I. Terms of Reference 278 II. Economic Data 279 III. Domestic Support 291 IV. Export Credit Guarantees 293

Individual Third-Country Analysis of Brazilian and US Export Prices

Download 3.13 Mb.

Page	8/38
Date	20.05.2018
Size	3.13 Mb.
	#49759

1 ... 4 5 6 7 8 9 10 11 ... 38

Domestic Prices for Brazil, US and China

Individual Third-Country Analysis of Brazilian and US Export Prices
123. When US and Brazilian prices in the 40 individual markets are examined, the evidence does not support the United States’ broad assertion that Brazilian prices are consistently lower than US prices. In some of the key textile producing countries where US exporters ship large quantities of US upland cotton – Hong Kong, Taiwan, India, Pakistan, and Vietnam – US prices during MY 1999-2002 were lower than Brazilian prices.^¹⁶²
124. In other large consuming (textile producing) countries, while US prices were somewhat higher than Brazilian prices, the volume of US exports vastly exceeded the volume of Brazilian exports, including China (120 times greater), Bangladesh (26 times greater), Colombia (9 times greater), Indonesia (17 times greater), Japan (35 times greater), Korea (3384 times greater), and Peru (631 times greater).^¹⁶³ Huge volumes of US exports in these 7 markets compared to the volumes of Brazilian exports suggest that US exports, not Brazilian exports, play a dominating role in the conditions of competition in those markets.
125. Indeed, overall, the cumulative volume of US exports to the 40 countries where Brazilian exports were shipped was 15 times greater than the volume of Brazilian exports.^¹⁶⁴ For Brazilian exports to secure any sales, let alone increase market share in competition with heavily subsidized US exports in these markets, it would not be surprising that textile consumers of upland cotton would expect Brazilian exporters to discount their prices. Andrew Macdonald confirmed that such discount pricing was necessary for non-Brazilian cottons to compete with US cotton.^¹⁶⁵ This is particularly true in Argentina, Bolivia, Chile, Colombia, Ecuador, Indonesia, Peru, South Korea, Turkey and Venezuela where US exporters received the benefit of US export credit guarantees.^¹⁶⁶ No exporter of non-US cotton could hope to compete with those export credits without discounting its upland cotton prices.^¹⁶⁷
126. But even where the volume of Brazilian exports was greater than US exports, the evidence does not suggest the absence of overall price suppression from the effects of US subsidies.^¹⁶⁸ Brazilian exports were greater than US exports in only 10 of the 40 countries (Argentina, Bolivia, Cuba, Poland, France, Germany, Greece, Portugal, South Africa, and Spain) over the MY 1999-2002 period. These 10 countries represented a tiny fraction – 0.49 per cent – of total US exports during MY 1999-2002 to the 41 countries, and only 37.5 per cent of Brazilian exports.^¹⁶⁹ Yet, even without significant US exports to these countries, the domestic prices of these countries still largely followed A-Index prices, as discussed in Brazil’s Answer to Question 233.
127. Further, in the 10 countries where the volume of Brazilian exports was larger than the United States, Brazilian prices were higher than US prices in 3 of the 10 countries – Poland, Greece, and South Africa.^¹⁷⁰ Further, US prices in 4 of the 10 countries – France, Germany, Spain, and Portugal – were much higher than Brazilian prices.^¹⁷¹ The very low volume of US exports to these same four countries suggests that far higher quality US cotton was sold in these countries which produce low-volume specialized cotton textile products.^¹⁷² Finally, US and Brazilian prices were very close in remaining 3 countries – Argentina, Bolivia, and Cuba.^¹⁷³ Thus, contrary to the US arguments, this evidence does not suggest that Brazilian prices are suppressing US prices, even in those markets where there are larger volumes of Brazilian imports.
128. Similarly, the unqualified US assertion that “the Brazilian A-Index quote is consistently below the US A-Index quote”^¹⁷⁴ is both untrue as well as irrelevant to the question of whether US production subsidies caused suppression of world and third country market prices. Comparing US Memphis A-Index prices with Brazil A-Index prices shows the following graph:^¹⁷⁵

129. As the Panel can see from the graph, since October of 2001, the US Memphis A-Index price has been slightly higher or slightly lower than the Brazilian A-Index price.^¹⁷⁶ But throughout the entire period, the movements of both US Memphis and Brazilian A-Index prices track each other very closely.^¹⁷⁷ This is the key fact illustrating the transmission of global price-suppressing effects as discussed in Answer to Question 233 supra.
Domestic Prices for Brazil, US and China
130. Nor does the record support the US claim that Brazilian prices undercut US export prices to the Brazilian market. Indeed, the evidence suggests the opposite. More than 90 per cent of Brazilian production of upland cotton was sold in Brazil during MY 1999-2002.^¹⁷⁸ Imports represented approximately 20 per cent of Brazilian consumption during MY 1999-2002.^¹⁷⁹ US exports to Brazil represented 21 per cent of Brazilian imports with a total of 306 million pounds of US upland cotton exported to Brazil^¹⁸⁰ in MY 1999-2002.^¹⁸¹ US annual weighted average export prices to Brazil were lower in every year except MY 2000 when US exports volumes fell to very low levels.^¹⁸² Significantly, in MY 1999, when US export volumes to Brazil represented 17.6 per cent of very large Brazilian imports of upland cotton, the US weighted average prices were 10.27 cents below Brazilian domestic prices.^¹⁸³ Similarly, in MY 2002, when US exports represented 50 per cent of total Brazilian imports, US export prices were 4.57 cents per pound lower than Brazilian domestic prices.^¹⁸⁴ This evidence confirms the testimony of Andrew Macdonald that Brazilian textile purchases of cotton used, inter alia, US imports – or the threat of US imports – to negotiate lower Brazilian domestic prices during MY 1999-2002.^¹⁸⁵ It may also be an effect of US GSM 102 export credit guarantees. And this evidence further confirms the price suppressing effects of US subsidies in the Brazilian market.
Conclusion
131. The evidence set out above refutes the inappropriate non-weighted average analysis in Exhibit US-75, and confirms the effects of the US subsidies in each market where Brazil and the US upland cotton were found in MY 1999-2002. Any “competition” between US and Brazilian cottons took place against the background of subsidy-distorted marketplaces. In the final analysis, the fact that selected US prices are higher^¹⁸⁶ or lower in the A-Index or in individual third country markets is largely irrelevant to the question whether US subsidies are causing significant price suppression “in the same market” where US and Brazilian cotton is marketed. Rather, as Brazil has established, it is the impact of the US subsidies on the global supply and demand factors that suppresses prices in the world market as well as in the third countries where upland cotton is marketed. This means that the prices in each of those 40 third country markets – as well as the Brazilian and US markets – were already suppressed before any cotton was shipped by US or Brazilian exporters.
132. Finally, the generalized price-suppression effect manifests itself in all 42 individual country markets^¹⁸⁷ in which Brazilian upland cotton is consumed. Professor Sumner estimates that over 40 per cent of US upland cotton exports during MY 1999-2002 would not have been made but for the US subsidies.^¹⁸⁸ Forty per cent fewer US exports necessarily means lower, or even non-existent, US exports in many of the 40 markets where Brazil exported its upland cotton. Had significant volumes of US upland cotton not received Step 2 export payments, US exports and, thus, the amount of US upland cotton competing with Brazilian cottons would have been lower. Similarly, far fewer bids of US upland cotton financed by the GSM 102 programme may well have meant higher prices for Brazilian exporters and/or an expanded market share in each of these markets.

