3. Potential Competition

41. The European Communities stresses that Article III:2, second sentence, is concerned not only with tax differentials between products that are already actually competitive or substitutable on a given market, but also with tax differentials between products that are potentially competitive or substitutable. Furthermore, the notion of potential competition must be deemed to include not only competition that would exist "but for" the tax measures at issue, but also competition that could be reasonably expected to develop in the future having regard, for example, to existing trends in the market concerned or to the situation prevailing in other markets.
42. The European Communities explains that the relevance of potential competition, as defined above, flows from the well-established principle that GATT Article III does not protect export volumes but competitive opportunities. As stated by the Appellate Body in Japan � Taxes on Alcoholic Beverages II:

[I]t is irrelevant that the "trade effects" of the tax differential between imported and domestic products, as reflected in the volumes of imports, are insignificant or even non-existent. Article III protects expectations not of any particular trade volume but rather of the equal competitive relationship between imported and domestic products. ⁵²

43. The European Communities states that the recognition of the relevance of potential competition for the purposes of Article III:2, second sentence, is of particular importance in the present dispute. In the first place, pisco has benefited from protective taxation for a long time. As a result, the current level of actual competition between pisco and other spirits is necessarily less than the level that could have developed under equal tax conditions. Secondly, distilled spirits, like many foods and beverages, are so-called "experience goods", i.e., goods which must be purchased and consumed in order to appreciate their aptitude to satisfy the consumers' needs. Furthermore, consumption of spirits is largely based on habits, which only change gradually. For those reasons, market penetration is generally slow and short-term reactions to price changes tend to be relatively low.
44. Chile replies that the European Communities offers as an excuse that cross-elasticities are often low for products that are new to the market. Such a comment hardly seems appropriate to the case of whisky, which has a very long history in the Chilean market.

4. Relevant Factors and Their Evidentiary Weight

(a) Relevant Factors

45. The European Communities points out that in Japan � Taxes on Alcoholic Beverages II, the Appellate Body noted that "how much broader [the] category of 'directly competitive or substitutable products may be in a given case' is a matter for the panel to determine based on all relevant facts in that case".⁵³
46. Also, according to the European Communities, in the same report, the Appellate Body referred specifically to the following factors as being relevant for assessing whether two products are "directly competitive or substitutable"⁵⁴:

1. the physical characteristics of the products;
2. their end-uses;
3. their tariff classification; and
4. the "market place", and in particular the degree of elasticity of substitution between the products concerned.

47. In response, Chile states that the question whether products are like, directly competitive or substitutable under various GATT rules has provoked many disputes over the years, but has not led to the development of any particularly clear or objective standards to guide WTO member governments or their legislatures. The case by case approach of the panels to date may be thought desirable, insofar as it leaves panels considerable discretion for judgement as to what are permissible and impermissible distinctions, but the cost of that discretion is a loss in predictability and certainty for national governments and legislatures.
48. Thus, Chile argues that panels have taken the approach of considering a variety of factors (none of which are considered individually dispositive) rather than applying any readily apparent formula. In reading the various decisions, it is impossible to avoid the sense that there is as much intuition as science lying behind past GATT and WTO decisions. Under such a standard, Chile disagrees with the EC claim that its finest single malt whiskeys and its rarest cognacs must be viewed as directly competitive or substitutable with Pisco corriente.
49. Chile states that the EC argument in this case nominally tracks the factors that the Appellate Body endorsed in the Japan - Taxes on Alcoholic Beverages II case. Those factors include such matters as physical characteristics, common end-uses, and tariff classifications, but also the "market place".
50. Chile argues that in sum, each supposed element of the EC case appears ill-founded. While perhaps every factor need not be present with equal force to find two products to be directly competitive or substitutable, every element should not be so weak as the evidence presented by the European Communities. Accordingly, the Panel should find that the European Communities has not shown that pisco and other distilled spirits are directly competitive or substitutable, and that the Transitional System is consistent with Article III:2, even for the brief period it remains in force.

(b)Evidentiary Weight

(i) Physical Characteristics

51. The European Communities points out that as illustrated by the drafting history of Article III:2 and past panel reports, two products need not have the same physical characteristics in order to be "directly substitutable and competitive". As noted by the Panel Report on Japan - Taxes on Alcoholic Beverages II :

[C]ompetition can and does exist among products that do not necessarily share the same physical characteristics. In the Panel's view, the decisive criterion is whether they have common end uses ... ⁵⁵

52. According to the European Communities, at the same time, however, it is obvious that if two products have sufficiently similar physical characteristics, such similarity may in and of itself be sufficient to conclude that the products in question are apt to serve for the same end-uses and, therefore, that they are "directly competitive or substitutable".
53. The European Communities argues that in order to be "directly substitutable and competitive" two products need not have the same physical characteristics. In fact, if there were no differences in physical characteristics between two products, they would be "like" instead of "directly competitive or substitutable". The decisive criterion for determining whether two products are "directly competitive or substitutable" is whether they have common end-uses. At the same time, it seems obvious that if two products have sufficiently similar physical characteristics, such similarity may in and of itself be sufficient evidence to conclude that the products in question are apt for the same end-uses and, therefore, that they are "directly competitive or substitutable".
54. According to the European Communities, the existence of differences in physical characteristics between Japanese shochu and the other distilled spirits did not prevent the two panels on Japan � Taxes on Alcoholic Beverages I and II from concluding that they were directly "competitive and substitutable" in the Japanese market. Likewise, in Korea � Taxes on Alcoholic Beverages, the panel concluded that soju was "directly competitive or substitutable" with a number of other distilled spirits, despite the existence of differences with respect to factors such as alcohol content, ageing, colour or flavouring. ⁵⁶ The differences between pisco and other distilled spirits are similar to those existing between shochu or soju and other distilled spirits. Therefore, those differences cannot exclude of themselves a finding that pisco is "directly competitive or substitutable" with other distilled spirits.
55. In rebuttal, Chile states that not all physical characteristics of the products involved need to overlap for the products in question to be considered "directly competitive or substitutable". However, Chile does consider that the European Communities needs to demonstrate more than a coincidence of one physical characteristic, in this case the alcohol content of the beverage, before claiming that products are directly competitive or substitutable in the sense required by Article III.

(ii) End-uses

56. The European Communities states that as observed by the panel on Japan � Taxes on Alcoholic Beverages II in the above quoted passage, commonality of end-uses is the "decisive criterion" for establishing whether two products are "directly competitive or substitutable". In fact, the other criteria are relevant only to the extent that they can provide an indication as to the existence of such commonality of end-uses.
57. According to the European Communities, in order to be "directly competitive or substitutable" for the purposes of Article III:2, second sentence, two products need not be competitive or substitutable with respect to all their possible end-uses. This was made clear by the Appellate Body in Canada � Periodicals. ⁵⁷ In that case, Canada argued that US magazines were not "directly competitive or substitutable" with Canadian magazines because, while they provided a reasonable substitute as an advertising medium, they were poor substitutes as an entertainment and communication medium. Thus, according to Canada, US and Canadian magazines were only "imperfect substitutes". The Appellate Body dismissed this argument by pointing that a case of "perfect substitutability" would fall within Article III:2, first sentence.
58. The European Communities goes on to state that similarly, in Japan - Taxes on Alcoholic Beverages I, the panel based its conclusion that Japan's Liquor Tax Law violated Article III:2, second sentence, on the finding that there existed direct competition or substitutability among the liquors concerned, "even if not necessarily in respect of all the economic uses to which the products may be put".⁵⁸
59. Further, the European Communities contends that from the principle that Article III is concerned with the protection of competitive opportunities, it follows that the end-uses to be taken into account include all the objective (or functional) end-uses to which the products may be put, irrespective of whether the products are currently being employed for those end-uses in the market concerned. Similarity of actual end-uses may, of course, provide further evidence of substitutability, but is not required for a finding that products are "directly competitive or substitutable".

(iii) Tariff Classification

60. In the view of the European Communities, tariff nomenclatures classify products in accordance with their physical characteristics and end-uses. For that reason, as noted by the Appellate Body in Japan � Taxes on Alcoholic Beverages II, "if sufficiently detailed, tariff classification can be a helpful sign of product similarity".⁵⁹
61. The European Communities points out that tariff classification has already been relied upon by several previous panel reports in order to make a "like" product determination. ⁶⁰ A fortiori, tariff classification may also be relevant for the purposes of determining whether two products fall within the broader category of "directly competitive or substitutable" products.
62. According to the European Communities, two products falling within the same tariff position may be covered by different bindings and vice-versa. Tariff nomenclatures such as the Harmonized System classify products according to their objective characteristics. For that reason, they may provide useful guidance in order to determine whether products are "like" or "directly competitive or substitutable".
63. Chile concedes the obvious point that all distilled spirits share a common tariff category; however, according to Chile, the point is of virtually no legal significance, because there are 4-digit HS categories which include products which are obviously not "directly competitive or substitutable"; for example, aviation gas and vaseline white oil (HS 2710), mackerel and caviar (HS 1604), lobster and crabmeat meal (HS 0306), and ivory and nails (HS 0507).

(iv) Market Place

64. The European Communities notes that in Japan � Taxes on Alcoholic Beverages II⁶¹, the Appellate Body approved the Panel's decision to look at the "market place" and in particular to the cross-price elasticity between the products concerned. At the same time, however, the Appellate Body emphasised that cross-price elasticity is "not the decisive criterion". According to the Appellate Body, cross-price elasticity is but one of the means of examining the relevant market. In turn, looking at the relevant market is but "one among a number of means" of identifying the products that are directly competitive or substitutable in a particular case.
65. According to the European Communities, a high rate of cross-price elasticity may be sufficient to establish that two products are directly competitive or substitutable for the purposes of the second sentence of Article III:2. The opposite, however, is not necessarily true. A relatively low rate of cross-price elasticity does not of itself exclude the possibility that two products may be "directly competitive or substitutable". Such a low rate may simply be the result of the tax measures at issue. Or it may reflect the fact that imports are new entrants in the market concerned, a hypothesis that may be of particular significance in the case of experience goods such as those at issue in this dispute.
66. The European Communities further points out that as noted by the Panel Report on Japan � Taxes on Alcoholic Beverages II:

... a tax system that discriminates against imports has the consequence of creating and even freezing preferences for domestic goods, In the Panel's view, this meant that consumer surveys in a country with such a tax system would likely understate the degree of potential competitiveness between substitutable products. ⁶²

67. According to the European Communities, when "looking at the market-place," it may also be relevant to consider, in addition to cross-price elasticity, other factors such as the actual end-uses of the products (as opposed to their objective end uses), their availability in different sales channels or the way in which they are advertised. Nevertheless, when assessing the evidence related to these factors, it must always be borne in mind that Article III:2 is concerned with the protection of competitive opportunities and not of actual competition. Evidence that two products are sold in the same channels and used for the same purposes can be taken as evidence that they are "directly competitive or substitutable". The absence of such evidence, however, does not, in and of itself, lead to draw the opposite conclusion.
68. Chile replies that the European Communities offers as an excuse that cross-price elasticities are often low for products that are new to the market. Such a comment hardly seems appropriate to the case of whisky, which has a very long history in the Chilean market. Further, Chile argues that the regression analysis submitted by the European Communities has serious methodological flaws. Chile further states that the Search Marketing surveys submitted by the European Communities have inconsistent results.
69. Chile further argues that common distribution channels are not strong evidence, and multi-purpose channels are weaker still than dedicated channels.
70. The European Communities responds that the regression analysis it has submitted to the Panel was conducted by a consultation firm at the request of the Chilean pisco industry. Further, it claims that the regression analysis Chile submitted to the Panel has a more serious flaw of multicolinearity.

5. Product Categories

71. The European Communities argues that all pisco must be considered as a single product for the purposes of Article III:2. Accordingly, the existence of direct competition or substitutability must be assessed with respect to the category of pisco as a whole and not with respect to each of the four types of pisco.
72. The European Communities claims that, according to the regulations in force, the four different varieties of pisco are distinguished solely in terms of their alcohol strength. ⁶³ This means that, for example, any pisco with 43° or more is entitled to use the designation gran pisco, irrespective of the varieties of muscat grape from which it is made, the length of its ageing period or the type of container in which it has been matured. As acknowledged by Chile during the consultations⁶⁴, all that is required in order to produce pisco especial or pisco corriente instead of gran pisco or pisco reservado is adding some more water before bottling the product.
73. The European Communities then argues that this difference does not warrant treating each of them as a distinct product for the purposes of Article III:2, second sentence, especially since, as will be shown below, most pisco is consumed mixed with other non-alcoholic beverages or "on the rocks". In Japan � Taxes on Alcoholic Beverages II, the panel noted that "a difference in the alcoholic strength of two products dos not preclude a finding of likeness, especially since alcoholic beverages are often drunk in diluted form".⁶⁵
74. Further, the European Communities points out that as shown in Tables 7 and 8 below, most other distilled spirits are also bottled at different alcohol strengths. Japanese shochu, for instance, is bottled in a range from 20° to 45° . Yet, the two panels on Japan � Taxes on Alcoholic Beverages I and II have treated all shochu as a single product, irrespective of its alcohol content.
75. The European Communities disagrees with Chile's claim that gran pisco and pisco reservado are products of higher quality and, for that reason, are sold at higher prices than the other types of pisco. ⁶⁶ There is no clear correlation between quality/prices and alcohol strength. The European Communities submitted a survey of retail prices conducted in May 1998 by Search Marketing S.A. at five big super-markets of Santiago (hereafter, "the 1998 SM price survey"). ⁶⁷ This survey shows that price differentials among brands of the same alcohol strength are often larger than price differentials among strength categories within the same brand. Thus, for example, pisco especial of high quality brands such as "Alto del Carmen", "Control de Guarda" or "Mistral" is more expensive than pisco reservado or even gran pisco of average quality such as the basic "Capel" and "Control" brands.
76. The European Communities further argues that, with respect to the samples of pisco provided to the Panel⁶⁸, the bottles and labels of all piscos of the same brand tend to be identical, except for the mandatory indication of their respective alcohol strength and denomination. Promotional claims are generally made with respect to all products of the same brand, rather than with respect to specific alcohol strengths.
77. Further, according to the European Communities, the pisco industry has pursued a marketing strategy apparently aimed at segmenting the pisco market in order to expand its consumer base and raise unit profitability. This policy has taken two different courses. The first of them was to develop product categories based on alcohol strength, a strategy that was officially sanctioned by Decree-Law 78/1986. The second and more recent trend is to extend the line of brands of each producer. Thus, in addition to its basic brand "Capel", Capel is now selling the "prestige" brands "Artesanos de Cochiguaz" and "Alto del Carmen", as well as less expensive brands such as "Limarí" and "Límite". In turn, Control's brand portfolio includes, besides the brand "Control", the brands "Control de Guarda", "La Serena", "Mistral", "Tres Erres" and "Sotaqui". The identity of these new brands has been built upon motives such as the place of origin of the grapes, traditional manufacturing methods, product age claims, fancy packaging and differentiated advertising. ⁶⁹
78. The European Communities further argues that most brands are sold in a wide range of strengths, with the consequence that the number of brands/strengths now available in the market may be close to 100, despite the fact that the two main producers account for more than 90 % of the sales. The proliferation of new brands, each with a different identity, has blurred the segmentation based on alcohol content which the industry had attempted to build on the basis of the legal categorisation established in Decree 78/1986. As already mentioned above, inter-brand differences in quality and price may now be greater than intra-brand differences between strength categories.
79. The European Communities then concludes that in any event, pisco is by no means unique in being sold in different types/qualities. If anything, pisco is a much more homogeneous product than other spirits. For instance, to mention but the best known types of whisky, one may distinguish between malt Scotch, grain Scotch, blended Scotch, Canadian, Irish, Bourbon and Rye, each with its own specific characteristics. Moreover, within each of those types of whisky, it is still possible to draw further distinctions. Thus, in the case of blended Scotch, the trade makes a distinction between "premium" blended Scotch (12 or more years old) and "standard" blended Scotch (less than 12 years old). The differences in terms of quality and price between the four varieties of pisco alleged by Chile are not greater than those that exist between different types/qualities of whisky.
80. In response to a question posed by the Panel about product categories and comparisons, Chile contends that there are both similarities and differences between, on the one hand, the cases of Boss Suits/Discount Store Suits, Mont Blanc Pens/Bic Pens and, on the other hand, pisco and whisky or other distilled spirits. In the cases of the suits and the pens, one product is essentially a luxury version of the other, with corresponding differences in price and presumably quality. As a result, in most markets there is probably little substitutability between the products, as poor consumers cannot afford Boss suits or Mont Blanc pens, while rich ones will not buy discount suits or pens where a choice is available (although Chile thinks that even rich consumers may be more likely to use Bic Pens for some purposes than to buy cheap suits). A hypothetical "average" consumer of suites might see some substitutability, but more likely would find neither a satisfactory alternative. Chile believes that the differences in market segment and price are such that a higher tax on the luxury product would not infringe Article III:2, even if all Boss suits were imported and all discount suits were domestically made.
81. Chile goes on to argue that the case of pisco and other distilled spirits has some similarities, but also some important differences. The analogy to the pens and suits cases would be more precise if one compared a cheap, 30� pisco corriente to a high quality gran pisco, since one is a luxury version of the other.
82. In Chile's view, the more appropriate analogies when comparing pisco corriente and, for example, a good Scotch whisky or French cognac would be the difference between a cheap apple and a top quality orange; Belgian endives and cabbage, or a cheap soybean oil and a high quality extra virgin olive oil from Umbria. In each of these examples, the two different products are capable of being used for the same purposes and, at some level of price and consumer desperation, might be used interchangeably. However, their markets are in practice very different and they would not be considered directly competitive or substitutable by consumers in most markets. Indeed, except when pursuing the lowest possible taxes for its products, the Scotch Whisky Association itself would doubtless wholeheartedly reject the idea that its products were directly competitive or substitutable with pisco corriente.
83. The European Communities notes that Chile's changes of analogies from vehicles to vegetables does not help its case. Pisco corriente and "a good Scotch whisky" are compared by Chile, respectively, to a "cheap apple" and a "top quality orange". However, apples and oranges were precisely one of the examples of "directly competitive and substitutable products" cited by the drafters of GATT. The qualification that the apple must be a "cheap" one and the orange "top quality" is inappropriate. The present case is not just about expensive "good" Scotch and cheap pisco corriente. It is about all kinds of whisky and of pisco. The European Communities has shown that the prices of pisco and of other spirits already overlap, despite differences in taxation.
84. According to the European Communities, the example of "cheap soybean oil" and "high quality extra virgin olive oil from Umbria" is flawed for similar reasons. The relevant comparison would be with olive oil and soybean oil, without any further qualifications.
85. Also, the European Communities states that endives are not a "luxurious" product in Belgium. Over time, they could become less expensive also in Washington DC, provided that they are not subject to a "luxury" tax. In Europe, avocados or kiwis could have been considered as a "luxury" item product 20 years ago. They would have remained so, had they been subjected to protective taxation.

