U.H.F.C. Company appeals from the final judgment of the Court of International Trade in an antidumping proceeding, U.H.F.C. Co. v. United States, 706 F. Supp. 914 (Ct.Int'l Trade 1989) (Musgrave, J.). Animal glue imported from the Netherlands has been subject to an antidumping order imposed by the Treasury Department on December 22, 1977. This appeal concerns the second of three review periods conducted since 1980 by the International Trade Administration (ITA), which resulted in a final determination imposing a dumping duty rate of 24.60 percent payable by U.H.F.C., the United States importers of glue sold by the Netherlands manufacturer, B.V. Lijmfabriek C. Trommelen.1 In arriving at this rate, the ITA concluded that sales in the Netherlands were the appropriate basis for determining foreign market value (FMV); denied U.H.F.C.'s request for price adjustments based on certain physical differences between the glues sold in the United States and in the home market; and calculated the dumping margin based on the FMV of a glue grade less similar to the U.S. glue than other glues sold in the home market. In making its determination, the ITA invoked its authority under 19 U.S.C. Sec. 1677e(b) (redesignated in 1988 as Sec. 1677e(c)2) to use the "best information otherwise available" because, in ITA's view, Trommelen had failed to supply certain information as requested by ITA.

After the ITA published its final determination, U.H.F.C. brought suit in the Court of International Trade challenging (1) the selection of home market sales as the basis for determining FMV; (2) the denial of certain price adjustments; and (3) the selection of sales of a less similar home market glue in calculating the dumping margin. The Court of International Trade affirmed the methodology used by the ITA to determine the dumping margin in all respects. U.H.F.C. appealed the trial court's decision to this court pursuant to 28 U.S.C. Sec. 1295(a)(5) (1988). We now reverse and remand.

* BACKGROUND

This is a trade case involving the importation of animal glues from the Netherlands which have been sold at a price less than their fair value. Animal glues are produced and sold in different, identifiable grades. The glue grade is determined on the basis of comparative jelly-strength and viscosity value. The glue is generally priced based on its jelly strength, with weak jelly, 30 bloomgrams, being the lowest priced and very strong jelly, 500 bloomgrams, being the highest priced. Glues of different strength can be blended in linear mathematical proportions to yield a glue of desired strength. For example, an order for 200 pounds of glue at 300 bloomgrams could be filled by mixing 100 pounds of 250 bloomgram glue with 100 pounds of 350 bloomgram glue. Animal glues may be used in widely varying applications such as general adhesives, abrasives or sizing agents.

Representatives of the United States animal glue industry filed a petition with the Secretary of the Treasury seeking the imposition of antidumping duties against animal glue imports from the Netherlands. After an affirmative less-than-fair-value determination by the Department of the Treasury3 and an affirmative injury determination by the International Trade Commission,4 Treasury published a dumping order against Animal Glue and Inedible Gelatin from the Netherlands. See 42 Fed.Reg. 64115 (1977). Thereafter, authority for administering the antidumping laws was transferred to the Commerce Department, and ITA has conducted three administrative reviews of the animal glue dumping duty.

This action concerns the second of these review periods, covering the animal glue entries made by U.H.F.C. that were purchased from Trommelen spanning December 1, 1980 through November 30, 1981. See 48 Fed.Reg. 45583-84 (1983). During this review period, Trommelen sold to U.H.F.C. for import into the United States glue grades having strengths of 250, 350, 365, 380 and 450 bloomgrams. At the same time, Trommelen sold glue grades in the Netherlands having bloomgram strengths of 150, 170, 190, 210, 220, 230, 260, 290, 300 and 400. Trommelen made no sales to the United States of glues identical in grade to those sold in the Netherlands. The grades are normally a function of customer requirements.

Administrative reviews of antidumping duty orders are conducted pursuant to 19 U.S.C. Sec. 1675 (1982).5 To determine the dumping margin, the statute instructs ITA to separately determine the United States price (USP) and the foreign market value (FMV) and then subtract the USP from the FMV to arrive at the dumping margin which becomes the duty rate. See 19 U.S.C. Sec. 1675(a)(2) (1988).

