CHAPTER 6
SECTOR-SPECIFIC MULTILATERAL TRADE AGREEMENTS
The GATT was conceived as a general agreement that would apply to all merchandise trade. In principle, therefore, GATT rules on nondiscrimination, transparency, tariff bindings and so forth, apply to all sectors. In practice, however, industry-specific pressures for protection in major trading nations created strong incentives for governments to grant 'special' treatment to 'special' sectors. This chapter focuses on two key sectors where pressures for protection proved too strong for the trading system to handle—agriculture, and textiles and clothing. Over time these sectors were gradually removed from the reach of many GATT 1947 disciplines. It was only during the Uruguay Round that it proved possible to reintegrate them into the trading system. However, progress on agriculture since then proved very difficult to achieve.
This chapter also discusses the Information Technology Agreement (ITA). Although negotiated among a subset of WTO members during 1996, the ITA is applied on a MFN basis. The ITA is noteworthy in that it is sometimes regarded as proof that single sector agreements are feasible in the WTO context, and it illustrates that it may not be necessary to engage in broader negotiating rounds and cross-issue linkage to achieve liberalization of trade. We argue below that the ITA experience does not support this line of reasoning.
270
the political economy of the world trading system
6.1. Agriculture
Poor agrarian economies frequently tax agriculture relative to other tradable sectors. As nations become richer, their policy regimes change from taxing to assisting farmers (Lindert, 1991). The post-1950 period saw substantial growth in agricultural protection and insulation in the advanced industrial economies and its spread to newly industrializing economies (Johnson, 1973). That tendency accelerated in the 1980s to the point where some protectionist countries went beyond self-sufficiency to generate surpluses. These could only be disposed of with the help of export subsidies. This led to serious budgetary pressures and increasing opposition to the cost of agricultural support policies. It also led traditional agricultural-exporting countries to insist that MTNs focus on reducing agricultural protection.
Historically, agricultural trade policy has tended to be been driven by short run phenomena. Food crises led to export controls, whereas gluts led to import barriers. Protectionist measures in agriculture of a permanent nature became increasingly prevalent at the end of the nineteenth century (Findlay and O'Rourke, 2007: 396 ff) and have been a persistent feature of global trade policy ever since. One cause of the protectionist policies that emerged in the late nineteenth century was the steady expansion of American production and the resulting fall in world prices. Although some nations reacted to the resulting change in incentives by adjusting—for example, the Netherlands became more specialized in livestock as the price of feed grains fell—others, including France, Germany and Austria-Hungary, reacted by protecting existing producers and subsidizing exports if domestic output exceeded consumption. Between the First and Second World Wars agricultural protection and domestic market regulation increased further. After the Second World War, agricultural ministries in OECD countries exempted agriculture from key GATT disciplines and recurrent MTNs. The US led the way with its request for a waiver in 1955. With the creation of the EEC and its Common Agricultural Policy (CAP) in 1957, European countries also insisted on special treatment for this sector.
The CAP is the pre-eminent example of how farmers can be insulated from foreign competition. It provided for an intervention or support price at which the Community guaranteed to purchase the agricultural output from farmers, and a threshold price (above the internal support price) below which no imports were allowed. In order to isolate the EU market from international competition, a variable levy equal to the margin between the threshold price and the lowest representative offer price on world markets was imposed on imports. Moreover, a restitution amounting to the difference between the average world price and the internal EU price was granted to European exporters. The programme was extremely costly. Domestic support to agricultural producers averaged more than US$92 billion per year in the EU during 1986-90. However, the EU was not alone.
sector-specific multilateral trade agreements
271
Domestic support in Japan and the US over the same period averaged US$35 and US$24 billion respectively. In Japan, rice was produced at a cost four times that of competitive producers elsewhere. The same applied to Swiss meat and butter. Budgetary support in the US, the EU and Japan accounted for some 15 per cent of government spending; a figure comparable to what was spent on education. In the post-1990 period the EU, Japan, the US and other OECD nations continued to support farmers, collectively transferring some US$300 billion a year to farmers. This was done through a variety of instruments, but mostly through border protection.
Patterns of intervention in agriculture
Why do governments intervene in agriculture? And why are the policies that are observed so different across countries? Rationales for intervention include: (1) to stabilize and Increase farm incomes; (2) to guarantee food security; (3) to improve the balance of payments; (4) to support the development of other sectors of the economy; and (5) to increase agricultural output (Fitchett, 1987). These reasons are in part noneconomic or driven by special interest politics. The political influence of the agricultural sector is substantial in many countries. Agriculture produces food, and food is often very political. President Nyerere of Tanzania once said that if he needed shoes and apartheid South Africa was the only place to get them he would go shoeless, but if he needed corn and the only source available was apartheid South Africa he would buy there. Food shortages can lead to riots, revolutions and wars.
There is a striking difference between the way agriculture is treated in poor and rich countries. In many developing countries, policies tax agriculture and subsidize food consumption of the urban population. In industrialized countries exactly the opposite pattern can be observed: an urban population that is taxed to support farm production and incomes. But in both cases, governments use subsidies, trade barriers, state trading and public purchasing to regulate production and trade (Box 6.1).
Relative to other sectors of the economy, as of the early 1990s agriculture in many industrialized countries was regulated, subsidized and insulated from market forces to an exceptional degree. Production quotas, state purchasing and distribution, subsidies and administered pricing often worked at cross-purposes. In the EU, support programmes were so effective in stimulating output that they had to be complemented by production quotas and incentives to take land out of production (so-called set asides). Matters were not much better in many developing countries. Marketing boards—monopoly buyers and distributors of food—were often established that set prices for farm products below world market levels in a deliberate attempt to lower the cost of subsidizing the prices of basic foodstuffs for the urban
272
the political economy of the world trading system
Box 6.1. Why poor countries tax and rich countries subsidize agriculture
Average rates of protection for industries tend to decline across countries as capital-labour ratios increase. Industrialized countries with large capital stocks—both physical plant and equipment and human capital—relative to labour are more open to trade than countries with large stocks of labour relative to capital (mostly developing countries) (Rodrik, 1995). However, rich countries tend to be much more protectionist towards agriculture (supporting domestic production and closing off markets against import competition). In contrast, poor countries tend to promote imports, either explicitly through import subsidies, or implicitly by taxing domestic production.
Anderson and Hayami (1986) argue that this can be explained as follows. In a poor country, food accounts for a large share of total household consumption, whereas in rich countries food accounts for only a small share of expenditure. Moreover, agriculture is the main source of employment in a poor country, whereas it typically accounts for less than 5 per cent of the labour force in a rich one. In poor countries agriculture is also much less capital-intensive than in rich ones. These stylized facts do much to explain the different policy stances that are observed. If agriculture is protected in a poor nation, the resulting increases in food prices have a large impact on the demand for labour (given the size of the agricultural sector) and thus on economy-wide wages (because labour is mobile). The wage rise will be offset to a greater or lesser extent by the rise in food prices, food being so important in consumption. At the same time the wage increase puts upward pressure on the price of nontradables (services) and has a negative impact on industry by lowering profits. As the gains per farmer of protection are low, and the loss per industrialist is high, the latter will be induced to invest resources to oppose agricultural support policies. Supporting agricultural production in a poor country therefore may not make political sense. The converse applies to rich nations, where agricultural support has much less of an impact on' wages (the sector being small), on the prices of nontradables and on industrial profits.
A simulation model developed by Anderson (1995) that incorporates these basic differences between poor and rich countries reveals that a 10 per cent rise in the relative price of manufactures in a poor nation (that is, a tax on agriculture) will reduce farm incomes by only 2 per cent, whereas raising those of industrialists by 45 per cent. In contrast, a 10 per cent tax on industry in a rich country (that is, a policy of supporting agriculture) raises incomes of farmers by over 20 per cent, but reduces those of industrialists by only 3 per cent. These differences in costs and benefits for different groups in society—in conjunction with the differences in sizes of the various groups— help explain why farmers in rich countries are willing to invest substantial resources to obtain and maintain protection, and why industrialists and urban populations in developing countries are able to benefit at the expense of farmers.
Honma (1993) empirically investigates whether agricultural protection is determined according to the Anderson-Hayami (1986) framework of endogenous protection. Using panel data on 14 industrial countries between 1955 and 1987, Honma finds that the nominal rate of protection: (1) declines the higher the ratio of labour productivity in : agriculture to that in industry; (2) rises as the share in agriculture increases to 4.5 per cent and falls beyond thereafter; and (3) increases as the terms of trade of agricultural products (relative to manufactured goods) decline. He also concludes that the EU is an outlier, with higher levels of support than similar countries. One reason for this may be the role that the CAP played in the formation of the EEC and the role it has had in sustaining European cooperation (Messerlin, 2001).
sector-specific multilateral trade agreements
273
population or to tax tradable commodities and generate revenue for the government. The result was often a drop in agricultural output, migration to the cities and rising imports of food. This pattern was complemented by the effect of OECD countries' food aid, which further reduced the incentive to adopt a more economically rational agricultural policy.
As a consequence of agricultural intervention, countries with no comparative advantage in agriculture not only became major producers but also net exporters. Production support policies had to be complemented by export subsidies to allow surpluses to be sold. These in turn led to numerous trade conflicts. As farm surpluses were dumped at subsidized prices in international markets, agricultural trade increasingly became managed trade. During the 1986-90 period, OECD economies annually subsidized exports averaging 48.2 million tons of wheat, 19.5 million tons of coarse grains, 1.8 million tons of sugar, 1.2 million tons of beef and 1.2 million tons of cheese and butter. Average annual export subsidies in the EU during 1986-90 were more than US$13 billion, with most of the money allocated to exports of bovine meat, wheat and coarse grains, butter and other milk products (GATT, 1994c). The loss of developing country export revenue resulting from agricultural protectionism in the US, the EU and Japan was significant. For sugar and beef alone, it was estimated to be the equivalent of about half of total international development aid (World Bank, 1986). Policies in developing countries tended to make the situation worse for farmers as they often discouraged farm production through a variety of agricultural and nonagricultural policies. The former included state control of inputs and prices, the latter included high import barriers for manufactures and overvalued exchange rates, both of which reduced the incentive to invest in agriculture (Krueger, Schiff and Valdes, 1988).
A 2008 World Bank research project has generated annual time series estimates of rates and values of assistance/taxation over the past half century for around 75 countries—together accounting for 90 per cent of global population, GDP and agricultural production. For each country, nominal rates of assistance (NRAs)1 are calculated for key products (Anderson, 2008). The growth of agricultural production support in high-income countries began to reverse in the 1990s, although if 'decoupled' income support for farmers is included, the rate of support has not declined substantially since the Uruguay Round was completed (Figure 6.1). Until recently developing country governments effectively taxed their farmers, imposing an effective tax rate on the order of 20 per cent from the mid-1950s to the mid-1980s. Since then it has diminished, and, on average, become slightly positive.
The US dollar value of the gross subsidy equivalents of the NRAs are shown in Table 6.1. These estimates suggest that, from the mid-1950s through to the
1 The NRA includes the effects of both farm output and farm input price distortions and is expressed as a percentage of total farm production valued at undistorted prices.
274
the political economy of the world trading system
Fig. 6.1. Nominal Rate of Assistance to farmers, high-income and developing countries, 1955 to 2004 (per cent)
Note: Averaged using weights based on the gross value of agricultural production at undistorted prices. High income
includes Republic of Korea and Taiwan, Province of China.
Source: Anderson and Valenzuela (2008).
mid-1970s, assistance to farmers in high-income countries almost exactly offset taxation of farmers in developing countries. Until the late 1980s, farmers in developing countries were at a double disadvantage in terms of competitiveness: farmers in OECD countries benefitted from significant levels of support, whereas agricultural production in developing countries tended to be taxed. Since the early 1980s, the gradual decline in taxation of farmers in developing countries and the growth in
Table 6.1. Gross subsidy equivalents of assistance to farmers by region, 1955-2004 [current US$ billion per year)
Country Groups	19G0-4	1965-9	1970-4	1975-9	1980-4	1985-9	1990-4	1995-9	2000-
High-income	30.3	43.1	49,0	96.4	133.0	174.1	216.3	200.2	190.9
Developing	-18.7	-21.9	-46.7	-69.0	-92.0	-35.8	0.8	47,9	65.2
Africa	-0.6	-1.3	-3.3	-5.9	-4.0	1.5	-6.3	-6.0	-7.9
Asia	— 17.5	-19.0	-37.8	-55,1	-70.8	-28.9	. -1.8	27.1	48j0
Latin America	-0.3	-0.7	-4.9	-6.7	-10.4	-9.3	4.7	6.5	5.3
European Transition	-0.2	-0.9	■  -0.6 '..	- 1.3	-6.8	0.8	4.2	20.3	19.7
Note: 'High-income countries' is a subset of OECD countries (Western Europe, Japan, United States, Canada, Australia, New Zealand.and Republic of Korea) and Taiwan, Province of China for the periods after 1995.
Source: Anderson and Valenzuela (2008).
sector-specific multilateral trade agreements
275
assistance to high-income country farmers have combined to see the net global transfer to farmers increase to more than US$250 billion per year. Regionally, outside the high-income group, it is Asia where the payments are largest on aggregate. On a per farmer basis, however, they are now largest in Europe's transition economies. In Africa, meanwhile, farmers still confront discrimination relative to other forms of economic activity.
The data on gross subsidy equivalents of support to farmers show that the overall level of assistance in high-income countries has been virtually constant for the last 15 years, and has been rising in developing countries, perhaps in part as a response to the example set by high-income nations. The high level of production support in high-income countries distorts domestic and world market prices and is detrimental to producers in developing countries and consumers in the high-income countries themselves. Historical policies in many developing countries of taxing agriculture have also been detrimental to farmers, and the more recent trend towards a more neutral policy stance for agriculture relative to other sectors of activity is a positive development from an economic policy perspective.
An important question confronting voters and exporters in foreign countries is how best to try and change the political equilibrium in countries that protect agriculture. The work by Anderson and Hayami mentioned earlier suggests that the observed patterns of intervention can be explained by economic 'fundamentals' such as the share of agriculture in output, per capita incomes, etc. But these factors simply help to understand the forces that lead to agricultural taxation or support over time in the absence of countervailing pressures. Analysis of the political economy forces that generate a given set of policies can help policymakers and other groups seeking to change a given equilibrium. An interesting finding of the empirical literature on this subject is that matters may not be as bad as they could be, in that governments appear to put a much higher value on economic welfare than on assisting farm groups (that is, on responding to political contributions made by such groups) (Box 6.2).
From GATT 1947 to the WTO
The rules applying to agricultural trade under GATT 1947 were weaker than those for manufactured goods because many nations regarded agriculture as a sector of economic activity that deserves special treatment. This attitude manifested itself during the post-war negotiations on the ITO in US insistence that the ITO not affect its agricultural policies. Although the ITO was never ratified, GATT rules on agriculture were in part written to fit existing US agricultural policies. Disciplines for agriculture differed in two major respects from those on trade in manufactures. First, quotas were allowed for agricultural commodities if concurrent measures were taken to restrict domestic production or used to remove a temporary
276       the political economy of the world trading system
Box 6.2. Empirical political economy research on agricultural protection
Empirical research on agricultural protection is plentiful (see, for example, the survey by de Gorter and Swinnen, 2002). The literature tends to find that support increases under adverse market conditions for the farming industry; that countries with a comparative disadvantage in agriculture have greater protection (supporting the view that losers from liberalization organize politically), and that a high budget share for food consumption reduces protection. Gardner's (1987) is a classic study in this literature. Gardner assumes governments maximize the weighted sum of consumer surplus (C) and producer rents (R): W= C + 0R. In turn, C and R are functions of farm output quantities (i.e. policy therefore targets production). Unlike Grossman and Helpman (1994) who provide microfoundations for their objective function by including a formal model of the lobbying process, Gardner attributes the value of the parameter 0 to the forces that determine lobbying effectiveness. These were measured for 17 farm commodities by the number of producers, their geographical dispersion, the return to agricultural support (determined by output per farm) and the stability of the industry (variability of production patterns) for the period 1912-80. Gardner found that the lower the demand elasticity of a commodity, the greater the level of intervention. This is consistent with the intuition that it is most efficient to tax commodities with the lowest price elasticities of demand, and is the basis for Gardner's conclusion that interventions in US agriculture have been relatively efficient.
Gawande and Hoekman (2006) apply the Grossman-Helpman model—discussed in . Chapter 1—to agricultural protection in the US. As is the case for applications of this model to trade in manufactures, their estimates of the government preference (or weight) on welfare is very high: government places at least 40 times as much weight on a dollar of welfare as on a dollar of political contributions (lobbying) by farm interests. That is, the US government is approximately a welfare-maximizer.
These results suggest that things are not as bad as they could be—or as bad as economic models of endogenous protection predict they should be. However, the results are difficult to square with the fact that the deadweight losses of agricultural protection are many billions of dollars. One way of reconciling the estimates with political reality is to recognize that government is not a singular entity with the power to supply protection with certainty. In practice the legislative process leading to protection is uncertain. Contributions are made before the award of protection or subsidy, not on delivery of protection or subsidy. Lobbies also make contributions to political agents who cannot guarantee a specific outcome. These real-world dimensions should reduce lobbying contributions, reflecting the much lower probability of success. A unitary government view of the world leads to the conclusion that the low tariffs are due to a welfare-loving government, when in reality it is a self-interested government that delivers tariffs in response to (lower) contributions under uncertainty
domestic surplus (Article XIGATT). Quantitative restrictions could also be used to deal with shortages of food or other essential exportables. Second, export subsidies on primary products were permitted, as long as these did not lead to more than 'an equitable share' of world trade for the subsidizing country. What equitable meant in practice was not clear, however, and differences in interpretation led to a number
sector-specific multilateral trade agreements
277
of GATT disputes. Over time, more flexibility in the use of QRs and other NTBs in the agricultural sector was introduced through special waivers (starting with the US in 1955); in protocols of accession (for example, Switzerland); through limited tariff bindings on agricultural imports (opening the way for the use of variable levies under the CAP); by allowing residual grandfathered restrictions on imports of agricultural goods to be maintained; and through a proliferation of various 'grey area measures' such as VERs and 'orderly marketing arrangements'.
By accepting—implicitly or explicitly—the notion that agriculture is unique, it proved to be virtually impossible to make cross-sectoral linkages or tradeoffs in MTNs. The establishment of separate negotiating groups for agriculture, staffed 'by civil servants experienced in the defense of domestic farm-support policies... [was] a way of avoiding a tradeoff between agriculture and industry' (Josling, 1977: 11). Problems were compounded by the commodity-specific approach that was pursued in MTN talks on agriculture.
Throughout the 1960s and 1970s, agricultural discussions between the two major players—the EU and the US—were based on two totally different conceptions. The EU favoured the development of a system to manage world trade so as to facilitate the functioning of the CAP. The US, in contrast, supported by countries such as Australia, Canada and New Zealand, sought significant liberalization, while safeguarding national policies that protected powerful agricultural interests (such as dairy and sugar producers). As a result, very little progress was made in the Kennedy and Tokyo Rounds on agriculture. In both MTNs, the basic premise of the EU was that the CAP was non-negotiable, and that the focus of discussions should be on stabilizing world agricultural markets. The Community proposed that international commodity agreements be negotiated for products such as cereals, rice, sugar and dairy. The US in contrast emphasized the need to expand agricultural trade and to end the special status of agriculture in the GATT (which was somewhat ironic given that the US had started the process of hollowing out multilateral disciplines for agriculture).
In the Tokyo Round these incompatible positions blocked the negotiations for a long time. The deadlock was broken only after the Carter Administration was inaugurated and Robert Strauss was appointed as Special Trade Representative (in 1977). President Carter put greater weight on the successful conclusion of the round than his predecessor and was willing to give in on agriculture. Subsequent bilateral bid-offer negotiations resulted in the reduction of certain tariffs and an increase in various quotas, but did little to achieve general US objectives.
Two sectoral agreements were negotiated: an Agreement on Bovine Meat and an International Dairy Arrangement. Neither was far-reaching. The agreement on meat ostensibly was aimed at increasing trade and the stability of the world market. It implied no binding obligations, however, and in practice had little effect, if any. The dairy agreement was more substantive in that it set minimum prices for major dairy products. However, these prices proved to be unenforceable in practice.
278       the political economy op the world trading system
A number of disputes occurred regarding circumvention of the minimum prices by certain signatories (especially the EU), which led the United States to withdraw from the agreement in February 1985. Both arrangements were brought into the WTO as Plurilateral Agreements in 1994, but were dissolved in 1999 (see Chapter 11).
In the early 1980s a constituency emerged in the EU that favoured a reduction in agricultural support. Agricultural subsidies were a significant burden for heavily strained government treasuries, and became increasingly difficult to defend as the ideological balance swung towards greater reliance on markets, competition and deregulation. Two successive oil shocks had led to large fiscal deficits, compounding the pressure on government finances. A decision by the US to engage in a subsidy war with the EU in the 1980s—partly driven by a decline in international food prices, which raised opposition to EU export subsidization—also helped to increase the financial pressure. At the same time, agricultural disputes became more intense, and further enhanced the incentive for dealing with agriculture in the GATT.
These factors allowed the ministerial meeting that launched the Uruguay Round to put agriculture on the table in a comprehensive manner for the first time. The Punta del Este negotiating mandate broke new ground in that there was an explicit reference to liberalization, with all policies affecting agricultural trade to be discussed, including domestic and export subsidies. This contrasted with the Kennedy and Tokyo Round ministerial declarations, which emphasized the status of agriculture as a special (unique) sector and were oriented towards the negotiation of commodity-specific agreements. However, as the negotiations commenced, it rapidly became clear that discussions would continue to be dominated by transatlantic ping pong between the two largest agricultural traders—the EU and the US—which together accounted for about 40 per cent of international trade in food. Any agreement required a deal that they could live with. But they were by no means the only players. Other significant actors included the European Free Trade Association (EFTA) countries and Japan (with highly protectionist systems and basically in the EU camp) and a group of 14 agricultural exporters that sought significant liberalization. This coalition was called the Cairns Group (after the Australian city where the group was formed) and was an ally of the US. It included Argentina, Australia, Brazil, Canada, Chile, Colombia, Fiji, Hungary, Indonesia, Malaysia, New Zealand, the Philippines, Thailand and Uruguay.
The Cairns Group objective was to graduaDy attain free trade in agricultural commodities, eliminate production distortions, and ensure that binding undertakings to this effect were made. The US initially sought the complete liberalization of trade in agriculture. It was particularly concerned about export subsidies, and sought their rapid and unconditional elimination. The US also insisted on the need to introduce a clear-cut separation between income support for agricultural producers and policies that affected the level of farm production. Income support
sector-specific multilateral trade agreements
279
could be accepted, but only if decoupled from production. The EU initially proposed that negotiations first concentrate on emergency measures for certain sectors, including cereals, sugar and dairy products, to remove structural disequi-libria on world markets, followed by liberalization of trade and a reduction of support policies. The EU argued that the goal should not be free trade, but achieving stability on world agricultural markets. It proposed to follow a bid-offer process for specific products along the lines of previous MTNs. It also argued that existing zero (or low) tariff bindings on oilseeds led to severe distortions in the EU market and sought to negotiate a 'rebalancing' of its agricultural protection to make it more uniform. This desire for rebalancing became one of the more contentious issues of the negotiations.2 Japan supported the idea of a freeze on export subsidy expenditures as a short run step, to be followed by a gradual phase-out, but suggested that domestic subsidies be permitted to maintain a minimum (unspecified) level of self-sufficiency for national security reasons.
Bridging the gap between the EU and the US-Cairns positions proved extremely difficult, not only because of fundamental, substantive differences, but also because of the negotiating strategies that were pursued. Although clearly unacceptable to the EU, for the first two years of the Uruguay Round the US insisted the objective should be the total elimination of trade-distorting support policies within ten years. The resulting standoff led to the breakdown of the Montreal mid-term review of the round in December 1988. After a four-month period of informal consultations it was agreed that the long run objective in the agricultural area was to be progressive reduction in agricultural support, not elimination. This compromise allowed negotiations to continue. In the final phase of the round, discussions remained very contentious, with serious differences of opinion emerging within the EU as well as between the EU and other GATT contracting parties.
At the December 1990 ministerial meeting that was supposed to conclude the round, no agreement could be achieved on agriculture, leading to a breakdown of talks on all the issues on the agenda. In effect, the EU refused to accept the compromise text that was proposed by the chairman of the negotiating group— which would have averaged a cut of about 25 per cent in bound protection levels— as going too far in disciplining export subsidies and the use of specific policies. The proposal would have had significant implications for the CAP—the reform of which was under active discussion at the time. The EU needed to settle its internal debates on agriculture first—in particular to placate the French, who opposed any
2 In earlier MTNs predating the formation of the European Community, a number of European countries bound tariffs on cotton, soybeans (oil, meal and seeds), vegetables and canned fruit at low or zero levels. When these countries joined the EEC, these bindings were incorporated into the common external tariff of the Community. As the CAP led to higher prices of grains, European producers began to import large quantities of soybeans and related products, on which tariff bindings were low. This was a major source of irritation for the EU Commission, which unsuccessfully attempted to close this 'gap in the CAP' in subsequent years.
280      the political economy of the world trading system
significant move towards meeting US-Cairns Group demands. Latin American members of the Cairns Group played a major role in opposing any significant weakening of the chairman's proposed text. Argentina, supported by Brazil, made it clear that they would refuse to accept the proposed deal, and stood ready to scuttle the Uruguay Round over the issue (Ricupero, 1998).
An agreement between the EU and the US was eventually reached, after much brinkmanship, with the so-called Blair House Accord in November 1992. By that time internal CAP reform proposals had been developed by the European Commission, allowing a deal to be struck. The EU obtained agreement that its compensation payment policies—under which farmers were paid to take land out of production—would not be included in the definition of the Aggregate Measure of Support (discussed below). It was also agreed that this measure would not be product-specific, and that the extent of liberalization would be limited to a cut of about one-sixth over six years, or less than 3 per cent per year. Although French farmers in particular continued to oppose the deal, the Commission contained this by arguing that the agreement did not go beyond the internally agreed reform of the CAP.
The Uruguay Round Agreement on Agriculture
The Agreement on Agriculture (AoA) that emerged from the Uruguay Round has four main parts dealing with export competition, market access, domestic support and SPS measures (the latter is discussed in Chapter 5).
Export competition. All existing export subsidies had to be scheduled and bound. No new export subsidies were permitted, i.e. any subsidies not scheduled became illegal. By 2000 scheduled export subsidies were to be reduced by 36 per cent in value terms and 21 per cent in volume terms, relative to a 1986-90 base period, in both cases on a commodity-by-commodity basis. For some commodities only the agreed 21 per cent cut in the volume of subsidized exports was actually achieved, because international food prices in the late 1990s were higher than in the late 1980s, so that exportable surpluses could be disposed of with lower subsidy outlays.
Market access. On market access it was agreed that NTBs would immediately be converted into tariffs and that industrial countries reduce these tariffs by an average of 36 per cent over six years (24 per cent for developing countries). All agricultural tariffs were bound, an advance over the situation applying to other merchandise tariff lines (see Chapter 5). In practice the cut in tariff bindings could be less than one-sixth as a weighted average, as each tariff item needed to be reduced by only 15 per cent of the claimed 1986-8 tariff equivalents (10 per cent for developing countries). There was also considerable scope to concentrate tariff reductions in commodity groups with relatively little effect on trade (Josling, 1994).
sector-specific multilateral trade agreements
28l
The tariff bindings that were implemented by WTO members were in many cases far higher than the actual tariff equivalents of NTBs that applied in the 1986-8 base period. The EU, for example, set bindings about 60 per cent above the actual tariff equivalents of the CAP in the late 1980s, whereas the US set bindings about 45 per cent higher (Ingco, 1996). Many developing countries chose to bind their tariffs on agricultural imports at more than 50 per cent and some as high as 150 per cent— far above the tariff equivalents of restrictions actually in place in the early 1990s. This 'dirty' tariffication implied that actual tariffs at the beginning of the twenty-first century provided no less protection than did the NTBs of the late 1980s.
The so-called binding overhang that resulted was significant. Binding tariffs at such high levels allowed countries to set the actual tariff below the ceiling but to vary it so as to stabilize the domestic market, analogous to the earlier EU system of variable import levies and export subsidies. Such 'made to measure' tariffs (Corden, 1974) are often driven by the seasonal calendar—high rates of protection are imposed during periods when locally produced commodities are available. The high bindings implied that the reduction in fluctuations in international food markets that tariffication was expected to deliver would not necessarily be attained (Goldin and van den Mensbrugghe, 1996).
In recognition of the fact that applied tariffs for some products were set at prohibitive levels, minimum market access commitments were negotiated. These required that the share of imports in domestic consumption for products subject to prohibitive import restrictions increase to at least 5 per cent by 2000 (8 per cent in the case of rice in Japan in lieu of tariffication, less in the case of developing countries).3 The vehicle used to ensure this minimum market access is generally a tariff rate quota (TRQ), under which a certain volume of imports (the quota) enters at a lower tariff, and out-of-quota imports are subject to a much higher tariff. Special safeguard mechanisms are available to protect domestic producers if imports exceed specific trigger quantities or are priced below trigger price levels. There is also scope to minimize the impact of those imports on the domestic market. For example, a country's required rice imports could be of low feed quality or could be re-exported as food aid.
The market access rules formally introduce scope for discriminating in the allocation of TRQs between countries. The administration of such quotas tends to legitimize a role for state trading agencies. When such agencies have selling rights on the domestic market in addition to a monopoly on imports of farm products, they can charge excessive mark-ups and thereby distort domestic prices easily and relatively covertly—just as such agencies can hide export subsidies if they
3 Countries seeking to delay tariffication were permitted to do so for six years (ten for developing countries) if imports were below 3 per cent of domestic consumption in the 1986-8 base period, no export subsidies were granted and measures to restrict output were implemented. In such cases the minimum market access requirement was higher, increasing from 4 per cent in 1995 to 8 per cent in 2000.
282
the political economy of the world trading system
are given that monopoly. Elements of quantitative management of both export and import trade in farm products were therefore legitimized under the WTO.
Domestic support. The third major element of the agreement is a set of disciplines on domestic production support to agriculture. Negotiators defined three categories of subsidies and other types of support, the so-called Green, Amber and Blue 'boxes'. The first comprises instruments that are permitted and unconstrained; the latter two include policies that affect production. The agreement requires high income countries to reduce an Aggregate Measure of Support (AMS) by 20 per cent by 2000 (again relative to a 1986-8 base period). The policies covered by the AMS are considered to distort production and trade and constitute what is often called the Amber Box. This is defined in Article 6 AoA, and includes measures to support prices and subsidies that are tied to, are conditional on, or affect agricultural output. De minimis supports are allowed (no more than 5% of agricultural production for developed countries, 10% for developing countries), but the 30 WTO members that had subsidies exceeding de minimis levels at the beginning of the post-Uruguay Round reform period were to reduce them by 20 per cent.
World Trade Organization members were required to calculate and enter their base period AMS in their schedules, as well as the 'final bound commitment level' for the AMS. The AMS includes expenditures on domestic subsidies as well as market price support policies such as administered prices, and therefore captures both border and nonborder policies. In principle it covers all support policies that affect trade. However, EU compensation payments and US deficiency payments were excluded from the AMS. Instead, they were put into in a separate 'Blue Box' that was a key component of the Blair House deal between the US and the EU mentioned above. Any support that in principle would be included in the Amber Box is mapped to the Blue Box if it requires farmers to limit output. In contrast to the tariff reduction obligations, which apply at the tariff line, the AMS reduction requirements pertain to the agricultural sector as a whole.4 That is, the AMS is aggregated over commodities and programmes.
Given the goal of reducing the trade-distorting effects of agricultural policies, the AMS excludes instruments that in principle have minimal effects on production and trade. These so-called Green Box support instruments are defined in Annex 2 AoA. They span subsidies that do not distort trade, are fiscal in nature (are financed by the government budget) and do not involve price support. Green Box measures include programmes that support agriculture generally and do not involve direct transfers to farmers; income transfers that are decoupled from production; and
4 There is a similarity between the AMS and the montant de soutien concept, which was introduced by the EU during the Kennedy Round (see Evans, 1972). The montant de soutien was defined as the difference between the world price of a product and the price received by a domestic producer. In other words, it was the nominal rate of protection taking into account all instruments affecting producer prices. However, in the Kennedy Round the intention was that support measures would be calculated per commodity.
sector-specific multilateral trade agreements
283
policies that contribute less than 5 per cent of the value of production. There are no restrictions on the use of Green Box measures.
A so-called peace clause was negotiated regarding the use of countervailing duties and dispute settlement actions to contest the effects of subsidies: members agreed to refrain from new CVD actions for a six-year period and disputes until the end of 2003 (Article 13 AoA). The latter ensured that agricultural subsidies could not give rise to claims of serious prejudice under the SCM Agreement (see Chapter 5) during the implementation period. With expiry of the peace clause in 2004, WTO members can initiate dispute settlement proceedings claiming that subsidy programmes have caused serious prejudice. Before 2004 they needed to make a case that the specific (and very restrictive) conditions laid out in Article 13 had been satisfied. With the expiry of the peace clause all agricultural subsidies—whether they are Green Box, Amber Box, Blue Box or export subsides—can be challenged if a WTO member deems to have suffered serious prejudice or adverse effects as a result.
Developing countries only needed to reduce tariffs, support and export subsidies by two-thirds of the levels mentioned earlier, and had until 2005 to implement this. They were also exempted from the tariffication requirement for products that are primary staples in traditional diets, as long as imports are at least 4 per cent of consumption by 2005. Only production support that exceeds 10 per cent was subject to AMS reduction. Input subsidies for low-income farmers are permitted, as are generally available investment subsidies and export subsidies related to export marketing and internal distribution and transport. It is unclear what the tariff reductions imply in terms of effective liberalization of developing country agricultural markets, as they were not committed to use a particular base year for tariffication. In effect developing countries were granted the freedom to impose tariffs at whatever level they chose to.
The AoA was expected to reduce agricultural protection by about one-fifth or more in industrialized and emerging market economies. An aggregate measure of protection compiled by the OECD Secretariat—the Producer Support Estimates (PSE)—reveal that direct payments and agricultural market price support policies fell in some OECD countries after 1999-2001—e.g. Japan and the US— but increased in the EU, Republic of Korea and a number of other countries (Figure 6.2). All in all, total support to farmers in OECD countries did not decline much (Figure 6.1; Table 6.1), in part as a result of the dirty tariffication that occurred. Anderson, Martin and van den Mensbrugghe (2006) estimate that abolition of agricultural tariffs, subsidies and domestic support programmes would boost global welfare by nearly US$300 billion per year by 2015. Developing countries would receive around half of the global gains from completely freeing all merchandise trade, two-thirds of which would come from global agricultural liberalization. Thus, the level of post-Uruguay Round protection remained high. What the AoA did was somewhat similar to the GATS (Chapter 7) in that the main achievement
284       the political economy of the world trading system
160
140
120
5 100
S 80 <& 1/1
I
Yearly Average 1999-2001
Yearly Average 2004-6
I European Union    ■ Japan   BUS    □ Korea   □ All-Others 0ECD
Fig. 6.2. Producer support estimates for OECD members, 1999-2006
Source: http://www.oecd.org/document/0/0,3343,en_2649_33773_39508672_1_1_1_1,00.html
was that agriculture was brought into the fold and a structure was created for future negotiations to lower applied levels of protection.
The Doha Round: pursuing the built-in agenda
Article 20 AoA called for new negotiations to be launched in 2000. These were duly started and subsequently morphed into the Doha Round. Pressure to continue liberalization was strong. Agricultural exporters continued to push for the elimination of trade-distorting farm policies. As, if not more, important, the planned expansion of EU membership required further reform of the CAP, as the status quo regime was not financially sustainable if it were to be extended to Central European countries. Even in the absence of additional liberalization, locking in such reforms through the WTO would be of great value.
The agenda for negotiations on agricultural trade liberalization was conceptually straightforward, centring on the 'three pillars' of agricultural interventions (export competition, market access and domestic support), improving provisions for special and differential treatment for developing countries, and clarifying the scope to pursue noneconomic (nontrade) objectives—e.g. EU arguments that agricultural support programmes were 'multifunctional.'5
5 The WTO website provides an excellent overview and detailed information on the state of play and the process of the Doha negotiations on agriculture.
sector-specific multilateral trade agreements 285
Much of the negotiations focused on agreeing on so-called modalities to reduce bound tariffs and support programmes. These modalities were supposed to establish a framework for WTO members to make commitments on the three pillars on a 'formula' basis. Rather than pursue request-offer negotiations, members agreed to establish specific targets for reductions in export subsidies, domestic support and cuts in tariffs. The talks were characterized by a North-South divide. Developed countries' farm subsidies were major bones of contention, in part because in addition to the traditional set of interventionist OECD countries the US became a target for demands by exporters to open up its market and reduce support. A prominent focal point was US cotton policy, which was detrimental to West African producers. US resistance to making this a priority area for reform played an important role in souring the atmosphere.
The Doha Declaration had envisaged that countries would submit comprehensive draft commitments, based on the 'modalities' in 2003 at the Cancun ministerial conference. A joint EU/US paper drafted at the urging of a mini-ministerial meeting in Montreal in July 2003, proposed a framework for agricultural modalities. With slight modifications, this became the basis of the text on agriculture in the first draft declaration issued on the eve of the Cancun ministerial. This draft was widely seen as enshrining the more protectionist elements of both the US (e.g. maintaining ample scope for countercyclical subsidy payments) and the EU (e.g. maintaining significant barriers to imports). In response Brazil organized a group of developing countries that united around the common cause of reducing trade-distorting policies maintained by the EU and US.
The resulting G20 negotiating group—which included Argentina, Brazil, Chile, China, Egypt, India, Mexico, Nigeria, Pakistan, Philippines, South Africa and Thailand—became a powerful force in the Doha agricultural talks. The creation of the G20 led to a sharp reduction in the visibility and influence of the Cairns Group, as its developing country members were also in either the G20 (Argentina, Brazil, Chile, South Africa) and/or the G33 (e.g. Indonesia, Philippines) and the US ceased to be aligned with other traditional exporters. The G20 comprised a mix of countries with a much broader set of objectives than increasing exports (the main focal point of the Cairns Group). The bargaining coalitions that emerged during the Doha Round were more diverse than those that operated during the Uruguay Round. In addition to the G20—which became a key negotiating group—the G33, a group of developing countries with 'defensive' interests (many were net importers and sought to limit liberalization and expand SDT provisions), played a prominent role.6 Another defensive group was the G10, comprising countries that provided high levels of support to agriculture (non-EU European countries, Israel, Japan, Korea and Taiwan).
6 Members included many Caribbean and Sub-Saharan African countries, as well as India, Indonesia, Korea, