What's New?
 - Sitemap - Calendar
Trade Agreements - FTAA Process - Trade Issues 

espa�ol - fran�ais - portugu�s
Search

Korea - Taxes on Alcoholic Beverages

Report of the Panel

(Continued)


    B. United States

  1. In response to a question whether if two products have a very low cross-price elasticity that is sufficient to consider them directly competitive or substitutable products, the United States asserts that:
  2. (a) The standard that should be used to establish that two products are "directly competitive or substitutable" within the meaning of Article III:2 is a case-by-case examination that may consider a number of factors. There is no one standard that can operate across all cases. Cross-price elasticity is but one factor that may be helpful in conducting an analysis of whether products are directly competitive or substitutable. It is unlikely that a very low cross-price elasticity in and of itself will be sufficient to make a determination in any particular case.

    (b) The quotation from the second US submission expresses the economic point that any shift away from one product to another in response to a relative rise in the first product's price is a sign of cross-price elasticity and therefore substitutability. However, the US response did not purport to establish what degree of substitutability was "direct" within the meaning of Article III:2. Clearly the word "directly" in the Note Ad Article III:2 places some limits on the scope of Article III:2. However, this case does not present a situation where it is necessary to test those limits. In this case, the products are physically similar, and several other factors support the fact that the products are directly competitive or substitutable.

    (c) The Panel should be in a position to determine substitutability largely on the basis of common physical characteristics, which are reflected in a common HS heading, 302 and the Dodwell study presents evidence of a supplementary nature. The Appellate Body in Japan - Taxes on Alcoholic Beverages II did not, as Korea has claimed incorrectly, require a market analysis or prescribe the use of a cross-price elasticity study as the sole means of establishing whether products are directly competitive or substitutable. To the contrary, the issue only arose because some parties argued it was inappropriate to place undue emphasis on the market place generally or such studies specifically. The Appellate Body first determined that it "did not seem appropriate" to look at the actual market in addition to physical characteristics, common end uses and tariff classification. In approving the Panel's use of the ASI cross-price elasticity study (which the Dodwell study emulated) the Appellate Body underscored that cross-price elasticity of demand was not the "decisive criterion" for its determination that the products were directly competitive or substitutable. The Appellate body Report finds no support for the suggestion that a failure to show a positive cross-price elasticity in general or a particular cross-price elasticity can be the basis for concluding that products are not directly competitive or substitutable under GATT Article III:2.

  3. In response to a question about whether there is a de minimis standard in assessing the question of "so as to afford protection", the United States argues that:
  4. (a) the criterion for meeting the third element of the analysis under the second sentence of Article III:2 is set out by the Appellate Body in Japan - Taxes on Alcoholic Beverages II:

    "[P]rotective application can most often be discerned from the design, the architecture, and the revealing structure of the measure. The very magnitude of the dissimilar taxation in a particular case may be evidence of such a protective application, as the Panel rightly concluded in this case. Most often, there will be other factors to be considered as well. In concluding this inquiry, panels should give full consideration to all the relevant facts and all the relevant circumstances in any given case".303

    (b) In concluding that the panel had not erred in determining that the tax was protective solely on the basis of the large differentials between tax rates applied to shochu and Western spirits, the Appellate Body appears to have reasoned that the large differences between such physically similar products could not be otherwise explained. The United States considers that the magnitude of the differences in tax rates between such similar products in this case compel the same inference. On the other hand, however, the Appellate Body did not determine that panels must find a large differential in rates in order to find protective application; the differentials in rates is addressed by the second element of the Article III:2 analysis, whether the products are not "similarly taxed".

    (c) It would not be appropriate to determine that a measure does not have a protective application simply on the basis of significantly different pre-tax prices. A number of factors can affect pre-tax prices, such as exchange rates, and recent experience has shown how quickly those factors can change. It would not be appropriate to base determination of protective effect on a "snapshot" of pre-tax prices. Instead, the Appellate Body Report makes it clear that the panel must examine all the facts and circumstances concerning the structure of the law. In this instance, the structure of Korea's Liquor Tax Law distinguishes between products on the basis of arbitrary physical characteristics in a way that can only be explained by an intention to identify and define products that happen to be imported. Combined with Korea's long history of protecting soju from imports of western spirits, the differential tax rates then further support the conclusion that the law is structured to protect imports.

    (d) There is no de minimis interpretation of the concept "so as to afford protection to domestic production." The prohibition is absolute - any measure applied so as to afford protection of a directly competitive or substitutable import that is not "similarly" taxed is providing too much protection.

