9.111 In interpreting the term "benefit" in Article 1.1(b) of the SCM Agreement, we recall that Article 31.1 of the Vienna Convention on the Law of Treaties provides that a treaty shall be interpreted "in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose." We note that Canada purports to respect Article 31.1 of the Vienna Convention in its interpretation of the term "benefit". However, for the following reasons, we are not persuaded that the application of Article 31.1 results in the interpretation of "benefit" advocated by Canada.

9.112 First, in our opinion the ordinary meaning of "benefit" clearly encompasses some form of advantage. We do not consider that the ordinary meaning of "benefit" per se includes any notion of net cost to the government. As Canada itself has noted, the dictionary definition of "benefit" refers to "advantage", and not to net cost.⁵³³ In order to determine whether a financial contribution (in the sense of Article 1.1(a)(i)) confers a "benefit", i.e., an advantage, it is necessary to determine whether the financial contribution places the recipient in a more advantageous position than would have been the case but for the financial contribution. In our view, the only logical basis for determining the position the recipient would have been in absent the financial contribution is the market. Accordingly, a financial contribution will only confer a "benefit", i.e., an advantage, if it is provided on terms that are more advantageous than those that would have been available to the recipient on the market.

9.113 We note that the relevant context supports our interpretation of the ordinary meaning of "benefit". In particular, we note that Article 14 provides guidelines for calculating "the benefit to the recipient conferred pursuant to paragraph 1 of Article 1." These guidelines employ a commercial benchmark, whereby a financial contribution "shall not be considered as conferring a benefit" unless that financial contribution is made on terms that are more advantageous than would have been available to the recipient on the commercial market. Although Article 14 applies expressly for the purposes of Part V of the SCM Agreement (countervailing measures), and although the commercial benchmark approach of Article 14 is not the only test for calculating the amount of a subsidy, we recall that Article 14 refers expressly to commercial benchmarks for identifying explicit situations in which an Article 1.1 "benefit" shall not arise. We see no reason why the commercial benchmarks applied in Article 14 for the purpose of determining when an Article 1.1 "benefit" does not arise should not serve as relevant context for determining when an Article 1.1 "benefit" does arise.

9.114 Second, we do not accept Canada's argument that "benefit" must be interpreted more narrowly than its ordinary meaning of "advantage". Canada asserts that a narrow interpretation is necessary in order to exclude "normal commercial activity" such as commercial contracts (by a government) that accord advantages to firms relative to their competitors. In light of the preceding paragraph, however, we consider that using a commercial benchmark to identify the existence of a "benefit" will not create the problem alluded to by Canada. That is, under such a benchmark, "normal [governmental] commercial activity" will not be deemed to confer a "benefit", i.e., an advantage, provided it really is "normal commercial activity". In other words, such activity will not be deemed to confer a "benefit" provided the terms of the contract are not more advantageous than those that would have been negotiated on the market for an equivalent transaction.

9.115 Third, if "benefit" were to include the notion of net cost to government, it could exclude from the definition of "subsidy" situations explicitly identified in Article 1.1(a)(1) itself as constituting government financial contributions even though no cost to the government might be involved. Specifically, Article 1.1(a)(1)(iv) identifies as a "financial contribution" the situation where a government directs a private body to make "financial contributions" within the meaning of Article 1.1(a)(1)(i)-(iii). In such a situation, the net cost could be incurred entirely by the private body rather than the government. Canada's interpretation of "benefit" (i.e., to include net cost to government) would render Article 1.1(a)(1)(iv) meaningless, since a form of "financial contribution" explicitly included in Article 1.1(a) would automatically (i.e., because it would never meet the net cost to government test) be excluded by Article 1.1(b).

