CHAPTER 11
PLURILATERAL AGREEMENTS
The GATT 1947, a treaty between contracting parties that functioned on the basis of consensus, was very difficult to amend and expand. As discussed in Chapters 1 and 2, in an effort to circumvent this problem, in the 1960s and 1970s groups of like-minded countries that sought to agree on more specific rules for policies covered by the GATT negotiated so-called codes of conduct. These codes bound only signatories, and were mostly applied on a MFN basis. Most of the existing codes were mapped into the WTO during the Uruguay Round, and their disciplines became binding upon all WTO members. They are discussed in Chapters 5 and 9. However, four Tokyo Round codes were not converted into multilateral agreements. Instead they became so-called plurilateral agreements. These bind only signatories and do not have to be applied on a MFN basis. As the WTO has no general MFN obligation—nondiscrimination requirements are contained in each of the various multilateral trade agreements—the plurilateral agreements contained in Annex 4 of the WTO are examples of what has been termed conditional MFN agreements.
Although members are free to discriminate against nonsignatories of a plurilateral agreement, subsets of WTO members cannot simply get together and form a club without the permission of other members. A plurilateral agreement can only be appended to the WTO on the basis of consensus (Article X:9 WTO). The plurilateral option therefore offers a mechanism for groups of WTO members to agree to rules in a policy area that is not covered by the WTO or goes beyond existing disciplines, as long as the membership as a whole perceives this not to be detrimental to their interests.
Of the four Tokyo Round agreements included in Annex 4 of the WTO in 1995, two ceased to apply in September 1997. The International Bovine Meat Agreement
plurilateral agreements 513
and the International Dairy Agreement were terminated after their governing bodies decided that the objectives of the agreements could more effectively be pursued through other WTO bodies, including the Committees on Agriculture and Sanitary and Phytosanitary Measures (they are discussed in Hoekman and Kos-tecki, 1995). As of 2008, only two plurilateral agreements were operational, the Agreement on Government Procurement and the Agreement on Trade in Civil Aircraft. Of the two, the government procurement agreement is by far the more important.
11.1. Government Procurement
All over the world government agencies procure goods and services as inputs into the production of public goods and services—education, defence, utilities, infrastructure, health and so forth. The size of the associated public procurement market is often very large, depending on the economic system of a nation and its GDP. Governments concerned with maximizing the use of scarce financial resources have developed procedures and mechanisms to attempt to ensure that public entities procure goods and services efficiently. A common element is to mimic the working of the market by requiring that public entities seek competitive bids from potential suppliers of goods and services.
Starting in the 1980s, many governments began to pursue more far-reaching efforts to enhance the efficiency of public services by directly subjecting production units to competitive forces. Examples include privatization of state-owned enterprises, permitting entry of privately owned firms into sectors traditionally reserved for public entities (such as utilities) and contracting out activities to independent suppliers. These developments changed the public procurement market significantly. What was once produced 'in-house' by government entities began to be supplied by private operators—shrinking the public procurement market But at the same time the overall market for goods and services did not decline—there was simply a shift towards real market-based contracting as opposed to one where the focus was on mimicking the role of the market through competitive bidding for government contracts.
The cost-minimizing goal underlying competitive bidding requirements for purchases by public entities is frequently not attained because legislation requires procuring entities to pursue other objectives as well. These may include a desire to promote the development of domestic industry or technology, to support particular types of firms (such as small- and medium-sized enterprises) and to safeguard national security. Often, such objectives are pursued by requiring procurement
514
the political economy of the world trading system
practices to be used that discriminate against foreign suppliers. Examples of discriminatory policies pursued by governments include outright prohibitions on foreign sourcing (civil servants must fly national airlines), threshold criteria for foreign sourcing to be permitted (minimum cost or price differentials compared to local suppliers), or satisfaction of offset or local content requirements.
As a result, government procurement policies are often a somewhat schizophrenic exercise in 'constrained cost minimization'. The basic goal is value for money, subject to the other policy goals that need to be taken into account. In practice, these other goals often imply that for all, or part, of any contract the scope of competition is reduced to the set of firms that meet specific criteria laid out in legislation or procurement regulations. In some instances there may not be any competition at all. Governments may use selective or single tendering procedures under which a procuring entity directly approaches a specific firm for a bid. Discriminatory public procurement practices are a major market access issue on the WTO agenda, given that procurement markets account for 5-10 per cent of GDP in most WTO members.
If many countries pursue discriminatory procurement practices, the end result for the world as a whole is likely to be inferior in welfare terms to a cooperative outcome where governments agree to refrain from using procurement as a tool to protect national industries or to pursue noneconomic objectives. After all, governments have other policies that can achieve the other objectives that are pursued through procurement policy—such as subsidies and the tax system. Recognizing this, governments have attempted to negotiate multilateral rules of the game for public procurement. The Tokyo Round Government Procurement Agreement (GPA) was the reflection of such an effort. Liberalization of procurement markets has also been pursued in a regional context. Procurement disciplines are prominent in the EU, where member states are prohibited from discriminating against tenders from foreign firms (be they located in other EU member states or outside Europe). As is the case with IPRs, international cooperation has been driven in part by bilateral pressures. A number of US laws require USTR to monitor foreign procurement policies that impact negatively on US firms, and foreign government procurement figures prominently in the USTR's annual Trade Practices report (see the USTR homepage). The same is true for similar reports issued by the EU (under the Trade Barriers Regulation), Japan and other countries. This pressure is not only focused on discriminatory practices; the US in particular has stressed issues related to transparency and corruption.
