or, (B) for the foreign government, in a CVD case where the administering authority has determined a single country-wide rate; and,

the number of exporters or producers who have submitted such information is not so large that individual examination of such exporters or producers would be unduly burdensome and inhibit the timely completion of the investigation.

Canfor argues that the only reasonable interpretation of the countervailing duty statute is that it compels the agency to accept voluntary responses in aggregate cases unless the number of exporters or producers who submit such information is so large that individual examination of them would be unduly burdensome and inhibit the timely completion of the investigation. Canfor claims that the number of respondents is not excessive and that Commerce acted unreasonably in not responding to its request for a questionnaire.

Tembec claims that it provided Commerce with the information necessary to calculate a company-specific rate three months before the final determination was due, and two months before Commerce undertook examination of some 30 companies requesting exclusion. Tembec asserts that it was the only company to voluntarily submit information to calculate a company-specific CVD rate and that the number of exporters seeking individual rates was not so large so as to be unduly burdensome. Tembec requests that the case be remanded to Commerce with instructions to calculate a company-specific rate for Tembec based on the evidence of record.

The Canadian Parties add that, in relation to Tembec and other producers or exporters who may request a company-specific rate, Commerce had an obligation under 19 U.S.C. § 1677m(a)(2) to provide company-specific rates when the number of exporters or producers concerned is not too large. The Canadian Parties agree with Tembec that its request was timely and that the Panel should provide relief.

Commerce asserts that in a country-wide aggregate investigation, calculation of individual rates is not an option. Commerce reasons that, in an aggregate country-wide countervailing duty investigation, all exporters or producers are assigned the same countervailing duty rate, even if some, had they been investigated individually, would have received higher or lower rates. The very nature of the concept of aggregate rate investigations inheres in not determining individual rates because of the extraordinary burden of doing so for a large number of respondents. The regulation discussing voluntary respondents (19 C.F.R. § 351.204(d)) references other situations where voluntary responses are applicable, but does not mention its application to investigations conducted on an aggregate basis.

The Panel is mindful of the standard of review applicable to this case and states again that an agency’s statutory interpretation is to be upheld if it is “sufficiently reasonable,” even if it is not “the only reasonable construction or the one the court would adopt had the question initially arisen in a judicial proceeding.” American Lamb, supra, 785 F.2d at 1001 (citing Chevron). The Panel is of the view that Commerce’s interpretation of the statute and its own regulation is not unreasonable and that it did not have a statutory or regulatory obligation to consider Canfor and Tembec’s request for company-specific duty rates. Thus, the Panel finds that Commerce’s decision to decline to consider Canfor and Tembec’s request for a company-specific rate to be supported by substantial evidence and to otherwise be in accordance with the law.

XII. CALCULATIONS

Claim I: The Department failed to determine the exact amount of the benefit that is attributable to softwood lumber products

The Canadian Parties argue that the Department did not determine the exact portion of the benefit that can properly be attributed to softwood lumber. Apart from softwood lumber, processing of logs obtained through stumpage programs yields co-products such as chips and sawdust that have commercial value. In their brief, the 75 Canadian Parties cite numbers in the range from 31-56 as the percentage of the wood volume used to produce softwood lumber. They contend that by failing to apportion the appropriate amount of the benefit to co-products and subtract that amount from the amount apportioned to softwood lumber products, the Department inflated the benefit received and assessed a duty higher than the subsidy received. At the hearing, counsel for the Canadian Parties cited Kajaria Iron Castings Ltd v. United States, 156 F.3d 1163 (Fed. Cir. 1998), as authority that requires the Department to net out the benefit of the subsidy provided that the necessary data was available. According to Canadian counsel, the necessary data was available.

The Canadian Parties argue that the Department’s calculation methodology was in violation of §701(a) of the Tariff Act and Article 19.4 of the SCM Agreement, both of which require that countervailing duties shall not exceed the value of the subsidy. They claim that as a consequence of this method of calculation, the CVD rate was increased by 8 percentage points.

Commerce points out that the statute does not provide specific guidance on the method of calculating subsidy rates. Consequently, Commerce claims it has broad discretion to devise a methodology consistent with the statute. The Department relies on its regulations, specifically 19 C.F.R. § 351.525(a), which provides that in calculating the ad valorem subsidy rate, the total benefit shall be divided by the total sales to which the benefit applies.

In this case, the Department argues that the subsidy provided is used to produce both subject and non-subject merchandise. The regulations address this issue by allocating the benefit to all products to which the subsidy applies. The Canadian Parties claim that the Department should have determined what portion of the logs are used to produce subject merchandise and calculated the benefit based on that volume is contrary to the Department's regulations. Commerce points out that sawmills must process the entire log in order to produce lumber. Any unfairness that might result by calculating the benefit based on the total volume of logs that entered sawmills is eradicated by allocating the total benefit not solely to sales of subject merchandise but to sales of all products that originate from processing of logs.

In examining this question, the Panel must assess whether the Department’s method of determining the numerator of the subsidy rate fraction is consistent with the governing statute. More precisely, the issue is whether the Department should have calculated the amount of the benefit by multiplying the price differential by the total volume of logs that entered sawmills or just the volume of wood fiber that eventually ended up as subject merchandise.

