4.88 The expert for the Community further challenged the Japanese econometric study. According to this expert, time series which are used in Japan study, are inherently difficult to analyze as they tend to be beset by at least three types of problems that make a statistical analysis difficult. Concerning the trends, this means that the development of the different variables is driven by factors that cannot be explained by an econometric model, but happen due to autonomous factors. For instance, consumption of a product may change simply due to fashion, economic growth, population growth, etc. When a large part of the changes in consumption is influenced by factors of this nature, it will be difficult to separate statistically the influence of price movements from that of trends. Concerning autocorrelation, this means that extraordinary influences on the variables in one year are likely to be present in the following year. An advertising campaign in one year that pushed up consumption of the product will also have an impact on consumption of the following year. If it is not corrected, autocorrelation decreases the precision of the econometric estimates, if it is corrected (e.g. by using the Cochrane-Orcutt method), it will reduce the number of available data points. Concerning multicollinearity, this means that for the case at hand, changes in the consumption of one type of liquor are related to changes in the consumption of another liquor. A hot summer, for instance, will increase the consumption of all beverages. Again, the results show how difficult it is to statistically separate the influence of one variable from that of another. According to the Community expert, the data set used for the Japanese statistical analysis is fraught with all three problems, which would make it difficult to prove any statistical connection between shochu and whisky consumption. In addition to these problems that occur naturally in time series analysis, the date set suffers from two further limitations: first, there are breaks in the data series. During the time period analyzed there have been several tax reforms that affected the structure of liquor prices and consumption. Such breaks in the data series have, as a consequence, that what is measured at the beginning of the time period is not identical with what is measured at the end of a time period. Second, there is illegitimate aggregation of product groups. There are major shifts within a product group such as whisky. Although domestic and foreign whisky are inherently close products, the aggregation into one category hides the fact that up to 1989 the tax system treated them differently. Because the tax reform has affected domestic and foreign whisky in an opposite direction, an aggregation of the two types will lead to misleading statistical results. The Community expert submitted also that the importance of the tax reform of 1989 can be seen graphically. As the reform made domestic whisky more expensive and imported whisky less expensive, for 1989 and 1990 the statistics show a major increase in imported and a decrease in domestic whisky. After this period, the two whisky varieties start moving in line with each other. This is to be expected as they are identical products and are affected by the same external factors. For the Community, the inspection of the graph therefore leads to two evident conclusions: Firstly, it can be seen very clearly that a reduction in tax for imported whisky has a substantial impact on its consumption. Secondly, any statistical analysis that fails to take account of the structural break in the time series will come to misleading conclusions, because what is measured with the aggregate whisky consumption before and after 1989 is evidently not the same. Spurious statistical results are the natural outcome. According to the Community expert, it will be difficult for a statistical analysis to establish the impact of the whisky market on the market for shochu, because the influence here tends to be overwhelmed by other factors that happened at the same time. On the other hand, when shochu consumption increased, that of whisky fell and vice versa. While certainly such a graph does not provide an ultimate proof, it appears, however, highly suggestive of a substitution relationship between the two types of products.

4.89 Japan reiterated that it applied the statistical method used in the Bossard study commissioned by the European Commission, which used time series similar to Japan's household survey statistics. The Commission submitted a report citing the study to the European Council and to the European Parliament last September, and convened a conference of manufacturers and consumers in Lisbon last November to discuss the report. In Japan's view, the Community's criticism of the Japanese study applies equally to the Community's study. Moreover, although the Commission presented a graph of yearly consumption amount of shochu and whisky to assess the impact of the 1989 tax reform, Japan also presented a graph indicating the annual changes in the consumption amount of shochu and former second grade whisky. Japan submitted a chart of the "Annual Changes in Sales of Former Second Grade Whisky and Shochu" (See Annex II). In 1989 and 1990, in the wake of the 1989 reform, consumption of both former second grade whisky and shochu showed a marked decline. The issue to be examined here, for Japan, is how those consumers who drank second grade whisky in 1988 behaved in 1989 and in 1990. Unless they cease to drink alcoholic beverages (Japan argued that Canada's lagged response theory means several years of temperance, which Japan regarded as implausible), the decrease in the absolute amount of former second grade whisky consumption should have worked to increase the absolute amount of other alcoholic beverages. Thus the decline in the absolute amount of consumption of both former second grade whisky and shochu led Japan to conclude that it was more reasonable to assume that consumption shifted from former second grade whisky to alcoholic beverages other than shochu. For Japan, the share of shochu in the total amount of distilled liquor consumption in Japan increased in 1989 and 1990, only because the decrease of absolute amount of shochu consumption was smaller than that of former second grade whisky and had nothing to do with the shift between the two products.