Discriminatory procurement policies are often considered prima facie evidence of protectionism—-governments explicitly favour domestic suppliers of goods and services. However, procurement policy that gives a price preference of 10 per cent to a local industry or to firms from a specific region or below a certain size is not equivalent to a 10 per cent tariff. This is because demand by the private sector for imports may not be affected by the preference policy. As long as the government
plurilateral agreements
515
market is only a fraction of total demand for a product, as is often the case, the tariff equivalent will only be a fraction of 10 per cent. Indeed, as pointed out by Baldwin and Richardson (1972) in a seminal analysis, if domestic and foreign products are good substitutes and government demand is less than the initial domestic supply capacity, discrimination will have no effect on prices, overall output and welfare. The increased demand by the government for domestic output will be exactly offset by greater private sector imports, so that the policy has no effect on equilibrium prices and production of the domestic industry.
This result continues to obtain if there is imperfect competition (oligopoly) as long as goods are perfect substitutes.1 Imports might actually increase as a result of discriminatory policy if domestic firms are induced to cut back sales to the private sector in an attempt to raise prices. However, if government demand exceeds local production, discrimination will result in domestic prices being bid up, and output of the domestic industry will expand to meet demand. If allowance is made for the fact that in the longer run firms will enter into markets where there are excess profits, over time prices will fall and at the end of the day the discrimination policy may again have no negative implications for welfare (see Annex 2).
Matters are different if the procurement is for products for which there are just a few suppliers. In such situations there may be potential economic rationale for discrimination. McAffee and McMillan (1989) show that if domestic firms have a competitive disadvantage in producing the product (are higher cost producers compared to foreign firms), and only a limited number of firms (foreign and domestic) bid for the contract, a price preference policy may induce foreign firms to lower their bids. If the products procured are intangible (services) or there are problems in monitoring and enforcing contract compliance, discrimination can increase the likelihood of performance. Problems of asymmetric information and contract compliance may give entities a natural preference to choose suppliers located within their jurisdictions as this can reduce monitoring costs. Such proximity incentives will make it more difficult for foreign firms to bid successfully, even in the absence of formal discrimination. The policy issue that then arises is whether there are barriers against establishment (FDI) by foreign suppliers (Evenett and Hoekman, 2005).
Although discriminatory procurement may enhance national welfare by lowering procurement costs in small numbers settings, simulation studies suggest that welfare gains are likely to be modest at best. Greater profits of domestic firms or cost-savings to public entities will tend to offset by increased prices. As a result, the potential cost-savings are reduced (Deltas and Evenett, 1997). Given that in most
1 As noted by Deardorff and Stern (1998), domestic middlemen will always have an incentive to import a good and resell it to the government after processing it enough to qualify as domestic. The same forces therefore apply if goods are imperfect substitutes. Difficulties in determining the origin of products will always reduce the effectiveness of discriminatory policies, such as procurement preferences, that are not enforced by customs officials. The level of tariffs and NTBs applied at the border will be the main constraint on such arbitrage activities.
5l6       the political economy of the world trading system
instances the optimal discriminatory policy will be difficult to determine (it generally will vary depending on the specifics of the situation), in practice favouritism can be expected to be more costly than a policy of nondiscrimination. In many situations the information required to judge if diverging from nondiscrimination is beneficial will not be available. Nondiscrimination has therefore been argued to be a good rule of thumb (Hoekman, 1998).
If account is taken of the rent-seeking distortions that may be induced by discriminatory policies and the social cost of corruption and bribery, the case for nondiscrimination is substantially strengthened. All of the above arguments regarding the economic pros and cons of discrimination cease to apply if government entities do not maximize social welfare. Nondiscrimination will generally reduce discretion and enhance transparency of the procurement process and thus reduce the scope for rent-seeking. Most important in this connection is transparency and a system of rules to impede corruption. Open and competitive bidding, whether or not there are preferences for domestic industry, is a key instrument in this regard.
The WTO Government Procurement Agreement
As noted in Chapter 5, Article 111:8 GATT excludes procurement from the national treatment obligation. Article XIII GATS does the same for services. The 1979 Tokyo Round GPA extended basic GATT obligations such as nondiscrimination and transparency to the purchases of goods by selected government entities. The GPA has been re-negotiated three times. The second, 1996 version of the Agreement, extended its reach to services. The most recent revision of the GPA was provisionally agreed in December 2006 (GPA/W/297). The periodic re-negotiations are mandated by the Agreement itself. Although they have coincided with the Uruguay and Doha Rounds, as a plurilateral agreement, the (re-)negotiations were not formally part of the rounds.
The aim of the latest revision—which significantly re-wrote and re-organized the 1996 version—was to make the GPA more attractive to nonsignatories by simplifying the rules, to reflect advances in information technology and to expand the coverage of the agreement (Anderson, 2007). Final agreement on the 2006 text is conditional on a mutually satisfactory outcome of parallel (and ongoing) negotiations to open up additional government procurement to international competition. A purported objective of the re-organization and redrafting of the text of the GPA was simplification and making it reflect better the process that procuring entities go through. It is not clear that this objective was achieved—in our view in some respects the 1996 text was more transparent and easier to understand.