The calculation of the subsidy rate is governed by the statute and the regulations promulgated by Commerce. The Panel agrees with Commerce that the statute does not provide detailed guidance in this matter but supplies a general principle that the CVD rate should not exceed the benefit provided. See 19 U.S.C. § 1671 (providing that "there shall be imposed . . . a countervailing duty, in addition to any other duty imposed, equal to the amount of the net countervailable subsidy) (emphasis added). The Panel thus turns to the regulations and must determine whether the Department properly applied its regulations in this matter. The Panel notes that the Department is entitled to deference in interpreting its own regulations unless such an interpretation is clearly erroneous. Thomas Jefferson, supra, 512 U.S. at 512. The Panel concludes that Commerce properly adhered to the regulations and affirms.

The Department relies on 19 C.F.R. § 351.525(b) for the principle that a subsidy will be attributed to the products that it would most likely benefit from it. The nature of the subsidy will in most cases determine to which products it will be attributed. An export subsidy by its nature benefits the producer only when it exports its products and thus will not be attributed to the producer’s domestic sales. See 19 C.F.R. § 351.525(b)(2). Similarly, subsidies that are conditional on the sale of products on a certain market will be attributed only to products sold on that market and not to products sold on other markets. See 19 C.F.R. § 351.525(b)(4). Commerce points out that the provision of timber by the Government of Canada and the provincial governments through stumpage programs is not tied to any specific product. Thus, any subsidy that might arise from the operation of these programs must be attributed to all products that result from the processing of logs that enter sawmills.

The Canadian Parties do not contest this principle but argue that it is possible to determine the benefit that is attributable solely to the subject merchandise. In their view, once this apportioning of the benefit is carried out, there would be no need to attribute the entire benefit to all products that result from the processing of logs, but instead to attribute to the subject merchandise only that portion of the benefit that is properly attributable to the subject merchandise. According to the Canadian Parties, this would result in a more accurate subsidy rate.

As Commerce points out, sawmills purchase logs for the sole reason of processing them into lumber. The fact that discarded material in the form of chips and sawdust has some commercial value does not negate the fact that the value of the log is determined by its conduciveness to be processed into lumber. Thus, the price of a log is driven by the ultimate volume of lumber that will be derived from it. If genetic engineering could produce a log of such shape that nearly 100% of its volume could be processed into lumber, that log would surely be much more valuable than a conventional log of the same total volume that can yield only 40-50% of its volume in lumber.

To illustrate by way of an example with hypothetical values, when a sawmill pays $10 for a log that is 10 cubic meters in volume and ends up with 4 cubic meters in lumber by processing the log, the $10 initial sales price is determined to a large extent by the value of 4 cubic meters of lumber and only to a very small extent by the value of 6 cubic 77 meters of chips or sawdust. The Canadian Parties’ approach would have us believe that the sawmill paid $4 for the portion of the log that it used to make the lumber. This would be true only if the 6 cubic meters of chips or sawdust were equal in value to 6 cubic meters of lumber. Even the Canadian Parties concede that this is not the case.

Assume that lumber is valued at $5 per cubic meter and that by-products sell at 0.50 cents per cubic meter. The sawmill would thus earn $20 from the sale of lumber and $3 from the sale of by-products. To tailor this analogy to the factual scenario in this investigation, assume that the same log is sold in the United States for $15 and that the sawmill thus receives a $5 benefit. Under the Department’s method, the subsidy rate would equal 5/23 or 21.7%. The Canadian Parties’ proposed calculation would require that the benefit be discounted at 60% as that represents the volume of wood fiber that became by-products. The subsidy rate fraction thus becomes 2/20 or 10% -- a marked difference. The Canadian Parties’ methodology would result in a government subsidy to by-products of 100% ($3 benefit divided by $3 in sales) -- a commercially improbable result.

The Canadian Parties’ argument rests on the premise that every cubic centimeter of wood in a log processed in a sawmill is of equal value. As counsel for the Department pointed out at oral argument, the subsidy rate is determined by value and not volume. Even if it were possible to separate the volumes that went into the production of lumber and by-products, the correct amount of benefit that was received for the production of subject merchandise can never be determined by reliance on volume alone. The relative values of such volumes would have to be factored into the calculation in order to achieve a proper result. Keeping with the above hypothetical figures, the total volume of wood fiber that is used to produce lumber is worth 10 times more than the volume used to derive by-products. In other words, the 60% of the volume that ended up as by-products accounted for only 13% ($3 is 13% of $23) of the sales value. If any discounting were to be done, it would have to reflect the ratio between the sales values of lumber and byproducts. Under this method, the subsidy rate for the lumber would equal 4.35/20 ($5 benefit discounted at 13% equals 4.35) or 21.7% -- the same rate achieved by the Department’s calculations.