4.90 On the issue of whether liquors are directly competitive and substitutable products in Japan, Canada argued that the market confirms the cross-price elasticity between Canadian whisky and shochu. Since 1987, on-shelf retail prices for distilled liquors sold in Japan have undergone significant price movements. During this period, prices of imported whisky have declined considerably, while prices for some domestic whisky, particularly formerly Second Grade whisky, have increased significantly. Sample on-shelf retail prices listed in surveys covering each of the years 1992 through 1994 conducted by the Japan External Trade Organization ("JETRO") ⁶⁹ make clear the rapid decline in prices for imported whisky. For example, the reported average retail price of the bourbon whisky "Four Roses" declined from �2,740 in 1992 to �2,090 in 1994 while "Ballantines" Scotch declined in the same period from �2,480 to �2,195. Sample on-shelf prices for Canadian whisky, like the imported whisky prices quoted in the JETRO surveys, have similarly declined. For example, the price for Canadian Club 6 Year Old fell from �2,500 during the years 1991-1993 to �1,900 in 1995, a decline of 24.0 per cent. During the same period, prices for Canadian Club Classic 12 Year Old fell from a median price of �3,750 between 1991-1993 to �2,900 in 1995, a decline of 22.7 per cent. Prices for Seagrams V.O. also fell from a median price of �2,680 in 1990 to �1,730 in 1995, a decline of 35.4 per cent. Seagrams Crown Royal declined from a median price of �3,330 in 1990 to �2,230 in 1995, a decline of 33.0 per cent.

4.91 Canada submitted another on-shelf retail price survey for the years 1987, 1989 and 1995 of sales of former Special Grade whisky, former Second Grade whisky and shochu Group A which it claims confirms the rapid decline in the retail price of imported whisky during this time. Pursuant to this survey, retail prices in Japan for imported former Special Grade whisky fell during the period of the study. This trend applied to both premium and standard labels. Retail prices for formerly Special Grade imported whisky (standard labels) declined from a range of �2,800-3,200/750ml in 1987 to �2,400-2,800 in 1989 and to �1,700-2,100 in 1995. Retail prices of formerly Special Grade imported whisky (premium labels) declined from a range of �4,000-5,000/750ml in 1987 to �3,500-4,500 in 1989 and to �2,500-3,500 in 1995. On the basis of the median price for these ranges, retail prices for formerly Special Grade imported whisky (standard) declined 36.7 per cent during this period (11.5 per cent after 1989); and formerly Special Grade imported whisky (premium) declined 33.3 per cent (25.0 per cent after 1989). In contrast, retail prices for domestic formerly Second Grade whisky increased from a range of �1,600-1750/1.8 litre in 1987 to �3,100-3,270 in 1989 and to �2,900-3,270 in 1995, resulting in a median price increase of 84 per cent during this period. At the same time, retail prices for shochu Group A increased from a range of �580-600/720ml in 1987 to �620-640 in 1989 and to �650-700 in 1995. Thus, median prices for shochu increased, but at a relatively more modest 14.3 per cent. Thus, between 1987 and 1995, the price differential between shochu and formerly Second Grade whisky significantly increased, while the price differential between shochu Group A and imported whisky decreased. The JETRO survey of 1993 states that there are several reasons for the retail price decline on imported whisky, including a decline in the specific tax rates for this product:

