dissenting.
I respectfully dissent.
Affirming this final countervailing duty determination, as my two esteemed colleagues have voted to do, threatens to revive an old International Trade Administration (“ITA”) interpretation of the statute’s specificity requirement previously abrogated under settled law. Indeed, ITA itself abandoned this misinterpretation five years ago after its “general availability test” was rejected by the Court of International Trade (“CIT”) in Cabot Corp. v. United States, 9 CIT 489, 620 F.Supp. 722 (1985). Ever since, both the administering agency and the specialized trade court have agreed that the former ITA interpretation was incorrect, as well as on what the correct test is. The CIT has remanded cases in which ITA applied its old interpretation. Finally, this old interpretation clearly contravenes Congressional purpose as expressed by Congress itself, which explicitly disap*1580proved it in 1988, while approving the new test ITA uses today.
In this case, however, instead of exercising discretion to determine if the discount gas sales to the defendant float glass manufacturers constituted a bounty or grant by applying the correct test to the facts, ITA declined to investigate the relevant facts and simply defined the subsidy out of existence with its old, statutorily incorrect general availability test. Therefore, this is not an appeal that can properly be resolved by deference to agency discretion.
While I believe this appeal is resolved with the wrong result, my greater concern is that this affirmance threatens to unsettle the law. Since this is our first decision in an appeal on the question, its unsettling effect may be compounded. Nor is concern lessened because there is no opinion of the court and hence no opinion with prece-dential force. Since Senior Judge Smith, while voting to affirm the CIT result, did not join in Chief Judge Nies’ opinion, she speaks only for herself, just as I speak only for myself. This jurisprudential truth, however, likely will be lost when the bar reads and cites that opinion. After all, facts and reasons are described and a two-judge result is noted; together, they imply a precedential holding.
The result here conflicts not only with the “specificity” requirement of the 1979 statute as expressly and definitively construed by ITA and CIT, but also with the result in every other ITA investigation and CIT decision since 1985. It is a result based on an agency analysis that clearly did not measure the “case by case effect of benefits provided to recipients” — here, the discount natural gas sales — as Cabot and the other authorities explicitly require. See 620 F.Supp. at 732.
On the contrary, ITA applied only a “general availability” of benefits test, which Chief Judge Nies agrees is statutorily incorrect. Although there is a suggestion it relied on a “nominal general availability” analysis, clearly, ITA relied only on price, determining it was the same for all. That is the very definition of a generally available benefit. Nominal availability was never an issue in this case as all other industries not only can but actually do buy the discount gas at the uniform, published price. Nor, in any event, has the Cabot test been limited by CIT to piercing nominal general availability. Certainly, Judge Carman in PPG Industries itself does not rely on any such limitation, as he did, properly, in Cabot where only three companies could even use the heavy fuel nominally available to all companies.
Similarly, there is a suggestion that this case presents the dichotomy between de jure and de facto analysis. Since under the program the gas sales are made to all industries, however, this case involves only de facto analysis. The true issue here is not whether de facto analysis is required but what such a case by case analysis includes.
The CIT has answered that question. However, by affirming in this appeal, our court weakens a six-year consistent line of CIT cases, starting with Cabot, to which PPG Industries is the sole exception. These cases hold that ITA must look beyond general availability and must assess the “effect of benefits” on the “recipients” on a “case by case” basis, particularly whether the subsidies disproportionately benefit a foreign company or industry compared to others there, thereby creating a significant “competitive advantage” over companies here. The mandate of Cabot is clear: ITA is required to undertake a factual investigation comparing the effects of the benefit on the foreign party named in the petition with those on other recipients of the same subsidy. Cabot, 620 F.Supp. at 732 (“The appropriate standard focuses on the de facto case by case effect of benefits.”).
The logic of this reasoning is undeniable.
Even in PPG Industries, the sole aberration, the CIT (Carman, J.) started from the same premise, i.e., that to avoid
the absurdity of a rule that transforms an obvious bounty into a non-countervail-able benefit by making the program “generally available” ... the appropriate standard [must focus] on the de facto case by case effect of benefits provided *1581to recipients rather than on the nominal availability of benefits.