As a result of the periodic re-negotiations, the coverage of the GPA has expanded substantially over time to include services and more government entities, and its
plurilateral agreements
517
disciplines clarified and updated to reflect new technologies and procurement practices. Membership of the GPA is limited to mostly OECD countries. As of 2008 it comprised Canada, the European Communities, the 27 EU member states, the Netherlands with respect to Aruba, Hong Kong, Iceland, Israel, Japan, Liechtenstein, Norway, the Republic of Korea, Singapore, Switzerland and the United States.2
The GPA applies to any measure regarding covered procurement, whether or not it is conducted exclusively or partially by electronic means' (Article II:i). The concept of procurement covers all contractual options, including purchase, leasing, rental and hire purchase, with or without the option to buy.3 A positive list is used to determine what procurement is covered. The GPA applies only to entities listed in Appendix I of the agreement. There are three 'entity annexes': Annex 1 lists covered central government entities; Annex 2 lists subcentral government entities; and Annex 3 lists all other entities that procure 'in accordance with the provisions of this Agreement'. Annex 3 is a catch-all category that includes bodies such as utilities. Entities listed in Annex 3 may be partially or totally private. What constitutes a government entity is nowhere defined in the agreement, reflecting a lack of consensus on what constitutes a public undertaking—more specifically, whether a former state-owned or controlled enterprise that has been privatized or that is subject to competition should be required to follow GPA procurement practices. Instead a pragmatic approach is taken—governments negotiate which entities are listed.
The entities listed in the three annexes are subject to the rules and disciplines of the GPA with respect to their procurement of goods and services if the value of the procurement exceeds certain specified thresholds (see Table 11.1) and the goods or services involved are not exempted from the coverage of the agreement. As far as goods are concerned, in principle all procurement is covered, unless specified otherwise in an annex. Procurement of services is subject to a positive list: only the procurement by covered entities of services explicitly scheduled in Annexes 4 and 5 are subject to the GPAs rules, and then only insofar as no qualifications or limitations are maintained in the relevant annexes. To give an indication of the orders of magnitude involved in the 1995 extension of the GPA's coverage to subcentral entities and services—the offers made by the US and the EU covered some US$100 billion of purchases, with care being taken to maintain reciprocity through addition of removal of specific entities and sub-national authorities (Schott and Buurman, 1994: 74).
The primary obligation imposed by the GPA on covered entities is nondiscrimination—national treatment and MFN (Article V). This extends not only to
2 The European Communities refers to the Community"s institutions. Formally, there were 40 signatories to the GPA in 2008, as each EU member state has signed the agreement individually in addition to the European Communities. This is because in some dimensions of procurement EU member states retain competence.
3 The GPA applies to purchases of goods and services that are not intended for resale. If government entities engage in trade (buying and selling) Article XVII GATT applies (see Chapter 5).
518      the political economy of the world trading system
Table 11.1. GPA thresholds for coverage of procurement contracts (SDRs)
Category of Procurement Threshold
Central government entities
Goods 130,000
Services except construction services 130,000
Construction services 51000,000t
Annex 2: Subcentral government entities
Goods 200,000*
Services except construction services 200,000*
Construction services 5,000,000§
Annex 3: All other entities whose procurement is covered by the Agreement
Goods -100,000*
Services except construction services 400,000^
Construction services 5,000,000§
In general public enterprises or public authorities such as utilities.
I Israel: 8.5 million; Japan 4.5 million {with architecture services: 45-0,000). i US and Canada: 355,000; Israel: 250,000.
5 Israel: 8.5 million;.Japan and Korea: 15 million.
II Canada and Israel: 355,000; Japan: 130,000.
Source: WTO Government Procurement Agreement.
imports but also to subsidiaries of locally established foreign firms. The GPA thus goes beyond the GATT, which does not extend national treatment to foreign affiliates, and the GATS, which does so only if specific commitments to that effect have been made. Under the GPA all foreign affiliates established in the country are to be treated the same as national firms. Local content, price preferences and similar discriminatory policies are prohibited. Moreover, signatories may not discriminate against foreign suppliers by applying rules of origin that differ from those they apply in general to MFN-based trade.
Most of the provisions of the GPA concern transparency broadly defined. Thus, much attention is given to requiring signatories to specify where information on procurement systems and opportunities will be published (including through electronic means). These must be listed in Appendices II through IV to the GPA. There are detailed requirements for publication of notices of intended procurement, the conditions for participation and permitted systems to ascertain that suppliers are qualified, technical specifications and tender documentation, minimum time periods to allow bids to occur and regular reporting of statistics on procurement activities of covered entities,
The GPA does not explicitly require that procurement be competitive or that certain procurement methods be used. In this regard it is quite different from the procurement guidelines that other international organizations and national governments apply. Regarding conduct of procurement, signatories are simply
plurilateral agreements
519
required to '... conduct covered procurement in a transparent and impartial manner that is consistent with this Agreement, using methods such as open tendering, selective tendering and limited tendering; avoids conflicts of interest and prevents corrupts practices' (Article V:4). Open tendering is any method that allows any supplier to bid (i.e. competitive tendering). Selective tendering is a method where only suppliers that satisfy specific criteria for participation may bid (usually prequalified suppliers). Limited tendering is noncompetitive and usually involves a procuring entity approaching one or more potential supplier of its choice.
The rules in the GPA regarding selective tendering are basically aimed at ensuring that foreign suppliers can demonstrate they qualify and are not discriminated against in this regard (e.g. have the information needed). Limited tendering may only be used if no tenders were received or they were not responsive, only one supplier can provide the good or service (e.g. artwork, products protected by IPRs), for additional, follow-on deliveries, in situations of extreme urgency, for commodities (the presumption being that there is a world price for standardized, homogenous goods) or for prototypes.