    (e) In light of the Appellate Body's emphasis on the structure of the measure, it would not be appropriate to limit the inquiry to the tax differentials and require a complaining party to determine the extent to which in the current market, with respect to current products at current prices, the taxes are effective in reducing sales of imports. Such an approach would, at a minimum, be at odds with the basic principles of the WTO Agreement. As Korea's law applies obviously punitive taxes on Western spirits not on the basis of price differences, but on the basis of arbitrarily physical criteria, the higher rates adversely affect all products within the criteria - including products in a wide range of prices available in the United States.

  5. In response to a question concerning the price mix of exports to Korea compared to other markets, the United States argues that:
  6. (a) The pre-tax prices of imports reflect costs such as transportation and tariffs (20% ad valorem in Korea), costs that are not reflected in prices of domestic products. The punitive taxes on Western spirits likely contribute to higher pre-tax prices than would otherwise exist under a neutral tax structure".

    (b) The majority of imports of distilled spirits into Korea are from the European Community, and accordingly, the United States is not in a position to draw any general conclusions about marketing of obvious brands in the Korean market at this time. Aside from some bulk shipments, the overwhelming majority of US exports are of Jack Daniels and Jim Beam Whisky, brands which enjoy particular international name recognition. Such name recognition is of particular value in developing a presence in new markets due to the high costs of exporting.

    (c) The US Government does not have access to the pre-tax prices charged by distilled spirits around the world by particular brand, which, according to industry representatives, is considered confidential information by the respective companies.

  7. In response to Panel questions and arguments raised by Korea with respect to competition issues, the United States notes that it did not consider it would be appropriate to borrow market analysis under national competition laws in analyzing what products are "directly competitive or substitutable" for purposes of Article III:2. The United States considers it important to bear in mind the different objectives of antitrust law, on the one hand, and of GATT Article III:2 on the other. Article 3.2 of the DSU makes clear that the provisions of the GATT 1994 and other WTO agreements are to be construed "in accordance with customary rules of interpretation of public international law." As the Appellate Body has noted on several occasions, Articles 31 and 32 of the Vienna Convention of the Law of Treaties embody the customary international law of interpretation. Article 31 sets forth the basic principle that treaties are to be interpreted "in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose."
  8. According to the United States, consistent with this principle, GATT Article III:2 should be interpreted in the light of the overall purpose of Article III, which is "to avoid protectionism in the application of internal tax and regulatory measures." 304 Article III is an anti-discrimination provision aimed at ensuring that government measures do not skew competitive conditions in favor of domestic products. Competition laws, by contrast, address privately-created threats to market competition, regardless of whether the competing producers or products are domestic or foreign. Because the object and purpose of the two sets of rules are quite different, it would not be appropriate to borrow and apply a competition analysis in deciding whether a tax measure is consistent with GATT Article III:2.
  9. C. Korea

  10. In response to a question concerning the ingredients and comparability of pre-mixes, Korea states that:
  11. (a) Pre-mixes first became available in May 1994 with a view to attracting women consumers and consumers in their twenties who preferred low alcoholic beverages. 305 Pre-mixes have 10 per cent to 15 per cent alcoholic content.

    (b) Standard soju has 25 per cent alcoholic content. The salient features which distinguish premixes from standard soju are that the former contains scent, coloration and more than 2 per cent extract. To be precise, pre-mixes do not contain standard soju. The mixture is a combination of various ingredients and joojung (ethyl Alcohol) which is the raw material from which standard soju is produced.

    (c) To make the pre-mixes appeal to women, the producers first eliminated the pungent odour and taste of joojung by adding fruit scents (such as lemon and cherry). Thus, the producers add lemon/cherry juice concentrate, acerola juice concentrate, and lemon/cherry spices which cannot be added to standard soju pursuant to the law. Sweeteners such as stevioside, sugar and fructose are added in higher dosage than standard soju to give the pre-mixes a sweeter taste.

    Pre-mixes
    (Lemon/cherry remixes)

    Standard soju
    (Green, Chungsaek soju)
    SugarSugar
    Citric AcidStevioside
    Co2 GasCitric Acid
    Lemon/Cherry concentrateMineral Salt
    Acerola juice concentrateAmino acid
    Lemon/cherry spicesSolbitol
    Food colours
    Fructose
    Stevioside

    (d) Korea further stated that standard soju is served in a typical small glass, and is rarely if ever drunk mixed. Standard soju is served "straight", and commonly drunk with meals.

    (e) Pre-mixes, otherwise known as soju-based cocktails, have quite a different, much sweeter taste than standard soju (they are also classified as liqueurs in the liquor tax law). They have a lower alcohol content as well. Their composition is different, as indicated above. Soju-based cocktails are not suited for consumption with meals.

    (f) It is inappropriate to associate pre-mixes with standard soju, in the same way that it would be inappropriate to associate Bailey�s (a blend of, inter alia, fresh cream and whisky) with whisky. Bailey�s and soju-based cocktails are classified under the same heading, with other liqueurs, in the Korean liquor tax law.

    (g) Finally, it is easy to overestimate the popularity of soju-based cocktails, as the European Communities has done. As their novelty has worn off, the increases in sales of them have tapered off.

  12. In response to a question concerning whether Article III:2 covers potential or future competition, Korea states that:
  13. (a) If "potential competition" refers to competition that would exist "but for" an allegedly discriminatory tax, Korea could imagine that potential competition falls within the scope of Article III:2. Korea�s discussion of "pre-tax" prices addressed this argument and shows that even if the effect of the tax were eliminated, the products at issue would not be in direct competition. The pre-tax prices of the products at issue, according to the complainants� own figures, range from 400 per cent more expensive than soju before tax to more than 1 800 per cent more expensive.

    (b) If by "future competition", the panel means competition that would appear at some point in the future if, for example, consumers changed their habits, or if the pre-tax price of whisky fell to the level of soju, then Korea considers that "future competition" is not covered by Article III:2. Complainants cannot base Article III:2 allegations on speculations about future changes in the market. Rather, complainants must wait to bring a WTO case if and when relevant changes appear.

  14. In response to a question about the possible explanations of apparent inconsistencies in the Dodwell study, Korea states that:
  15. (a) Choice subject to large random elements

    (i) If one were seeking an explanation of ostensible inconsistencies of choice in real life, the idea that choice is subject to a large random element would be an obvious first hypothesis. It leads to a view of buyers choosing good X one day or hour and good Y the next, depending on mood or a host of other circumstances. The relative price of X and Y might affect the frequency with which each is chosen. In a short enough period of observation, however, the random element might dominate, so that some consumers will appear to respond to a rise in the price of X by buying less Y.

    (ii) As an explanation of inconsistencies in answers to a questionnaire based upon hypothetical prices, however, this hypothesis is problematic. An interview will normally be short, so that mood or other factors that might drive the purchase of one good rather than another might reasonably be assumed to be constant throughout the interview. If that assumption is correct, however, random elements affecting demand cannot be called upon to explain inconsistencies in results.

    (iii) Of course, the assumption that mood is constant through the interview may be false. If it is false, however, interpretation of the results of the survey faces a different problem. In the present context, for example, the purpose of the Dodwell survey is to isolate the effects of changes in prices from other factors that might affect demand. But if mood or other factors change within the interview, or if respondents are allowed to imagine themselves making one choice in one mood, and another choice in another mood, the interview has failed to isolate changes in prices from other factors affecting demand. Its results will give a false picture of the effect of prices on demand.

    (b) Mistakes in reporting responses

    A simple explanation of the inconsistencies is that respondents are being consistent, but that interviewers are mis-reporting their responses. This hypothesis is included for completeness only.

    (c) Possible explanation of unexpected responses

    Were the responses in unexpected directions, but consistent, the facts to be explained would be different. We offer below two hypotheses that might in principle explain unexpected results, and also comment on why these seem incapable of explaining inconsistencies in result.

    (d) Gifts and prices

    (i) Respondents who think they are being asked about a single bottle purchase, may answer questions with the purchase of a bottle of spirits as a gift in mind. In that case, however, they might respond to price changes in ways that appear perverse. A reduction in price might make a spirit less desirable as a gift, and an increase might make it more desirable. The position would be further complicated if respondents answered some questions thinking in terms of gifts, and some questions thinking in terms of personal consumption.

    (ii) But while the gift motive might in principle explain ostensibly perverse reactions to changes in price, it lacks explanatory power in the present case. That is because, first, it is inconsistent rather than perverse reactions that are primarily at issue here. Second, the inconsistent reactions in chart 2 are responses to changes in the price of soju, not whisky, and standard soju is not usually given as a gift: it is too cheap to play that role.

    (e) Whisky and soju are complements in consumption

    (i) There is no great difficulty in imagining circumstances in which two alcoholic beverages are complements, at least for some drinkers. In some communities, for example, whisky is typically drunk with a beer "chaser". Alternatively, drinkers might follow a ritual of drinking two rounds of whisky and then two rounds of beer. In either case, beer and whisky might behave as complements rather than substitutes - rather than a rise in the price of whisky increasing the quantity consumed of beer, which is what would happen if beer and whisky were substitutes, the rise in the price of whisky might reduce the quantity consumed of beer.

    (ii) Drinkers may act as if whisky and a beer chaser, for example, is a single drink. Thus, an increase in the price of either beer or whisky will reduce the number of drinks taken. It will therefore reduce the quantity consumed of the drink whose price has remained constant.