The Panel notes that the determination of the value of the benefit cannot be subjected to one principle but will depend on the facts of each individual case, the nature of the benefit being provided, and the manner in which the benefit is being provided to the recipient. In the case relied upon by the Canadian Parties, Kajaria Iron Castings, a complex tax subsidy case, two classes of merchandise were involved: (1) certain castings (the “subject merchandise”) which received tax rebates termed “CCS rebates”, and (2) certain non-subject castings which also received tax rebates termed “IPRS rebates”. Two types of “subsidies” were claimed: (a) an over-rebate of both types of tax rebates, and (b) the government’s failure to tax the subsidy thus received. Commerce countervailed both the value of the over-rebates and the value of the tax that should have been levied on the over-rebate amounts, and did so both with respect to the “CCS rebates” on the subject 78 merchandise and to the “IPRS rebates” on the non-subject merchandise. Commerce included the “IPRS rebates” as countervailable benefits on the theory that these rebates were an “untied” subsidy that “was not particular to only specific merchandise (either subject or non-subject merchandise).” 156 F.3d at 1176 (quoting Commerce Remand Results). The Federal Circuit held the latter ruling improper: “The entire benefit received [in respect of the IPRS rebates], i.e., the subsidy and the nontaxation of the subsidy, was related solely to non-subject castings.” Material to our inquiry here, the Federal Circuit ruled: . . .

[W]hen the party under investigation provides documentation that allows Commerce to separate the portion of [the subsidy] related to non-subject merchandise from the remainder of a countervailable [subsidy], Commerce should not countervail the portion of [the subsidy] tied to non-subject merchandise. [156 F.3d at 1176.]

The Court also held it improper for Commerce to countervail both the subsidy and the nontaxation of the subsidy as benefits because, in so doing, Commerce in effect double-counted the over-rebates: “Commerce imposed a countervailing duty that was not ‘equal to the amount of the net subsidy’ in contravention of 19 U.S.C. §1671(a).” 156 F.3d at 1175.

Thus, in Kajaria Iron Castings, the apportioning of the benefit to subject merchandise did not present legal or administrative difficulties, and the Federal Circuit correctly required the Department to parcel out the exact amount attributable to the subject merchandise.

The Panel concludes that Commerce’s methodology of accounting for the sales of by-products in the denominator of the subsidy rate calculation adequately implements the governing statute and the regulations. Accordingly, the Panel does not find to be unreasonable the manner in which Commerce determined the numerator of the subsidy rate calculations.

Claim II: The Department failed to conduct a pass-through analysis

The Canadian Parties argue that the Department was required to conduct a pass-through analysis as the record evidence indicated that numerous sawmills purchase logs from independent loggers through arm’s length transactions. Since the stumpage benefit was calculated based on the volume of all Crown logs entering sawmills, it must be determined that all of the logs were provided to the sawmills at the subsidized price. The Canadian parties request that the Panel determine whether the statute and the regulations require the Department to conduct a pass-through analysis so as to determine the total benefit used as the numerator in the subsidy rate calculation.

As discussed above in the Upstream Subsidies section, the Panel affirms the Department’s determination not to conduct an upstream subsidy analysis. Thus, the Commerce’s calculation in this respect will not be disturbed. Claim

III: The Department failed to include sales of residual products in the denominator of the subsidy rate calculation

The Canadian Parties claim that the Department should have included the sales of residual products from logs, such as poles, posts, and railway ties, in the denominator of the subsidy rate calculation. “Residual products” are defined as products resulting from production processes other than the lumber production process. Instead, the Department included only sales of softwood lumber products and co-products, i.e., products that result from the lumber production process.

The Canadian Parties rely on the Department's regulation, 19 C.F.R. § 351.525(b)(3), which requires the Department to allocate the benefit of a subsidy to all the products of a firm that produced subject merchandise. The Canadian Parties assert that sawmills that receive the benefit of stumpage programs commonly produce residual products as well as softwood lumber products. Since the Department failed to confine the numerator of the subsidy calculation to the exact amount of benefit properly attributable to softwood lumber products and co-products, it should have included in the denominator all sales of all products that are made from logs that are included in the numerator. According to the Canadian Parties, sales of residual products represent about 12% of all softwood shipments, amounting to a total value of C$1,529,051,100. They regard the Department’s exclusion of sales of residual products as an application of the adverse facts available doctrine, which is properly triggered only when a party fails to cooperate. The Canadian Parties assert that there was no allegation that Canada failed to cooperate to the best of its ability, and no corroborating evidence to show that sales of residual products were actually $0.00.

Commerce argues that merchandise that is to be included in the denominator of the subsidy rate calculation must have resulted from the processing of subsidized logs. According to Commerce, the Canadian Parties failed to show that residual products were products of timber entering sawmills.

The Department relied on information from Statistics Canada ("StatCan") which divides lumber shipments into four categories: (1) softwood lumber; (2) hardwood lumber; (3) co-products; and (4) residual products. According to StatCan, residual products are those that do not fit in the other three categories. The Department examined the products in this latter category and concluded that it included products, such as logs, pulpwood that was harvested by sawmills and sold to other manufacturers, and particle and waferboard, that were not a result of lumber processing and were never processed by sawmills. Even if the Statistics Canada residual products category contained some products that resulted from lumber processing, the Department claims it had no way of segregating such products due to the failure of the Canadian Parties to provide the necessary information to enable the Department to do so. Commerce claims that the burden of production is on the party with knowledge and control of the information. According to the Department, the exclusion of residual products from the denominator was therefore appropriate under the law.

