Report of the Appellate Body

(Continued)

B. Assessment of Chile's Price Band System in the Light of Article 4.2 and Footnote 1

218. We turn now to the Panel's finding that Chile's price band system is a border measure similar to a variable import levy and a minimum import price within the meaning of footnote 1 to Article 4.2 of the Agreement on Agriculture .¹⁹³

219. Footnote 1 lists six categories of border measures and a residual category of such measures that are included in "measures of the kind which have been required to be converted into ordinary customs duties" within the meaning of Article 4.2.¹⁹⁴ The list is illustrative, and includes "quantitative import restrictions, variable import levies, minimum import prices, discretionary import licensing, non-tariff measures maintained through state-trading enterprises, voluntary export restraints, and similar border measures other than ordinary customs duties". These kinds of measures were identified by the negotiators of the Agreement on Agriculture as measures that had to be converted into ordinary customs duties in order to ensure enhanced market access for imports of agricultural products.

220. Before the Panel, Argentina alleged that Chile's price band system is a "minimum import price" or a "variable import levy" system or, in any event, a "similar border measure other than ordinary customs duties", and that, because of the prohibition of such measures in Article 4.2, Chile's price band system must not be maintained.¹⁹⁵

221. A plain reading of Article 4.2 and footnote 1 makes clear that, if Chile's price band system falls within any one of the categories of measures listed in footnote 1, it is among the "measures of the kind which have been required to be converted into ordinary customs duties", and thus must not be maintained, resorted to, or reverted to, as of the date of entry into force of the WTO Agreement.¹⁹⁶ Therefore, we will examine whether Chile's price band system falls within one or more of the categories of measures that are prohibited by Article 4.2 and footnote 1.

222. It must be emphasized that the Panel did not find that Chile's price band system constitutes a "variable import levy" or "minimum import price" system per se. Rather, the Panel found that Chile's price band system:

� is a hybrid instrument, which has most, but not all, of its characteristics in common with either or both a variable import levy and a minimum import price. After careful assessment of the evidence before us, however, we consider as a factual matter that the Chilean PBS shares sufficient fundamental characteristics with those schemes for it to be considered similar to them, and that the observed differences between the Chilean PBS and either of those schemes are not of such a nature as to detract from this similarity.¹⁹⁷ (original emphasis, underlining added)

223. Chile argues, on appeal, that the Panel erred in finding that Chile's price band system is a border measure similar to a variable import levy or a minimum import price within the meaning of footnote 1 to Article 4.2.

224. At the outset, we stress that, as Argentina argues¹⁹⁸, the Panel's characterization of its finding "as a factual matter" does not mean that the issue whether Chile's price band system is a border measure similar to a variable import levy or a minimum import price is shielded from appellate review. This is a question of law, and not of fact, and thus is clearly within our jurisdiction under Article 17.6 of the DSU.¹⁹⁹ As we said in our Report in EC - Hormones, the assessment of the consistency or inconsistency of a given fact or set of facts with the requirements of a given treaty provision is an issue of legal characterization.²⁰⁰ The mere assertion by a panel that its conclusion is a "factual matter" does not make it so. Here, the Panel's interpretation of the terms "variable import levies", "minimum import prices", and "similar border measures other than ordinary customs duties", as these terms are used in footnote 1, constitutes, not a factual determination, but rather a legal interpretation of the words of Article 4.2. Hence, these interpretations are within the purview of appellate review under Article 17.6 of the DSU. Moreover, the Panel's appraisal of Chile's price band system in the light of its legal interpretation is an application of the law to the facts of the case. All the same, in reviewing the Panel's assessment of Chile's price band system, we are mindful of the need to give due deference to the discretion of the Panel, as the "trier of fact", to weigh the evidence before it.