PPG Indus., Inc. v. United States, 11 CIT 344, 662 F.Supp. 258, 265 (1987) (quoting Cabot Corp., 620 F.Supp. at 732) (emphasis added in part). As a result, at least where, as here, benefits are not just nominally available to but are actually received by many, other factors must be considered:
Although general availability may be a manifestation that a program has not conferred a benefit upon a specific recipient, general availability is not the statutory test. It is merely one of several relevant factors to be considered in determining whether or not a benefit or competitive advantage has been conferred upon a “specific enterprise or industry, or group of enterprises or industries.”
Id. (citation omitted) (emphasis added). This multi-factor test, based on Cabot, also written by Judge Carman, has been consistently followed by the CIT. See, e.g., Roses, Inc. v. United States, 743 F.Supp. 870 (Ct.Int’l Trade 1990); Armco Inc. v. United States, 733 F.Supp. 1514 (Ct.Int’l Trade 1990); Comeau Seafoods Ltd. v. United States, 724 F.Supp. 1407 (Ct.Int’l Trade 1989).
In Cabot itself, Judge Carman explained that even
generally available benefits, when actually bestowed, may constitute specific grants conferred upon specific identifiable entities, which would be subject to countervailing duties.
Cabot, 620 F.Supp. at 731 (emphasis added in part). Gauging whether countervailable benefits are “actually bestowed” and “grants conferred” requires first, an inquiry whether the benefits are “specific [and] quantifiable” and can be traced to “specific identifiable entities,” “specific industries” or a “specific class.” Id. at 731-32. The Cabot test next requires “determining if [the subsidy] amounts to an additional benefit or competitive advantage.” Id. at 732. If so, the program likely is countervailable under the statute. Id. But none of these Cabot inquiries was made here.
Nevertheless, the government argues that this test “has always been applied [by ITA] in the same manner.” Indeed, as ITA itself acknowledged in 1986 (post-Cabot) in Certain Softwood Lumber Products from Canada, 51 Fed.Reg. 37453, 37456 (1986) (Preliminary Affirmative Countervailing Duty Determination), it must conduct a multi-factor test of effects, which may include an examination, as PPG insists was necessary here, of disproportionate beneficiaries or dominant users. In 1989, ITA further acknowledged that in determining whether a domestic program was counter-vailable, consideration must be given to “whether certain enterprises, industries, or groups thereof receive disproportionately large benefits under a program.” Countervailing Duties, 54 Fed.Reg. 23366, 23379 (1989) (Proposed Rules). Answering that question can only require comparing the effects of the subsidy on one foreign company or class with those on another. Indeed, it is clear that in some cases a program may so disproportionately benefit a single company or class that on that factor alone, a countervailable subsidy could be found. Thus, in individual cases, disproportionate benefit can indeed be “in and of itself controlling.” In its 1984 final determination, on which this appeal is based, however, ITA explicitly failed to make such a comparison, or, indeed, to inquire beyond general availability:
We verified that the float glass companies paid the published price for natural gas which was available to all industries, and therefore received no benefit.
Unprocessed Float Glass From Mexico; Countervailing Duty Determination, 49 Fed.Reg. 23097, 23099-100 (1984) (emphasis added) (Final Affirmative Countervailing Duty Determination). Thus, contrary to the government, this test has not “always been applied in the same manner." In fact, before Cabot in 1985, it was never applied. This investigation was pre-Cabot.
It is suggested that ITA needed only to investigate PPG’s charge of preferential pricing and when that charge was shown to *1582be untrue, needed do nothing more. But PPG also charged that published price was a countervailable subsidy to these defendant float glass manufacturers as individual companies and also as members of a larger class. PPG then does not argue that the subsidy is countervailable “based merely on the recipients of the subsidy being identifiable.” Thus PPG does indeed seek reversal as a matter of fact as well as law. Moreover, under the statute, PPG has no burden of proof. Therefore, once PPG’s petition was accepted, ITA was obliged to investigate the facts that are relevant under the statutory test. Here, it declined to do so.