Determining FMV first requires the ITA to establish which foreign market sales to use. To this end, Congress has mandated a statutory hierarchy aimed at achieving the best approximation of the value of the good in the foreign market. The statute prescribes, in pertinent part:

(a) Determination; fictitious market; sales agencies

For purposes of this subtitle--

(1) In general

The foreign market value of imported merchandise shall be the price, at the time of exportation of such merchandise to the United States--

(A) at which such or similar merchandise is sold or, in the absence of sales, offered for sale in the principal markets of the country from which exported, in the usual wholesale quantities and in the ordinary course of trade for home consumption, or

(B) if not sold or offered for sale for home consumption, or if the administering authority determines that the quantity sold for home consumption is so small in relation to the quantity sold for exportation to countries other than the United States as to form an inadequate basis for comparison, then the price at which so sold or offered for sale for exportation to countries other than the United States, ... [Emphasis added.]

Thus, where there are sales of "such or similar merchandise" in the home market which satisfy subparagraph A, such sales are to be used as the basis for determining FMV provided they are not too "small in relation to the quantity sold for exportation to countries other than the United States as to form an adequate basis for comparison."

To aid in determining whether the "quantity [of such or similar merchandise] sold for home consumption" forms a sufficient basis for comparison, the ITA promulgated regulation 19 C.F.R. Sec. 353.4(a) (1988):6

(a) In General. If it is established ... that during the representative period chosen for investigation, the quantity of such or similar merchandise sold for consumption in the country of exportation is so small in relation to the quantity sold for exportation to countries other than the United States (normally, less than 5% of the amount sold to third counrties) as to be an inadequate basis for determining the foreign market value of the merchandise imported into the United States, the foreign market value of the imported merchandise shall be determined either by reference to the price at which such or similar merchandise is sold or offered for sale for exportation to countries other than the United States or by reference to its constructed value. [Emphasis added.]

The products that are to be considered "such or similar merchandise" in determining which market is to serve as a basis for FMV pursuant to section 1677b(a)(1) are defined in a hierarchical fashion in 19 U.S.C. Sec. 1677(16) (1988). Section 1677(16) provides:

(16) Such or similar merchandise

The term "such or similar merchandise" means merchandise in the first of the following categories in respect of which a determination for the purposes of part II of this subtitle can be satisfactorily made:

(A) The merchandise which is the subject of an investigation and other merchandise which is identical in physical characteristics with, and was produced in the same country by the same person as, that merchandise.

(B) Merchandise--

(i) produced in the same country and by the same person as the merchandise which is the subject of the investigation,

(ii) like that merchandise in component material or materials and in the purposes for which used, and

(iii) approximately equal in commercial value to that merchandise.

(C) Merchandise--

(i) produced in the same country and by the same person and of the same general class or kind as the merchandise which is the subject of the investigation,

(ii) like that merchandise in the purposes for which used, and

(iii) which the administering authority determines may reasonably be compared with that merchandise. [Emphasis added.]

In response to the ITA's 1981 review period antidumping questionnaire, Trommelen requested that sales to other countries, namely Rumania and the United Kingdom, be used for calculating foreign market value because Trommelen's sales in the Netherlands accounted for only 4.9 percent of its total sales (including the United States and the Netherlands) during the review period.

The ITA instead found that Trommelen's home market sales sufficed to serve as the basis for establishing FMV. The ITA reasoned that all animal glue sales in the Netherlands were "such or similar merchandise" under section 1677(16)(C) because the range of glue grades sold in the Netherlands "may reasonably be compared" with the range of glue grades sold in the United States as a result of the many "common uses" to which the glues might be put. Taking all of Trommelen's sales in the Netherlands into account, the ITA determined that home market sales were not too small under the statute and implementing regulation, because these sales amounted to 6.07 percent of the quantity sold to countries other than the United States during the period.

Once the ITA selected the Netherlands as the foreign market, Trommelen submitted a supplemental questionnaire which claimed price adjustments based upon the differences between the quantity of sales in the Netherlands as compared with the U.S. sales and upon differences in jellystrength. To support these price adjustments, Trommelen submitted price lists of other companies which showed approximate costs across many jelly-strengths and another company, Eschem, Inc., submitted a method for computing the cost of glue at any jelly strength by blending prices. The antidumping statute requires, where properly established, adjustments to the foreign market value, inter alia, for "the amount of any difference between the USP and the FMV that is wholly or partly due to ... the fact that [similar merchandise] under [19 U.S.C. Sec. 1677(16) ] (B) or (C) is used in determining FMV." See 19 U.S.C. Sec. 1677b(a)(4) (1988).