At the outset, the Panel notes inconsistencies in the briefs and oral argument presented by the Department of Commerce regarding the standard upon which to base inclusion or exclusion of products in the denominator of the subsidy calculation. In their brief, the Department argues that Canadian Parties failed to show that residual products were products of timber entering sawmills. At a different point in the brief, however, the Department states that the denominator should include merchandise that “resulted from the processing of subsidized logs”. And, at oral argument, counsel for the Department admitted that products such as shakes and shingles legitimately belong in the denominator. These statements imply that as long as a product is derived from logs that entered sawmills, regardless of what production process is used, the sales value attributable to the product should be included in the denominator. This is the position advocated by the Canadian Parties. On the other hand, the Department’s brief faults the Canadian Parties for submitting a list of residual products during the investigation that included products “that did not result from the production of lumber” (italics supplied). At oral argument, counsel for the Department claimed that the Canadian list “included products that were not part of the lumber-manufacturing process”. These statements indicate that the Department considers only products that result from the lumber-manufacturing process to be properly part of the denominator.

To add to the confusion, the petitioner claims that the numerator includes only the logs that were used in the lumber production process and that, consequently, none of the logs from the numerator were used to produce residual products. The petitioner asserts that during the period of investigation, both the Department and the Canadian Parties operated under the assumption that only logs used to produce lumber would be considered under the category “all logs entering sawmills” and that the data that the Canadian Parties provided conformed to this assumption. The petitioners cite to record evidence that demonstrates that the data submitted for each province reflected only that portion of the harvest that was used for the production of lumber. Thus, according to petitioners, since no logs included in the numerator were used to produce residuals, there is no need to include any sales of residual products in the denominator.

In an effort to sort out the various claims, the Panel observes that all parties at least implicitly recognize that had the numerator included logs used in the production of residual products, the sales of residual products should be included in the denominator. This principle was recognized as valid by the Department in the case of accounting for sales of co-products. In the Panel’s view, the fact that residual products employ a different production process should not alter the application of this basic principle. However, both the Department and the petitioner state in their briefs that the subsidy in this case is attributable to “all products produced during the softwood lumber production process”. This is in effect a statement that the subsidy is tied to the softwood lumber production process; as such it directly contradicts the Department’s assertion that this case involves subsidies that are not tied to lumber production.

The Panel recognizes the deference owed to the Department in the implementation of its own regulations. However the Department must support its decision with a reasoned explanation. As the Department has recognized in its brief:

While [a reasoned explanation] is not formally part of the standard of review that this Panel must apply, the statute does require the Department to publish a notice supporting its final determination in the Federal Register. This notice must provide “an explanation of the basis for its determination that addresses relevant arguments . . . concerning the establishment of . . . a countervailable subsidy . . . .” [Citing 19 U.S.C. § 1677f(i)(1)&(3)(A).] With respect to the interested parties’ legal arguments, the statute merely “requires that issues material to the agency’s determination be discussed so that the ‘path of the agency may reasonably be discerned’ by a reviewing court.” [Citing the URAA Statement of Administrative Action, H.R. Doc. No. 103-316, at 892 (1994).] [DOC Response Brief, vol. II, A-10.]

In the present case, inconsistent claims advanced by the Department render it not possible for this Panel to “reasonably . . . discern[]” the “path” followed by the Department. In the view of the Panel, the Department may not claim that the stumpage subsidy is not tied to lumber production in the context of explaining its decision to include in the numerator the timber volume attributable to co-products, while at the same time it argues the opposite in an effort to exclude from the denominator all products that did not result from the lumber production process.

The Canadian Parties’ position has its own inconsistencies. At oral argument, counsel for the Canadian Parties claimed that all of the products on the list submitted to the Department during the investigation were derived from logs entering sawmills. In their brief, however, although not conceding error, the Canadian Parties allow for the possibility that some of the products on the list to which the Department objected, such as particle and wafer board, may not in fact have been processed by sawmills. The Canadian Parties further claim that a “trifling error” is no reason to exclude the entire category. The Department, on the other hand, counters that “many of the products contained in the residual products category were never processed by the sawmills (emphasis supplied).” DOC Br. I-11.

Thus, the Panel is faced with three claims. The petitioner claims that the logs included in the numerator were not used to produce residual products. The Department concedes that some residual products on Canada’s list could have been included in the denominator, but refuses to do so claiming that it received insufficient data from the Canadian Parties to enable it separate out residual products that qualify for inclusion in the denominator. The Canadian Parties assert that all products on the list they provided to the Department are produced by sawmills and should therefore be included in the denominator.

The Panel finds that the Department did not provide a reasoned explanation for its decision to categorically exclude all residual products from the denominator. The Panel is aware that the problem essentially involves the expertise of the agency; the Panel may not substitute its judgment for that of the agency provided the agency is supported by substantial evidence on the record. In this instance, however, the Department’s explanation is inconsistent with its prior practice in this very investigation. The Department justified its exclusion of residual products by stating that “they were not the result of the lumber production process”. As discussed above, the Department’s justification implies the premise that the benefit is tied to the lumber production process – a premise that the Department itself vigorously opposed in the context of the numerator calculation. As shown in the discussion regarding the calculation of the numerator, the Department’s regulations clearly require it to attribute the benefit of a subsidy to all the products of a firm that receives the subsidy. See 19 C.F.R. § 351.525(3) (requiring the Secretary to “attribute a domestic subsidy to all products sold by a firm”). The benefit in this case was calculated on the basis of all logs entering sawmills. The Panel notes that although the Department’s explanation indicates an implicit reliance on this argument, the Department never explicitly adopted the petitioner’s assertion that no logs in the numerator were used to produce residual products.