Despite quoting Cabot’s requirement that ITA assess the “case by case effect of benefits provided to recipients,” PPG Indus., 662 F.Supp. at 265, the CIT, in reviewing ITA’s final determination in PPG Industries, concluded, illogically, that “[sjince the ITA determined [that the float glass manufacturers] paid the published price for natural gas which was available to all industries, ... ITA’s determination was in substantial accordance with the statutory standard.” Id. at 272 (emphasis added). Yet, Judge Carman himself says earlier in PPG Industries that general availability “is not the statutory test.” Id. at 265 (emphasis added). Thus, like Chief Judge Nies, Judge Carman addressed only the preferential pricing allegations of PPG’s ITA petition and CIT complaint, while ignoring the allegation that sales to these defendants at the published price were themselves a countervailable subsidy.
Nor may ITA properly rely in this case on its earlier determination in Anhydrous and Aqua Ammonia from Mexico, 48 Fed. Reg. 28522 (1983) (Final Negative Countervailing Duty Determination). There, ITA determined that no countervailable subsidy existed solely on the ground that:
[sjince all industrial users of natural gas can obtain this good at the same price, gas is not provided to a “specific enterprise or industry, or group of enterprises or industries” under section 771(5)(B) of the Act.
Id. at 28524 (emphasis added). Thus, there too ITA plainly failed to look beyond general availability, but merely confirmed that the defendant manufacturers paid the same energy price as all other industries in Mexico. That analysis was as defective as this one and for the same reason: Instead of investigating, ITA defined the issue away. Nor, as the other opinion suggests, did PPG have some burden to come forward with “additional information ... to cause ITA to review” its original determination that the natural gas program was “not countervailable as a general proposition,” since that determination was based entirely on an incorrect statutory interpretation, and since PPG as petitioner has no burden to plead specific effects on defendants anyway. Similarly, PPG had no burden to prove that defendants were part of a “class” of “energy-intensive” industries, for it had alleged that these defendants, as individual enterprises, received a counter-vailable benefit. Nor was PPG required to somehow equate the defendant float glass manufacturers with “de facto sole recipients” in other countervailing duty cases. Under Cabot, all PPG was required to do was identify the company or class allegedly receiving the countervailable benefits. Whether or not, under a “case by case” analysis, a class including defendants is “comparable” (in size) with the three-company class in Cabot is irrelevant. Moreover, to reason, as the other opinion does, that ITA need not review a program because in a prior determination the program was found “not countervailable as a general proposition,” clearly contravenes the rationale behind Cabot’s “case by case” requirement.
Since ITA plainly did not apply to defendants here a “specificity analysis,” measuring the “case by case effect of benefits provided to recipients,” to affirm its determination is wrong as a matter of law. Moreover, our affirming the CIT’s erroneous holding in PPG Industries undermines subsequent CIT holdings, which uniformly and correctly applied the Cabot test.
For example, in Roses, Judge Restani specifically applied Cabot to overturn the ITA determination, holding that:
*1583[T]he appropriate test in the current investigation should have been whether a competitive advantage in fact was bestowed on a specific enterprise or industry, or a group thereof, by the program at issue. Thus, the general availability rule under which ITA conducted the investigation was flawed.
Roses, 743 F.Supp. at 879 (emphasis added). Similarly, in Armco, the court, applying Cabot, reversed ITA’s determination that a depreciation program was noncountervaila-ble because 'it was available to all industries depreciating similar assets. Such an analysis, the court explained, “misses the point”:
[I]f a particular benefit did constitute an unfair, countervailable subsidy when granted to a certain company or companies, the fact that other companies were also eligible for the benefit, or even that other companies also utilized the benefit, would not, as the Cabot Court pointed out, change the fact that the benefit was unfair, and thus counter-vailable, when granted to the first company.