In implementing the determination of this statutory adjustment, Commerce promulgated regulation 353.16, see 19 C.F.R. Sec. 353.16 (1988),7 which provides direction regarding price adjustments based on differences in physical characteristics between similar merchandise:

Sec. 353.16 Differences in Physical Characteristics

In comparing the United States price with the selling price in the home market ... in the case of similar merchandise, due allowance shall be made for differences in the physical characteristics of the merchandise in the markets being compared. In this regard, the Secretary will be guided primarily by the differences in cost of production, to the extent that it is established to his satisfaction that the amount of any price differential is wholly or partly due to such differences, but, when appropriate, the effect of such differences upon the market value of the merchandise may also be considered. In the case of merchandise which does not lend itself to comparison with other merchandise for the purpose of this section, any method reasonably calculated to reflect the impact on cost or value of any differences in the merchandise under consideration may be used. Differences in costs of producing merchandise with identical physical characteristics as end products will not be considered appropriate adjustments. [Emphasis added.]

Based on this regulation, the ITA requested that the Dutch manufacturers supply cost of production information on each of the glue strengths sold in the home market so that ITA could determine whether such price adjustments based on physical differences could be made. Trommelen responded that it could not provide such specific information, stating:

Every glue batch produced in our plant gives another quality. We produce qualities in the jellystrength range of 160 up to 520 bloomgrams. In order to blend a lot of 20 tons you need at least 4 different batches of 5 tons. Consequently it is not possible to give firm blending formulas for each grade sold in the USA and Holland. The final result depends on several factors....

We already informed that apart from volume of the transaction the jellystrength expressed in bloomgrams is determining the price. In order to give you a guideline to convert the prices to comparable qualities, we suggest that you take our standard prices for the supplied qualities, which were valid during the major part of the period of investigation....

The ITA denied the adjustments requested by Trommelen, and rejected Eschem's submission, responding that "section 353.16 of the Commerce Regulations provides that adjustments for physical differences in merchandise must be based on the differences in the cost of production" (emphasis supplied).

After making these determinations, the ITA made its preliminary determination of the dumping margin. In the absence of foreign sales of identical glue grades, the ITA was placed in the position of having to arrive at a dumping margin by subtracting the USP from the FMV of "similar" glue grades sold in the Netherlands. In such cases, according to the ITA, it has adopted a policy of using the home market merchandise "most similar" to the merchandise sold in the United States. For U.S. glue grades 350, 365, 380 and 450, the ITA chose home market glue grade 400. For U.S. grade 250, the ITA stated that it could not determine the "most similar" glue and instead invoked its authority under section 1677e(b), the "best information" rule, to choose grade 300, a relatively higher priced glue, bypassing use of home market grades 230, 260 and 290. The ITA then calculated the dumping margin to be 23.7 percent and published its preliminary results. See 48 Fed.Reg. 2030 (1983).

After the preliminary results were published, the ITA received comments from Olympic Adhesives, Inc.,8 another former U.S. manufacturer and then importer, who objected to comparing 300 strength glue in the Netherlands with 250 strength glue sold in the United States, claiming that 230 strength glue sold in the Netherlands was more similar than 300 in accordance with the ITA's policy of selecting the "most similar" merchandise for comparison, that 230 strength glue had been used as the basis for comparison with 250 strength glue in the previous review period, and that the ITA was constrained by the statute, 19 U.S.C. Sec. 1677(16), to follow a hierarchy in choosing merchandise for comparison. See 48 Fed.Reg. 45583-84 (1983). The ITA acknowledged that 230 strength glue "appears to be more similar to quality 250," and that 230 glue had been used in the previous review but justified using 300 strength glue due to Trommelen's failure to provide adequate cost of production data which required the ITA to invoke their statutory authority under 19 U.S.C. Sec. 1677e(b) to use the "best information otherwise available." Id. at 45584. The ITA then published its final results, which set the dumping margin for the period at 24.60 percent. Id.

U.H.F.C. then commenced the instant suit in the Court of International Trade pursuant to 19 U.S.C. Sec. 1516a (1988), seeking review of the final determination made by the ITA for the 1981 review period. U.H.F.C. contested the choice of home market sales in the Netherlands as the basis for determining FMV, the denial of price adjustments attributable to differences in physical characteristics between similar merchandise, and the choice of the FMV of 300 quality glue as the glue to be compared with the USP of 250 quality glue for purposes of calculating the dumping margin.