9.116 Fourth, we do not accept Canada's reliance on Annex IV.1 of the SCM Agreement as contextual support for considering net cost to government in determining "benefit". Annex IV.1 concerns the calculation of the amount of a subsidy for the purpose of Article 6.1(a) of the SCM Agreement. Canada effectively argues that the "cost to government" referred to in paragraph 1 of Annex IV constitutes the definition, at least for the purpose of determining the existence of serious prejudice under Article 6.1(a), of "benefit" in the sense of Article 1.1(b). However, Canada's is neither the only interpretation of this provision, nor the most convincing. In our opinion, the need to calculate the value of a subsidy only arises once the existence of the subsidy, and therefore the "financial contribution" and "benefit", have been established. Because "benefit" must be established before the value of the alleged subsidy may be considered, provisions concerning the valuation of subsidies are not necessarily relevant for the purpose of establishing the existence of a subsidy (and therefore "benefit"). The contextual relevance of Annex IV.1 differs significantly from that of Article 14, referred to in para. 9.113 above, since the latter provision deals expressly with the calculation of "benefit" "pursuant to paragraph 1 of Article 1". Annex IV.1, by contrast, refers only to the calculation of the amount of a subsidy, and does not apply expressly for the purpose of calculating "benefit" within the meaning of Article 1.1 of the SCM Agreement.

9.117 Fifth, we are unable to accept Canada's argument that item (k) of the Illustrative List of Annex I of the SCM Agreement constitutes contextual guidance for determining the existence of "benefit" in the specific context of government credit under Article 1. In our view, item (k) of the Illustrative List applies in determining whether or not a prohibited export subsidy exists. We do not consider, and the parties have not argued, that item (k) determines whether or not a "subsidy" exists within the meaning of Article 1 of the SCM Agreement. Whereas Canada might have argued that item (k) could be relied on for the purpose of determining when an Article 1 subsidy does not exist, Canada expressly declined to rely on this a contrario reading of item (k). We therefore take no view on the a contrario approach as such. ⁵³⁴

9.118 Furthermore, whereas Article 31.1 of the Vienna Convention requires "the ordinary meaning to be given to the terms of the treaty in their context", we do not consider that this requires us to ignore the ordinary meaning of the term "benefit" simply to accommodate a conflicting contextual interpretation. As noted above, we consider that the ordinary meaning of "benefit" refers to advantage, to the exclusion of any notion of net cost to the government.

9.119 Sixth, we note that the SCM Agreement does not contain any express statement of its object and purpose. We therefore consider it unwise to attach undue importance to arguments concerning the object and purpose of the SCM Agreement. In our view, however, the avoidance of net cost to government is not the object and purpose of the multilateral disciplines contained in the SCM Agreement. Rather, as suggested by Canada itself at para. 96 of its first submission, we consider that the object and purpose of the SCM Agreement could more appropriately be summarised as the establishment of multilateral disciplines "on the premise that some forms of government intervention distort international trade, [or] have the potential to distort [international trade]".

9.120 For all the above reasons, we reject Canada's 'net cost to government' interpretation of the term "benefit" in Article 1.1(b) of the SCM Agreement. Thus, leaving aside situations of alleged "income or price supports" within the meaning of Article 1.1(a)(2), we consider that a "financial contribution" by a government or public body confers a "benefit", and therefore constitutes a "subsidy" within the meaning of Article 1 of the SCM Agreement, when it confers an advantage on the recipient relative to applicable commercial benchmarks, i.e., when it is provided on terms that are more advantageous than those that would be available to the recipient on the market.

D. Export Development Corporation

9.121 Brazil challenges the EDC programme as a per se export subsidy prohibited by Article 3 of the SCM Agreement. Brazil also challenges the EDC as applied, including funding, support, funds and benefits granted under the auspices of the EDC programme to the regional aircraft industry.⁵³⁵

9.122 Brazil submits that the EDC is an agency of the Government of Canada which was established by the Export Development Act "for the purposes of supporting and developing, directly or indirectly, Canada's export trade and Canadian capacity to engage in that trade and to respond to international business opportunities."⁵³⁶ Brazil refers to EDC materials to the effect that "EDC's mandate is to help Canadian business compete and succeed in the global marketplace", and to attempt to satisfy "the seemingly endless appetite of Canadian exporters for financial support"⁵³⁷ through a variety of financial and risk absorption services, including export trade insurance, sales, financing, loan guarantees, and equity investments. Brazil asserts that these benefits are not available from private financial institutions. Brazil asserts that all assistance granted to the regional aircraft industry under the EDC programme is contingent on export.

9.123 Brazil asserts that EDC provides financing assistance to Canadian regional aircraft exporters in a variety of ways, four of which are specifically challenged by Brazil in these proceedings: debt financing, loan guarantees, residual value guarantees, and equity financing. Before addressing Brazil's specific arguments concerning these EDC activities, it is first necessary to examine Brazil's claim regarding the EDC programme per se.

1. Is the EDC programme per se a prohibited export subsidy?

9.124 To assess Brazil's claim against the EDC programme, we must first determine whether the EDC programme per se mandates the grant of prohibited export subsidies in a manner inconsistent with Articles 3.1(a) and 3.2 of the SCM Agreement. In this regard, we recall the distinction that GATT/WTO panels have consistently drawn between discretionary legislation and mandatory legislation. For example, in United States - Tobacco the panel "recalled that panels had consistently ruled that legislation which mandated action inconsistent with the General Agreement could be challenged as such, whereas legislation which merely gave the discretion to the executive authority ... to act inconsistently with the General Agreement could not be challenged as such; only the actual application of such legislation inconsistent with the General Agreement could be subject to challenge".⁵³⁸

9.125 During the Panel proceedings, we addressed the following question to Brazil:

9.126 In response, Brazil stated that:

9.127 Leaving aside the issue of export contingency, we thus understand Brazil to argue that EDC is effectively required to grant subsidies. However, we find nothing in Brazil's various submissions in support of this argument. The only factual evidence proffered by Brazil in support of its argument is the quote from EDC's mandate that EDC was established "for the purposes of supporting and developing, directly or indirectly, Canada's export trade and Canadian capacity to engage in that trade and to respond to international business opportunities."⁵⁴⁰ This statement by itself clearly cannot be viewed as a requirement to provide prohibited export subsidies. Nor has Brazil demonstrated otherwise that such support and development necessarily involves subsidization. Although such support and development might conceivably take the form of subsidization, there is nothing to suggest that this will necessarily be the case. In our view, a mandate to support and develop Canada's export trade does not amount to a mandate to grant subsidies, since such support and development could be provided in a broad variety of ways.

9.128 We consider that Brazil effectively concedes that the EDC mandate does not require the grant of export subsidies when it states that the EDC mandate has been interpreted to require the EDC to fund projects that give "Canadian exporters an edge when they bid on overseas projects."⁵⁴¹ For Brazil, this "edge" necessarily refers to subsidization. Even if the grant of an "edge" did imply the grant of subsidies,⁵⁴² and even if in practice the EDC programme were applied so as to grant subsidies, this would not mean that, in law, the EDC mandate requires the grant of subsidies. Rather, in such circumstances the grant of subsidies would be the result of the exercise of the administering authority's discretion in interpreting its mandate. We again recall that the panel in US - Tobacco recollected "that panels had consistently ruled that legislation which mandated action inconsistent with the General Agreement could be challenged as such, whereas legislation which merely gave the discretion to the executive authority ... to act inconsistently with the General Agreement could not be challenged as such..."

9.129 For these reasons, we find that Brazil has failed to demonstrate that the EDC programme as such mandates the grant of subsidies. Rather, the EDC programme constitutes discretionary legislation. In light of the distinction that GATT/WTO panels have consistently drawn between discretionary legislation and mandatory legislation, we find that we may not make any findings on the EDC programme per se. We therefore confine our analysis to Brazil's claims concerning the actual application of the EDC programme in the regional aircraft sector.

2. Does the EDC programme as applied provide prohibited export subsidies?

9.130 We recall that Brazil challenges four types of EDC financing assistance allegedly provided to the Canadian regional aircraft industry: debt financing, loan guarantees, residual value guarantees and equity financing. We shall address each of these alleged forms of EDC financing assistance in turn, for the purpose of determining whether they constitute "subsid[ies]" within the meaning of Article 1 of the SCM Agreement. Only if we make affirmative determinations in this regard will we consider whether such subsidies are "contingent ... upon export performance" within the meaning of Article 3.1(a).

(a) EDC debt financing

(i) Arguments of the parties

9.131 Brazil asserts that Canada grants subsidies in the form of direct financing at concessionary rates for up to 90 percent of the cost of an aircraft. According to Brazil, EDC financing of up to 90 percent (or more) of an aircraft's cost constitutes a direct transfer of funds by grant or loan, within the meaning of Article 1.1(b) of the Agreement. Furthermore, "EDC's provision of financing of up to 90 percent (or more) of an aircraft's cost over a 15-year or 15-year-plus period at concessionary rates confers the obvious benefit, within the meaning of Article 1.1, of lowering the price of an exported aircraft for the purchaser. No private financial institution or investor would provide this degree of financing on concessionary terms ..."⁵⁴³ Brazil's claim against EDC debt financing is therefore based on its view that a "benefit" is conferred when a financial contribution is provided at terms that would not be available to the recipient on the market.

9.132 Brazil refers to a statement by Mr. Labbé, a former EDC President, to demonstrate the "benefit" afforded to Canadian exporters by EDC debt financing:

9.133 According to Brazil, "EDC's help - to the tune of a 'few basis points' - must be better than that which would otherwise be commercially available, or an EDC financing package would not, in EDC's former President's words, 'give Canadian exporters an edge.' This 'edge' is a 'benefit' conferred upon exporters, within the ordinary meaning of Article 1.1 of the Subsidies Agreement."

9.134 Brazil also asserts that "one of the benefits of EDC financing is that a Canadian exporter can 'advis[e] potential foreign buyers that Canadian financing may be available for their purchase,' thereby 'enhanc[ing] the competitiveness of [the exporter's] sales proposal.'⁵⁴⁵ The Canadian regional aircraft exporter's product is more attractive to a purchaser, in turn, for the simple reason that it costs less than it would without the Canadian Government's help."⁵⁴⁶

9.135 Brazil argues that EDC debt financing confers a "benefit" because it does not provide for the collection of any risk premium. Brazil notes statements by EDC personnel to the effect that EDC "absorb[s] the risk on behalf of Canadian exporters, beyond what is possible by other financial intermediaries."⁵⁴⁷ Brazil cites EDC's net interest margin (3.03 percent in 1996, and 2.82 percent in 1997) to establish that it is not being compensated for this extra risk and, conversely, that recipients of EDC funding are paying less for that funding than they would from commercial sources. Brazil asserts that, to accept a loan portfolio of EDC's risk class, private investors would demand a spread of 1,242 basis points over riskless US Treasuries of 15 plus years maturity, or 17.73 percent. Brazil submits that the difference between this figure and EDC's net interest margin demonstrates that EDC's lending activities do not effectively cover the real and quantifiable costs associated with the riskiness of EDC's loan portfolio. Brazil considers that EDC's net interest margin should be higher than that of commercial banks, since EDC's loan portfolio is higher risk than commercial banks' loan portfolio. According to Brazil, it is a well-known and accepted financial concept that investments with higher risks demand higher returns.

9.136 Brazil refers to EDC's loan portfolio to establish that EDC's debt financing is riskier than that undertaken by commercial banks. Brazil notes that impaired or non-performing loans comprise 14.4 percent of gross loans receivable. Brazil asserts, on the basis of Standard & Poor's analysis, that an equivalent ratio of 3 percent "can cause concern", and even a ratio of 7 percent would indicate that the future of the relevant institution "may be doubtful".⁵⁴⁸ Brazil asserts that EDC's 14.4 percent ratio of gross impaired loans to gross loans receivable compares to ratios of approximately one percent for The Royal Bank of Canada, the Bank of Montreal, Canadian Imperial Bank of Commerce, and the average of all US FDIC-insured commercial banks. In addition, Brazil notes that approximately 57.6 percent of EDC's performing loan portfolio has been classified by EDC itself as "below investment grade" or "speculative grade". Brazil also refers to EDC's ratio of allowance (or reserve) for losses on loans to gross loans receivable. Brazil asserts that banks are required to maintain a reserve for loan losses in order to cover possible future loan losses. Brazil argues that the level of reserve maintained by a bank therefore reflects management's judgment regarding the quality of its loan portfolio. In this regard, Brazil notes that EDC's allowance for loan losses as a percent of total loans is 13.2 percent, compared with between one and two percent for each of the three major Canadian banks and the US industry average. Brazil also refers to EDC's ratio of net write-offs to gross loans receivable. Brazil asserts that EDC's ratio of 0.15 percent is "absurdly low", and questions both the criteria applied by the EDC in determining when loans are written off, and whether the criteria are similar to those applied by commercial banks.

9.137 Furthermore, Brazil relies on the following statement by an EDC official before the Canadian Parliament to demonstrate that a risk premium is not collected by EDC:

To continue with Brazil submits that EDC cannot be collecting the requisite

⁵³³ See para. 5.30 above.

⁵³⁴ In this regard, we note an apparent inconsistency in Canada's arguments as we see no difference between Canada's reliance on item (k) as specific contextual guidance for the interpretation of "subsidy" under Article 1, and the very a contrario reading of item (k) that Canada has expressly rejected for the purpose of defining when government credit constitutes a "subsidy" within the meaning of Article 1 of the SCM Agreement. That is, we consider that Canada is effectively seeking to rely on an a contrario reading of item (k), while at the same time asserting that it rejects this approach.

⁵³⁵ We note that Brazil's request for establishment of a panel (WT/DS70/2) refers to various alleged prohibited export subsidies provided to the "Canadian industry producing civil aircraft". Our terms of reference, therefore, relate to assistance provided to the Canadian civil aircraft industry. With the exception of its first written submission, however, all findings requested by Brazil concerning assistance provided under the EDC programme refer to the "regional aircraft industry". Similarly, all arguments adduced by Brazil concern the regional aircraft industry. We consider that the Canadian regional aircraft industry is a sub-set of the Canadian civil aircraft industry. For these reasons, we only examine Brazil's claim against assistance provided under the EDC programme insofar as such assistance relates to the Canadian regional aircraft industry. Brazil stated that "[r]egional aircraft generally range from 30 to 70 seats, and serve markets up to about 1,600 kilometres apart". Brazil noted in addition that customized regional aircraft may occasionally include up to 78 seats, and that Bombardier and Fairchild Dornier USA are currently developing regional aircraft that may range up to 90 seats (see para. 1.2 of Brazil's first written submission).

⁵³⁶ Export Development Act, Section 10(1).

⁵³⁷ Export Development Corporation, 1995 Chairman and President's Message ("Message"), pages 2 and 4.

⁵³⁸ United States - Measures Affecting the Importation, Internal Sale and Use of Tobacco, BISD 41S/131, para. 118, adopted on 4 October 1994 ("United States - Tobacco").

⁵³⁹ CanadExport On-Line, Focus on Export Development Corporation, pg. 2 (statement of Mr. Paul Labbé, former President of the EDC).

⁵⁴⁰ Export Development Act, Section 10(1).

⁵⁴¹ CanadExport On-Line, Focus on Export Development Corporation, page 2 (statement of Mr. Paul Labbé, former President of the EDC).

⁵⁴² Canada expressly denies that giving an "edge" implies the grant of subsidies. Canada asserts that the "edge" given to Canadian exporters derives from EDC's experience and expertise.

⁵⁴³ See para. 6.60 above.

⁵⁴⁴ CanadExport On-Line, Focus on EDC, page 2.

⁵⁴⁵ Infoentrepreneurs website, page 13.

⁵⁴⁶ See para. 6.59 above.

⁵⁴⁷ Export Development Corporation, 1995 Chairman and President's Message, page 2.

⁵⁴⁸ See para. 6.13 above.