(ii) The 1998 Search Marketing Survey

186. The European Communities further claims that another consumer survey performed by Search Marketing S.A. at the request of the EC spirits industry (the "1998 SM survey") ¹²⁴ further supports a finding that pisco and all other distilled spirits are directly competitive and substitutable.
187. According to the European Communities, the surveyors asked two questions to a representative sample comprising over 400 consumers who had purchased both pisco and at least one other spirit during the last six months.
188. The European Communities explains that the purpose of the first question was to measure the substitutability between pisco and other distilled spirits under current taxation and pricing conditions. To that effect, the question was drafted in the following terms: "if you wanted to buy a bottle of pisco, but pisco was not available, what of the following beverages would you buy instead?". Possible answers included, in addition to other types of distilled spirits, wine, beer, non-alcoholic beverages and "nothing". The same question was then repeated for each of the other types of distilled spirits covered by the survey.
189. The European Communities notes that Table 21 below summarises the answers given by the respondents in the situation where they wanted to buy pisco but that spirit was not available. It indicates that a large majority of consumers regard other distilled spirits as the closest substitute for pisco.

Table 21 ¹²⁵

Response to the Question: "What would you buy if pisco was not available?"

Other spirits	70 %
Wine/Beer	17 %
Non alcoholic beverages	0 %
Nothing	13 %

190. The European Communities points out that in Japan � Taxes on Alcoholic Beverages II, Japan submitted a consumer survey showing that, in case of unavailability of shochu, "only" 10 % of consumers would switch to whisky and other spirits, whereas the remaining 90 % would turn to beer or other beverages. The panel was of the view that a 10 % switch was "proof of significant elasticity of substitution" between shochu and other spirits. ¹²⁶
191. In the view of the European Communities, equally significant, as shown in Table 22 below, pisco was mentioned as the main alternative to each of the other spirits covered by the survey in the hypotheses where those spirits were not available.

Table 22 ¹²⁷

Response to the Question: "What would you buy if [whisky/tequila/brandy/rum/vodka/gin] was not available?"

	Whisky	Tequila	Brandy	Rum	Vodka	Gin
Pisco	50 %	56 %	52 %	43 %	48 %	45 %
Other spirits	25 %	17 %	15 %	19 %	21 %	19 %
Wine/beer	3 %	2 %	2 %	2 %	2 %	2 %
Non alcoholic beverages	0 %	0 %	0 %	0 %	0 %	0 %
Nothing	22 %	25 %	31 %	36 %	29 %	34 %

Basis: % of all respondents

192. The European Communities further explains that the second question asked by the surveyors aimed to measure the respondents' reaction to changes in the relative price of pisco and the other distilled spirits. Respondents were initially asked to make purchase choices between all products at current prices. ¹²⁸ They were then shown an estimate of the prices that would prevail if all distilled spirits were taxed at the rate of 27 % ad valorem (i.e., the rate that will apply to most pisco as from 1 December 2000), and asked to make a choice at those prices. For these purposes, it was estimated that the envisaged tax change would bring about an increase in the prices of pisco of 1.7 % and a simultaneous reduction in the prices of whisky and of "other spirits" of 25.3 % and 2.3 %, respectively.
193. According to the European Communities, the responses to this second question indicate the existence of a significant degree of cross-price elasticity between pisco and other spirits. As shown in Table 23, the share of respondents choosing whisky and other spirits instead of pisco would increase from 17.7 % to 30.5 %, i.e., by as much as 72 %. The increase is particularly large in the case of whisky, which would benefit from the largest price reduction in the event that all spirits were taxed at 27 %. The share increase is also substantial, even if less marked, in the case of "other spirits", which would benefit from a smaller price reduction than whisky. Finally, it is worth noting that the increase in the share of whisky and "other spirits" takes place at the expense of all the categories of pisco, thus refuting any possible claims to the effect that whisky and "other spirits" compete only with certain types of pisco.

Table 23 ¹²⁹

	At current prices	At prices with 27 % tax	Variation (%)
Pisco tradicional	12 %	9.7 %	- 19.2 %
Pisco especial	47.2 %	42.3 %	- 10.4 %
Pisco reservado + Gran pisco	23 %	17.5 %	- 23.9 %
Whisky	6.3 %	14.1 %	+ 124 %
Other spirits	11.5 %	16.4 %	+ 43 %