The Court of International Trade agreed with the ITA's decision to use home market sales, concluding that the ITA did not abuse its discretion in finding the entire range of glue grades sold in the Netherlands to be similar to the range of glues sold to the United States, and that the ITA's reliance on home market sales was based on a reasonable interpretation of the statute. The court also agreed with the ITA's denial of price adjustments concluding that the ITA reasonably invoked its authority to resort to the "best information" rule, given that Trommelen had failed to supply requested cost of production data and that the data supplied was not a reliable indicator of value. Finally, the court declined to disturb the ITA's choice of 300 grade glue for comparison to U.S. grade 250 glue in arriving at the dumping margin, concluding that the ITA properly invoked the "best information" rule in rejecting Netherlands sales of grade 230 glue and that grade 300 glue was "similar" to U.S. grade 250 under 19 U.S.C. Sec. 1677(16)(C).

1. Whether the ITA correctly concluded that sales in the Netherlands were sufficient to serve as the basis for determining FMV under 19 U.S.C. Sec. 1677b(a) and 19 C.F.R. Sec. 353.4(a)?

2. Whether the ITA's denial of price adjustments for differences in physical characteristics under 19 U.S.C. Sec. 1677b(a)(4) and 19 C.F.R. Sec. 353.16 was supported by substantial evidence and otherwise in accordance with law?3. Whether the ITA's choice, as the "best information otherwise available" under 19 U.S.C. Sec. 1677e(b), of home market sales of grade 300 glue as the basis for FMV for comparison with the USP of 250 grade glue when calculating the dumping margin is supported by substantial evidence?

U.H.F.C. brought this action pursuant to 19 U.S.C. Sec. 1516a(a)(2)(B)(iii) (1988) which permits any interested party to commence an action in the Court of International Trade to review final determinations of administrative reviews of antidumping duty orders. See Olympic Adhesives, Inc. v. United States, 899 F.2d 1565, 1570 (Fed.Cir.1990). "We review that court's review of an [agency] determination by applying anew the statute's express judicial review standard." Atlantic Sugar, Ltd. v. United States, 744 F.2d 1556, 1559 n. 10, 2 Fed.Cir. (T) 130, 133 n. 10 (Fed.Cir.1984). Thus, we must determine whether the Court of International Trade correctly concluded that the ITA's determination was supported by substantial evidence on the record and was otherwise in accordance with law. See 19 U.S.C. Sec. 1516a(b)(1) (1988).

The government argues, and the court agreed, that the ITA's selection of sales in the home market was proper because the ITA reasonably concluded from the evidence that the range of glue strengths sold in the Netherlands, were "such or similar merchandise" under definition (C) of section 1677(16) and that the total amount of these sales was sufficient to serve as the basis for FMV per statute and regulation. Appellant counters that the conclusion that all Netherlands sales are "such or similar merchandise" when compared with the glue strengths sold in the United States is unsupported, and that, in any event, only those sales of glue grades actually used by the ITA in calculating the dumping margin are sufficiently similar to be used to determine whether the amount of home market sales is sufficient to serve as the basis for FMV. We are unpersuaded, however, that the ITA's selection of home market sales is unsupported or that the ITA's interpretation of the statute governing the selection of the basis for FMV is unreasonable.

The antidumping statute sets up a mandatory hierarchy for determining which geographic market is to be used as the basis for establishing FMV. See 19 U.S.C. Sec. 1677b(a) (1988). The statute dictates that the FMV be established using the first market that qualifies, and expresses a preference for home market sales (country of manufacture). Id.; see also Washington Red Raspberry Comm'n v. U.S., 859 F.2d 898, 902 (Fed.Cir.1988); Ansaldo Componenti, S.p.A. v. United States, 628 F. Supp. 198, 204 (Ct.Int'l Trade 1986). One condition the antidumping statute places on the use of the preferred geographic market, the home market, is that sales of such or similar merchandise in this market not be "so small in relation to the quantity sold for exportation to countries other than the United States as to form an inadequate basis for comparison." 19 U.S.C. Sec. 1677b(a)(1)(B) (1988); see also Motorcycle Batteries From Taiwan; Final Determination of Sales at Less Than Fair Value, 47 Fed.Reg. 9264, 9264, 9266 (1982). The implementing regulation for this statutory limitation, found at 19 C.F.R. Sec. 353.4(a) (1988), among other things, sets forth a figure of 5 percent as the lower threshold below which sales in the home market, when compared to sales to all other countries excluding the United States, will normally be considered inadequate to permit home market sales to serve as the basis for FMV. See, e.g., Elemental Sulphur from Canada; Results of Administrative Review of Antidumping Finding, 47 Fed.Reg. 14507, 14508 (1982).