There is no explicit hierarchy of the three tendering methods mentioned in Article V and governments are free to use others. The preference for competitive procurement methods is implicit in the agreement, reflected in requirements that notices of intended or planned procurement be published (including information on the mode of procurement, its nature and quantity, dates of delivery, economic and technical requirements, and amounts and terms of payment), in the conditions that must be satisfied if governments use limited tendering, and in the disciplines on treatment of tenders and contract awards. Article XIII on limited tendering makes it clear that competition is preferred by making use of this method conditional on it not being used to avoid competition among suppliers, to discriminate or protect domestic suppliers. Article XV requires that entities award contracts to the supplier 'determined to be fully capable of undertaking the contract' and who is either the lowest tender (if price is the sole criterion) or the tender that is most advantageous (in terms of the evaluation criteria set out in the notices or tender documentation). It is rather surprising that the objective of competitive procurement is not embedded in the preamble of the agreement. The 'fuzziness' as regards the preference for competitive bidding may reflect the desire of signatories to see membership of the GPA expand to include developing countries.
Price-preference policies, offsets and similar policies that are widely used by governments are in principle prohibited for covered procurement as a result of the national treatment rule (Article V). This has been a problem for developing countries, as these countries use procurement as an instrument to achieve objectives other than 'value for money'. Article IV of the 2006 GPA permits developing countries to negotiate the right to adopt or retain price preference policies and offset requirements on a transitional basis, and delay the implementation of any
520
the political economy of the world trading system
and all provisions other than MFN for up to three years (five years for a LDC). Moreover, after accession the GPA Committee may extend the transition periods or approve the use of new transitional price preferences or offsets if there are 'special circumstances that were unforeseen during the accession process' (Article IV7). Existing signatories also commit themselves to give due consideration to any request by a developing country for technical cooperation and capacity building' (Article IV:8). Some scope therefore exists for maintaining a price preference or offset policy—but it is time-limited.
The nature of procurement is such that most of the time, unless rapid action can be taken, firms will not have an interest in bringing cases contesting violations of the rules of the game. Accordingly, the GPA supplements the right of signatories to invoke the WTO DSU—which is too slow to be relevant for many real-world procurement situations—with a requirement that members establish domestic review procedures. These bid-protest or -challenge mechanisms should provide for rapid interim measures to correct breaches of the agreement or a failure of a government entity to comply with the measures implementing the GPA (Article XVIII). Measures to preserve commercial opportunities may involve suspension of the procurement process, or compensation for the loss or damages suffered. This may be limited to the costs for preparing the tender or the costs relating to the challenge, or both.
Articles IX and XVII GPA require each signatory, on request from another party, to promptly provide pertinent information concerning the reasons why the supplier's application to qualify was rejected, why an existing qualification was terminated, and information necessary to determine whether a procurement was conducted in accordance with the GPA, including pertinent information on the characteristics and relative advantages of the tender that was selected. The 2006 provisions in this area are weaker than those in the 1996 agreement.
Operation of the GPA
The GPA requires signatories to report annual statistics on procurement by covered entities to the Committee on Government Procurement. Such data reporting was intended to help parties determine how well the agreement was functioning, in part by providing comparable cross-country information on sourcing practices. Signatories began reporting statistics for the year 1983. Unfortunately, there has been very little empirical research using these data, and it does not appear that signatories to the GPA have used the statistics as a way of monitoring the operation of the agreement.
Data reported in Hoekman (1998) for the 1983-92 period—when only central government procurement of goods was covered by the GPA—revealed that the largest procurement market, by a substantial margin, opened up under the GPA
plurilateral agreements
521
was that of the United States, which accounted for almost half of the total procurement reported. Smaller countries, on average, procured much more on international markets than did large countries. If Canada, the EU, Japan and the US were excluded, about 60 per cent of purchases by covered entities exceeding the threshold went to national suppliers. This compared to more than 90 per cent for the large players. As EU statistics defined 'domestic' as including intra-EU sour-cing, reported self-sufficiency ratios for the EU-12 were above 90 per cent on average. In interpreting these statistics it should be noted that no distinction is made between domestic firms proper and foreign firms that have established a local presence. To the extent that large countries attract a greater amount of FDI, higher self-sufficiency ratios are not indicative of discriminatory policies.
In the EU, Japan and the US, the share of domestic firms in total above threshold procurement by covered entities remained virtually unchanged during 1983-92. For the smaller countries, however, with the exception of Singapore and Switzerland, the share of procurement from national sources declined over time. It is impossible to attribute such changes in sourcing patterns to the GPA—regional developments also played a role, such as the NAFTA in North America and efforts to liberalize EU procurement markets. Unilateral deregulation and privatization policies also must have had an impact. Nonetheless, the finding that smaller GPA members became less nationalistic in their purchasing decisions suggests that practices did become more open.
During the same period, the share of contracts that exceeded the threshold tended to increase. In 1983-5, some 39 per cent of all procurement by covered entities fell above the threshold. By 1991-2, it had risen to 49 per cent. This can be explained in part by a reduction in the threshold in 1988, from SDRi50,ooo to SDRi30,ooo. As of the early 1990s, the share of above threshold contracts for both EU and US entities averaged around 60 per cent.
Under the GPA, open competitive tendering procedures are, in principle, to be used for all contracts that exceed the relevant threshold. As noted earlier, limited tendering procedures involving an entity negotiating with potential suppliers individually is only allowed under certain conditions and members are required to report data on their use of this method. The issue became important in US-Japan trade relations in the 1980s, following US complaints that the use of limited tendering was excessive (Stern and Hoekman, 1987). In the period investigated byHoekman (1998), the use of limited tendering varied from a reported low of zero (Singapore) to a high of over 30 per cent on average for France, Italy, Switzerland and Hong Kong. Across all signatories the average share of limited tendering was about 13 per cent. As of 1992, both the EU and the US used limited tendering for about 10 per cent of contracts. Japan s use of limited tendering rose from around 12 per cent during 1983-5 to 21 per cent during 1990-2.