4.92 For Canada, as the JETRO 1993 survey implies, the narrowing price differential between imported whisky and shochu Group A has improved the price competitiveness of imported whisky vis-à-vis shochu. On the other hand, the median retail price ranges it had submitted previously indicate that retail prices for imported whisky continue to remain well above retail prices for shochu and thus hinder whisky's ability to compete with shochu. This is made clear in the Japan Spirits and Liquors Makers' Association submission to the Japanese Ministry of Finance of October, 1995:

4.93 In addition to their arguments and counter arguments suggested in the present section, the Community, Canada, the United States and Japan submitted further evidence on cross-price elasticity of shochu and other imported distilled spirits in their discussions of the application of the aim-and-effect test, when arguing, more specifically, whether the Liquor Tax Law has the effect of distorting the competitive relationship between shochu and other imported liquors so as to afford protection (see Sections F and G below).

4.94 As to the second step of the legal test it suggested for the second sentence of GATT Article III:2 in assessing whether a measure imposed on substitutable or directly competitive products is "so as to afford protection", the Community reiterated that the following criteria may be relevant in order to determine whether a difference in taxation is "so as to afford protection" to domestic production: 1) The level of the tax differential (but contrary to the first sentence of Article III:2, a tax difference does not lead automatically to a violation of the second sentence of Article III:2); 2) The degree of substitutability and competition between the two products; 3) Whether the less taxed product is produced in other countries. In this context the Community recalled the conclusion of the 1987 Panel Report which found that the following factors were sufficient evidence of fiscal distortions between imported distilled liquors and domestic shochu affording protection to the domestic production of shochu:

The Community also recalled that since Article III:2 protects trade expectations on the competitive relationship between imported and domestic products rather than expectations on trade volumes, it is not necessary, in order to establish a violation of Article III:2, second sentence, to show that the difference in taxation has had an actual effect on the volume of trade.

4.96 Japan responded to the Community's arguments on the three criteria. First, concerning the potential protective effect, Japan submitted that the tax differential should be measured on the basis of the tax/price ratio, as it is a criterion to judge whether or not a tax affords protection, and for Japan, there is no differential in the tax/price ratios. Secondly, for Japan, shochu and other distilled liquors do not show the aptitude of the two products to serve the same uses, and differ in the extent and the form in which the two products are available to the public, beyond what would apply to all alcoholic beverages. Cross-price elasticity of demand does not, therefore, exist. If a directly competitive or substitutable relationship were to be found in this case, it would have to be found between all alcoholic beverages, and, consequently, any liquor taxation currently in force would become inconsistent with Article III, unless all products show the same tax/price ratio. The degree of substitutability and competition between the products is minimal at best. Third, shochu is widely produced in Asian countries, and the third criterion is not met. Thus, Japan concluded that if the Community's interpretation is applied to the facts, one inevitably reaches a conclusion that Japan�s liquor tax is consistent with Article III:2, second sentence.

4.97 Japan argued that the Community is criticizing Japan's tax distinction among distilled liquors while dividing wine into six categories in its liquor tax directive and legitimizing Germany�s application of four completely different rates to categories of wines. For Japan, a position which holds that champagne and sherry may be distinguished from other wine while shochu and whisky should be treated alike, is equal to turning Article III into an instrument of harmonization of internal taxes with a system of a particular group of countries. Japan reiterated that the purpose of Article III is not to require Members to adopt a particular system of taxes or regulations, nor to harmonize taxation systems. Japan argued that only a small number of WTO Members apply a flat rate to all categories of distilled liquors and a larger number of Members apply more than one rate in one way or another. In Japan's view, the conclusion advocated by the Community in the present case would substantially affect other countries as well. Japan referred the Panel to its chart on worldwide "Taxation on Distilled Liquors" (see Annex III).