Applying these statutory sections and regulations to the 1981 review period, the ITA concluded that home market sales were sufficient to serve as the basis for FMV. The ITA had found that sales of all home market glues regardless of grade were "such or similar merchandise" with respect to the glue grades sold in the United States under section 1677(16)(C). In accordance with the statute and regulation, the ITA then totalled the home market sales of all glue grades and compared that amount with the total glue sales to other countries, excluding the United States, and found home market sales to represent 6.07 percent of sales to third countries. Thus, the sales to the home market exceeded the 5 percent regulatory threshold and glue sales in the Netherlands were sufficient to serve as the basis for FMV.

U.H.F.C. first asserts that ITA's determination that home market sales are sufficient as a basis for FMV must fail because it rests on an erroneous subsidiary finding that all glues sold by Trommelen in the Netherlands, regardless of glue grade, are "such or similar merchandise" with respect to all sales to the United States, regardless of glue grade. Specifically, U.H.F.C. contests, as unsupported by substantial evidence, the ITA's finding that all glues, regardless of grade, "may reasonably be compared" with glues sold to the United States under section 1677(16)(C). The ITA found that the glues "may reasonably be compared" regardless of grade based on the evidence showing many common uses to which the glues may be put. The government points to a Trommelen brochure that states:

Our product is exported to more than 35 countries all over the world. It is essential in the production of the following articles. Abrasives--Matches--Composed Cork Products--Currency and Security Papers--Musical Instruments--Textiles, etc.

The brochure goes on to provide instructions on how to process glues of different grades to obtain glue of any desired strength. The brochure can reasonably be understood to demonstrate that any glue grade, either with or without processing, might be used in any of the mentioned applications depending on the glue strength needed for the particular application, i.e., that the glue grades have many common uses.

Thus, we conclude that the record contains substantial evidence that the glue grades have many common uses. We further agree with the government that substantial evidence supports the conclusion that home market glues regardless of grade "may reasonably be compared" based on their many "common uses." And since the other criteria of section 1677(16)(C) are not in dispute, the conclusion necessarily follows that the ITA's finding that all home market glues are "such or similar merchandise" under section 1677(16)(C) is supported by substantial evidence.

U.H.F.C. next asserts that glue sales in the Netherlands did not meet the 5 percent threshold of 19 C.F.R. Sec. 353.4(a) because the ITA found only two grades, 300 and 400, to be sufficiently "similar" to serve as the basis for FMV, and that these grades amount to only 2.42 percent of the amount sold to third countries. We agree with the government that this argument confuses the difference between "similar merchandise" and "most similar" merchandise. The determination of the dumping margin requires calculation of an FMV and a USP, and subtracting the USP from the FMV. See 19 U.S.C. Sec. 1675(a)(2). The ITA found, and that finding is supported by substantial evidence, that all glues sold to the home market were "similar merchandise" for purposes of determining the market which serves as the basis for FMV. Once the home market is found to be a viable basis for FMV, the ITA has adopted a policy of matching the USP of merchandise with the FMV of the merchandise "most similar" to it in an effort to obtain the most accurate dumping margin using whatever market has been found to be the basis for FMV.9 See, Smith-Corona Group, Consumer Prods. Div., SCM Corp. v. United States, 713 F.2d 1568, 1578, 1 Fed.Cir. (T) 130, 140-41 (Fed.Cir.1983) (antidumping law administered to achieve fair "apples to apples" comparison); see also, Certain Unfinished Mirrors from Portugal, 51 Fed.Reg. 43409 (1986) (two-part methodology of determining dumping margin followed); Shop Towels of Cotton From the People's Republic of China; Final Results of Administrative Review of Antidumping Duty Order, 50 Fed.Reg. 26020, 26021 (1985) (same).