To continue with Arguments on Each Factor

⁵² Appellate Body Report on Japan - Taxes on Alcoholic Beverages II, supra., p. 16. See also Working Party on Brazilian Internal Taxes, adopted 30 June 1949, II/181, 185, para. 16; Panel Report on United States - Taxes on Petroleum and Certain Imported Substances, supra., para. 5.1.9; Panel Report on United States - Measures affecting Alcoholic and Malt Beverages (hereafter, "United States � Malt Beverages"), BISD 39S/206, para. 5.6; and Panel Report on United States - Measures Affecting the Importation, Sale and Use of Tobacco, BISD 41S/131, paras. 99-100.

⁵³ Appellate Body Report on Japan - Taxes on Alcoholic Beverages II, supra., p. 25.

⁵⁴ Ibid.

⁵⁵ Panel Report on Japan - Taxes on Alcoholic Beverages II, supra., para. 6.22.

⁵⁶ Panel Report on Korea � Taxes on Alcoholic Beverages, 17 September 1998, WT/DS75/R and WT/DS84/R, para. 10.67.

⁵⁷ Appellate Body Report on Canada - Periodicals, supra., p. 28.

⁵⁸ Panel Report on Japan - Taxes on Alcoholic Beverages I, supra., para. 5.7.

⁵⁹ Appellate Body Report on Japan � Taxes on Alcoholic Beverages II, supra., p. 21.

⁶⁰ Panel Report on Japan - Taxes on Alcoholic Beverages I, supra., para. 5.6. See also the Panel Report on EEC � Measures on Animal Feed Proteins, supra; and Panel Report on United States � Standards for Reformulated and Conventional Gasoline, WT/DS2/R, adopted on 20 May 1996.

⁶¹ Appellate Body Report on Japan � Taxes on Alcoholic Beverages II, supra., p. 25.

⁶² Panel Report on Japan - Taxes on Alcoholic Beverages II, supra., para. 6.28, citing the Panel Report on Japan - Taxes on Alcoholic Beverages I, supra., para. 5.7.

⁶³ Article 13 of Decree 78/1986 (EC Exhibit 13). In this connection, the European Communities notes that Chile claimed in Chile's answers to questions from the EC, dated 24 February 1998 (EC Exhibit 2) that each of the four types of pisco is made from different varieties of grapes and according to a different manufacturing process, in particular as regards the maturation phase.

⁶⁴ Chile's answers to questions from the EC, dated 24 February 1998 (EC Exhibit 2).

⁶⁵ Japan - Taxes on Alcoholic Beverages II, supra., para. 6.23.

⁶⁶ Chile's answers to questions from the European Communities, dated 24 February 1998 (EC Exhibit 2).

⁶⁷ A copy is attached as EC Exhibit 23. The European Communities notes that additional confirmation is provided by the price data shown in the 1997 report of The International Wine & Spirits Record (hereafter the "IWSR Report"), pp. 80-81, Table G (EC Exhibit 19).

⁶⁸ EC Exhibits 62-71.

⁶⁹ See e.g. the description of Control's different brands contained in Control's Internet home page (EC Exhibit 51).