Choi (2003) and Evenett and Shingal (2006) have undertaken country-specific studies of the operation and impact of the GPA, focusing on Korea and Japan
522
the political economy of the world trading system
respectively. Choi finds that accession to the agreement by Korea was followed by a reduction in the share of procurement using limited tendering (which fell from 27 per cent in 1993-5 to 22.5 per cent in 1996-8. However, the share of foreign supplied goods during this period fell. Evenett and Shingal conclude that in 1999 in Japan more contracts fell below the GPA thresholds than in earlier years, and that of the contracts that exceeded the threshold—-and thus were covered by the GPA—a smaller share was awarded to foreign suppliers in 1998-9 than in 1990-1. Thus, during the 1990s, it appears that the GPA did nothing to increase the market access for foreign suppliers.
The data reporting requirements of the GPA are not as useful or informative as they might be because most signatories do not report on a timely or comprehensive basis. One reason the studies just mentioned focus on Korea and Japan is that these countries report regularly. More regular reporting—and analysis of the reports by the WTO Secretariat—would do much to improve knowledge regarding implementation of the agreement. That said, what matters from an economic point of view is primarily the size of government demand for a good or service relative to total domestic supply. As discussed previously, it is particularly in cases where the government is a big player relative to domestic supply that there can be significant effects on national welfare and foreign suppliers. Multilateral scrutiny will have potentially the largest payoff if it focuses on such situations. As the GPA reporting requirements are quite burdensome, an added benefit of a more focused approach to data collection would be a reduction in the costs of surveillance.
There have been only three disputes under the WTO on procurement that have led to invocation of the DSU. All have involved the EU or the US. In 1996, the EU objected to a law of the Commonwealth of Massachusetts that prohibited public authorities in Massachusetts from procuring goods or services from persons who do business with Myanmar. The EU (joined in 1997 by Japan) argued that this violated the GPA, as Massachusetts is covered under the US schedule to the GPA. A panel was established, but proceedings were suspended at the request of the complainants, following a US Supreme Court decision that struck down the ban as being unconstitutional because it infringed upon the authority of the President of the United States to set foreign policy. In 1997, the EU raised the procurement of a navigation satellite by the Japanese Ministry of Transport, arguing that the technical specifications in the tender were not neutral because they referred explicitly to US specifications. Here also a MAS was found (a panel was never established).
A third case concerned US allegations that the Korean Airport Construction Authority's practices relating to qualification for bidding as a prime contractor, domestic partnering and the absence of access to challenge procedures violated the GPA. In the Inchon airport dispute the US argued that Korea had failed to comply with the GPA by imposing bid deadlines and domestic partnerships and by awarding the construction contract to the Korean Airport Authority. The WTO
plurilateral agreements
523
panel ruled that the Korean Airport Authority was not covered by the GPA (it was not listed in Korea's annexes) and therefore was outside the scope of the agreement (Matsushita, 2006).
There have also been cases involving developing countries that are not members of the GPA. For example, in 1991, two Brazilian firms won an international tender for electric power transformers issued by the Federal Electricity Commission of Mexico. Subsequently, three of the Mexican firms that lost the tender brought an antidumping petition against the Brazilian firms. Antidumping duties ranging from 26 to 35 per cent were imposed in September 1993. Brazil requested the GATT Antidumping Committee to conciliate, arguing that the AD duty was calculated by comparing prices bid by the different firms for the original tender, and not by comparing prices charged in the home and the export market. This dispute was also resolved bilaterally (Hoekman and Kostecki, 1995).
As is the case regarding trends in procurement sourcing by signatories, little is known about the extent to which the GPA domestic challenge mechanisms have been used in practice. Matsushita (2006) discusses one instance that arose in Japan. In this case, Motorola, a US company, brought a complaint against a Japan Railway procurement tender, arguing that it did not use an ISO standard whose enactment was imminent. The review body ruled that the GPA did not require procuring entities to take such expected changes in international norms into account—what matters are the standards that prevail at the time a tender is issued.
The challenge of expanding membership
Public procurement markets are too big to be left beyond the reach of the multilateral trading system. As mentioned above, membership of the GPA is quite limited. Indeed, not all OECD countries have signed it, e.g. Australia and New Zealand are not members. The US has made expansion of membership a priority issue, linking this to the broader issue of combating corruption.4 Despite long standing efforts to expand the number of signatories, very little progress in that direction has been achieved in the 30 years that the GPA has existed. The increase in membership since the early 1980s has predominantly been driven by the expansion of the EU from 12 to 27 members. Given that the EU is a signatory to the GPA and imposes procurement disciplines that are much more detailed and
4 This has mostly been reflected in unilateral decisions to make bribery and corrupt procurement a criminal offence in the United States and efforts to obtain agreement from OECD members to do the same. In April 1996, largely at the insistence of the US, OECD members agreed not to allow firms to write off bribes against tax obligations (Oxford Analytica, 18 April 1996). In 1997 the OECD adopted the Convention on Combating Bribery of Foreign Public Officials in International Business Transactions, which made bribery a criminal offence and required all signatories to pass legislation to that effect.
524
the political economy of the world trading system
prescriptive than those of the GPA, the expansion of membership to date does not reflect very well on the agreement.