225. The Panel described its approach to assessing whether Chile's price band system is similar to "variable import levies" and/or "minimum import prices" within the meaning of footnote 1 as follows:

First, as regards the term "similar", dictionaries define this term as "having a resemblance or likeness", "of the same nature or kind", and "having characteristics in common". Two measures are in our view "similar" if they share some, but not all, of their fundamental characteristics. If two measures share all of their fundamental characteristics, they are identical rather than similar. A border measure should therefore have some fundamental characteristics in common with one or more of the measures explicitly listed in footnote 1. It is then a matter of weighing the evidence to determine whether the characteristics are sufficiently close to be considered "similar".²⁰¹ (emphasis added, footnotes omitted)

226. We agree with the first part of the Panel's definition of the term "similar" as "having a resemblance or likeness", "of the same nature or kind", and "having characteristics in common".²⁰² However, in our view, the Panel went unnecessarily far in focusing on the degree to which two measures share characteristics of a "fundamental" nature. We see no basis for determining similarity by relying on characteristics of a "fundamental" nature. The Panel seems to substitute for the task of defining the term "similar" that of defining the term "fundamental". This merely complicates matters, because it raises the question of how to distinguish "fundamental" characteristics from those of a less than "fundamental" nature. The better and appropriate approach is to determine similarity by asking the question whether two or more things have likeness or resemblance sufficient to be similar to each other. In our view, the task of determining whether something is similar to something else must be approached on an empirical basis.

227. As suggested by Argentina, the Panel decided to assess Chile's price band system by comparing it to several individual categories of measures listed in footnote 1. Before looking at these categories of measures, we note that all of the border measures listed in footnote 1 have in common the object and effect of restricting the volumes, and distorting the prices, of imports of agricultural products in ways different from the ways that ordinary customs duties do. Moreover, all of these measures have in common also that they disconnect domestic prices from international price developments, and thus impede the transmission of world market prices to the domestic market. However, even if Chile's price band system were to share these common characteristics with all of these border measures, it would not be sufficient to make that system a "similar border measure" within the meaning of footnote 1. There must be something more. To be "similar", Chile's price band system-in its specific factual configuration-must have, to recall the dictionary definitions we mentioned, sufficient "resemblance or likeness to", or be "of the same nature or kind" as, at least one of the specific categories of measures listed in footnote 1.

228. Before addressing the issue of how much or what kind of "similarity" Chile's price band system must display to be a measure prohibited by Article 4.2, it is necessary for us to identify with what that system is required to be similar. Any examination of "similarity" presupposes a comparative analysis. Thus, to determine whether Chile's price band system is "similar" within the meaning of footnote 1, it is necessary to identify with which categories that system must be compared. The Panel compared Chile's price band system with the same categories as those identified by Argentina. While Chile disagreed with the conclusions the Panel reached in relying on that comparison, Chile does not dispute the choice of those categories.

229. In assessing whether Chile's price band system is a "similar border measure", the Panel compared Chile's system to "variable import levies" and "minimum import prices" within the meaning of footnote 1. WTO Members have not chosen to define any of these "terms of art" in the Agreement on Agriculture or anywhere else in the WTO Agreement. The Panel concluded that it could not develop an interpretation of the term "variable import levies" solely on the basis of the methods of interpretation codified in Article 31 of the Vienna Convention.²⁰³ The Panel decided, therefore, to have recourse to "supplementary means of interpretation" within the meaning of Article 32 of that Convention. This led to the Panel's identification of what the Panel described as "fundamental characteristics" of "variable import levies" and "minimum import prices".²⁰⁴

230. In response to our questioning at the oral hearing, the participants said that they agree with these characteristics, although Chile believes that the Panel's list is incomplete.²⁰⁵ However, we do not believe that the Panel properly applied Article 32 of the Vienna Convention in its analysis;²⁰⁶ nor do we find it useful to endorse the characteristics identified by the Panel through this process as being of a "fundamental" nature.

231. Instead, we proceed to interpret the terms "variable import levies" and "minimum import prices, using the customary rules of interpretation as codified in the Vienna Convention. As always, in employing these rules, we discuss the ordinary meaning of these terms in their context, and in the light of their object and purpose.

232. We begin with the interpretation of "variable import levies". In examining the ordinary meaning of the term "variable import levies" as it appears in footnote 1, we note that a "levy" is a duty, tax, charge, or other exaction usually imposed or raised by legal execution or process.²⁰⁷ An "import" levy is, of course, a duty assessed upon importation. A levy is "variable" when it is "liable to vary".²⁰⁸ This feature alone, however, is not conclusive as to what constitutes a "variable import levy" within the meaning of footnote 1. An "ordinary customs duty" could also fit this description. A Member may, fully in accordance with Article II of the GATT 1994, exact a duty upon importation and periodically change the rate at which it applies that duty (provided the changed rates remain below the tariff rates bound in the Member's Schedule).²⁰⁹ This change in the applied rate of duty could be made, for example, through an act of a Member's legislature or executive at any time. Moreover, it is clear that the term "variable import levies" as used in footnote 1 must have a meaning different from "ordinary customs duties", because "variable import levies" must be converted into "ordinary customs duties". Thus, the mere fact that an import duty can be varied cannot, alone, bring that duty within the category of "variable import levies" for purposes of footnote 1.