It is well settled that an agency's interpretation of the statute it has been entrusted by Congress to administer is to be upheld unless it is unreasonable. Zenith Radio Corp. v. United States, 437 U.S. 443, 451, 98 S. Ct. 2441, 2445, 57 L. Ed. 2d 337 (1978); Consumer Prods. Div., SCM Corp. v. Silver Reed America, Inc., 753 F.2d 1033, 1038-39, 3 Fed.Cir. (T) 83, 89-90 (Fed.Cir.1985); Smith-Corona, 713 F.2d at 1582, 1 Fed.Cir. (T) at 142-43. U.H.F.C. points to nothing in the statute or regulation that transforms the merchandise ITA considers "most similar" for purposes of computing the dumping margin under section 1675(a)(2) into the "such or similar merchandise" that must be considered to establish which market serves as the basis for FMV under section 1677b(a)(1). Indeed, the statute expressing a preference for home market as the basis for FMV requires that the sufficiency of sales in the home market be determined by reference to quantities of "such or similar merchandise" as defined in section 1677(16). As previously shown, the ITA's finding that all sales of glue in the Netherlands are "such or similar merchandise" is supported by substantial evidence. The statute and implementing regulations require no more.

We, therefore, uphold as a reasonable interpretation of the statute ITA's two-step policy of first determining which merchandise is "such or similar" in order to establish which market is to serve as the basis for FMV and then selecting which "such or similar" merchandise is the "most similar" in order to use the FMV for the latter in quantifying the dumping margin.

U.H.F.C. concludes that assuming the home market sales surpass the 5 percent threshold, that this 5 percent mark is not a hard and fast rule, but merely a rule of thumb that can be overcome and is overcome by the evidence it presented in this case. U.H.F.C. cites to prior determinations published by ITA as an indication that the ITA has "itself noted many factors" to be evaluated when judging whether the home market furnishes an adequate basis for FMV, and reasons that consideration of relevant factors mandates a finding that the home market was not a viable market. See, e.g., Certain Steel Wire Nails from the Republic of Korea, 45 Fed.Reg. 34941, 34942 (1980). U.H.F.C. points to evidence which shows that sales of identical glues in third countries are of grades identical to those sold to the United States; that the third country markets are larger, primary markets in contrast to the Netherlands which is a subsidiary market; and that sales to third countries were verified and made in quantities more comparable to those sales made to the United States. Coupled with the goal of the antidumping duty law to seek comparison of transactions in different markets on the most nearly comparable terms, U.H.F.C. argues that this evidence "compel[s] the use of third-country sales as the basis for FMV." The short answer to this argument is that the goal of the antidumping duty laws is carried forth in the statutory and regulatory provisions, which express a preference for the home market as the basis for FMV and here have been properly interpreted and applied by the ITA to support the use of home market sales as the basis for FMV. Evidence of the structure of third-country markets, as compared with the home market (size of markets, quantities of goods sold, identity of goods sold with those sold to the United States) even if considered relevant by the ITA, are simply insufficient to overturn a supported finding by the ITA that the statutorily preferred market, the home market, establishes a viable basis for determining FMV.

B. Denial of Price Adjustments for Physical Differences

The ITA denied Trommelen's request for adjustments to the FMV under 19 U.S.C. Sec. 1677b(a)(4) and 19 C.F.R. Sec. 353.16 based on acknowledged physical differences between the glue strengths found to be similar merchandise. Appellant urges that the ITA erroneously based its denial of price adjustments on the reasoning that price adjustments for physical differences must be made from cost of production information which it was impossible for appellant to supply. Since the glue making process is a continuous one which yields batches of various strengths, cost of production information on a per grade basis simply does not exist and therefore could not be supplied. On appeal, the government acknowledges, as the court recognized, that 19 C.F.R. Sec. 353.16 does not require price adjustments to be based only on cost of production information, and that a contrary interpretation would contravene that regulation. The regulation provides, in pertinent part:

[D]ue allowance shall be made for differences in the physical characteristics of the merchandise ... guided primarily by the differences in cost of production ... but, when appropriate, the effect of such differences upon the market value of the merchandise may also be considered. In the case of merchandise which does not lend itself to comparison with other merchandise for the purpose of this section, any method reasonably calculated to reflect the impact on cost or value of any differences in the merchandise under consideration may be used. [Emphasis added.]

19 C.F.R. Sec. 353.16. Per its express terms, the regulation expresses a preference for adjustments to be based on cost of production information, but the regulation specifically contemplates and authorizes adjustments based on other methods as well, including the effect of a physical characteristic on the market value of the merchandise. Thus, an interpretation of this regulation restricting price adjustments to differences in cost of production would render nugatory a considerable portion of the regulation's express terms. If, as U.H.F.C. suggests, the ITA's decision is premised on such an interpretation, it clearly cannot stand. See, e.g., Barber v. United States, 676 F.2d 651, 658, 230 Ct. Cl. 287, 298 (1982); Aparacor, Inc. v. United States, 571 F.2d 552, 557, 215 Ct. Cl. 596, 605 (1978).