Armco, 733 F.Supp. at 1530 (emphasis added). To the same effect is Comeau Seafoods Ltd., 724 F.Supp. at 1414 (“[T]he test is a de facto case by case analysis of the benefits conferred.”).
What vitality will those and similar CIT decisions have in view of our court’s result in PPG Industries?
Finally, ITA’s incorrect determination, based on a statutorily incorrect test, contravenes the clear intent of Congress, which has adopted the Cabot “effect of benefits” test as reflective of its original intent. Congress amended section 1677(5)(B), originating in 1979, in the Omnibus Trade and Competitiveness Act of 1988, adding a “Special Rule”:
In applying subparagraph (A) [which includes the specificity language], the administering authority, in each investigation, shall determine whether the bounty, grant, or subsidy in law or in fact is provided to a specific enterprise or industry, or group of enterprises or industries. Nominal general availability, under the terms of the law, regulation, program, or rule establishing a bounty, grant, or subsidy, of the benefits thereunder is not a basis for determining that the bounty, grant, or subsidy is not, or has not been, in fact provided to a specific enterprise or industry, or group thereof. (Emphasis added).
The legislative history, moreover, clearly indicates that Congress explicitly endorsed the test required in the Cabot case, H.R. Rep. No. 40, Comm. on Ways and Means, 100th Cong., 1st Sess. (H.R.Rep.No. 40), at 123-24 (“This [amendment], in effect, codifies the finding of the [CIT] decided in Cabot Corp. v. United States____ The test ... is whether there is a sufficient degree of competitive advantage in international commerce being bestowed upon a discreet class of beneficiaries that would not exist but for government action.” (emphasis added)). The purpose in enacting the Special Rule was not to change existing law, but to ensure that ITA properly applied the 1979 statute. According to the House Committee Report, ITA “had been interpreting the concept of ‘specificity’ and ‘general availability’ in an unduly narrow manner.” Id. at 123. The Cabot test “provide^] a sound interpretive rule to be applied in those cases where broadly available benefits are at issue.” Id. at 124. The Special Rule, by endorsing the Cabot test, therefore, simply “clarifies the application of the countervailing duty law to domestic subsidies.” Conf.Rep. No. 576, Omnibus Trade and Competitiveness Act of 1988, 100th Cong., 2d Sess. at 587. Thus, as of 1989, both Congress in its Special Rule and ITA in its Proposed Rule set forth the correct test, which is presently applied. But the correct test was not applied here, so our decision here needlessly casts doubt upon whether it is indeed correct or required.
It is suggested that Congress, in adopting the clarifying Special Rule, only adopted a “portion” of Cabot. Nothing could be more clearly not so, as the above quotations demonstrate. It is also suggested that defeat in 1984 of an amendment to make natural energy subsidies per se countervailable proves that Congress’ *1584intent in 1979 on specificity could not cover energy subsidies even when the facts of the case warrant. The illogic of equating a per se rule of countervailability and a fact-based discretionary rule needs no elaboration. In any event, in adopting the Special Rule, Congress explicitly envisioned it being used in situations where natural energy, natural gas in particular, was the generally available benefit in question. H.R. Rep. No. 40 at 124.
Rather than upholding this incorrect determination, this court should insist that ITA apply the correct statutory test. Moreover, since the failure to assess the effects of these gas sales or look at anything more than general availability was clearly not harmless error, remand is our only proper course. Indeed, remand has been ordered in every CIT case where ITA had failed to apply the Cabot test, except in PPG Industries. See, e.g., Roses, 743 F.Supp. at 881 (on remand, ITA cannot rely on “its flawed ‘general availability’ test”); Armco, 733 F.Supp. at 1529-30 (remand necessary because ITA conducted only a general availability inquiry); Cabot, 620 F.Supp. 722 (remand required for the same reasons). The result here should likewise be a remand.
At the very least, I would have hoped our decision would not have confused the correct interpretation of the specificity requirement. Regrettably, its clarification will have to await another day and perhaps another case.