A number of countries that have acceded to the WTO since 1995 made commitments to negotiate accession to the GPA, with China being the most notable and important economy to have done so. Time will tell whether these accession promises will be realized. China has a clear interest insofar as membership of the GPA will give it access to markets of signatories. One reason low-income countries have been reluctant to accede to the GPA is that they do not see reciprocal concessions from GPA members as having value. Negotiating leverage is a function of the size of the incremental market access that aspiring members can offer, and for most developing countries this tends to be small. Moreover, many developing countries are not players on the international procurement market, implying that the standard mercantilist bargain is not available to move things along.
Even if one abstracts from the skewed nature of the bargain from a (political) mercantilist perspective, what matters is whether adherence to the rules of the GPA will improve welfare. Taking a more 'rational' economic perspective, the question confronting developing countries is whether the eventual loss of the ability to use procurement policy as an instrument of industrial, regional or social policy is a good or bad thing. And, if it is a bad thing, whether there are other sources of gain that offset the loss. There is not a lot of empirical research on the effectiveness of procurement discrimination in achieving the industrial and other policy objectives. Tax/subsidy instruments are likely to be more efficient in assisting domestic target groups than procurement favouritism, but governments may confront fiscal constraints that impede the use of such policies. Moreover, an advantage of procurement that favours specific domestic groups is that it can help the most efficient firms in that group (as they must compete for the contracts) (Watermeyer, 2004), whereas a subsidy to a region or a minority will be less selective. This is a policy area that should be the focus of much more targeted research in developing countries—most of the literature pertains to the OECD countries. Whatever the case may be, in practice a large share of the procurement market where discrimination is now used to pursue equity and social objectives is unlikely to be of great interest to foreign suppliers: the average size of contracts is likely to be relatively small.
Price preferences have the advantage of being transparent and less distortive than other types of discriminatory policies that are often pursued (such as bans on participation by foreign bidders or local content and offset requirements). Tariffying such policies through an agreement permitting the maintenance of price preference schemes by developing countries would provide a focal point for future multilateral negotiations to reduce discrimination. Such preferences are allowed subject to certain conditions and limits by multilateral development banks. Provisions for their use are also included in the UNCITRAL Model Law on Procurement. Many developing countries have incorporated such preferences
plurilateral agreements
525
into their legislation. Recognition of the legitimacy of price preferences and offsets for developing countries—without time limits and transition periods—could help alter the incentives for accession.
Many of the purchases by government entities comprise services or products where economic forces favour procuring from local suppliers. In such cases, procurement preferences will only be binding if foreign firms cannot contest the market through FDI, or if government entities differentiate across firms on the basis of their nationality. Outright market access restrictions that take the form of a ban on FDI are costly to the economy as a whole, and policy efforts that focus on elimination of such bans are likely to have a greater payoff than attempting to outlaw discrimination.
The Working Group on Transparency in Procurement
Discrimination is just one, albeit important, dimension of possible multilateral disciplines for government procurement. It is widely believed that there are significant potential gains from disciplines that promote transparent procurement mechanisms, thereby reducing the scope for corruption and rent seeking. The 1996 WTO Singapore ministerial conference established a working group to study 'transparency in government procurement practices' and ways to develop elements for inclusion in an appropriate agreement'. Many developing countries perceived this to be a Trojan horse (a vehicle to start discussing discrimination and extend the coverage of GPA disciplines). However, given that discrimination is probably a second-order issue in comparison with corruption, there was a strong prima facie case to focus on transparency first and foremost.
The WTO working group proceeded by addressing the 'Items on the Chairman's Checklist of Issues' relating to a potential agreement on transparency in government procurement. The checklist comprised such broad issues as the definition of government procurement and the scope and coverage of a potential agreement, the substantive elements of a potential agreement on transparency, including various aspects of access to general and specific procurement-related information and procedural matters, as well as compliance mechanisms of a potential agreement and issues relating to developing countries, including the role of SDTand technical assistance and capacity-building. Signatories of the GPA not surprisingly strongly supported negotiations on transparency, as did a number of non-GPA countries such as Australia. Several draft proposals for an agreement were submitted in November 1999, including by the EU, Japan, Australia and a joint submission by Hungary, Korea, Singapore and the US. These countries sought to conclude an agreement at the Seattle ministerial. Many developing countries emphasized that much more discussion was needed on the implications of transparency obligations in the procurement area.
526      the political economy of the world trading system
At the Doha ministerial conference in 2001, ministers agreed that negotiations of transparency in procurement would take place after the fifth ministerial conference 'on the basis of a decision to be taken, by explicit consensus, at that Session on modalities of negotiations'. The ministerial declaration emphasized that negotiations shall be limited to the transparency aspects and therefore will not restrict the scope for countries to give preferences to domestic supplies and suppliers'. In line with the Doha Ministerial Declaration, which highlighted the need to 'take into account participants' development priorities, especially those of least-developed country participants', the Working Group on Transparency in Government Procurement discussed extensively the development implications of a possible agreement in this area. The Doha mandate also recognized the need for enhanced technical assistance and capacity-building and contained a commitment to provide such assistance both during any negotiations and after their conclusion.