236. The Panel notes Exhibit US-47 (and the chart in paragraph 13 of the US 2 December oral statement). Please provide a conceptually analogous chart to Exhibit US-63 with respect to data relating to the US interpretation of "world market share". USA
237. Could a phenomenon that remains at approximately the same level over a given period of time be considered a "consistent trend" within the meaning of Article 6.3(d)? Do parties have any suggestions as to how to determine a "consistent trend", statistically or otherwise? BRA, USA
Brazil’s Answer:
133. In responding to this question, Brazil starts from the ordinary meaning of the term “consistent trend.” A “trend” means “a general course, tendency or drift.”^¹⁸⁹ “Consistent” means “congruous, compatible with, not contradictory, marked by uniformity or regularity.”^¹⁹⁰ A consistent trend within the meaning of Article 6.3(d) therefore means a non-contradictory tendency marked by regularity.
134. Generally, a phenomenon that remains at approximately the same level could be considered a consistent trend. However, read in the context of Article 6.3(d), which speaks of an “increase that follows a consistent trend over a period when subsidies have been granted,” the consistent trend must reflect the increase in the world market share of a Member over its previous three-year average. Thus, for a given period of time, the world market share of a Member could remain at approximately the same level and be compatible with a finding of a consistent trend. Yet, there will not be a consistent trend, within the meaning of Article 6.3(d), if during the final year there is no increase of a Member’s world market share over its previous three-year average. It follows that over the period when subsidies have been granted there must be a regular tendency of an increase in the world market share of a Member – although no increase in each and every year is required for the conditions of Article 6.3(d) to be fulfilled.
135. This interpretation is confirmed by the scope of Article 6.3(d), which applies to primary products or commodities. The parties agree that this includes agricultural products that necessarily will be affected by favourable or adverse weather conditions. The United States, for example, repeatedly asserts that the weather-related problems of MY 1998 mean that data for that year should be disregarded. These weather-related effects alone could cause in particular years the world market share of a Member to increase or decrease. This may well be the reasoning behind requiring examination of a longer trend in Article 6.3(d). But to require a constant increase in the world market share during each year “when subsidies have been granted” would render the disciplines of Article 6.3(d) largely inutile for agricultural products faced with weather-related production fluctuations.
136. Brazil has offered data on the US world market share over the period since MY 1996 as well as since MY 1986.^¹⁹¹ The data shows that the trendline for the US world market share increases whether based on MY 1996 or MY 1986 as the starting point. The Panel’s question asks for guidance on the conditions for a finding whether this trend is consistent.
137. Brazil cautious that statistical methods may not be helpful in analyzing the consistency of a trend within the meaning of Article 6.3(d) for two reasons: first, Article 6.3(d) is concerned with yearly data over a relatively short period. That means that only very few data points will be available, which will affect the statistical significance of any results. Second, as discussed above, it is the nature of world market shares in agricultural products to fluctuate due to weather-related effects. This phenomenon tends to further weaken any statistical significance of methods analyzing the consistency of a trend.
138. With these considerations in mind, Brazil suggests that the Panel analyze the trendlines offered by Brazil^¹⁹² and decide whether on their face they demonstrate – as Brazil has argued – a trend consistent with a finding that the effect of the subsidies is an increase in the US world market share for upland cotton. Brazil believes they do, particularly if the severe MY 1998 decline is disregarded.
139. However, Brazil has also run a regression analysis for the trends over the period MY 1986-2003, MY 1996-2003 and MY 1996-2002, with the results being reproduced in the graphs below.^¹⁹³

The R² – the measure for the fit of a regression analysis – varies between 0.44 for MY 1986-2003 and 0.60 for MY 1996-2003. The R² demonstrate that there is a positive relationship between the US world market share and its increase over time. In view of the limited number of observations and the nature of the commodity market in question, this is a very high number.

Directory: english -> tratop e -> dispu e
dispu e -> Annex g list of exhibits submitted by the parties
dispu e -> World Trade Organization
dispu e -> World Trade Organization

Download 3.13 Mb.

Share with your friends:

1 ... 4 5 6 7 8 9 10 11 ... 38