Owing to these inconsistencies, the Panel remands the issue of the inclusion of residual products in the denominator of the subsidy calculation to the Department for further consideration. Specifically, the Department should address whether all logs included in the numerator were used solely to produce softwood lumber products and co-products. If so, the Department may not include sales of residual products in the denominator. The Department should clarify the basis for its rejection of residual products on the list supplied by the Canadian Parties. Did the Department reject residual products because “many” of such products did not result from the lumber production process, or because the logs used were never processed by sawmills? In both its Final Determination and its brief, the Department relies on both justifications. Since the Panel considers the two grounds inconsistent, the Department has not provided the Panel with a clear path of the Department’s reasoning. To the extent that sawmills produce residual products using material from logs that are included in the numerator, sales of such products should be included in the denominator, regardless of the process employed in their production.

The Panel recognizes that agency expertise is involved in this exercise and affirms that the Department has broad discretion in the application of its expertise. The Panel also recognizes that inclusion of residual products in the denominator could pose analytical difficulties. For example, the Canadian Parties claimed at oral argument that residual products such as particleboard and fiberboard are made by sawmills from sawdust. Sales of sawdust, as a co-product, are already included in the denominator. Using co-products as inputs for residual products may result in double counting thereby inflating the denominator. In addition, processes used in the production of residual products may add considerable value to the product, giving rise to the question of how to account for that added value and still derive a subsidy rate that adequately reflects the benefit received.

Claim IV: The Department failed to include sales of excluded companies and sales from the Maritime Provinces in the denominator

The Canadian Parties argue that it is a “well-established principle” that the Department must include sales of all producers of subject merchandise of the country under investigation in the course of establishing a country-wide rate. They claim that the Department violated this rule by excluding from the denominator both sales of excluded companies and exports from the Maritime Provinces.

The Canadian Parties rely on IPSCO v. United States, 899 F.2d 1192, 1197 (Fed. Cir. 1990) and Kajaria Iron Castings, supra, 156 F.3d. 1163, for the proposition that the proper procedure in determining a country-wide rate is for the "ITA to calculate a weighted-average net subsidy rate by dividing the sum of all benefits provided in subsidy of the subject goods by the total value of export sales of the goods in the United States."

As far as the sales from the Maritime Provinces are concerned, the Canadian Parties point out that Maritime timber is not subsidized and hence companies from the Maritime Provinces should be excluded from the CVD order. However, they say, exports from all provinces must be included in the calculation of a country-wide rate. The Department's prior practice affirms this principle: In Fresh Atlantic Salmon from Chile, the Department investigated subsidies from only two regions but used sales of all salmon from Chile in the subsidy rate denominator.

The Department’s response is that the statute allows the Department to calculate the subsidy rate on a country-wide basis if it determines that calculation of company-specific rates would be impracticable. Commerce claims that it properly excluded companies that did not benefit from the subsidies from the calculation of the countrywide rate as well as sales of companies from the Maritime Provinces as these provinces provided no subsidy.

As to the Canadian Parties’ claim that sales of these producers should nevertheless have been included in the denominator, the Department argues that the Canadian Parties are relying on legal standards that are no longer in effect. Prior to the URAA, the statute contained a preference for country-wide rates. In this period, a country-wide rate was calculated through a weighted average of company-specific rates. The current law, however, expresses a preference for company-specific rates, and, when a country-wide rate becomes necessary, the rate is calculated based on the aggregate use of the subsidies. The Department points to the statute that now provides that the “estimated country-wide rate . . . shall be based on industry-wide data regarding the use of subsidies determined to be countervailable.” 19 U.S.C. § 1671d(c)(5)(B). Therefore, sales of companies that did not receive or use a subsidy, the Department argues, should not be included in the denominator of the country-wide subsidy rate calculation.

The Panel must determine whether the statute authorizes the Department to exclude from the denominator of the subsidy calculation sales of companies found to be subsidized at a zero or de minimis rate. The relevant section of the governing statute provides that the “estimated country-wide rate . . . shall be based on industry-wide data regarding the use of subsidies determined to be countervailable.” 19 U.S.C. § 1671d(c)(5)(B).

The Panel notes that the phrase "industry-wide data regarding the use of subsidies" purports to guide the Department in determining the numerator of the subsidy rate calculation. In other words, it requires the Department to collect information about the amount of benefit that is provided to the industry, in this case the lumber manufacturing and remanufacturing industry, as a whole. This is in contrast to the methods used to calculate the company-specific and all-others rates as such rates are based on investigations of individual companies and thus rely on company-specific data. The problem in this instance, however, is not the amount in the numerator but rather the correct sales value to which the numerator should be allocated.

In a paradigm case involving a country-wide investigation, the numerator, calculated on basis of industry-wide data, would be allocated to industry-wide sales, and the resulting CVD rate would be applied to all producers. See 19 U.S.C. § 1677f- 1(e)(2)(B). The Department's regulations clearly contemplate that individual companies may be excluded from the CVD order if their company exclusion requests are granted. See 19 C.F.R. § 351.204(e)(4) (providing that in an aggregate investigation, the Secretary shall consider requests for exclusion to the extent practicable). Neither the statute nor the regulations address how the Department should treat sales of excluded companies.

The Panel understands the Department’s argument to assert that the "use" language in the statute mandates that sales of companies that have been found not to use the subsidy must be disregarded in the denominator. Since the numerator must match the 85 denominator, counting sales of excluded companies would place these values in the denominator without the benefit to the companies being present in the numerator.

The Panel is persuaded by this argument. The Panel finds additional support for this principle in the statutory provision governing the calculation of the all-others rate. See 19 U.S.C. § 1671d(c)(5)(A)(i). The functional similarity of the all-others and the country-wide rates, and the fact that the all-others and country-wide rates are always grouped together and contrasted with the company-specific method in the governing statute, indicates that principles derived from one of these two methods are relevant in interpreting the other.

The two methods are juxtaposed against the company-specific method in § 1677f- 1(e)(2), where the statute affords the Department a choice between calculating an all-others rate or a country-wide rate in the event that the preferred method, the company-specific rate, is not practicable owing to the large number of producers involved. See 19 U.S.C. § 1677f-1(e)(2). Second, the Panel notes that the statute describes both methods under a common provision entitled "Methods for determining the all-others rate and the country-wide subsidy rate". See 19 U.S.C. § 1671d(c)(5). Finally, an additional indication of the two methods' functional similarity is the fact that under the pre-URAA statute, the country-wide rate was calculated in the same way as the all-others rate is calculated in the present statute. See IPSCO, supra, 899 F.2d at 1196.

More specifically, an "all-others" rate is a combination of company-specific rates and the country-wide rate. When many producers are involved in an investigation, the Department may determine individual rates for some producers either by relying on a statistically valid sample or by selecting those producers that account for the largest volume of exports of subject merchandise. See 19 U.S.C. § 1677f-1(e)(2)(A). That means that the selected producers will be countervailed at company-specific rates. The calculation does not stop there, however. The Department must calculate an all-others rate which will be applied, as the name suggests, to all other producers not individually investigated. The all-others rate is equal to weighted average of rates established for individually investigated producers. See 19 U.S.C. § 1671d(c)(5)(A)(i). The statute specifically provides that such an average shall not include any zero or de minimis subsidy rates or rates determined under the “facts available” in situations in which an interested party fails to cooperate. Ibid.; see also 19 U.S.C. § 1677e(a)(2)(D) (providing that when an interested party fails to provide the requested information, the Department may "use facts otherwise available in reaching the applicable determination"). In other words, where any of the producers that the Department individually investigates is subsidized at a rate of zero or at a de minimis rate, that rate is not included in the weighted average that produces the all-others rate.

To use the aid of an example, assume that the Department individually investigates companies A, B, and C, and determines that company A is subsidized at a rate of 15%, company B at a rate of 20%, and company C at a rate of 0%. For purposes of calculating the all-others rate, the Department would average only the rates for companies A and B, and not the rates for all three producers. This reflects the principle that only those producers that use the subsidy may provide the basis for the all-others rate calculation. This position is further bolstered by the fact that rates calculated on the basis of adverse facts available are also not used in calculating the all-others rate. Thus Congress wanted to arrive at a methodology that would allow the Department to determine rates that most appropriately reflect the level of subsidization in investigations in which it is not possible to use the most reliable method – the company-specific rate. By excluding zero and de minimis rates, Congress wanted to avoid artificial deflation in the all-others rate. Similarly, excluding rates based on adverse facts available, which would most likely exceed the actual level of subsidization, prevents inflation of the all-others rate.

As the above discussion shows, in the all-others rate provision Congress has affirmed that the calculation of subsidy rates should be based on the rates of companies that actually use the subsidy. The Panel notes the similarities to calculation of all-others and country-wide rates. The Panel also notes that the calculation of a country-wide rate does not involve averaging of company-specific rates but rather an allocation of the industry-wide data on the use of the subsidy to the relevant sales. Even so, when faced with the problem of how to account for sales of an excluded company, the Department is faced with essentially the same problem as when it must average several company-specific rates that include zero or de minimis rates. In both situations, inclusion of the zero-rate company would deflate the subsidy rate. The Panel finds that the statute clearly expresses the command that such a deflation is not authorized.

This approach is reasonable. The basic purpose of countervailing duty discipline is to reach a CVD rate that will adequately offset the benefit provided without exceeding the actual level of the benefit received. The Canadian Parties' approach could result in a CVD rate inadequate to implement the statutory purpose. Since unsubsidized producers are excluded from the numerator, leaving their sales figures in the denominator could provide an unwarranted reduction in the subsidy rate calculated for other producers. Producers of subject merchandise could be assigned a subsidy rate that is less than the benefit they are receiving.

Application of the methodologies proposed by the parties in this proceeding may be illustrated as follows. Assume that there are only two groups of producers of the subject merchandise in an exporting country, and assume that the government provides a subsidy benefit of $100. Assume further that group A receives no benefit while group B receives the whole $100. Total sales in the investigating country amount to $1,000, of which $200 are sales by producers in group A, and $800 are sales by producers in group B. Thus, group A is subsidized at a rate of 0% while group B is subsidized at a rate of 12.5%. Companies in group A are granted exclusion orders.

According to the Department's approach, sales of group A would be excluded from the denominator, and thus the country-wide CVD rate would be set at 12.5%. Group A would pay no countervailing duties while the Customs Service would collect $100 in CVD from group B. This procedure would successfully offset the benefit of the subsidy.

Under the Canadian Parties' approach, sales of group A would be included in the denominator, but the products of group B would not be subject to CVD. The countrywide CVD rate would be set at 10%. Customs would collects $80 in countervailing duties on imports from group B and nothing on imports from group A. Under this approach, group B would retain $20 of the $100 benefit received. In different terms, the Canadian Parties' approach would allow group B to remain subsidized at a rate of 2.5%.

The Panel holds that excluding from the denominator sales of companies that were found to receive a zero or a de minimis subsidy is consistent with the statue and affirms the Department's determination.

Claim V: The Department used unreliable data to calculate final mill values

The Canadian Parties claim that the Department abandoned its prior practice that favored the use of first mill data in calculating the subsidy rate but included final mill data as well. "First mill" data represents sales of lumber coming out of sawmills while "final mill" data refers to sales of remanufactured products.

The main problem with this approach, according to the Canadian Parties, is the lack of reliable data upon which to determine final mill values. Canada maintains rigorous and reliable statistics on first mill values, but not on final mill sales. However, most of the data the Department required to assess final mill sales is confidential and cannot be disclosed according to Canadian law. There are two possible sources of data for final mill sales: the 1997 Statistics Canada report (called "Exhibit 36" in this proceeding) and a report compiled by the Pacific Forestry Center ("PFC report"). Canada worked with economists from the Pacific Forestry Center to estimate the value that remanufactured products add to the shipments of softwood lumber. The Canadian Parties contend that the Department relied on Exhibit 36 although the data used in Exhibit 36 is far less accurate than the data used in the PFC report. According to the Canadian Parties, Exhibit 36, prepared in 1997, does not contain data relating to the period of investigation nor does it contain price indices that could be used to project “final mill” sales values to the period of investigation. An additional problem with the Exhibit 36 data is that it consists of sales of in-scope remanufactured products that do not originate from sawmills. Finally, the Canadian Parties assert that Exhibit 36 is based on understated sales figures from only two provinces, which sales figures the Department used to extrapolate to a country-wide "final mill" value for six other provinces and territories. Hence the Canadian Parties argue that the Department's constructed value based on Exhibit 36 is not supported by substantial evidence on the record.

As an illustration of the inaccuracy of the Department's constructed data, the Canadian Parties point to the discrepancy between the PFC report and Exhibit 36 regarding the value of final mill sales from British Columbia. According to the PFC report, which the Department verified, in-scope value-added shipments in British Columbia alone amount to C$1,366,109,265. Yet, the value the Department constructed in the final determination for six provinces and two territories, including British Columbia, was only C$272,895,026 – less than one fifth of the verified amount for British Columbia alone.

Canada also complains that the Department named new products as subject merchandise late in the investigation. The data from Exhibit 36 does not contain the value of shipments for those new products since the Department did not request that information at the time Exhibit 36 was provided.

Both the Department and petitioner agree that to arrive at an accurate final mill shipment value, there can be no double-counting. When assessing the value of remanufactured products, the Department must deduct the value of the input lumber that is already counted in the first mill value. Statistics Canada had no way to perform this deduction in preparing Exhibit 36.

The Department also conceded that in order to use the 1997 Statistics Canada data in Exhibit 36, an inflation adjustment should have been performed. The Department failed to perform the inflation adjustment.

The Canadian Parties argue that, by using the 1997 Statistics Canada data and rejecting PFC data, the Department in effect applied an "adverse inference". The Department may do this only if the interested party fails to cooperate. However, the Government of Canada cooperated fully. Nevertheless, the Department based its adverse inference on the fact that the lack of data for every province in Exhibit 36 is the "direct result of Statistics Canada's refusal to provide the data on a Province-specific basis." H- 48 (quoting Commerce Decision Memo). But provincial information on "final mill" sales could not be disclosed as it is illegal to do so under Canadian law that protects competitive information. The United States has similar laws.

The Canadian Parties argue that the Department's final mill values bear no rational relationship to country-wide sales of remanufactured products. The Department purported to calculate the rate on final mill values but effectively did so on first mill data. If the Department intended to use a final mill calculation, it was required to use an estimate that reasonably reflects the value of Canadian final mill sales.

Canada also argues that the Department's past practice in Lumber III is to prefer final mill values unless first mill data is much more reliable and Canada has made every effort to provide accurate first mill values. In the latter case the Department will calculate the subsidy rate using first mill data. U.S. Customs must then collect duties on that value as well. In the present case, however, the subsidy rate was calculated using a sawmill shipments denominator ("first mill") but that rate was applied to the total value of shipments by value-added remanufacturers.

The Department argues that since the scope of this investigation included both lumber and certain remanufactured products, the Department was correct in using final mill data. This method is in accordance with the Department's practice of using final mill values in the calculation of subsidy rates. In spite of the preference for final mill values, the Department will use first mill data if final mill values are not available or are not reliable.

Commerce points out that the problem in this case is that the U.S. Customs Service assesses the amount of duty to be paid based the entered value of a product. Under the Canadian Parties’ proposed methodology, duties should be assessed on the value of the lumber after being processed only by the first mill. But, this method would ignore the value added by the remanufacturing process and would depart from the Department practice of assessing the countervailing duty based on entered value.