No agreement on modalities for negotiations could be reached at the fifth ministerial conference, held in Cancun in September 2003. On 1 August 2004, the WTO General Council adopted a decision that removed this subject, as well as competition and investment, from the Doha Work Programme. The decision did not indicate what might occur, if anything, following the completion of the Doha Round. It is difficult to understand why there was resistance to negotiating an agreement on transparency in procurement. The most compelling explanation is that many developing countries were not convinced that this would not end in a future discussion on market access. From a systemic perspective that is indeed what would be logical—given the size of procurement markets this is clearly an area where there is a rationale for WTO members to agree to mutual disarmament. In the absence of market access incentives and disciplines it is also not obvious what the rationale is for discussing transparency in the WTO. Some of the economic dimensions of enhancing transparency of procurement processes are discussed in Annex 2. It is shown that there is not necessarily a one-to-one mapping between more transparency and more market access.
11.2. The Civil Aircraft Agreement
The Agreement on Trade in Civil Aircraft aims to reduce both tariffs and trade-distorting NTMs affecting production and trade in civil aircraft. It was the only sector-specific agreement covering a manufactured product that was successfully negotiated in the Tokyo Round. Members to the agreement include most leading civil aircraft exporters (the Russian Federation, not a WTO member, is an exception). A Committee on Civil Aircraft oversees the agreement.
plurilateral agreements
527
The agreement is to a large extent a zero-for-zero tariff agreement, as signatories agree to eliminate import duties on civil aircraft and the bulk of aircraft parts. The agreement also includes disciplines on TBT and subsidies. A major reason it was negotiated was because little progress could be made during the Tokyo Round on subsidies. The demandeur for the agreement was the US, which sought to constrain the ability of European governments to subsidize Airbus, as it was increasingly becoming a competitor for US companies producing large civil aircraft (Boeing, Lockheed, McDonnell Douglas). US efforts to ban the use of subsidies failed, however, and no binding disciplines were introduced that went beyond then-prevailing rules on subsidies—which were quite weak (see Chapter 5).
Disputes regarding trade in civil aircraft have been a recurring element of trade relations between the US and the EU, reflecting the battle for market share fought between Boeing and McDonnell Douglas on the one hand (now merged), and Airbus on the other. The Civil Aircraft Agreement has not succeeded in reducing the sources of tension between these two dominant players, nor has it been effective in addressing conflicts between smaller aircraft producing nations such as Canada and Brazil. The latter became embroiled in disputes regarding the alleged use of export subsidies in the late 1990s, but the Agreement on Civil Aircraft did not play a role in the various panel cases, which are discussed in Chapter 3.5
The EU and the US instead pursued a bilateral track. In the mid- to late 1980s tensions flared up between the US and the EU following a decision by Air India to cancel an order for Boeing 757 jets after Airbus offered big discounts on its new A320 plane. This led to the negotiation of a bilateral agreement between the two parties that incorporated more specific disciplines. The 1992 EU-US Agreement on Trade in Large Civil Aircraft (denned as planes with a capacity of 100 seats or more) required that the parties provide each other with data on financing of new aircraft; banned new production subsidies outright; limited the amount of so-called launch aid (support for the development of a new type of aircraft) to one-third of total development costs; and constrained indirect support (such as R&D funding) to 4 per cent of any recipient firm's annual turnover and 3 per cent for the industry as a whole. Moreover, launch aid must be fully repaid within 17 years with an interest rate that reflects the government's costs of funds.
These disciplines, although still permitting a significant amount of support to be given to airplane producers, did bite: Airbus had received assistance that exceeded the limits laid out in the 1992 agreement. Irwin and Pavcnik (2004) estimate that implementation of the agreement increased global aircraft prices by some 4 per cent, and the marginal costs of production by 5 per cent. But the agreement did not result in a cessation of the commercial rivalry between the EU and the US. Matters
5 Canada—Measures Affecting the Export of Civilian Aircraft, complaint by Brazil (WT/DS70, WT/DS71); Brazil—Export Financing Programme for Aircraft, complaint by Canada (WT/DS46).
528       the political economy of the world trading system
came to a head in 2004 as a result of US objections to what it regarded as excessive launch aid and other support that the EU agreed to provide Airbus for the development of the A380, the double-decker jumbo jet. The US withdrew from the agreement and initiated DS proceedings in the WTO. The EU immediately retaliated by bringing a case of its own, alleging that Boeing was benefitting from large-scale subsidies in the form of investment incentives (including tax concessions and infrastructure) provided by state-level governments in the US, as well as prohibited export subsidies and excessive R&D financing (much of which was for military contracts but benefitted the civilian production lines). The US invoked the SCM Agreement, arguing serious prejudice and adverse effects (loss of markets, price suppression, etc.). The EU argued before the panel that the Civil Aircraft Agreement and the subsequent bilateral extension should be applied, and that many of the measures the US had challenged were not subsidies as defined in the SCM Agreement (because they were repaid by Airbus).
An interesting feature of this case is that until 2004 neither party had invoked WTO mechanisms, not just DS but also CVDs. One reason for the absence of CVDs was that the US industry did not want to bring cases forward for fear of retaliation. As in other areas this is one where the changing structure of the global production process has greatly affected the incentives confronting both the aircraft producers and their governments. An increasing share of the components that make up an airplane are produced outside the EU and US respectively. Airbus sources from US suppliers and Boeing from EU-based firms. Both source from the rest of the world.
How the dispute will be resolved remains to be determined. Given the long standing pattern of recurring intervention by both parties in the industry, the linkages to national security and the high-tech nature of aircraft production, the very large fixed costs of aircraft development and the effective duopoly nature of the global industry, it is very unlikely that any WTO ruling will be implemented. It is also clear that neither party will be found to be in compliance with WTO rules. What will be required is a new agreement that imposes stricter disciplines but recognizes that governments will continue to support the industry. Huf-bauer (2007) recommends re-negotiating the plurilateral agreement, building on the OECD experience with disciplining export credit subsidies. This could involve a mix of minimum standards for subsidies, combined with a 'peace clause' (agreement not to file cases—as was done in the Uruguay Round Agreement on Agriculture for subsidy disputes for a period of time—see Chapter 6) and effective notification, monitoring and surveillance by a WTO body.