4.98 The Community agreed that Article III does not require Members to adopt a particular system of taxes or regulations, or to harmonize with a system of a particular Member. Nonetheless, the Community believed that the fiscal autonomy of Members is limited by their obligation under GATT to afford imported products equal conditions of competition, inter alia, with respect to internal taxes. The Community was not requesting Japan to introduce any particular system of taxation or to set the tax rates for distilled spirits at any particular level. All the Community was asking from Japan was that shochu and all other like and substitutable or directly competitive distilled spirits not be taxed in a discriminatory or protectionist manner. For the rest, the Community submitted that Japan retains full autonomy to choose its own tax system. Following the adoption of a panel report upholding the Community�s claims, Japan would enjoy complete discretion to decide whether to maintain the current system of specific taxes or to replace it by, for example, a system of ad valorem excise taxes or by a system of ad valorem consumption taxes or by a mixed system combining specific and ad valorem taxes. The Community reiterated that Japan also retained complete freedom to decide not to impose any tax at all on distilled spirits. Japan would also be able to choose the level of the tax rates. If, for example, Japan decided to maintain a system of specific excise taxes, the rates for all distilled spirits could be set at or above the current level for "whisky/brandy" (the highest) or at or below the current level for shochu B (the lowest), as well as at any level in between. As it stands, the Liquor Tax Law is inconsistent with the second sentence of Article III:2. The Community also added that the taxation systems of countries other than Japan are not covered by the terms of reference of the present dispute.

4.99 In its discussion of its legal test for the second sentence of Article III:2, Canada also referred the Panel to the 1987 Panel Report which set out four factors that established "sufficient evidence of fiscal distortions of the competitive relationship between imported distilled liquors and domestic shochu" so as to "afford protection to the domestic production of shochu": (1) the considerably lower specific tax rates on shochu than on imported whisky; (2) the almost exclusive production in Japan of shochu; (3) the mutual substitutability of distilled liquors like whisky and shochu as illustrated by the increasing imports into Japan of "Western-style" distilled liquors; and (4) the imposition of an ad valorem tax on imported whisky but not on shochu.

4.100 For Canada, the application of the three relevant criteria (omitting the fourth one) mentioned above to the current Liquor Tax Law confirms its inconsistency with the second sentence of Article III:2, in distorting the competitive relationship between Canadian whisky and domestically produced shochu:

4.101 Canada argued that an examination of the relationship between the differential specific tax rates on whisky and shochu and the resulting effects on price competitiveness between the two products makes it manifestly clear that the Liquor Tax Law is inconsistent with the observation of the Working Party Report on Border Tax Adjustments and confirmed in the 1987 Panel Report, that internal taxes must be trade neutral. The Liquor Tax Law is not trade neutral. It distorts the relative prices of whisky and shochu. It, thereby, distorts the competitive relationship between these two products and consequently affords protection to shochu production in contravention of Article III:2, second sentence. Canada submitted that price is a crucial element in determining the competitive relationship between shochu and whisky. Thus, it is significant that retail prices of imported whisky are responsive to changes in liquor tax rates. Accordingly, in Japan, the specific tax rates imposed pursuant to the Liquor Tax Law have a direct effect on the prices of imported whisky and, consequently, on the competitive relationship between this product category and shochu.

