(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) |
Sugar | Sugar |
Citric Acid | Stevioside |
Co2 Gas | Citric Acid |
Lemon/Cherry concentrate | Mineral Salt |
Acerola juice concentrate | Amino acid |
Lemon/cherry spices | Solbitol |
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.
(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.
(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.
(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.