According to the court, however, Trommelen's submission to ITA showing how price data could be used to make adjustments for differences in glue strength was deficient because it showed that the difference in prices was not "dispositive of value, and by extension, physical difference." U.H.F.C. Co., 706 F. Supp. at 922. To support this statement, the court relied upon evidence showing that Trommelen's price data indicated that the same strength was sold at different prices in the home market and that Trommelen had indicated that there is in fact an inverse relationship between the strengths of glues and their cost of production. Id. The court thus concluded that:

Given the fact that Trommelen failed to provide the requested cost of production information and that the price information in question was not a reliable indicator of value, the ITA used other information, i.e., the "best information available." ... On this basis the ITA declined to make an adjustment to foreign market value for physical differences in the merchandise, and the Court finds that this denial is supported by substantial evidence and is otherwise in accordance with law. Id.

It is clear that the court was of the view that the ITA's reason for denying price adjustments did not lie solely in the failure to provide cost of production information, but also rested on the ITA's determination that the alternative information actually submitted for adjusting differences was insufficient or unreliable.

Appellant argues that the latter rationale set out by the court to sustain the agency's denial of price adjustments represents an impermissible post hoc rationalization. The agency did not articulate this position until trial so that the court, per U.H.F.C., decided the alternative data was inadequate and not the agency. According to appellant, the agency's decision rested solely on the agency's unreasonable requirement that price adjustments in this case be calculated on the basis of differences in cost of production data. We agree.

On review, a court, including the Court of International Trade, is constrained to apply the applicable standard of review to the rationale put forth by the agency in support of its action. Citizens to Preserve Overton Park, Inc. v. Secretary of Transp., 401 U.S. 402, 413, 91 S. Ct. 814, 822, 28 L. Ed. 2d 136 (1971). Post-hoc rationalizations of agency actions first advocated by counsel in court may not serve as the basis for sustaining the agency's determination. See, Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168, 83 S. Ct. 239, 245, 9 L. Ed. 2d 207 (1962); Atcor, Inc. v. United States, 11 C.I.T. 148, 152, 658 F. Supp. 295, 299 (1987).

In its final determination, the agency stated that:

The department requested cost of production data ... so that we could make adjustments for physical differences in merchandise. The producer failed to provide adequate data, and as a result, we make [the] comparison ... (without adjustments for differences).

48 Fed.Reg. at 45584. Further, Eschem, Inc., another Dutch glue manufacturer involved here, submitted a lengthy correspondence explaining a method for calculating price adjustments based on the blending prices of individual glue strengths. The ITA's response leaves no doubt of ITA's view that no price adjustments could be made here without cost of production data:

Section 353.16 of the Commerce Regulations provides that adjustments for physical differences in merchandise must be based on the differences in the cost of production.

Whether this was due to a legally erroneous interpretation of the regulation by the staff at the time or simply bureaucratic rigidity is immaterial. The record consistently demonstrates that the only basis for ITA's decision to deny price adjustments was the failure of the manufacturers to supply cost of production information for individual grades which would establish that higher grades cost more to produce.

In any event, if the agency did imply somewhere in the record that the data actually submitted was inadequate, the agency's denial of price adjustments for that reason would be unsupportable. ITA did not establish that it was entitled to use the "best information otherwise available" by reason of the failure or refusal of the Dutch manufacturers to submit requested information.10

Both the ITA and the court acknowledged that price is generally determined by glue strength, with higher bloomgram strength glues generally commanding higher prices. Thus, price differences for each bloomgram strength generally reflect the value attached to that physical difference. Moreover, there is an inverse relationship between glue strengths and cost of production. Further, the evidence showing that some glues of lesser strengths were sold at higher prices does not refute that the price information provided by Trommelen on each glue strength was indicative of the value of this difference in glue strength.