4.102 Canada added that given the findings in the 1987 Panel Report that even small tax differences can influence the competitive relationship between directly competing distilled liquors, the significant differences in the tax burdens imposed by the tax differentials under the current Liquor Tax Law affecting the price of imported whisky, a fortiori magnify the distortions in the competitive relationship between shochu and whiskies. Indeed, given the nexus between specific tax rates, the price of imported whisky and, ultimately, the competitive relationship between shochu and whisky, the significance of these large tax differentials in distorting the competitive relationship between shochu and whisky is made clear when considered against the retail price of imported whisky and shochu and thus, the domestic market "value" ascribed to these products. Since 1989, the tax rate on whisky has remained constant. During this period of time, and indeed since 1987, the retail price on imported whisky has rapidly declined. Thus, for whisky, the tax rate has not responded to changes in domestic market value. By contrast, Canada submitted that in 1994 there was a modest increase in the tax rate levied on shochu corresponding to a small increase since 1987 in the retail price of this product. Unlike shochu, the net result of the specific tax rates imposed on whisky pursuant to the Liquor Tax Law is that as the retail price, and thus the domestic market value, of imported whisky has fallen, the tax portion of the retail price has increased. Canada argued that, since 1989, the liquor tax on imported whisky is consuming an ever-increasing portion of the retail price. As described in one survey of retail prices in Japan, even though the retail price of formerly Special Grade imported whisky has proportionately decreased by an amount greater than the increase in the retail price of shochu, the tax burden levied pursuant to the Liquor Tax Law on this whisky product has proportionately increased by an amount greater than that imposed on shochu. Given the fact that price is a crucial element in determining the competitive relationship between whisky and shochu, the tax burden levied on imported whisky, not reflecting its market "value", hinders the price movement of imported whisky and hence the ability of imported whisky to compete with shochu. Canada submitted that to render neutral the proportionate differential tax burdens imposed on shochu and whiskies, the retail price of imported whisky would have to increase by a proportionate amount thereby decreasing the ability of imported whisky to compete with shochu in Japan's domestic market. In short, Canada argued that by imposing differential tax burdens on shochu and whisky that do not reflect the market value of imported whisky, the tax rates imposed on shochu and whisky pursuant to the Liquor Tax Law distort the competitive relationship between these two product categories and thus, cannot be claimed to be trade neutral.

4.103 Japan responded that Canada did not satisfy the three last criteria for its legal test under Article III:2, second sentence. Japan admitted that the liquor tax is an internal tax (first criterion of Canada's test). For Japan, as argued with the Community, shochu and whisky are not directly competitive or substitutable (second criterion). In discussing whether or not the products are similarly taxed (third criterion), Canada compares the amount of the tax per one litre of the products or per alcohol contained. However, Japan argued that in light of the emphasis Canada repeatedly attached to the price as a crucial element in determining the competitive relationship between shochu and whisky, the comparison ought to be made on the basis of the tax burden in relation to the price. By this standard, Japan argued that shochu and whisky are similarly taxed. Canada's fourth criterion "of affording protection" is, again, not met, in Japan's view, since tax/price ratios are roughly equal between imported whisky and domestic shochu: In restoring the balance in tax/price ratio between whisky and shochu in the 1994 reform, Japan chose to increase tax on shochu, not to reduce tax on whisky, for fiscal reasons, but both methods are equally effective in restoring the balance. Japan stressed that what is relevant to the purpose of neutrality and equity is relative relations of tax/price ratios among categories, not the absolute level of the ratios per se.

4.104 In addition, Canada argued that the Superfund panel report has established the principle that a finding of fiscal distortion in the competitive relationship between imported and domestic products constitutes an "irrefutable presumption" of nullification and impairment of benefits. While the Superfund panel report spoke in terms of Article III:2, first sentence, the 1987 Panel Report and the 1992 Malt Beverages panel report established that the same principle applies equally to Article III:2, second sentence. For Canada, this principle has been codified in Article 3:8 of the DSU, which provides:

4.105 Japan responded that the criterion unique to Canada's test is distortion of the competitive relationship and that it is noteworthy that Canada's argument rests on distortion of relative prices. Accurate data show, however, that the Liquor Tax Law does not distort relative prices between whisky and shochu, and, accordingly, this criterion of distortive effects is not met. In order to reach a finding of inconsistency with Article III, all of Canada�s criteria would have to be met, but it was not the case, since Canada had not proven that shochu and Canadian whisky are directly competitive, that shochu and Canadian whisky are not similarly taxed, or that the Liquor Tax Law affords protection to domestic production. For Japan, Canada had, therefore, not proven that the Liquor Tax Law is inconsistent with Article III:2, even under Canada's interpretation of the provision.

4.106 As argued in paragraphs 4.24 to 4.32 above, the United States submitted that the central concern of Article III is to prohibit the targeting of imports and suggested that application to the Liquor Tax Law of the aim-and-effect test of earlier panel reports would confirm the inconsistency of that measure with the provisions of Article III:2, second sentence, in that the regulatory distinctions made by the legislation are so as to afford protection.

