3.29 Canada explained that, in putting forward its case, it had borne in mind that the drafting history of Article I of the General Agreement suggested that the question of "like products" be examined on a case-by-case basis. Canada had reviewed the criteria used in previous Panel cases assessing "likeness". These criteria included: (a) practices of other contracting parties; (b) the physical origin and properties of the products; (c) treatment of the products in internal regulations by the importing country; and (d) the "end-use" of the product.

3.30 Canada considered that the obligation of Article I to accord equal treatment immediately and unconditionally applied both to "like products", as well as to the countries from which the like products originate. It was Canada's position that dimension lumber is a "like product", regardless of the species from which it is manufactured and Article I required equal tariff treatment.

3.31 Japan, pointing to the text of GATT Article I:1 itself, and to documents dealing with the drafting history and interpretation of the "like products" concept (UN -: EPCT/C/II/65, page 2; EPCT/C. II/PV.12, page 7 (1946); EPCT/C.II/36, page 8 (1946); E/Conf. 2/C.III/SR.5, page 4 (1947); GATT/CP/4/39, paragraph 8; IC/SR.9, page 2 (1953); and GATT:BISD 25S, pages 49-53 and BISD 28S, pages 92-98), stated that, in its view, Article I:1 was not intended to seek out an ideal tariff classification, but that, in cases of doubt, or disputes, concerning "like products", past deliberations had focused on whether, or not, most-favoured-nation treatment was extended to the products concerned, irrespective of countries of origin, and the tariff classification of the country in question had been examined to see if it was discriminatory. At the same time, the classification system of the country concerned was respected and, as Japan saw it, past Panel deliberations had resulted in judgments that two products were not "like products" if there were any practical differences between the products examined. Japan felt that attempts to determine likeness in an a priori manner, without adequate attention to tariff classification, would cause confusion in existing tariff classification systems and also in the context of tariff negotiations.

3.32 With reference to the latter two points, Japan feared that any moves to introduce tariff sub-classifications based on "end-use" criteria, would have the result that negotiators, when considering a concession-request on a given tariff position, would have to examine for "likeness", with the product covered by the requested position, all other products covered under any other tariff position, and, if there existed such "like" products, the negotiators would then have to decide whether, or not, they would be in a position, and willing, to grant the concession, bearing in mind reciprocity obligations and other relevant desiderata and requirements.

3.33 Canada explained that it could not accept that specific tariff classifications set out by the importing country should determine whether products were like or not. Acceptance of that position would imply a system whereby products which were not covered by the same tariff line could not be considered "like products". Applying the same logic, such an interpretation would also imply that goods covered by a single tariff line would be presumed to be like goods. The fact that goods were covered by a single tariff line most likely implied not more than that there was some sort of relationship amongst them. It would be no more logical to assume that goods covered under the same tariff line were "like products" than that goods under different tariff lines were not. This could be amply demonstrated, but it was also obvious from the case at hand, for example by reference to Japan's recent tariff reclassification of larch dimension lumber. Canada also could not accept assertions to the effect that separate classification for dimension lumber might threaten the stability of the tariff classification system, or that discussions regarding sub-divided items could impair the functioning of the entire tariff negotiation system in use based on tariff numbers. The greater danger, by far, to the trading system, in Canada's view, would be the acceptance of the Japanese argument that the tariff classification system should determine the "likeness" of products. This would provide countries with a means to avoid their Article I obligation with respect to like products, while leaving the affected exporting countries with no recourse.

3.34 Japan clarified in this connection that its stand was not that products classified in different tariff positions could not be "like" within the meaning of GATT. Obviously, different tariff classifications could be compared with a view to determining whether, or not, the products affected were like products within the meaning of GATT. Japan's position was that planed lumber, including dimension lumber, from SPF species, was not like planed lumber from other softwood species, in terms of product origin and -characteristics. In other words, it was not Japan's position that inclusion in different tariff classifications was a sufficient condition for "unlikeness", but that "likeness" under Article I:1 should be examined on the basis of tariff classifications. What was unacceptable to Japan was the "sub-classification" of Japan's Tariff, attempted by Canada, and the comparison for likeness of some of the products in different Japanese tariff classifications, chosen by the complainant.

3.35 As Japan understood the situation, Canada admitted that SPF planed lumber generally, and planed lumber of other coniferous species generally, were not like products. Yet, Canada was attempting to build a case by establishing within existing sub-positions of the Japanese Tariff sub-groups of goods with a degree of similarity (the so-called "dimension lumber"), so as to find allegedly "like products" that receive different tariff treatment, thereby forcing Japan into a concession that had not been negotiated. There was no precedent in GATT proceedings for such compulsory sub-classification. If a complainant in a GATT proceeding could engage in the type of sub-classification attempted by Canada, many nations' tariff schedules would be found to be rife with Article I violations. For example, various forms and shapes of silver, copper and aluminium would, it might be assumed, be considered as not being "like" products within the meaning of Article I:1. Consequently, contracting parties would be fully justified in applying different tariff rates to these products. If, however, tariff schedules could be sub-classified through GATT proceedings, a complainant might assert that silver-, copper- and aluminium wire, all of which conduct electricity (although with different levels of conductivity), must be treated the same for tariff purposes. Other examples could be imagined; for instance, while plywood and waferboard were not "like products", and Canada maintained different duty-rates on these products, a complainant might assert that plywood and waferboard used in the limited context of subflooring were "like" products. The fact that Canada applies a different tariff rate to oak flooring and other wood flooring might also be considered a GATT violation.