    (iii) The problem, though, is that in Dodwell Chart 2, whisky and soju act like substitutes when the price of soju rises from 1 000 to 1 100 won, but like complements when the price of soju rises from 1 100 to 1 200. It is not the latter fact that is hard to explain (at least in principle!) - it is the inconsistency between the two.

    (iv) One might think in terms of a population made up of some drinkers who regard scotch and soju as substitutes, and some who regard them as complements. For some price changes the first group dominates, while for others the second group determines the direction of the net change.

    (v) Before pressing along that theoretical path, however, it is well to recall what is driving the problem. At issue is the effect of a 100-won change in the price of a bottle of soju306. But for soju and scotch to be complements, they must be drunk in a tight combination with one another. What then counts is the price of the combination. But with the price of scotch so many multiples of the price of soju, a 10 per cent change in the price of soju will have only a very small effect on the price of a soju-whisky combination. For a 100-won rise in the price of soju to cause the number selecting premium whisky to fall from 41 to 36, and the number selecting standard scotch to fall from 56 to 49 requires a sensitivity to price that is nether plausible nor suggested by anything else in the Dodwell findings.

    (f) Korea concludes by noting that in its first submission, it described the inconsistencies as "troubling", but commented that the Dodwell Study "has much more serious problems". 307 Korea sees no reason to change that assessment.

    (g) Korea also continues to believe that the attention of respondents might have wandered during their progress though the 16 sets of hypothetical prices offered them by Dodwell interviews (to say nothing of that of the interviewers themselves). That hypothesis is the one that seems to best fit the facts.

  16. In response to a question concerning the water content of distilled and diluted soju, Korea states that:
  17. (a) For the benefit of the Panel Korea hereby provides an answer related to the water content of both standard and distilled soju and also explains briefly the manufacturing process to better understand the differences.

    (b) In case of standard soju, water is added before and after distillation. Prior to distillation, one steams tapioca and/or sweet potatoes so that they are in a mashed form. Second, water is added. Third, enzymes and yeast are added so that the mashed tapioca and/or sweet potatoes will ferment. This fermentation process will lead to a 10-11 per cent alcoholic content liquid which is in a sludge form. The ingredients constitute 20 per cent, water 79.9 per cent and yeast 0.1 per cent.

    (c) Then the material undergoes continuous distillation until one obtains as pure an alcohol as possible (95 per cent ethyl alcohol). After distillation, water is added and then six to seven additives are inserted. Ethyl alcohol (joojung) constitutes 26.4 per cent and water constitutes 73.6 per cent at this stage.

    (d) Contrary to the notion that standard soju is simply a diluted form of distilled soju, the latter uses different base materials, primarily rice and sometimes other grains. Water is added only prior to distillation. No water is added after distillation. Producing a 45 per cent distilled soju through single distillation is a know-how developed by Korean producers over several hundred years.

    (e) In case of distilled soju, one takes white rice and steams it. Afterwards, one adds water and yeast which acts as a catalyst to commence the fermentation process. The ingredients take up 40 per cent, water 59 per cent and yeast 1 per cent. After the material ferments, one has a product which has a low alcoholic content. Then the fermented product undergoes a single distillation so that the final product has 45 per cent alcoholic content. No water is added after distillation.

    The percentage of water added is illustrated in the following chart:

    Before distillationAfter distillation
    Standard soju79.9%73.6%
    Distilled soju59.0%0%

To continue with Korea


302 According to the United States, contrary to Korea's allegation, the Appellate Body in Japan - Taxes on Alcoholic Beverages II did not reject the existence of a single tariff heading as significant guidance in examining substitutability. The Appellate Body distinguished between tariff nomenclature and tariff bindings, which it said could include a wide range of products and therefore must be viewed with caution in examining the term "like". Appellate Body Report, supra., pp.21-22.

303 Appellate Body Report, supra., p. 29.

304 Appellate Body Report on Japan � Taxes on Alcoholic Beverages II, supra., p. 16.

305 See the Sofres report: "In the past Korean women had negative sentiments towards alcohol., However, the current generation of women is drinking more frequently each year. The Korean distillers and producers reflect this trend by offering low alcoholic content drinks like ... "Lemon Soju". Please note that Korean producers do not use the term "soju" in the brand name for these pre-mixes. Some examples of names are "Lemon 15", "Cherry 15", "Lemon Remix" and "Cherry Remix".

306 At the current exchange rate, 100 won equals US$ 0.0691 and ECU 0.0637 (as of 23 March 1998).

307 See Attachment 2 of Korea, p.6.