4.107 The United States argued that the protective aim of the Liquor Tax Law structure is apparent from (1) the stated policy objective and whether it was known at the time the legislation was enacted that it would draw a line between one group of products that would be foreign and another group that would be domestic (ex-ante knowledge), (2) the internal inconsistencies of the legislation and its structural incentives, (3) legislative statements and the preparatory work, as well as from (4) the arbitrary and irrational categories of the legislation under scrutiny. The United States continued by stating that:

4.109 Japan continued its counter argument on the aims of the legislation, in stating that it is not arbitrary to apply the same tax rate to pre-mixes, while subjecting unmixed original products to different rates; Canada while applying different rates on distilled liquors and wine, applies the same rate on distilled liquor based pre-mixes and on wine based pre-mixes. The recharacterization of the "Juhyo" brand vodka into shochu by the manufacturer is, in essence, marketing of two different products under the same brand name, and does not indicate arbitrariness in categorization. The shochu product is produced from a different set of materials and is marketed as a New Juhyo. For Japan, this should be looked at as an attempt to take advantage of a popular brand name. Suntory, the manufacturer of Juhyo brand, sells Reserve whisky and Reserve wine. The Liquor Tax Law is not "the legislation that draws a line between one group of products that would be foreign and another group that would be domestic". Whisky, spirits and liqueur have been produced in large quantities in Japan, and, therefore, are not foreign, while shochu is produced in large quantities in the Asian region, and is, therefore, not domestic. Finally, the Liquor Tax Law is designed to be neutral and does not have incentives in any direction. Thus, none of the factors supporting a protective aim have been shown. For Japan, all of the evidence supplied by the United States on arbitrary or exceptional categorization is, in fact, evidence of the difficulty common to all legal definitions of social concepts.

4.110 For the United States, the tax rate applicable to any particular alcoholic beverage in Japan is a function of its classification, the applicable tax rate, and any available exemptions or reductions. The definition of "shochu" was devised in 1962 when all Japanese import trade was subject to BOP quotas and imported shochu was nonexistent, in order to reserve low tax rates for an exclusively-domestic product category. In the US view, it was absolutely clear in 1962 that the definition of shochu would exclude imported distilled spirits, and would form part of a system favouring shochu through the tax rate. This categorization was designed to perpetuate the market situation of 1962. Ever since 1962, shochu A and particularly shochu B have benefited from the lowest tax rates on distilled spirits. The effect has been to cement into place distinctions made under conditions of perfect protection, and to perpetuate the closed market of the pre-liberalization period. The discrimination in tax rates was reduced but not eliminated after the 1987 Panel Report. Shochu could still be characterized as a Japanese product benefiting from discriminatory low tax rates. Indeed, the United States argued, when a system of tax classifications had been designed during such a period of absolute protection, the categorization of products based on their status during the quota period had to be re-evaluated after balance-of-payments quotas had been lifted. The tax discrimination still remained: �982.3 per litre for whisky and brandy versus �102.1 per litre for shochu B, and �367.3 per litre for spirits versus �155.7 per litre for shochu A. Japan has offered no convincing policy rationale for this differential. The United States cited a statement by the Japanese Government�s own Administrative Reform Council admitting that there is at present no logical rationale for the Liquor Tax Law, and no plausible explanation for the liquor tax system including the segmentation of categories and tax rates.