3.36 Canada explained that it was not attempting to create the "like product" obligation in Article I:1; it already existed. That obligation clearly referred to products, and more recent panel rulings had given less weight to tariff classification and greater consideration to other criteria. Support of the claim that the placement of a product in a tariff line should predominate over the actual characteristics of the product would create a dangerous precedent. While tariff classification was a criterion that had been examined by panels in relation to the "like product" concept in Article I, and in Article III, Canada considered that tariff classification was only one criterion, out of many, and that it should not be considered a definitive test. Similarly, while tariff classification practices of other countries was a criterion to be examined in reaching a judgment on whether or not products were "like", this was not synonymous with accepting the tariff classification system of a given country as a determinant of likeness, or otherwise.

3.37 Canada felt that in several of the past "like product" cases, the respective panels, and notably the Animal Feed Proteins Panel and the Alcoholic Beverages Panel, had considered the interpretation of the "like product concept" to be the same in both Articles. Therefore, in Canada's view, just as Article III of GATT imposed an obligation concerning the competitive conditions between imported and domestic products inside the market, so the like-product concept of Article I imposed an obligation concerning the competitive conditions between like products at the border. In the Canadian view the duty on SPF dimension lumber was an example of "tariff specialization". The duty applied to SPF dimension lumber was in stark contrast to the equal treatment of dimension lumber of all other species in the Japanese Tariff, and the equal treatment of all species (including SPF) in the internal laws and regulations of Japan. Canada recognized that, in general, equal treatment under domestic regulations would not necessarily imply that equal tariff treatment should be accorded, but, in this case, the equal treatment applied to dimension lumber was another part of the overall picture of dimension lumber being considered a single product, regardless of species. In addition, in this case, internal regulations governed end-use, which is a criterion in assessing likeness.

3.38 As a general principle, Canada supported the view that exceptions to GATT obligations should be narrowly construed, while the obligations themselves should be more broadly interpreted, to protect the rights of contracting parties. Canada stated that Articles I and III set out obligations intended to be trade-creating. Thus, in that context, a narrow definition of like products would not be appropriate. The concept itself would have to allow for a degree of difference while maintaining the sense of likeness.

3.39 Japan considered that the meaning of the "like products" concept stipulated in Article I and that of Article III were not the same. The objective of the "like products" obligation in Article I stipulated most-favoured-nation treatment, while in Article III the objective and the proviso related to national treatment. Japan considered that if the terms of reference of the Panel were to cover also Article III, there would be reason for referring to internal regulations, but they would be irrelevant with respect to the classification of "like products" within the meaning of Article I:1. Japan pointed out that in Japan's internal regulations dimension lumber had been recognized as comprising different products, based on species. It had also to be borne in mind that internal regulations had been established not so much on the basis of economic factors, as in the case of tariffs, but with regard to other objectives, such as safety. In other words, the purposes of such domestic regulations were completely different from the most-favoured nation principle in Article I:1, which obliged a country to provide equal treatment at the national border, irrespective of national origin. Japan could not accept the argument that the same tariff rate must necessarily be applied to products which are given equal treatment under domestic regulations. Japan also pointed out that Japan's laws and regulations, in fact, distinguished between different species in a number of ways.

3.40 Canada explained that "dimension lumber", of whatever species, had the same physical origin. More specifically , all dimension lumber manufactured in North America and imported into Japan was softwood of the order Coniferales, and virtually all of it was of the family Pinaceae. It was at the genus and species level that the Japanese tariff discriminated among dimension lumber. Thus, the tariff of 8 per cent applied to the genus Pinus, but only to certain species of the genera Abies and Picea; other species of these groups were exempt and entered duty-free. In Canada's view, it would be misleading to search further for differences in physical origin by references to "genera", since genera distinctions were not as clear-cut as might be assumed. Species from the genus Abies were found in both the SPF and the Hem-Fir groups (the latter duty-free), despite being very closely related biologically. Canada felt that the distinctions in Japan's Tariff were based more on geography than on biology.

3.41 As regards physical properties, Canada observed that Japan had contended that differences in strength between lumber of different species' groups dictated different end-uses in construction, including in platform -frame construction. Following that line of reasoning, one might speculate that the different strength-ratings of Douglas-Fir and Hem-Fir lumber had motivated their exemption from duty, as compared with SPF. Yet, certain dimension lumber which had lower "strength" than SPF, was also duty-free. Canada also doubted that possible differences in domestic availability of different lumber species could explain the species' differentiation in tariff treatment. Generally speaking, there existed a shortage of virtually all species of lumber in Japan. Approximately 60 per cent of Japan's total coniferous wood consumption was met through imports. Each year Japan imported large volumes of pine-, larch- and white-fir logs from the USSR, and hemlock-and Douglas-fir logs from the United States. Japan's domestic supply of silver-fir, white-fir and Japanese-larch derived from plantation forests was limited and, basically, it was not manufactured into dimension lumber, nor did it compete with the SPF dimension lumber required by Japan. Canada's case before the Panel dealt with dimension lumber, of which the domestic production, in any of the species, was negligible in Japan. Canada did not wish to suggest that there existed no differences in the physical characteristics of lumber of different species, for instance, in respect of such elements as colour, weight/density, handling and processing characteristics, resistance to decay, tendencies to warp, or shrink and, perhaps, still other elements. Certain differences in physical properties being acknowledged, it was, however, important, in any assessment of likeness, to take into account not only such observable differences, but to asses the importance of observed differences, the practical significance of which might be minor.