The Department claims that it repeatedly asked the GOC to provide reliable information on final mill values. The GOC had available information compiled by Statistics Canada which collects information through three different procedures: (1) an Annual Survey of Manufacturers; (2) a Monthly Survey of Manufacturers; and (3) a Monthly Survey of Sawmills and Planing Mills. The Government of Canada's position had been that the Statistics Canada information was not reliable. On December 17, 2001, Canada submitted the Statistics Canada 1997 Annual Survey of Manufacturers ("Exhibit 36") for British Columbia and Alberta. As for the remaining provinces, the GOC indicated that it could not provide the requested information due to confidentiality restrictions. Then, on January 7, 2002, the GOC submitted a study from the Pacific Forestry Center ("PFC"), which estimated the total value of remanufactured lumber shipments in British Columbia.

Commerce claims it uncovered serious flaws in the PFC study and instead used the information contained in Exhibit 36 to calculate an estimate of the value of remanufactured products. According to Commerce, it may reject evidence that is demonstrably inaccurate. Commerce asserts that the flaws in the PFC report made it impossible for the Department to establish the reasonableness of the methodology used.

The Department concludes that its estimate of the value of remanufactured products was a reasonable exercise of administrative discretion and not an application of adverse facts available, as the Canadian Parties contend. The Department made a reasonable choice between two sets of data to use in the construction of this estimate.

Here, the Panel must review a factual determination made by the Department that the PFC report was not sufficiently reliable. Factual determinations by the Department must be upheld if they are supported by substantial evidence on the record. The Panel may not reweigh the evidence and must affirm the Department's finding even if a contrary result is possible from the record evidence.

The Panel finds that the Department provided a reasoned explanation of its finding that the PFC report is not reliable. Owing to constraints imposed by Canadian law on the release of the relevant data, both studies were limited to data from only one or two provinces. In fact, Exhibit 36, which the Department relied on, contained data compiled by official government agencies for two provinces – British Columbia and Alberta – while the PFC report was compiled by a private company and contained data only for British Columbia. The evidence on the record indicates that the Department considered the PFC report in detail and addressed the arguments made by the Canadian Parties. See 19 U.S.C. §1677f(i)(3)(A (3)(A).

The Panel did not consider it necessary or appropriate to examine all the technical aspects of the questions at issue, but is satisfied that the Department's decision to rely on Exhibit 36 is supported by substantial evidence. The Panel affirms that decision.

XIII. COALITION ISSUES

The Coalition, as noted, on May 5, 2002, filed a complaint in this action which alleged errors of both law and fact in the Department’s determination.

The legal arguments generally concern the manner in which the Canadian provinces administer the stumpage programs. In substance, they allege that timber is sold by the provinces at administered, non-competitive prices far below market value, thereby providing a subsidy. These issues are dealt with insofar as they are relevant to the panel’s decision, elsewhere in this opinion.

The Coalition also raises specific factual aspects of the Department’s decision which are claimed to be unsupported by substantial evidence or otherwise not in accordance with the law. These issues consist of claims that the Department incorrectly failed to make proper adjustments in comparing the prices of benchmark timber to subject timber, and issues concerning the calculation of the subsidy.

With regard to the first set of issues, the Coalition agrees that it was proper for the Department to adjust the stumpage prices upwards to account for certain cash payments required by the provinces to be made as a condition of harvesting timber. These payments represented the costs of such things as silviculture levies, road building, and payments to special forestry funds. It is contended that the Department failed, however, to adjust the prices of benchmark timber to account for similar payments made by U. S. timber harvesters. In addition, it is claimed that the Investigating Authority failed to properly value coastal red cedar, sitka spruce, and cypress in Western Washington. Also complained of was Commerce’s calculation of sawtimber prices in Michigan for purposes of comparison with Ontario, and the failure to use Montana timber prices in the benchmark for comparison with the Province of Alberta. In light of the Panel’s conclusions about the use of U.S. prices as benchmarks by the Department, it concludes that the issues raised by the Coalition are moot.

The other issues regarding calculations are dealt with in the “calculations” section of this report.

XIV. COMMERCE’S REQUEST FOR A REMAND

The Investigating Authority has requested relief in the form of a remand to address certain issues. These relate to adjustments regarding secondary road construction, the reimbursement of silviculture overhead costs, and forest management and planning costs in Ontario, the calculation of road construction and maintenance adjustment for Saskatchewan and Manitoba, and to explain its rationale for not granting an adjustment for supplemental fire protection costs in Manitoba.

The only objection by the Canadian Parties to remand for these reasons was raised by the Province of Saskatchewan. As the Panel understands the issue related to Saskatchewan, the Department allowed as an adjustment in the final decision, the costs for road building on the basis that the tenure holders were required to construct roads to access the forests. However, the adjustment may have included the building of tertiary roads, while the building of only primary and secondary roads was mandated by the province. The Panel is satisfied that, to the extent that this issue survives the remand, the province will have an opportunity to address this question with the Department. The balance of the issues appears to have been obviated by our decision.

XV. CONCLUSION

For the reasons set forth in the foregoing opinion, the Panel hereby remands this matter to Commerce for further proceedings consistent with this opinion. Commerce is directed to report its Determination on Remand within sixty (60) days from the date of this decision.