4.111 The United States also argued that if a tax strongly favours a product of which all, or almost all, the amount consumed is produced domestically, but not another which is directly competitive or substitutable, a protective purpose may be inferred. The 1992 Malt Beverages panel report, in discussing the special tax treatment accorded by Mississippi to wine made from scuppernong grapes, noted that even if this wine were considered unlike other wine, the two kinds of wine would nevertheless have to be regarded as "directly competitive" products in terms of the Interpretative Note to Article III:2, second sentence, and the imposition of a higher tax on directly competing imported wine so as to afford protection to domestic production would be inconsistent with that provision. The United States also pointed to the sudden or dramatic difference in rates at the margin under the Liquor Tax Law which resulted in large differences in treatment of products on either side of the line drawn between shochu and other distilled spirits. The rates on shochu A and B are still much lower than the rates on other distilled spirits. There is still a 9.6 to 1 tax differential between whisky and shochu B, a 6.3 to 1 differential between whisky and shochu A, and a 2.4 to 1 differential between spirits and shochu A, at the respective reference alcoholic strength for each. The United States stated that the Liquor Tax Law targets the inherent characteristics of the product so that foreign manufacturers of gin or rum cannot make a product that has access to the lower tax rate without changing the nature of their product.

4.112 Japan responded that the tax rates currently applied to distilled liquors are completely different from those in 1962. Japan also argued that the Administrative Reform Committee's comment concerns the taxation of alcoholic beverages as a whole, and contains no specific reference to distilled liquors. For Japan, their critique of shortcomings in Japan's taxation of alcoholic beverages as a whole would be in fact valid for any liquor tax in the world. It is true that there is room for improvement, but no tax is perfect. In response to the United States condemnation on "the sudden or dramatic difference in rates at the margin", Japan argued that such difference is common to most tax distinctions: the rate of the US tax on wine jumps from $1.07 to $3.40 when the carbon dioxide in the wine crosses the threshold of 0.392g per 100 ml. Japan further argued that 75 per cent of whisky, 72 per cent of brandy, 82 per cent of spirits and 97 per cent of liqueur consumed in Japan is produced domestically. It is, therefore, farfetched to assume an element of targeting of imports behind these categories or taxes. With respect to whisky, in particular, Japan is the fifth largest producer in the world. Japan suggested that it cannot possibly target imports by taxing the category. Japan submitted that imports are not targeted and if the central concern of Article III:2 is the targeting of imports, as the United States argues, the lack of targeting is enough to establish that Japan's Liquor Tax Law is consistent with Article III. In support of its argument Japan referred the Panel to a chart showing the "Share of Domestic Production of Liquors in Total Sales in Japan" (see Annex IV).

4.113 The United States went on to argue that the distinction drawn by the Liquor Tax Law also has the effect of affording protection to domestic production. In this regard, data on sales and trade flows are relevant to show changes in the conditions of competition favouring domestic products. Other factors, including the creation of inherently domestic products and foreign products, and whether there is a large difference in rates between categories, also support the conclusion of a protective effect. In this context, the United States referred the Panel to some of the conclusions of the 1987 Panel Report which found that: 1) consumer habits varied in response to the respective prices of shochu and other distilled spirits, their availability through trade and their other competitive inter-relationships (i.e., substantial cross-elasticity of demand existed between these products); 2) the increasing imports of Western-style alcoholic beverages into Japan bore witness to this lasting competitive relationship and to the potential product substitution through trade among various alcoholic beverages; and 3) there existed direct competition or substitutability among imported and Japanese-made distilled liquors, including all grades of whiskies/brandies, vodka and shochu A and B, in terms of Article III:2, second sentence. The 1987 Panel Report found as well that the following factors were "sufficient evidence of fiscal distortions of the competitive relationship between imported distilled liquors and domestic shochu affording protection to the domestic production of shochu": a) the large difference in specific tax rates between the taxes on shochu and the taxes on imported distilled spirits; b) the fact that shochu was almost exclusively produced in Japan and that the lower taxation of shochu afforded protection to domestic production, rather than to a product produced in many countries; and c) the mutual substitutability of distilled liquors, as demonstrated by the increasing imports into Japan of distilled spirits and the consumer use of shochu in mixed drinks. For the United States an examination of the effect of the Liquor Tax Law in this present case should focus on qualitative alteration of conditions of competition such as targeting of imports, and evidence of cross-elasticity of demand between the favoured and disfavoured categories.