3.42 Japan, on the other hand, held the view that the differences were significant in practice, not minor, and that, because of these differences, the demand for SPF was limited and the available resources excessive. With regard to the import of SPF logs, Japan pointed out that imports of logs had a different impact on its processing industry in low income regions. Japan, therefore, felt that the issue was not directly relevant.

3.43 Canada explained that, after review of the physical properties of various species, including specific gravity, decay resistance, treatability, nail-holding capability, allowable stress and of the span-tables, it was clear that there was a random mixture of differences in physical properties among species. The mixture and relationship between the various species changed, depending on the actual characteristics being measured. In the opinion of the Canadian expert there was no justification for the Japanese contention that SPF lumber is of inferior quality. Although there might occur specific circumstances in which a larger size of SPF dimension lumber might have to be used in place of hem-fir, exactly the same circumstances could arise in comparing dimension lumber made from Douglas-fir to that made of hem-fir. These were limitations on the size of lumber, not the species per se, and could easily be accommodated at the design state. Any differences that might exist for platform-frame construction uses between dimension lumber of different species were minor, so that, in practice, different species' of dimension lumbers were completely interchangeable.

3.44 Canada did not deny that minor differences exist as between different coniferous wood species, but it considered that the differences were not such as to deny a finding of "like product". The Coffee Panel, as well as the Alcoholic Beverages Panel, had examined the question of "like products" and had compared products which had differences in physical characteristics. In both cases the Panels had determined that the minor differences in physical characteristics were not sufficient reason to allow for different tariff treatment. Canada considered that the physical differences between SPF dimension lumber and other species of dimension lumber were akin to the differences between types of coffee, or between vodka and shochu, which had been found by the respective Panels to be insignificant in determining likeness.

3.45 Japan recognized that criteria concerning physical origin and physical properties were of importance in examining whether or not a tariff classification was arbitrary and, thereby, could constitute a case of discriminatory treatment. Japan believed, however, that it was universally accepted that spruce-pine-fir (SPF) lumber and "hemlock lumber" (hemlock used, in short, for non-SPF lumber) had different physical characteristics and properties. Canada, in fact, had not claimed that lumber of SPF, as such, and lumber of hemlock, as such, were like products. The situation should not be any different when the respective lumber was planed. Canada's contention was that "dimension lumber" of SPF was not a like product to SPF lumber, nor was dimension lumber of hemlock "like" to hemlock lumber. On the other hand, SPF dimension lumber and hemlock dimension lumber, Canada claimed, were like products.

3.46 Japan explained that it considered that specific gravity and hardness were important in determining lumber usage. According to Japan, (western) hemlock had a specific gravity almost one-fifth higher than SPF (Engelman spruce) and a hardness 60 per cent higher; Douglas-fir had a hardness rating about 85 per cent higher. Compared with lodgepole Pine (in the SPF group) hardness ratings for (western) hemlock were 20 per cent higher and for Douglas-fir 40 per cent higher. In terms of allowable stress loads, Douglas-fir lumber had a rating almost 50 per cent higher than SPF lumber of the same size and grade. As a result of their different cell-structure, Douglas-fir and hemlock provided greater shear-resistance and nail-holding capacity (e.g. fewer nails required) than SPF lumber, and had also better natural decay resistance. Hemlock was also better suited to industrial preservative treatment than SPF lumber. The test data collected by Japan showed that there were significant differences in physical properties between species, not just "organoleptic" differences, or minor differences in strength. As regards the point made by Canada, that some non-SPF softwoods had lower stress coefficients than those for SPF, Japan stated that it would be mistaken to assume that Japan's tariff rate differentials had been established on the basis of differences in stress coefficients. Japan, in setting its tariff classification and -rates, had also to take into account domestic circumstances, such as the need for imports and for the protection of domestic forestry-products industries.

3.47 Japan believed that it was generally recognized that in the construction industry the strength characteristics of lumber, including "dimension lumber", were very important, both in design and in actual construction. General practice in Japan was that, in platform-frame construction, SPF "dimension lumber", because of its lower strength, was used in structural members that were perpendicular to the ground, whereas (high strength) Hem-Fir dimension lumber was used for horizontal structural members. Ground sills were constructed from "non SPF" lumber. In traditional (post-and-beam) construction in Japan, SPF-type lumber was used for roof rafters (requiring little strength), while "non-SPF" lumber was used for (roof) purlins.