Economists differ in their view of whether the EU decision to support the entry of Airbus was beneficial for the EU or for the world as a whole. Early analysts emphasized that there was most likely a good case to be made for the EU policy as it increased competition on a highly concentrated market—in effect the US had a
plurilateral agreements 529
monopoly on wide-body civil aircraft. But the increased (subsidized) competition also led to substantially greater concentration of the US industry, with Lockheed-Martin exiting the sector and McDonnell Douglas taken over by Boeing. Although a textbook example of 'strategic trade policy,' whether or not the EU subsidy policies have improved welfare remains an open question.
11.3. Towards More Clubs in the WTO?
Plurilateral agreements are outliers in the WTO system. Although there are presently only two such agreements, only one of which has proven to be robust and fully effective, the fact that the WTO offers the flexibility of negotiating such agreements may prove important in the future. One of the benefits of the Tokyo Round 'codes approach' was that it allowed for a 'variable geometry'. The downside of the codes was that they addressed GATT issues and it was not completely clear whether the MFN rule applied to them. This was important because the codes dealt with matters that were subject to the GATT—in effect they were used to overcome the immense difficulty that existed in obtaining the required agreement to revise existing disciplines (amend the GATT). There also was no unified dispute settlement system, creating the potential for forum shopping and the development of diverging case law.
With the creation of the WTO and the DSU there is, in principle, greater scope for subsets of WTO members to use plurilateral agreements to move forward on specific topics. Such agreements are a vehicle for like-minded countries to cooperate in areas not (yet) addressed by the WTO. They allow countries not willing to consider disciplines in a policy area to opt out. Given that the Doha Round clearly revealed that it may not be possible to get consensus on launching a negotiation on a subject, let alone conclude the negotiation successfully, the plurilateral route offers a mechanism to introduce areas into the WTO without the disciplines applying to all members. Accommodating diversity in interests through greater use of plurilaterals was one of the recommendations of the Sutherland Report (Sutherland et al., 2004). Fears that a move down this road would result in a potential repeat of the Uruguay Round TRIPS experience—where negotiators started with a limited agenda centring on trade in counterfeit goods but ended up with an agreement that harmonized elements of domestic intellectual property legislation—has been a factor underlying resistance by many developing countries to this idea. Their concern is that accepting a plurilateral agreement to be brought into the WTO sets a precedent that they will be confronted with subsequently. However, the GPA illustrates that there is no presumption that nonsignatories will
530
the political economy of the world trading system
be 'forced' to sign a plurilateral deal as time goes by—the GPA has been in existence for decades and membership remains very limited.
Lawrence (2006) discusses what he calls the club-of-clubs option and argues that this approach can help the WTO address the diverging interests of its members in an efficient way. He suggests a number of criteria, including that: clubs be restricted to subjects that are clearly trade-related, any new agreement is open to all members in the negotiation stage, i.e. participation in the development of rules should not be limited to likely signatories; and that club members be required to use the DSU to settle disputes, with eventual retaliation being restricted to the area covered by the agreement (as is the case under the GPA). Although greater use of plurilateral agreements will result in a multi-tier system with differentiated commitments and some erosion of the MFN principle—as club members would have the right to restrict benefits to other members—there is already significant differentiation in the level of obligations across countries.
11.4. Further Reading
Little has been written on how plurilateral agreements fit into the WTO framework. The trade policy-oriented literature on government procurement and civil aircraft is also relatively sparse. Many of the contributions in Bernard Hoelcman and Petros C. Mavroidis (eds), Law and Policy in Public Purchasing: The WTO Agreement on Government Procurement (Ann Arbor: University of Michigan Press, 1997) discuss the genesis, operation and relevance of the GPA to countries at differing levels of development. Contributions also cover the UNICITRAL model law for procurement and procurement regimes of non-GPA members such as India and New Zealand. Sue Arrowsmith, Government Procurement in the WTO (London: Kluwer Law International, 2003) provides a comprehensive discussion of the GPA from a mostly legal perspective.
A classic study of the economics of discrimination in procurement is Robert Baldwin and J. David Richardson, 'Government Purchasing Policies, Other NTBs, and the International Monetary Crisis', in H. English and K. Hay (eds), Obstacles to Trade in the Pacific Area (Ottawa: Carleton School of International Affairs, 1972). R. Preston McAfee and John McMillan, 'Government Procurement and International Trade', Journal of International Economics, 26 (1989): 291-308, make the theoretical case for discrimination in markets characterized by imperfect competition and small numbers of bidders. Albert Breton and Pierre Salmon review the literature and question some of the conventional wisdom regarding the rationales for procurement policies in 'Are Discriminatory Procurement Policies Motivated By Protectionism?', Kyklos, 49 (1995): 47-68.
plurilateral agreements
531
The role of the Civil Aircraft Agreement and the rivalry between the EU and US is discussed in Steve McGuire, Airbus Industrie (London: McMillan, 1997) and Nina Pavcnik, 'Trade Disputes in the Commercial Aircraft Industry,' The World Economy (2002): 733-51.
Robert Lawrence, 'Rulemaking Amidst Growing Diversity: A Club of Clubs Approach to WTO Reform and New Issue Selection,' Journal of International Economic Law (2006) suggests that greater use be made of plurilateral agreements in the WTO.