3. Brazil’s Article 6.3(c) Price Suppression Claims and the Econometric Studies 11. Brazil has offered considerable evidence in the form of documents and witness statements demonstrating the existence and payment of subsidies, as well as the causal link between the subsidies and significant price suppression. In addition, Brazil presented the Panel with evidence of a number of different studies that show significant price suppressing effects. Notably among these are the two Westcott/Meyer USDA studies (referred to and commissioned by the Payment Limitation Commission) showing 10 per cent price suppression for MY 2000 and an estimated 33.6 per cent price suppression in MY 2001 from only the effects of removing the marketing loan subsidies. By contrast, Professor Sumner found that US prices were suppressed by 32.7 per cent by the effects of all US subsidies that applied during MY 2001. In light of the lower level of price effects found by Professor Sumner, the United States claim that Professor Sumner’s analysis is not “conservative” is curious. Other studies by the ICAC – of which the United States and Brazil are both Members – show increases in world prices from the removal of some US subsidies of 10.5 per cent for MY 2000 and 26.3 per cent for MY 2001. Professor Sumner found that the effects of a removal of all US subsidies that applied during MY 2000 and MY 2001 would have resulted in world price increases of 7.74 per cent and 17.7 per cent respectively. Brazil has presented many other studies as evidence. They all show significant price suppression.
12. What has been the US reaction to every one of these studies? As they indicated over the past two days, they have found many initial problems with all of them. But they reserved the broad scale attack for Professor Sumner’s FAPRI model that has been repeatedly relied on by the US Congress and USDA. The United States even identified flaws in the results of the Westcott/Meyer 2000 and 2001 marketing loan studies. And the United States promises they will be busy for the next six weeks in critiquing all the studies cited by Brazil.
13. But the Panel must ask whether all these economists, including some of USDA’s own leading economists, could be wrong although their results support USDA’s own Chief Economist’s views that US producers are insulated from market forces by these subsidies? Could these economists be wrong because they made the mistake of applying the fundamental notion that large production subsidies create larger supplies, and larger supplies result in significantly lower prices?
14. In the final analysis, these econometric studies are useful tools to confirm what common sense already tells us. That $12.9 billion in subsidies provided between MY 1999-2002 have production effects. That the National Cotton Council was correct when it argued that US upland cotton farmers could not exist without all of the cotton-specific subsidies. That many US producers needed subsidies to bridge the huge gap between their total costs and market revenue. That US acreage did not decrease as prices plummeted to record lows between MY 1999-2001 – rather planted US acreage increased. That US producers planted 14.1 million acres of upland cotton when prices were at record lows in the spring of 2002. That US exports did not decrease as prices plunged and the US dollar appreciated, rather they increased. And that the effects of US subsidies on suppressed prices are transmitted to the world and individual country markets, including Brazil.
4. Brazil has established a claim under Article 6.3(d)
15. With respect to Brazil’s claim under Article 6.3(d), Brazil demonstrated that the ordinary meaning of the term “world market share” is the world market share of exports. USDA, the EC and Canada all use this term to refer to export market share, not share of world consumption. This interpretation is consistent with the use of the term “trade” in footnote 17 of the SCM Agreement which means the “sale and distribution of goods and services across international borders”. It is also consistent with the object and purpose of Article 6.3(d) which is to prevent a Member from using its subsidies to increase its share of the world market for a particular product.
16. Undisputed facts show that the US share of world trade increased considerably from MY 1998 to MY 2002, and is projected to remain at very high levels in MY 2003. This increase follows a consistent trend from MY 1996. The “consistent trend” need not be an unbroken line of increases during the trend period examined, as the United States appears to argue. Because of severe weather problems, such as occurred in the United States in MY 1998, export market share in agricultural problems will always susceptible to some annual variations not caused by subsidies. Rather the trend must reflect an overall increase and not reflect a number of wide swings within the period examined. The trend for US world market share of upland cotton from MY 1996, and particularly the period from MY 1998 onward, shows a sustained and significant increase in the US world market share. And the undisputed facts show that the record US world market share reached in MY 2001 and 2002 occurred at the same time as record high levels of US subsidies.
5. Brazil has established that the US share of world export trade is not equitable
17. The notion of “equitable share of world export trade” necessarily depends on the facts of each case. The fact-intensive nature of each case is reflected in the text of Article XVI:3, which requires the Panel to examine “special factors”. Brazil suggests that examining whether there were any subsidy-induced increases in market share is one factor to consider. Another factor is the relative cost of production of the Members competing for world market share. The undisputed facts show (based on September 2003 ICAC data) that the US share of world exports of cotton more than doubled between MY 1998-2002 – from 18.7 to 39.3 per cent. At the same time, the African producers’ collective share of world exports decreased from 10.2 to 8.1 per cent of world trade. Figure 26 shows these trends. If a picture is worth a thousand words, it is this one.
18. Brazil submits it is not equitable for a heavily subsidized WTO Member to more than double its share of competitive world markets for upland cotton in only five years reflecting a significant contribution of subsidies. And it is not equitable for that Member to do so when its costs of production were double those of the poorest and neediest producers in the world. Yet, when faced with these facts, the United States’ only response is that the definition of “inequitable” is hopelessly vague. Brazil does not believe the inequity in this case is so difficult to determine.
1. Brazil’s Analysis Fails to Establish to Whom Certain Payments Go and Whether Certain Payments May Properly Be Attributed to Exported Upland Cotton. One of the fundamental elements of Brazil’s claims is that Brazil needs to identify the “subsidized product” that is causing the serious prejudice that Brazil claims its interests are suffering.89 Brazil has not even explained, however, what is the “subsidized product” for each of the types of subsidies from which it claims serious prejudice. Brazil appears to assume that the “subsidized product” is upland cotton in the form traded on the world market. Yet many of the subsidies at issue are paid to producers of cotton. Cotton is processed and sold before being traded. Brazil has made no showing of how the subsidy to the producer can be assumed to pass through to the exporter.
2. Brazil’s panel request identifies the challenged measures as “subsidies provided to US producers, users, and/or exporters of upland cotton”. However, it is for Brazil as the complaining party to establish who are the recipients of the subsidies and that the subsidies are properly attributed to upland cotton. Brazil’s failure to do so means that it has not carried its burden in demonstrating that cotton is subsidized for purposes of considering adverse effects.
3. In the Peace Clause portion of this dispute, the United States has discussed at length certain decoupled payments that are not linked to production of upland cotton. With respect to these decoupled payments, Brazil has failed to demonstrate who the recipients of these payments are in connection with any exported upland cotton. Brazil simply presumes that every upland cotton producer is an upland cotton base acreage holder and receives a decoupled payment. Brazil has brought forward no facts to demonstrate that this is the case.
4. Even if Brazil had brought forward evidence that the recipients of these payments were upland cotton producers, that would not be enough. Brazil would still need to allocate these payments, which Brazil concedes are not linked to current production of upland cotton, over total production on a recipient’s farm.90 5. Thus, Brazil assumes that the subsidies91 at issue are received by someone currently producing cotton, based simply on the fact that the subsidy is based on past production of cotton. Brazil has not explained how this makes upland cotton currently for sale on the export market the “subsidized product” with respect to these payments. Brazil has failed to demonstrate that the recipients of the subsidies are involved in current cotton production, nor has it demonstrated how much of the subsidy, even under Brazil’s approach, should be allocated to other products produced by the recipient, such as corn or soybeans.
6. Brazil Has Not Established that US Subsidies Have Suppressed or Depressed Prices in the Same Market. As noted above, Brazil has in fact not even demonstrated the subsidized product for each of the subsidies it challenges or the size of the subsidies to exported upland cotton. However, without relieving Brazil of its burden on these issues, we note that even Brazil’s overly simplified approach does not suffice to demonstrate causation. US subsidies largely resulted from low market prices, not the other way around.
7. This is nowhere more evident than in marketing year 2001, a year with historically low market prices. Brazil has failed to explain that market signals (futures prices) at the time when planting decisions were taken by US producers suggested prices would remain high. Thus, the large marketing loan payments ultimately made in marketing year 2001 do not demonstrate that marketing year 2001 payments had the effect of increasing US production. Brazil’s expert acknowledges this very point, but Brazil has not presented in its further submission any information on “the expectations about production incentives that growers hold at the time they make their planting decisions”, information on which its own expert has stated “cotton plantings depend”. Thus, Brazil’s simple explanation of the conditions in marketing years 1999 through 2002 ignores “the basic economic principles” its own expert says are relevant in this case.
8. The Sumner Model Presented by Brazil Is Inadequately Explained, Inappropriately Applied for a Retrospective Analysis, and Apparently Uses Faulty Assumptions and Estimations. In presenting this reaction to Brazil’s expert’s analysis, the United States notes that the use of a simulation model to explore the counter factual of removal of US subsidies cannot be made without answers to previous questions on the subsidized product and size of the subsidies. That is, the use of a simulation model cannot relieve Brazil of its burden of arguing the elements necessary to establish its claims. This critique of Dr. Sumner’s analysis is made to show that Brazil’s approach is fundamentally flawed in all aspects.
9. Since Brazil has not provided access to the model itself, one cannot say with certainty how the modelling affects the results.92 Nonetheless, based on what has been presented in Annex I, Brazil’s analysis appears flawed in several respects and as a result, the conclusions drawn are biased and misleading. While the modelling approach used is well accepted for forward looking projections, using a baseline model to simulate counterfactual outcomes over the historical period 1999 2002 is problematic because of the implicit assumption of perfect foresight by producers of actual conditions in the historical year. This potentially overstates the effects of the programme because the model assumes outcomes that were unanticipated by producers when they made their planting decisions. Also, it is not clear to what extent actual observed data enter into the solution process. The difference is not merely conceptual: the choice of values can potentially affect the reported results.
10. Brazil’s use of lagged prices as a proxy for expected prices is also problematic. Recent studies have criticized the use of lagged variables as substitutes for expectations, and numerous papers use the futures price for next year’s crop as the best proxy for expected price. The use of futures prices in a multi commodity modelling framework for extended time projection is cumbersome. Nonetheless, the use of lagged prices as a modelling convenience does not preclude the possibility of bias. In those years where there are large shocks, lagged prices are poor predictors of expected price. Futures prices, by contrast, are more efficient because they are based on more current information.93 11. Brazil’s expert’s estimates of US programme impacts after marketing year 2001 are further inflated by his choice of a low price baseline for the counter factual comparison.94 The low price baseline exaggerates the 2003 07 results and ensures projections of significant marketing loan payments throughout 2003 07.
12. The economic literature on decoupled payments acknowledges the programmes may have some impact on production, and that those impacts depend in part on farmer’ s expectations. However, the research concludes that the impact appears negligible. Brazil’s expert, on the other hand, uses a stylized logic to come up with the estimates for the impact of production flexibility contract (PFC) payments that have neither empirical nor theoretical grounding. It is widely accepted that these programmes have whole farm impacts rather than crop specific impacts. Furthermore, the impact is much smaller than Brazil has estimated; the whole farm impact is, at its upper estimate, perhaps one quarter to one fifth the impact Brazil’s expert cites for cotton alone.
13. Brazil argues that market loss assistance (MLA) payments have a larger effect on area than do PFC payments, despite the fact that MLA payments were paid on the identical payment base as the PFC payments. Supplemental legislation authorizing each of these MLA payments was passed several months after planting for the crop year in question had occurred. Brazil asserts that producers had expectations about MLA payments at the time of planting. However, if producers had expectations of payment, then they also knew that they would be eligible to receive such a payment whether or not they planted cotton. Indeed, they could choose not to plant any crop at all and still be eligible for the payment.
14. Brazil argues that counter cyclical payments “clearly provide more production incentive than the market loss or the direct payments,” yet offers no empirical evidence to justify such a claim. The claim, as well as Brazil’s expert’s treatment of decoupled payments in general, is particularly puzzling given a recent paper by Brazil’s expert in which he concludes that the 2002 farm bill would have a minimal effect on cotton area and world prices. Brazil’s expert also remarked that: “The impacts of the FSRIA will be hard to isolate amid the normal flux of world markets".95 We agree with Dr. Sumner’s previously published conclusions on these points.
15. Crop insurance subsidies are generally available for most crop producers and hence do not give a specific advantage to one crop over another. Thus, their effects are not commodity specific, and have no or minimal impacts on cotton markets. Moreover, crop insurance purchases by cotton growers have generally been at lower coverage levels than for other row crops. Over 2002 03, roughly 90 per cent of cotton acreage insured was at coverage levels at 70 per cent or less, consistent with the criterion under paragraph 8(a) of Annex 2 of the Agreement on Agriculture. This suggests that even if one were to consider cotton crop insurance subsidies as crop specific, over 90 per cent of insured cotton area would be exempt as having no or minimal trade distorting effects.
16. Lastly, while some studies like the ones cited by Brazil have suggested crop insurance subsidies may have a slight effect on acreage, the effects on production are less clear. Recent studies suggest that farms with more insurance tend to use less inputs like fertilizer and pesticides and vice versa. This demonstrates a potential moral hazard problem with crop insurance: a negative effect on yields, which may well offset any marginal effects on crop area.
17. The size of Step 2 payments under Brazil’s baseline appears to be biased upwards, in part, due to the low price baseline discussed earlier. Brazil’s results are inconsistent with other analyses of Step 2.96 Thus, contrary to the results of Brazil’s expert’s model, the benefits of Step 2 payments would appear to largely accrue to the producer, with only negligible effects on world markets. While Brazil’s model documentation is lacking, one explanation for the difference may be a more price responsive acreage equation by Brazil.
18. While Brazil has presented a modelling framework that is conventional, much of how Brazil’s expert has modelled US farm payments can be considered “unconventional”. Thus, the analysis presented by Brazil in Annex I is not “conservative”, but rather produces results that are inconsistent with a wider body of academic research.
19. Additional Legal Arguments. With respect to price suppression or depression under Article 6.3(c) of the Subsidies Agreement, Brazil believes that it is the effect on the producers of the complaining Member that must be “significant”. We find it implausible that the Subsidies Agreement was intended to create multiple standards for panels to apply: that is, what may be “significant” to one Member’s producers may be “insignificant” to another’s. Context for rejecting Brazil’s approach can be found in Article 15.2 of the Subsidies Agreement, which sets out for countervailing duty purposes the same effects found in Article 6.3. This text makes it even more clear that the analysis is whether the level of price suppression or depression itself is "significant".97 Brazil has not suggested that the analysis under Articles 15.2 and 6.3(c) should be different.
20. With respect to GATT 1994 Article XVI:3, Brazil appears to assume that it may advance a claim under this provision on all challenged US subsidies. However, Article XVI:3 only applies to export subsidies. Therefore, as Brazil has predicated its claim under Article XVI:3 on evidence relating to all challenged US subsidies and not only those subsidies it alleges are export subsidies, Brazil has failed to establish a prima facie case on its claims.
21. Finally, with respect to Brazil’s claims of a threat of serious prejudice, the United States notes Brazil’s failure to present recent market and futures price data, which belie the notion that there is a clearly demonstrated and imminent likelihood of future serious prejudice. In fact, prices are currently above the level at which the marketing loan programme confers any benefit on US upland cotton producers and are expected to remain so. If there is not a “clearly demonstrated and imminent likelihood” of serious prejudice in marketing year 2003, it follows that there cannot be a threat of serious prejudice for marketing years 2004 07, either.
EXECUTIVE SUMMARY OF THE CLOSING STATEMENT
OF THE UNITED STATES AT THE SECOND SESSION OF THE
FIRST MEETING OF THE PANEL WITH THE PARTIES
1. The US comments speak briefly to Brazil’s allegations regarding “the effect” of US subsidies. Brazil has not shown causation between the US subsidies and the effects Brazil attributes to those subsidies. The United States has pointed out the failure of Brazil to separate and distinguish evidence on the effect of other factors from the alleged effect of the challenged US subsidies. Ultimately, this issue goes to the quality of the evidence before the Panel and whether Brazil has established a prima facie case on its claims.
2. There are three main elements of Brazil’s argument. First is the “temporal proximity” argument – that is, that low world prices correspond in time with high US subsidies.98 Mr. Chairman, there are subsidies and there are subsidies. For example, there is a difference if I give you a $10 subsidy to produce versus $10 whether you produce or not. Depending on the nature of the payment, one would estimate different effects. Therefore, one cannot merely aggregate the value of all US payments and claim that those subsidies have had “an effect” on production and prices.
3. In this part of its argument, Brazil misuses the data on production by making comparisons using marketing years 1998 and 2001. In 1998, production was driven downward by drought and record crop abandonment. In 2001, production was driven upward by record yields. To use 1998 and 2001 as the beginning and end of a comparison therefore distorts a proper analysis.
4. Brazil stated yesterday that the increase in US production in marketing year 2001 was not solely due to record yields but also to an increase in acreage. That is true – there was some increase in acreage in 2001, but Brazil has failed to make the proper comparison to put that information in context. Brazil should have compared the US acreage increase between marketing years 2000 and 2001 with that in the rest of the world. The United States invites the Panel’s attention to Exhibit US 63 circulated today. This exhibit reflects, for marketing years 1996 2002, the percentage change in harvested acreage over the previous marketing year in the United States and the rest of the world.
5. In marketing year 1996, when the programmes challenged by Brazil were introduced, you see a large decrease in US acreage compared to the rest of the world. The United States draws your attention to marketing year 1998, in which there is a large decline in US harvested acreage due to drought, followed by a large increase in marketing year 1999, which largely cancel each other out. In marketing year 2001, we see that the increase in acreage in the United States corresponds to the increase in acreage for the rest of the world. In marketing year 2002, the percent decline in harvested acreage in the United States is greater than that observed in the rest of the world. Thus, the data do not support Brazil’s contention that US producers are insulated from market forces. In fact, US harvested acreage largely increases and decreases in line with the rest of the world.
6. (Yesterday Mr. Moulis asked about the data in the upland cotton fact sheet. The data in Exhibit US 63 does not come from that fact sheet but from the most recent US Department of Agriculture data base – the specific source is indicated on the second page of the exhibit. Brazil has used this same source for numerous exhibits in its submissions.)
7. The second element of Brazil’s arguments that the United States would like to address is its reliance on Mr. Sumner’s model. We first would like to comment on something Mr. Sumner said today in his statement to the effect that the United States does not object to the use of the FAPRI baseline model. In fact, as reflected in the portion of the US opening statement delivered by Dr. Glauber, we do criticize as inappropriate the use of a baseline simulation model for retrospective analysis, a type of analysis for which it is not designed and is poorly suited.
8. Mr. Sumner’s analysis also uses an inappropriately low baseline for his prospective analysis of future years. I noted with interest Mr. Sumner’s statement that he used the November 2002 preliminary FAPRI baseline because this was available when he ran his model and that the results would have been even more extreme had he used FAPRI’s published 2002 baseline “released the previous winter”. The United States realizes it would have been inconvenient for Mr. Sumner to re run his model, but FAPRI released a more recent baseline in January 2003 (published in March 200399), many months before Brazil submitted the results of its model to the Panel and the United States. We believe this more recent FAPRI baseline would have been a more appropriate baseline with which to do calculations, but Brazil has chosen not to do so, instead presenting to the Panel results based on more out of date and inaccurate data. We wonder what would arise from a prospective analysis using such more recent data.
9. The third issue concerns the allegations of high US costs, an issue we have touched on only briefly in this hearing and will return to in more detail in our submissions. Brazil asks: without subsidies how could high cost US producers have stayed in business? It is important first to point out that all of the cost projections by the US Department of Agriculture cited by Brazil are merely updates of a 1997 cost survey. In every year subsequent to 1997, the Department simply takes the results of the 1997 cost survey and updates it to reflect the general increase in prices according to the producer price index.
10. This approach assumes that the mix of inputs remains the same in 1997 as in subsequent years. However, this causes a presentation of inaccurate data on what costs are now. Brazil has several times in this hearing stated that it is not denying that factors reducing costs have occurred – for example, pest eradication bringing new, low cost areas of the United States into production or the adoption of biotech cotton which requires fewer pesticide applications. Brazil, however, has not updated the cost information it presents to the Panel to account for such new developments and information.
11. The United States also notes Brazil’s repeated references to a so called cost/revenue gap. In fact, Brazil presents another such comparison for marketing year 2002 at page 5 to the annex to its Closing Statement. However, Brazil’s so called “gap” is the difference between an inaccurate average total cost per pound and the average marketing year farm price. Mr. Chairman, this is a faulty comparison. Total costs are relevant over the long term, but Brazil uses this (inaccurate) number to compare to revenue in the short term – that is, the market price for one year. Such a comparison tells you nothing and does not establish that it is only the effect of US subsidies to keep US cotton farmers in business.
12. In fact, Brazil has apparently not listened to the testimony of its own farmer witness, Mr. Christopher Ward. In his statement during the first day of this hearing, he said the following (and I quote from paragraph 6 of his statement):
But even with these high yields and the excellent quality of our land, we were not able to fully recover all of our variable costs of production during the 2000/01 and 2001/02 seasons. These variable costs included depreciation and maintenance of equipment, seed and fertilizer, labor, insurance, and fuel. Nor were we able to meet our total costs which include the additional fixed costs.100
That is, Mr. Ward says he has not been able to cover either his variable costs or his total costs for a period of two marketing years, and yet he continues producing. Under Brazil’s analysis, he should be out of the business of producing cotton. He is not, and Brazil claims he is not subsidized, so how can Brazil claim that it is “the effect of the subsidy” to keep US farmers in business when they allegedly were not able to cover their total costs in marketing year 2002? What’s true for Brazil should also be true for the United States.
13. Mr. Chairman, members of the Panel, on the basis of these arguments and the evidence presented to date, the United States does not believe that Brazil has established a prima facie case that the challenged US subsidies have caused the effects complained of. We will continue to develop and provide our response to the voluminous submissions of Brazil in our answers to your questions and in our rebuttal submission. Thank you.