3.48 Japan reasoned that the attempted analogy to the Coffee Panel was not appropriate since there were significant differences in the respective facts and circumstances. In the Coffee Panel case, Brazil's coffee, unlike coffee from other countries, was "almost entirely" subject to a higher duty. In the lumber case, Canada exported large quantities of lumber to Japan duty-free. Spain was the only country the Coffee Panel found to operate duty-rate differentiation among types of unroasted coffee. In the lumber case Japan was not the only country which distinguishes in tariff classification, and tariff-rate treatment, between wood and lumber by species. In the Coffee Panel case there had been no attempt by the complainant to create "like products" by introducing tariff sub-classifications of its own making into the tariff schedule of the importing country and, unlike in the Coffee Panel case, differences between lumber species were not just minor, or organoleptic, but properties of, and end-uses for, wood of different species were significantly different.

3.49 Canada, referring notably to the Coffee Panel and the Alcoholic Beverages Panel, stated that the end-use criterion would be influential in the determination of likeness. It was not required that the products be dedicated to a single end-use, but they had to have a similar relationship to the defined end-uses, hence the references in the Alcoholic Beverages Panel to end-uses that were "substantially identical" and "virtually identical". The intended end-use of dimension lumber was the construction of a 2X4 type building, in fact, the end-uses of all species of dimension lumber in construction were not "substantially/virtually identical", they were exactly the same. There was, admittedly, a minor, subsidiary end-use for dimension lumber in Japan, for export packaging. Japan contented that Hem-fir was the preferred dimension lumber for such packaging uses, but this was due solely to market-and price considerations (SPF being normally higher priced, and the 8 per cent duty on SPF was no help in that respect).

3.50 With reference to the end-use criterion, Japan considered that Canada's argument concerning the "likeness" of SPF dimension lumber and dimension lumber of other species amounted to saying that different species of dimension lumber, if length, size, weight, etc. were ignored, were totally interchangeable in construction. Japan did not consider this to be an acceptable position, or, otherwise, all materials for platform-frame buildings should be considered like products. In Japan's view, lumber in house construction, including also, to some extent, different hardwoods, had a number of uses that varied significantly by species. In any event this was the practice in Japan. If a building was designed to be built with a given species of lumber, the use of a different species could force a complete redesign of plans, with repercussions on the volume of lumber required and, perhaps, construction-time, -effort and -cost. Canada contended that the intended end-use of dimension lumber was for platform-frame construction. Japan noted that although this might be the intent, it was not necessarily the practice. In wooden box construction for export packaging there was considerable reliance on higher-strength Hem-fir lumber, including "dimension lumber", making it possible to construct smaller-sized boxes, or crating, thus permitting cost-savings in ocean-transport. In addition, Hem-fir lumber was also being used in traditional housing, civil engineering, and other uses.

3.51 With regard to the use of imported dimension lumber in Japan, Japan, on the basis of a study done for the purpose of this case, explained that such lumber had a variety of uses, namely: in post-and-beam housing construction, about 34 per cent; platform-frame housing, 52 per cent; packaging, 10 per cent; and in other uses, such as laminated lumber and civil engineering, 4 per cent. Canada observed that it could not accept the figures cited. In Canada's view the listed percentage figures could only have been calculated by using sizes of planed lumber that are not defined as dimension lumber in Canada, or in Japan.

3.52 Japan felt that consumers' perceptions, one of the criteria identified in past Panel cases, were also of importance in any determination of "likeness", noting also that differences in consumers' perceptions were often reflected in marked price differentials. This was the case in respect of different lumber species, and was accompanied by different use- habits and trading-patterns, and this was widely recognized, including in trade publications and advertising copy. It was relevant to notice that in the United States and in Canada, where presently there was no longer a differentiation in tariff-rate treatment of different softwood lumbers, prices in the market for softwood lumber items differed significantly as between species and species-groups. Japan also recalled that Canada had argued, in a 1986 countervailing duty case between Canada and the United States, that softwood lumbers of different species were objectively different, on account of their different physical characteristics, because of consumer's perception, channels-of-trade and -use.

3.53 Canada, referring notably to the approach adopted by the Alcoholic Beverages Panel, thought that it was important to assess "likeness", as much as possible, on the basis of objective criteria, including, in particular, composition and manufacturing processes of the product, rather than consumption habits or price differences.

3.54 In reply to a question raised by the Panel on the interpretation of Article I:1, Canada stated, at the first Panel hearing, that: "In Article I.1, the first obligation, that no discrimination shall be made between contracting parties, and the second obligation, that no discrimination shall be made between "like products", are both stipulated. Although the Japanese tariff did not discriminate against Canada in terms of country-based discrimination, Japan nevertheless discriminated between like products ("dimension lumber" of SPF and that of other softwoods). Japan, therefore, did not meet its second obligation". Basically the same point was made by Canada in its second submission to the Panel, namely:

Canada also maintained that the Article I obligation does not deal with tariff classifications and does not aim to impose any particular system for grouping of products, but, whatever the system used, it does impose an obligation to apply the same tariff treatment to like products. In this context, Canada drew attention to the conclusion of the Coffee Panel that "... whatever the (tariff) classification adopted, Article I:1 required that the same tariff treatment be applied to like products (paragraph 4.4)". Canada noted that Canadian exports of dimension lumber were mainly SPF, which was subject to duty, while most United States exports entered duty free. This gave preference to United States suppliers and discriminated against Canadian suppliers. This amounted to de facto discrimination between contracting parties.