194. The European Communities alleges that the 1998 SM survey can show only the immediate reaction of consumers to price changes. Yet, as discussed above, the consumption of distilled spirits is based on habits, which only change gradually. As a result, short-run elasticities of substitution between distilled spirits are, as a general rule, much lower than long-run elasticities. This mean that, over a certain period of time, the price changes resulting from the elimination of tax differentials are likely to lead to a shift in consumption from pisco to other spirits even larger than that shown in Table 23 above.
195. In rebuttal, Chile argues that given the weakness of a complaint based on cross-price elasticity (as noted in paragraph 4.[219] below, a low coefficient of 0.26 was computed as cross-price elasticity for pisco and whisky), the European Communities tries to bolster its argument with an assortment of information based on a survey of 400 "representative" Chilean consumers. The EC survey does not logically make the case that pisco is directly competitive or substitutable with other spirits.
196. Chile further argues that the European Communities also attempts to use marketing surveys as a kind of substitute for econometric analysis. Specifically, the European Communities refers to marketing surveys in which various consumers were asked what they would buy if there were no pisco and how they would react based on assumed increases in pisco prices accompanied by decreases in whisky prices. Such surveys are inherently much less reliable that an econometric analysis based on 15 years of data. Further, even the responses the European Communities received do not establish a directly competitive or substitutable relationship.
197. The European Communities responds that Chile seeks to discredit the consumer surveys submitted by the European Communities, but fails to advance any specific ground to cast doubt on the reliability of those surveys. Thus, for example, Chile appears to consider that it is sufficient to describe the sample of one of those surveys as being composed of 400 "representative" Chilean consumers in order to dispose ipso facto of that survey. The European Communities asks whether and why Chile is suggesting that the sample was not statistically representative.
198. The European Communities further notes that Chile argues that consumer surveys are "inherently much less reliable" than econometric studies. Previous panels, however, have not hesitated to rely upon the findings of consumer surveys in order to establish that products were "directly competitive or substitutable". In Japan � Taxes on Alcoholic Beverages II, the panel discarded a flawed regression submitted by Japan in favour of a much more robust consumer survey presented by the complainants. ¹³⁰
199. Chile further responds that the last "evidence" provided by the European Communities, to support the idea that there is a significant cross-price elasticity between pisco and other distilled spirits, are the results of another survey conducted by Search Marketing, which is extremely weak.
200. Chile notes that the first piece of evidence cited by the European Communities, is the answer made by consumers to the question "what would you buy if pisco is not available". The conclusions to be drawn from this kind of question are highly questionable, because, as any marketing survey expert could tell, people facing situations that are not "normal" in their mind, tend to react in different and unpredictable ways. Also, without a very well designed system to avoid inconsistent answers (and the evidence provided by the European Communities does not lead to a conclusion as to whether those measures have been taken or not) the results are of very little significance. Last � but not least � the question was designed to actually "force" respondents to buy something else, or buy nothing, but would not provide one of the most logical alternatives "I will buy (pisco, whisky, gin, etc.) in another store".
201. Chile further points out that the other piece of the Survey is a "quantitative" analysis, in which the consumers, faced with a change in the prices of pisco, whisky and other liquors, show their preferences in the new situation (Table 23). A detailed analysis of this shows the inconsistency of the results, because they don't resist any serious analysis and, therefore must be discarded as evidence.
202. Chile also argues that in fact, Table 24 ¹³¹ shows that if the price of pisco increases 1.7%, the price of whisky is reduced by 25.3% and the price of "other spirits" by 2.3%, then the consumption of pisco will decrease by 15.5%, the consumption of whisky will increase by 123.8% and the consumption of other spirits by 42.6%.

Table 24 ¹³²

Substitution Elasticity

	PRICE CHANGE	DEMAND CHANGE
PISCO	+ 1.7 %	- 15.5 %
WHISKY	- 25.5 %	+123.8 %
OTHERS	-2.3 %	+42.6 %

203. Chile comments that some basic microeconomic discussion must be added here, in order to make its points clear. It discussed the relationship between price elasticity (that is, how much the price of the good itself influences the quantity sold of that good) and cross-price elasticity, that means, how much the price variation of competitive goods, influence the quantity of the good at stake.
204. In the view of Chile, in "normal" goods, that is, goods that are consumed less if the price rises and more if the price falls, the own-price elasticity must be greater or equal to the cross-price elasticity of any competitive or substitutable good. If two goods could be substituted one for the other without any cost for the consumer (that means, they are almost equal products), the cross elasticity will be the same as the price elasticity (with sign changed) that is, a 1% increase in the price of a good, has the same effect on the consumption as a 1% decrease in the price of its competitor. If there is a cost in substituting those goods, then the same percentage price variation in the competitive good price has less effect on the demand than that percentage price variation of the good at stake. Thus, normally the maximum cross elasticity is equal to the price elasticity.
205. With this concept in mind, Chile reviews the results of 1998 SM survey, shown at Table 23 above. The survey was conducted in such a way that the respondents faced an hypothetical "once and for all" price variation, and therefore changes in demand are not influenced by changes in personal income, tastes, habits, etc. The only variable that changes is the price of different spirits.
206. Chile goes on to state that if it takes the border condition, that is, that cross elasticities are equal (in absolute value) to price elasticity (something not seen in the real world, but that presents the best case to sustain the reliability of the survey), it may calculate (with the data included in Table 23) what is the underlying minimum price elasticity.
207. Chile presents the following results:

Whisky:	-5
Other Spirits:	+2 (that is to say, an increase in price will mean an increase in consumption)
Pisco:	-0.5

208. In the opinion of Chile, the findings for whisky and other spirits are absurd, and speak loudly about the lack of reliability of the survey.
209. Chile explains that in the case of the whisky, a price elasticity of -5, means that a 10% reduction in price (with the prices of every other spirit unchanged) will lead to an increase in terms of volume of 50%. This is absolutely out of range. The whisky industry will be very pleased to face such a price elasticity, but this is not the case. Consumer goods usually show price elasticity below 2. This "finding" is also rebutted by the experience in 1982 � 1986. By then, the price of whisky grew by 67%, and per capita consumption fell by 64%, therefore price elasticity is at the most 1 (if there is some cross elasticity with pisco, price elasticity will be even less than 1).
210. Chile concludes that no comments need to be made on the elasticities of other spirits, because it is clear that nobody could sustain that case. This exercise shows that the 1998 SM survey is far from being a robust consumer survey.
211. Chile also disagrees with the EC comment that the "decrease in pisco price may have had some influence in pisco consumption". It is easy to realize that, if a good has no response to its price (the quantity is the same no matter what its price is), its demand is said to be inelastic, and those goods has no competitors. Chile has developed at length the reason why such goods have no competitors. Therefore, the comment made by the European Communities in its rebuttal should be regarded as a plain economic error.
212. Chile claims that, in summary, the European Communities has failed to provide persuasive and sufficient evidence of its assertion that pisco and other spirits are directly competitive or substitutable, as requested by Article III:2, second sentence.
213. The European Communities contests Chile's argument that the 1998 SM survey is not reliable because it did not offer as a possible survey response the option to go to another shop when asking the question "what would you drink, if pisco were not available?". The hypothetical question of the best alternative drink is part of the standard repertoire of survey design. As the survey offers both the possibility to switch to another drink or to drink nothing at all, it provides a good indicative assessment of the level of substitutability between pisco and other types of alcoholic beverages. If substitutability were nil, no choice of an alternative drink would have been made. The hypothetical question is certainly not beyond the intellectual capacity of a consumer. The proposal that the respondent should have had the choice to go to another store is not a serious suggestion.
214. The European Communities also argues that it was stated that the elasticities implied by the survey lead to the impossible result that the elasticity of whisky consumption with respect to pisco prices (cross-price elasticity) is higher than the own-price elasticity. These results are certainly not in violation of any economic principle. Evidently, the survey choices made are plausible and intuitive. Furthermore, the statement may confuse cross-price elasticity with elasticity of substitution. Without going into the technical details, it has to be kept in mind that the market shares of whisky and pisco are very unequal. By way of example, the European Communities refers to the own-price elasticity the Chilean delegation appears to have derived from the survey, namely of 0.5 for pisco and 5 for whisky. If whisky consumption is 8,000 cases annually and that of pisco 100,000 (the ratio roughly represents the actual proportions), a 1 % price increase for pisco will lead to a reduced consumption of 500 cases, a 1% increase in whisky prices will affect 400 cases of alcoholic drinks. It can easily be seen that a price change in pisco can have more influence on whisky consumption than a price change in whisky itself.
215. The European Communities asserts that this result would even be stronger if one applies own-price elasticities that are less unequal than those derived by the Chilean delegation. The European Communities has used the Chilean values only as an example, because the European Communities has no means of assessing how the Chilean delegation might have calculated them.
216. In conclusion, the European Communities alleges that therefore, Chile's argument does not invalidate the technical correctness the survey. Instead it can be summarised as stating simply that the Chilean delegation believes that the substitutability is lower than those derived from the survey.

To continue with The 1995 Gemines Study

¹²⁴ EC Exhibit 22.

¹²⁵ EC First Submission, Table 17.

¹²⁶ Panel Report on Japan - Taxes on Alcoholic Beverages II, supra., para 6.31.

¹²⁷ EC First Submission, Table 18.

¹²⁸ EC Exhibit 23.

¹²⁹ EC First Submission, Table 19 (p. 60).

¹³⁰ The European Communities refers to Panel Report on Japan - Taxes on Alcoholic Beverages II, supra., paras. 6.28-6.32. According to the European Communities, Japan's regression had similar methodological problems to those of Chile's regression analysis.

¹³¹ According to Chile, Table 24 is elaborated from data submitted by the European Communities.

¹³² Chile Oral Statement at the Second Substantive Meeting, Table V.