The inverse relationship between cost of production and glue strength reflects the realities of the animal hide and bone glue manufacturing process. As explained to us, this process resembles the pressing of olives or grapes in that the first pressing yields the highest quality and successive pressings yield lesser qualities compared to the first. But because it takes more pressings to obtain the lesser quality, in a sense the lesser costs more. No one would dispute, however, that the juice or oil obtained in the first pressing, being higher quality, has a greater value. So it is with the manufacture of animal hide and bone glues. In this situation as well, the relationship between the cost of production and grades of product is "inverse" and, thus, cost of production would tell nothing about the relationship between physical differences and the value of the glue. In addition, the strengths obtained vary with starting materials. Thus, cost of production can only be related to batches not individual grades. ITA did not dispute this fact but resorted to the "best information" rule simply because the manufacturers did not, even though they could not, supply the requested information for cost differences in production of different grades.

As we held in Olympic Adhesives, Inc., 899 F.2d at 1571:

The basic error we perceive arises from the ITA's overly sweeping view of the authority it is granted under section 1677e(b). In essence, the ITA interprets the phrase "whenever a party ... refuses or is unable to produce information requested" to cover, in the ITA's discretion, any inadequacy or insufficiency of a reply to a request for any type of information. Indeed, even where a reply is complete, the ITA may, as it did in this case, conclude that the information does not answer a question it wishes to resolve, and for that reason the party is deemed to "refuse" or "be unable to supply" information within the meaning of the statute. We cannot agree that the ITA's authorization under section 1677e(b) extends so far.

In this case as well, the ITA may not resort to the "best information" where the party's failure to give information is because the information does not and could not exist.

With respect to the evidence that glues of lesser strengths were sold in the same quantities for higher prices, we agree with U.H.F.C. that to require absolute consistency of prices would be absurd and out of step with commercial realities. Indeed, many variables go into the pricing equation, such as, the immediacy of a customer's need, the quantity ordered, the presence of a fixed-price contract, or the availability at the time of competitive products. Given the dynamics of the commercial marketplace, one would expect to find, as appears in this case, variations in the pricing of the same bloomgram strength and some sales of higher bloomgram strength glues at prices lower than prices of lesser strength glues. The government's reference to one instance where 300 bloomgram strength glue was sold during the review period for a lower price than an identical quantity of 210 bloomgram strength glue is comparable to finding one bad apple and concluding all in the bushel are spoiled. The record of home market sales during the review period indicates that the majority of sales of 230 bloomgram glue sold at prices between 3,65 and 3,85 dutch florins; the majority of sales of 300 strength glue at 4,25 and 4,45 florins; and the majority of sales of 400 strength glue at 5,75 florins. Thus, the record reveals, as the ITA and court acknowledged, that the value of any glue is directly related to a physical characteristic of that glue, namely the glue strength. In light of the relationship between price and physical difference, we find it necessary to remand to the court with instructions to direct the ITA to make price adjustments based on differences in physical characteristics between the merchandise sold in the home market and in the United States.

C. Choice of "Similar Merchandise" for FMV to Use in Calculation of Dumping Margin

With respect to Appellant's challenge to the ITA's selection of 300 strength glue sold in the Netherlands as the FMV for comparison with the USP of 250 strength glue in calculating the dumping margin, our decision reversing and remanding for determination of price adjustments based on differences in physical characteristics between the similar glue grades requires a redetermination of FMV after adjustments are made. Thus, the issue whether the ITA's selection of nonadjusted 300 grade glue over nonadjusted 230 grade glue as the FMV for comparison with the USP of 250 grade glue is rendered moot. See Advance Transformer Co. v. Levinson, 837 F.2d 1081, 1084, 5 USPQ2d 1600, 1603 (Fed.Cir.1988); Fonar Corp. v. Johnson and Johnson, 821 F.2d 627, 634, 3 USPQ2d 1109, 1114 (Fed.Cir.1987), cert. denied, 484 U.S. 1027, 108 S. Ct. 751, 98 L. Ed. 2d 764 (1988).

That part of the judgment based on the Court of International Trade's decision affirming the ITA's use of home market sales in the Netherlands is affirmed because all home market glue grades were properly determined to be "such or similar merchandise" and the quantity of home market sales was properly found to be sufficient to serve as the basis for FMV. That part of the judgment based on the court's decision that the ITA properly denied price adjustments based on physical differences is reversed because the ITA was not justified in resorting to the "best information" rule. Further, because the record reflects that the value of glue is directly related to the strength of the glue, the case must be remanded to the court with instructions to direct ITA to recalculate the dumping margin after making price adjustments for differences in physical characteristics between each glue strength.

Costs are awarded to Plaintiff-Appellant.

AFFIRMED-IN-PART, REVERSED-IN-PART, AND REMANDED.