3.55 Japan stated that it was convinced that the concept of "like products" provided in GATT clauses should not be construed independently of the context. As Japan understood it, the "like product" provision in Article I:1 was intended to prevent a contracting party from discriminating against any other contracting party",and not of determining which similar products were to be included in the same tariff classification, and thereby be accorded equal treatment, nor did Article I:1 aim at establishing the perfect tariff classification.

3.56 As Japan saw it, in all cases where the concept of "like product" under Article I.1 had been taken up in the past in the GATT, de facto or de jure discrimination between contracting parties had always been the criterion for judgment. Yet, in the present case, Canada seemed to argue that the Coffee Panel case had somehow expanded the scope of coverage of Article I:1, to include discrimination among the products of one nation. In Japan's view the gravamen of Brazil's complaint had been that the differentiated tariff treatment by Spain in respect of unroasted coffee had the effect of discriminating against Brazil. Discrimination existed because Brazilian coffee, unlike coffee from other nations, was "almost entirely" subject to the higher tariff. In para. 4.10 of its report the Coffee Panel had stated that "the tariff regime as presently applied by Spain was discriminatory vis-à-vis unroasted coffee originating in Brazil". The Panels decision was not based simply on discrimination vis-à-vis unwashed Arabica beans. Without the finding of discrimination among countries there could have been no judgement on the violation of Article I:1.

3.57 Japan stated that the Inventories of Standing Timber, by species, in the United States and Canada, indicated that both countries had large resources of timber of species subject to duty in Japan and also large resources of species that would not be subject to duty. Japan considered it unlikely, just like Canada, that the species distribution of these forest resources would be subject to significant changes, even in the long-term. As regards production of softwood lumber of different species, both Canada and the United States produced substantial quantities of lumber that would be subject to duty in Japan and substantial quantities of lumber that would be duty-free. There was no intention, whatsoever, by Japan to discriminate. The species of lumber that were subject to duty when imported into Japan were produced in many countries, not just Canada. The United States, New Zealand, Chile, Norway, Sweden, the Federal Republic of Germany, Korea Rep. and the USSR, among others, exported lumber to Japan that was subject to the 8 per cent duty. While it was true that the proportion of planed softwood lumber imports subject to duties was higher for Canada than for the United States, the proportion of planed lumber imports from Canada that was dutiable was lower than that for many other supplying countries and, for several years, following the establishment of the tariff-rate differential, Canada had been the largest supplier of duty-free planed lumber.

3.58 Japan stated that its imports of planed softwood lumber from Canada had grown from 150 thousand cubic metres in 1977 to 1.14 million cubic metres in 1987. Imports from Canada of these same woods subject to duty over this 10 year period grew from 21.9 thousand cubic metres to 348.4 thousand cubic metres, while the woods admitted duty-free grew from 128.3 - to 793.6 thousand cubic metres. Data furnished to the Panel by Canada, indicated, in Japan's view, that Canadian SPF dimension lumber had not only not been deprived of its market by US Hem-Fir dimension lumber, but Canadian kiln-dried SPF dimension lumber had seen its market share increase, in a growing overall dimension lumber market. Canada held a share close to 3/4th of total estimated dimension lumber imports in 1987. While imports from Canada of kiln-dried SPF dimension lumber (despite, as Canada had explained, certain technical difficulties in regard to the economics of kiln-drying dimension lumber from old-growth SPF stands) had risen, both in terms of market share and total import volume, imports of green Hem-fir dimension lumber had fallen significantly. The observed shift in the pattern of imports was attributable, according to Japan, to changes in market preferences in favour of kiln-dried lumber. Imports of kiln-dried Hem-fir dimension lumber were also rising.

3.59 With reference, inter alia, to the explanations provided, Japan was convinced (i) that Article I of the GATT only addresses measures that discriminate among contracting parties and that Japan's lumber tariffs do not discriminate among countries, (ii) that the attempt by Canada to sub-classify Japan's Tariff Schedule for finding "like-products" was unprecedented and inappropriate and (iii) that, even if one were to accept Canada's other arguments, different species of "dimension lumber" are "not like" within the meaning of Article I.

3.60 Canada noted that the questions of trade performance and market-share were irrelevant to GATT obligations. For example, in the EC/Canada Liquor Boards Panel (L/6304), the Canadian measures were found to be inconsistent, despite the fact that EC exporters had 55 per cent of the Canadian market. Canada also drew on the result of the Canada/Japan Leather Panel (BISD 27S, page 118) and noted that the motive of a contracting party for maintaining a discriminatory measure also was not relevant. A finding of non-conformity with Japan's GATT obligation constituted prima facie nullification and impairment.

4.1 The EEC explained that its interest related primarily to the general question of the interpretation of the notion of "like product" within the sense of Article I:1. In the EEC's view, "like products" could be interpreted: (i) to refer to products which are "like" in the economic sense, i.e. directly competitive or substitutable, or (ii) to refer to products which are "like" in the physical sense, taking essentially the tariff classification as the basis. The Community, for a number of reasons, was in favour of the second alternative.

4.2 The EEC recalled that the term "like product", or "like products", was used in a number of provisions of the General Agreement. The GATT drafting history confirmed that the term "like product" had different meaning in different contexts of the Draft Charter. Subsequent GATT practice indicated, as confirmed in the 1987 Alcoholic Beverages Panel report, that neither the General Agreement, nor the settlement of previous cases, gave any definition of such concept. The CONTRACTING PARTIES had, indeed, never developed a general definition of the term "like products". Past decisions on this question had been made on a case-by-case basis, after examining a number of relevant factors.

4.3 In the view of the Community, the term "like product" in Article I:1 had to be interpreted in the light of the objective of this fundamental provision of the General Agreement, i.e. to guarantee most-favoured-nation treatment. The objective was to avoid discrimination among other contracting parties, but not to avoid protective measures or a difference of treatment between imported and domestic products. It was, therefore, necessary to avoid an interpretation which would make the distinction between Article I and other provisions, such as Articles II, III and XI, unclear.

4.4 The importance of an interpretation of the notion of "like products" in Article I:1 which followed essentially the tariff classification became even more obvious in the light of the relationship between Articles I and II.

4.5 As it was recognized that Article I:1 applied in the same manner to consolidated and non-consolidated tariff positions, the interpretation of the notion of "like products" beyond a specific tariff position under which a concession had been granted could have the effect of extending the scope of the concession to products with respect to which it had not been negotiated. This would risk seriously putting into question the basis of negotiating tariff concessions, and would make it more difficult, if not impossible, to negotiate concessions for specific positions, or sub-positions, in areas where the same concessions were not envisaged with respect to directly competitive or substitutable products.

4.6 The Community was aware of the fact that a somewhat broader interpretation of the notion of "like products" had been accepted in the GATT in relation to Article III. However, even in the Article III context, it had been found that products should be considered as "like products" in view of their similar properties, end-uses and, usually, uniform classification in tariff nomenclatures. Moreover, Article I did not contain any reference to directly competitive or substitutable products as did the Note Ad Article III:2. In the context of Article III there existed good reason for a rule which was sufficiently wide to prevent any form of de facto discrimination against foreign products. On the contrary, in the context of Article I:1 a wide interpretation of the notion of "like products" could make trade liberalization more, rather than less, difficult, and could thus run counter to the objectives of the General Agreement.

4.7 For the reasons outlined above, the Community's position was that, in the area of tariffs, the erga omnes application of a tariff based on a specific tariff classification should be presumed to be in conformity with the MFN principle, unless the classification was demonstrated to be arbitrary in a manner which amounted to a de facto discrimination among foreign products from different sources. The Community, therefore, considered that it would be appropriate to accept, in principle, tariff classifications as providing the basis for the definition of "like products" for the purpose of Article I.

4.8 Commenting on the Community's submission, Canada explained that it was not claiming "likeness" between dimension lumber of different species on the basis of the concept of "directly competitive or substitutable products". In fact, Canada's approach in the present case related to the physical properties and other characteristics of likeness, but, for reasons outlined earlier, did not consider that tariff classification provided a criterion sufficient by itself for the determination of "likeness", and Canada, in its presentation, had shown the arbitrariness of Japan's argument in this regard and the discriminatory effect on Canadian exports of SPF dimension lumber vis-à-vis shipments of other dimension lumber from other sources.

4.9 Referring to the wider background of the evolving GATT jurisprudence, New Zealand observed that the interpretation of "like products" had generally been construed in a way as to prevent, what the Alcoholic Beverages Panel had called, "tariff specialization", namely the elaboration of ever-more detailed tariff descriptions which would have the practical effect of discriminating against other third-country suppliers. While tariff classification had been, and would -no doubt - continue to be, regarded as one of a group of criteria to be considered in determining "like product" in the context of Article I:1, tariff classification should not be considered a necessary, still less a sufficient, condition for such a finding, because of the dangers of allowing widespread abuse of the MFN clause through "breaking out" a tariff line into numerous specialized and essentially arbitrary categories. In New Zealand's view, this was precisely what Japan had done in the case at hand, creating, by the selection of the lumber types set apart for privileged customs treatment, an "area preference", the advantage of which, as New Zealand explained, was mainly reserved to the United States, taking account of the natural stands of species concerned.

4.10 In New Zealand's view, "dimension lumber" made from softwoods should be seen, using the Coffee Panel terminology, as being a well-defined and single product in terms of its "end-use". In Australia and North America carpenters used dimension lumber of a whole range of softwoods species completely interchangeably. In Japan, although the effective tariff discrimination did not help, carpenters also used dimension lumber made from SPF.

4.11 New Zealand recognized that there are differences in the properties of SPF timbers, both with respect to the imported species and the Japanese softwoods used for dimension lumber. Indeed, even within the same log, from the same tree, there would be observable differences in properties. The fact of difference was thus a trivial observation in terms of the provisions of the General Agreement's treatment of "like products". The term did not mean "identical product"; what mattered, was the relative significance of such differences. And these, in the present case, were minor, and not important enough to withhold a judgement that the products were like products.