233. To determine what kind of variability makes an import levy a "variable import levy", we turn to the immediate context of the other words in footnote 1. The term "variable import levies" appears after the introductory phrase "[t]hese measures include". Article 4.2-to which the footnote is attached-also speaks of "measures". This suggests that at least one feature of "variable import levies" is the fact that the measure itself�as a mechanism�must impose the variability of the duties. Variability is inherent in a measure if the measure incorporates a scheme or formula that causes and ensures that levies change automatically and continuously. Ordinary customs duties, by contrast, are subject to discrete changes in applied tariff rates that occur independently, and unrelated to such an underlying scheme or formula. The level at which ordinary customs duties are applied can be varied by a legislature, but such duties will not be automatically and continuously variable. To vary the applied rate of duty in the case of ordinary customs duties will always require separate legislative or administrative action, whereas the ordinary meaning of the term "variable" implies that no such action is required.

234. However, in our view, the presence of a formula causing automatic and continuous variability of duties is a necessary, but by no means a sufficient, condition for a particular measure to be a "variable import levy" within the meaning of footnote 1.²¹⁰  "Variable import levies" have additional features that undermine the object and purpose of Article 4, which is to achieve improved market access conditions for imports of agricultural products by permitting only the application of ordinary customs duties. These additional features include a lack of transparency and a lack of predictability in the level of duties that will result from such measures. This lack of transparency and this lack of predictability are liable to restrict the volume of imports. As Argentina points out, an exporter is less likely to ship to a market if that exporter does not know and cannot reasonably predict what the amount of duties will be.²¹¹ This lack of transparency and predictability will also contribute to distorting the prices of imports by impeding the transmission of international prices to the domestic market.

235. We turn now to the interpretation of the term "minimum import prices". Argentina alleges, and the Panel found, that Chile's price band system is similar also to a "minimum import price"²¹², which is another prohibited measure listed in footnote 1 of Article 4.2.

236. The term "minimum import price" refers generally to the lowest price at which imports of a certain product may enter a Member's domestic market. Here, too, no definition has been provided by the drafters of the Agreement on Agriculture. However, the Panel described "minimum import prices" as follows:

[these] schemes generally operate in relation to the actual transaction value of the imports. If the price of an individual consignment is below a specified minimum import price, an additional charge is imposed corresponding to the difference.²¹³

237. The Panel also said that minimum import prices "are generally not dissimilar from variable import levies in many respects, including in terms of their protective and stabilization effects, but that their mode of operation is generally less complicated."²¹⁴ The main difference between minimum import prices and variable import levies is, according to the Panel, that "variable import levies are generally based on the difference between the governmentally determined threshold and the lowest world market offer price for the product concerned, while minimum import price schemes generally operate in relation to the actual transaction value of the imports."²¹⁵ (emphasis added)

238. In response to questioning at the oral hearing, the participants said they do not object to the Panel's definition of a "minimum import price". Their disagreement relates instead to whether Chile's price band system is similar to a minimum import price system prohibited by Article 4.2.

239. We turn next to the Panel's determination that Chile's price band system is a border measure similar to "variable import levies" and "minimum import prices". We must determine whether Chile's price band system�in its particular features�shares sufficient features with these two categories of prohibited measures to resemble, or "be of the same nature or kind" and, thus, also to be prohibited by Article 4.2.