4.114 The United States pointed out that shochu consumed in Japan continues to be made almost exclusively in Japan. In 1994, imports of shochu were 1.7 per cent of total sales and 1 per cent of total sales of distilled spirits and "authentic distilled spirits". Also, in 1994, imports from third countries accounted for 27 per cent of the total sales of whisky, 29 per cent of the total sales of brandy, 18 per cent of the total sales of spirits and 78 per cent of the total sales of "authentic liqueurs". At the same time, domestically-made shochu accounted for over 80 per cent of all domestic sales of distilled spirits and authentic liqueurs. Thus, the protection given to shochu has had the effect of protection for domestic production.

4.115 On the market shares of shochu and the price-cross elasticity of shochu, the United States, in addition to the arguments detailed in paragraphs 4.82 to 4.93 above, noted that there were clear indications that the demand for shochu is largely influenced by fluctuations in demand for other distilled spirits and liqueurs. This could be seen in the rearrangement of the market place for distilled spirits after the 1989 tax reform. The 1989 reform unified tax rates on whisky, abolished the classification of whisky into three classes, and consequently more than tripled the tax rate on second-class whisky while lowering the taxes on other whisky, authentic liqueurs and spirits. The 1989 law also raised the tax on shochu by a small amount. In particular the United States submitted that:

4.117 The United States further argued that the protective effect of the tax distinction can also be seen in the attempts of Japanese shochu producers to make their products resemble whisky or to emphasize the points of similarity between the raw materials, ingredients, manufacturing process, appearance and tradition of their shochu and whisky or brandy. In May 1988, the shochu manufacturer Takara began to market "Jun Legend", a light amber coloured brand of shochu produced by blending two types of alcohol distilled from barley and corn and aging them in charred oak barrels for one to five years. Takara�s claim was that "the most noticeable characteristic of this brand is a flavour and taste similar to whisky". When the new brand was launched, Takara announced its expectation that the new brand would appeal to former consumers of second-class whisky, the tax rates on which were expected to increase with unification of tax rates on all classes of whisky. The favourable treatment of shochu is a classic instance of use of a tax system to perpetuate existing consumer preferences. Foreign manufacturers of gin or rum cannot make a product that has access to the lower tax rate accorded to shochu unless they change the nature of the product. Therefore, the United States concluded that the regulatory tax distinction made by the Liquor Tax Law between shochu and other imported liquor targets imports and has the aim and effect so as to afford protection, in contravention of the second sentence of Article III:2 (as well as the first sentence of Article III:2 for which the United States suggested the same test).

TO CONTINUE WITH JAPAN - TAXES ON ALCOHOLIC BEVERAGES

⁶⁹ JETRO is, according to Canada, a Japanese government agency affiliated with the Ministry of International Trade and Industry (MITI). According to Japan, JETRO is not a Japanese government agency but rather a non-profit, government-related organization affiliated with the Ministry of International Trade and Industry (MITI).

⁷⁰ Thus, in the Superfund case, para. 5.19, the panel stated: "A demonstration that a measure inconsistent with Article III:2, first sentence, has no or insignificant effects would therefore in the view of the Panel not be a sufficient demonstration that the benefits accruing under the provision had not been nullified or impaired even if such rebuttal were in principle permitted".

⁷¹ The 1987 Panel Report stated in para. 5.16: " [A]n increase in imports could not refute the presumption that discriminatory or protective taxes inconsistent with Article III:2 had impaired the competitive benefits protected under Article III:2 because, inter alia, an increase in imports did say nothing about what the trade might have been in the absence of the inconsistent trade restrictions".

⁷² Tan Hirosho, "Shuzeiho to no ichibu o kaisei suru horitsu" (The Law Partially Revising the Liquor Tax Law), in Zeisei Tsushin (Tax Policy News), June 1962, p. 23ff. The article identifies the author as the Deputy Director of the Ministry of Finance, Second Tax Policy Division.

⁷³ The United States added that this is fundamentally different from, for example, a tax incentive for adding catalytic converters to automobiles: an automobile can qualify for the incentive without changing its nature.