4.12 Strength-capacities were, of course, crucial to dimension lumber. But the requirement essential to protect Japanese consumers was met by Japanese building standards, not the Japanese tariff. The product either met those standards set for strength properties, or it did not. If SPF dimension lumber was not as widely used at present in Japanese non-traditional construction as it could be, this suggested that the discrimination in the tariff treatment was tending to reinforce consumer preferences. As recent GATT secretariat work on tariff escalation had pointed out, even small differences in the tariff treatment could be a powerful disincentive to use the more highly dutiable product. The application by Japan of discriminatory duties on dimension lumber of the SPF species served as an area-preference. Both Canadian and New Zealand predominant timber exports of lumber were at a competitive disadvantage because of this discriminatory treatment.

4.13 The licensed Japanese carpenter constructing traditional post-and- beam houses had complete freedom in the selection of materials, framing methods, assembly and finishing methods. Species substitution was usually not a problem. On the other hand, tradition and custom were major factors acting against species other than those traditionally used. While this was a matter for market-education on the usages of species of "like woods", the ability to penetrate the market and provide such education was determined to a large degree by the competitiveness, in price, of the product. The discriminatory tariff applied to SPF and pinus radiata had a substantial initial inhibitory effect on entry into this particular market. The returns from outlays on market education would need to be potentially available before such outlays would take place. The discriminatory tariff was thus a substantial inhibiting factor to the marketing of Canadian and New Zealand timbers. The penalty tariff imposed on SPF and radiata dimensional lumber effectively gave the US a preferential supplier status.

4.14 With respect to the linkage between Article I:1 and Article II, New Zealand considered that a finding by the Panel that dimension lumber made from SPF was a "like product" would simply reinforce the critical relationship between the two Articles in a beneficial not harmful way. If the Panel agreed that the products are "like products", this would, of course, create an obligation to extend equivalent substantive tariff treatment to SPF dimension lumber as accorded-dimension lumber made from hemlock. It would not create an obligation to harmonize the classification of the tariff. Japan would continue to be free to classify softwoods imports how it wished, provided the imports of softwoods dimension lumber were introduced into the commerce of Japan at the same tariff rate.

4.15 The EEC, with reference to its own submission, explained why it could not agree with New Zealand's interpretation of the "like products" concept in Article I:1 and the way in which New Zealand considered it to interact with provisions of other Articles of the General Agreement, notably Articles II and III.

4.16 Japan recalled that, in refutation of the Canadian submission, they had explained that Japan did not claim that products in different tariff classifications could not be examined for "likeness", but that it was necessary to respect existing tariff classifications. Similarly, the question of likeness had to be considered in the appropriate GATT context, namely, in the present case, the respect of the MFN principle. Japan's species-based lumber tariffs were not intended to "specialize", with a view to eluding the like products obligations of Article I, but the tariffs had been determined in view of the needs for the protection of domestic lumber-related industries. Also, SPF lumber and lumber of other coniferous species were not "like" products in term of physical-origin and -properties. Regarding the limited use of radiata pine in housing-construction, attributed by New Zealand to the extra-cost element resulting from the imposition of a duty of 8 per cent, Japan felt that it was not so much the tariff, but builder's experience, which limited the use.

5.1 Under its terms of reference established by the Council the Panel had to examine a complaint by Canada that the application of a tariff duty of 8 per cent on imports of spruce-pine-fir (SPF) dimension lumber by Japan is not in conformity with the provisions of Article I:1 of the General Agreement and nullifies and impairs benefits accruing to Canada under the General Agreement, SPF dimension lumber being a "like product" as compared to other types of dimension lumber entering Japan with zero duty.

5.2 The Panel noted that Heading No. 4407.10 of the Japanese Tariff was defined in conformity with the International Convention on the Harmonized Commodity Description and Coding System of 14 June 1983 (hereinafter called the Harmonized System) as follows:

4407 "Wood sawn or chipped lengthwise, sliced or peeled, whether or not planed, sanded or finger-jointed, of a thickness exceeding 6 mm":

4407.10 "Coniferous"

5.3 The Panel further noted that the dispute before the Panel largely focused on the consistency with Article I:1 of the 8 per cent tariff imposed by Japan under sub-position 4407.10-110 of the Japanese Tariff:

4407.10-110 1. "Of Pinus spp., Abies spp. (other than California red fir, grand fir, noble fir and pacific silver fir) or Picea spp, (other than Sitka spruce), not more than 160mm in thickness, (1) planed or sanded."

5.4 According to Canada, Article I:1 required Japan to accord also to SPF dimension lumber the advantage of the zero tariff granted by Japan, under sub-position 4407.10-320 of its Tariff, to planed and sanded lumber of "other" coniferous trees, including the genera cedar and other Chamaecyparis, hemlock (Tsuga), and douglas-fir (Pseudotsuga), and five species excluded from sub-position No. 4407.10-110.

5.5 The Panel noted that the tariff classification for 4407.10-110 had been established autonomously by Japan, without negotiation.

5.6 The terms of Article I:1, as far as they are relevant to the issue, read as follows:

5.7 In view of analysing the factual situation submitted to it under its terms of reference, the Panel had first to consider the legal framework in which the Canadian complaint had been raised. In substance, Canada complains of the fact that Japan had arranged its tariff classification in such a way that a considerable part of Canadian exports of SPF dimension lumber to Japan was submitted to a customs duty of 8 per cent, whereas other comparable types of dimension lumber enjoy the advantage of a zero-tariff duty. The Panel considered it impossible to appreciate fully the Canadian complaint if it had not in a preliminary way clarified the bearing of some principles of the GATT-system in relation to tariff structure and tariff classification.

5.8 The Panel noted in this respect that the General Agreement left wide discretion to the contracting parties in relation to the structure of national tariffs and the classification of goods in the framework of such structure (see the report of the Panel on Tariff Treatment of Unroasted Coffee, BISD 28S/102, at III, paragraph 4.4). The adoption of the Harmonized System, to which both Canada and Japan have adhered, had brought about a large measure of harmonization in the field of customs classification of goods, but this system did not entail any obligation as to the ultimate detail in the respective tariff classifications. Indeed, this nomenclature has been on purpose structured in such a way that it leaves room for further specifications.

5.9 The Panel was of the opinion that, under these conditions, a tariff classification going beyond the Harmonized System's structure is a legitimate means of adapting the tariff scheme to each contracting party's trade policy interests, comprising both its protection needs and its requirements for the purposes of tariff- and trade negotiations. It must however be borne in mind that such differentiations may lend themselves to abuse, insofar as they may serve to circumscribe tariff advantages in such a way that they are conducive to discrimination among like products originating in different contracting parties. A contracting party prejudiced by such action may request therefore that its own exports be treated as "like products" in spite of the fact that they might find themselves excluded by the differentiations retained in the importing country's tariff.

5.10 Tariff differentiation being basically a legitimate means of trade policy, a contracting party which claims to be prejudiced by such practice bears the burden of establishing that such tariff arrangement has been diverted from its normal purpose so as to become a means of discrimination in international trade. Such complaints have to be examined in considering simultaneously the internal protection interest involved in a given tariff specification, as well as its actual or potential influence on the pattern of imports from different extraneous sources. The Canadian complaint and the defence of Japan will have to be viewed in the light of these requirements.

5.11 "Dimension lumber" as understood by Canada is defined by its presentation in a standard form of measurements, quality-grading and finishing. It appears from the information provided by Canada that this type of lumber is largely used in platform-house construction in Canada as well as in the United States and that it has found also widespread use in Japan, as is testified by the existence of a Japanese technical standard known under the name of "JAS 600".

5.12 Japan objected to this claim on different grounds. Japan explained that dimension lumber was only one particular type of lumber among many other possible presentations and that house-building is only one of the many possible uses of this particular kind of lumber. From the legal point of view, Japan contended that the concept of "dimension lumber" is not used either in any internationally accepted tariff classification, or in the Japanese tariff classification. In accordance with the Harmonized System, position No. 4407.10 embraces all types of coniferous wood "sawn or chipped lengthwise ... exceeding 6mm". Apart from the thickness and the grade of finishing, customs treatment of lumber according to the Japanese Tariff was determined exclusively on the basis of a distinction established between certain biological genera or species. Dimension lumber was therefore not identified as a particular category in the framework of the Japanese tariff classification.

5.13 The Panel considered that the tariffs referred to by the General Agreement are, quite evidently, those of the individual contracting parties. This was inherent in the system of the Agreement and appeared also in the current practice of tariff negotiations, the subject matter of which were the national tariffs of the individual contracting parties. It followed that, if a claim of likeness was raised by a contracting party in relation to the tariff treatment of its goods on importation by some other contracting party, such a claim should be based on the classification of the latter, i.e. the importing country's tariff.

5.14 The Panel noted in this respect that "dimension lumber" as defined by Canada was a concept extraneous to the Japanese Tariff. It was a standard applied by the Canadian industry which appeared to have some equivalent in the United States and in Japan itself, but it could not be considered for that reason alone as a category for tariff classification purposes, nor did it belong to any internationally accepted customs classification. The Panel concluded therefore that reliance by Canada on the concept of dimension lumber was not an appropriate basis for establishing "likeness" of products under Article I:1 of the General Agreement.

5.15 At the same time, the Panel felt unable to examine the Canadian complaint in a broader context, as Canada had declared expressly that the issue before the Panel should not be confused by broadening the scope of the Panel's examination beyond 'dimension lumber' to planed lumber generally. Canada's complaint was limited to the specific product known in North America, and also in Japan, as dimension lumber. Canada did not contend that different lumber species per se should be considered like products, regardless of the product-form they might take (see para. 3.15 above). Thus there appeared to be no basis for examining the issue raised by Canada in the general context of the Japanese tariff classification.

5.16 In these circumstances the Panel was not in a position to pursue further the questions relating to the concept of "like products" in the framework of Article I:1 of the General Agreement.

6.1 In the light of the considerations set out in Section V above, the Panel could not establish that the tariff treatment of Canadian dimension lumber applied by Japan under its tariff number 4407.10-110 was inconsistent with Article I:1 of the General Agreement.