240. The Panel described Chile's price band system as having an "intrinsically unstable, intransparent and unpredictable nature �".²¹⁶ Indeed, the Panel saw "a considerable lack of transparency and unpredictability" in the measure.²¹⁷ On appeal, Argentina emphasizes that the combination of a lack of transparency and a lack of predictability are the features of Chile's price band system that, most of all, make it "similar" to "variable import levies" within the meaning of footnote 1.²¹⁸

241. We note that it is undisputed between the participants that a formula inherent in Chile's price band system causes and ensures automatic and continuous variability of the duties resulting from that system. However, one of Chile's arguments on appeal relates to the particular formula used in establishing the price bands in Chile's system. Chile alleges that the Panel did not take sufficient account of the fact that the lower and upper thresholds of Chile's price bands vary in relation to "world prices", and not in relation to domestic prices, or to some Chilean target price.²¹⁹ Chile argues that its price band system compares "current world prices" with "historic world prices" over a five-year period, rather than comparing them with prices on Chile's domestic market. Chile maintains that the lower thresholds of Chile's price bands are different in this respect from the floor or minimum price that the Panel thought was one of the characteristics of both variable import levies and minimum import price systems.²²⁰

242. The Panel stated that:

the lower threshold of the Chilean PBS is not explicitly derived from, or linked to, an internal market-related price, as is often the case in variable import levy schemes.²²¹

The Panel thus recognized that Chile's price bands vary in relation to "world prices", and that, in this respect, Chile's price band system is not identical to a variable import levy or a minimum import price scheme. The fact that Chile's price bands vary in relation to�albeit historic-world prices, rather than in relation to domestic market or target prices, does not suggest-at first glance-that Chile's price band system effectively disconnects the domestic market from international price developments. We will, however, return to this issue later.

243. The Panel also stated that Chile's price band system need not be identical to variable import levies or minimum import prices to be considered similar to these prohibited categories of measures listed in footnote 1, provided that Chile's price band system bears sufficient resemblance to such measures. The Panel went on to examine whether the determination of the lower thresholds of Chile's price bands operates in such a way as to render it similar to a domestic target price or domestic market price. The Panel noted that:

on the basis of the evidence before us, it cannot be excluded that the lower threshold of the PBS, given the way in which it is designed, particularly with the many adjustments made by the administering agencies to the basic world market price quotations employed, including for inflation, operates in practice as a "proxy" for such internal prices.²²²

244. In Chile's view, the Panel erred in law in finding "similarity" based on what "cannot be excluded". We believe that Chile reads too much into the Panel's formulation. The Panel did not equate Chile's price band system with variable import levies or minimum import price systems that are related to domestic target prices. Rather, taking into account the evidence submitted, the Panel stated only that the lower thresholds of Chile's price bands may often, but not in all cases, be equal to or higher than the domestic price. This may be due�in part�to the way in which the price band thresholds, which are first calculated on the basis of monthly f.o.b. world prices over the last five years, are converted to a c.i.f. basis. As Chile points out, this may also be due�in part�to the way in which domestic prices to a certain extent reflect changes in world market prices.²²³ As we see it, the Panel found "similarity" based on actual evidence, and not, as Chile implies, on conjecture.

245. We also find merit in the Panel's finding that:

the PBS thresholds are determined, inter alia, after discarding 25 per cent of "atypical observations" at the bottom and at the top, hence substantially increasing the likelihood that the lower threshold of the PBS will equal or exceed the higher internal price.²²⁴

Based on this, the Panel concluded that the lower thresholds of Chile's price bands operate like substitutes for domestic target prices. Hence, the Panel was satisfied that this feature of Chile's price band system was also similar to the features of variable import levies and minimum import prices.

246. We agree with the Panel's view-to a point. But we believe that the Panel placed too much emphasis on whether or not Chile's price bands are related to domestic target prices or domestic market prices. In our view-even though Chile's price bands are set in relation to world prices from a past five-year period�Chile's price band system can still have the effect of impeding the transmission of international price developments to the domestic market in a way similar to that of other categories of prohibited measures listed in footnote 1. There are factors other than world market prices that are relevant to the assessment of Chile's price bands. The prices that represent the highest 25 per cent as well as the lowest 25 per cent of the world prices from the past five years are discarded in selecting the "highest and lowest f.o.b. prices" for the determination of Chile's annual price bands. Furthermore, we place considerable importance on the intransparent and unpredictable way in which the "highest and lowest f.o.b. prices" that have been selected are converted to a c.i.f. basis by adding "import costs". As Chile concedes, no published legislation or regulation sets out how these "import costs" are calculated.²²⁵

247. In addition to the lack of transparency and the lack of predictability that are inherent in how Chile's price bands are established, we see similar shortcomings in the way the other essential element of Chile's price band system�the reference price�is determined. As we have explained, the duties resulting from Chile's price band system are equal to the difference between the price band thresholds and the reference price. Chile sets the reference price on a weekly basis, and it does so in a way that is neither transparent nor predictable.

248. The Panel described the particular reference price used in Chile's price band system in the following terms:

The Reference Price used in the context of the Chilean PBS is clearly disconnected from the actual transaction value, unlike minimum import price schemes. It does use a lowest "market of concern" price, however, similar to the lowest market offer price generally used in variable import levy schemes.²²⁶

249. Under Chile's price band system, the price used to set the weekly reference price is the lowest f.o.b. price observed, at the time of embarkation, in any foreign "market of concern" to Chile for "qualities of products actually liable to be imported to Chile".²²⁷ No Chilean legislation or regulation specifies how the international "markets of concern" and the "qualities of concern" are selected.²²⁸ Thus, it is not by any means certain that the weekly reference price is representative of the current world market price. Moreover, the weekly reference price used under Chile's price band system is certainly not representative of an average of current lowest prices found in all markets of concern. As a result, the process of selecting the reference price is not transparent, and it is not predictable for traders.

250. Furthermore, under Chile's system, the same weekly reference price applies to imports of all goods falling within the same product category, regardless of the origin of the goods, and regardless of the transaction value of the shipment. Moreover, unlike with the five-year average monthly prices used in the calculation of Chile's annual price bands, the lowest "market of concern" price used to determine the weekly reference price is not adjusted for "import costs", and thus is not converted from an f.o.b. basis to a c.i.f. basis. This is likely to inflate the amount of specific duties applied under Chile's price band system, because these duties are imposed in an amount equal to the difference between Chile's annual price band thresholds, which are based on higher c.i.f. prices, and Chile's weekly reference prices, which are based on lower f.o.b. prices. Therefore, the way in which Chile's weekly reference prices are determined contributes to giving Chile's price band system the effect of impeding the transmission of international price developments to Chile's market.

251. Consequently, even if were to assume, for the moment, that one feature of Chile's price band system is not similar to the features of "variable import levies" and "minimum import prices" because the thresholds of Chile's price bands vary in relation to�-albeit historic�world market prices rather than domestic target prices, this would not change our overall assessment of Chile's price band system. This is because specific duties resulting from Chile's price band system are equal to the difference between two parameters�the annual price band thresholds and the weekly reference prices applicable to the shipment in question. Therefore, continuing with our hypothesis, even if we were to assume that one of the two parameters�Chile's annual price band thresholds�does not distort the transmission of world market prices to Chile's market, it would nevertheless remain that the other parameter--Chile's weekly reference prices�is liable to distort-if not disconnect-that transmission by virtue of the way it is determined on a weekly basis. Consequently, even in such a hypothetical case, the duties resulting from Chile's price band system, which are equal to the difference between these two parameters, would not transmit world market price developments to Chile's market in the same way as "ordinary customs duties".

252. Thus, although there are some dissimilarities between Chile's price band system and the features of "minimum import prices" and "variable import levies" we have identified earlier, the way Chile's system is designed, and the way it operates in its overall nature, are sufficiently "similar" to the features of both of those two categories of prohibited measures to make Chile's price band system�in its particular features�a "similar border measure" within the meaning of footnote 1 to Article 4.2.

253. However, Chile argues that, in making its finding, the Panel failed to take proper account of the fact that the total amount of duties that may be levied as a result of Chile's price band system is "capped" at the level of the tariff rate of 31.5 per cent ad valorem bound in Chile's Schedule. According to Chile, the existence of this cap differentiates Chile's price band system from a "variable import levy". Chile argues that Chile's price band system enables imports to enter Chile's market below the lower thresholds of Chile's price bands when world market prices drop below a certain level, while allowing imports to enter at duty rates that can be as low as zero when the weekly reference prices rise above the upper thresholds of Chile's price bands. Chile submits that the cap makes Chile's price band system less distortive and less insulating than if Chile simply levied duties at its bound tariff level.²²⁹

254. This argument by Chile compels us to consider whether Chile's price band system ceases to be similar to a "variable import levy" because it is subject to a cap. In doing so, we find nothing in Article 4.2 to suggest that a measure prohibited by that provision would be rendered consistent with it if applied with a cap. Before the conclusion of the Uruguay Round, a measure could be recognized as a "variable import levy" even if the products to which the measure applied were subject to tariff bindings.²³⁰ And, there is nothing in the text of Article 4.2 to indicate that a measure, which was recognized as a "variable import levy" before the Uruguay Round, is exempt from the requirements of Article 4.2 simply because tariffs on some, or all, of the products to which that measure now applies were bound as a result of the Uruguay Round.

255. The context of Article 4.2 lends support to this interpretation. That context includes the Guidelines for the Calculation of Tariff Equivalents for the Specific Purpose Specified in Paragraph 6 and 10 of this Annex ("Guidelines"), which are an Attachment to Annex 5 on Special Treatment with respect to Paragraph 2 of Article 4. Both the Attachment and the Annex form part of the Agreement on Agriculture. Paragraph 6 of the Guidelines²³¹ envisages that tariff equivalents resulting from conversion of measures within the meaning of Article 4.2 may exceed previous bound rates. This implies that, even if the product to which that measure applied was in fact subject to a tariff binding before the Uruguay Round, conversion of that measure may nevertheless have been required. Therefore, a measure cannot be excluded per se from the scope of Article 4.2 simply because the products to which that measure applies are subject to a tariff binding.

256. Relevant context can also be found in Articles II and XI of the GATT 1994. If Members were free to apply a measure with a "cap"�which, in the absence of that "cap", would be a prohibited "variable import levy"�Article 4.2 would, in our view, add little to the longstanding requirements of Articles II:1(b) and XI:1 of the GATT 1947. In fact, Chile concedes that the scope of measures prohibited by Article 4.2 extends beyond the tariffs in excess of bound rates that are prohibited by Article II and the "restrictions other than taxes, duties and charges" that are prohibited by Article XI:1.²³² In any event, it is difficult to see why Uruguay Round negotiators would "compensate" Members for converting prohibited measures by permitting them to raise tariffs on certain products, while permitting those Members to retain those measures and, at the same time, impose those higher tariffs on those same products. It is not clear why, if this were so, a Member would ever have converted a measure. All that a Member would have had to do to comply with Article 4.2 would have been to adopt a tariff binding�even at a higher level-on the products covered by the original measure. Had this been the intention of the Uruguay Round negotiators, there would have been no need to list price-based measures in footnote 1 among the categories of measures prohibited by Article 4.2. The drafters of the Agreement on Agriculture simply could have adopted a requirement that all tariffs on agricultural products be bound.

257. Contrary to Chile's view, we are not persuaded that the presence or the absence of a cap is essential in determining whether or not Chile's price band system is similar to a measure prohibited by Article 4.2. Chile's tariff binding will impose a limit on the total amount of duties that may be applied, and thus permit fluctuations in world market prices to be reflected in Chile's market, in cases when the duties resulting from Chile's price band system, when added to the applied ad valorem duty, exceed that tariff binding. However, the existence of the tariff binding will not eliminate the distortion in the transmission of world market prices to Chile's market in all other cases, where the combination of the duties resulting from Chile's price band system, when added to the applied ad valorem duty, remains below Chile's bound rate of 31.5 per cent ad valorem.²³³

258. Moreover, contrary to what Chile argues, Chile's price band system is not necessarily less trade-distorting Nor does it insulate Chile's domestic market less, than it would, if Chile simply imposed duties at the bound tariff level of 31.5 per cent.²³⁴ As Argentina stresses, the amount of a duty is not the only concern of Chile's trading partners. As Argentina argues, significant for traders, also, are the lack of transparency of certain features of Chile's price band system; the unpredictability of the level of duties; and the automaticity, the frequency, and the extent to which the duties fluctuate. These specific characteristics of Chile's price band system prevent enhanced market access for imports of agricultural products, contrary to the object and purpose of Article 4.

259. The fact that duties resulting from Chile's price band system are "capped" at 31.5 per cent ad valorem merely reduces the extent of the trade distortions in that system by reducing the extent to which those duties fluctuate. It does not, however, eliminate those distortions. Moreover, the cap does not eliminate the lack of transparency, or the lack of predictability, in the fluctuation of the duties resulting from Chile's price band system. Thus, the fact that Chile's price band system is subject to a "cap" may be said to make this system less inconsistent with Article 4.2. But this is not enough. Article 4.2 not only prohibits "similar border measures" from being applied to some products, or to some shipments of some products with low transaction values, or the imposition of duties on some products in an amount beyond the level of a bound tariff rate. Article 4.2 prohibits the application of such "similar border measures" to all products in all cases.

260. Therefore, contrary to what Chile contends, Chile's price band system does not simply ensure a reasonable margin of fluctuation of domestic prices.²³⁵ In our view, "such reasonable margin of fluctuation" would mean that duties resulting from Chile's price band system would ensure that declines in world prices would not be fully reflected in domestic prices. However, when international prices fall, and when the weekly reference prices are below the lower thresholds of Chile's price bands, the total duties applied to particular shipments will, in many cases, result in an overall entry price of that shipment that rises rather than falls.²³⁶ Therefore, Chile's price band system does not merely moderate the effect of fluctuations in world market prices on Chile's market because it does not ensure that the entry price of imports to Chile falls in tandem with falling world market prices�albeit to a lesser extent than the decrease in those prices. Nor does it tend only to "compensate" for these price declines. Instead, specific duties resulting from Chile's price band system tend to "overcompensate" for them, and to elevate the entry price of imports to Chile above the lower threshold of the relevant price band. In these circumstances, the entry price of such imports to Chile under Chile's price band system is even higher than if Chile simply applied a minimum import price at the level of the lower threshold of a Chilean price band. Therefore, we disagree with Chile that its price band system simply "moderates the effect of fluctuations in international prices on Chile's market".²³⁷ Chile's price band system tends to "overcompensate" for the effect of decreases in international prices on the domestic market when weekly reference prices are set below the lower threshold of the relevant price band-up to the level at which Chile's tariff binding imposes a limit on the amount of duties that can be levied.

261. We emphasize that we reach our conclusion on the basis of the particular configuration and interaction of all these specific features of Chile's price band system. In assessing this measure, no one feature is determinative of whether a specific measure creates intransparent and unpredictable market access conditions. Nor does any particular feature of Chile's price band system, on its own, have the effect of disconnecting Chile's market from international price developments in a way that insulates Chile's market from the transmission of international prices, and prevents enhanced market access for imports of certain agricultural products.

262. We, therefore, uphold the Panel's finding, in paragraph 7.47 of the Panel Report, that Chile's price band system is a "border measure similar to 'variable import levies' and 'minimum import prices'" within the meaning of footnote 1 and Article 4.2 of the Agreement on Agriculture .

263. We turn now to examine the Panel's findings regarding the meaning of the term "ordinary customs duties" under Article 4.2 of the Agreement on Agriculture. We look first to the Panel's consideration of that term, and then review the Panel's interpretation in the light of Chile's objections on appeal.

193 Panel Report, paras. 7.47, 7.65 and 7.102.

194 Footnote 1 exempts "measures maintained under balance-of-payments provisions or under other general, non-agriculture-specific provisions of the GATT 1994 or the other Multilateral Trade Agreements in Annex 1A to the WTO Agreement." In their responses to questioning at the oral hearing, the participants agreed that these "measures" are not relevant in this appeal.

195 Panel Report, para. 7.20.

196 Provided such measure is not exempted under the latter part of footnote 1.

197 Panel Report, para. 7.46. The Panel also concluded that Chile's price band system applies exclusively to imported goods and is enforced at the border by Chile's customs authorities and that it is, therefore, clearly a border measure. We agree. Panel Report, para. 7.25.

198 Argentina's appellee's submission, para. 141.

199 Article 17.6 of the DSU provides: "An appeal shall be limited to issues of law covered in the panel report and legal interpretations developed by the panel."

200 In our Report in EC - Hormones, we held that:

Under Article 17.6 of the DSU, appellate review is limited to appeals on questions of law covered in a panel report and legal interpretations developed by the panel. Findings of fact, as distinguished from legal interpretations or legal conclusions, by a panel are, in principle, not subject to review by the Appellate Body. � Determination of the credibility and weight properly to be ascribed to (that is, the appreciation of) a given piece of evidence is part and parcel of the fact finding process and is, in principle, left to the discretion of a panel as the trier of facts. The consistency or inconsistency of a given fact or set of facts with the requirements of a given treaty provision is, however, a legal characterization issue. It is a legal question. (emphasis added)

Appellate Body Report, supra, footnote 69, para. 132.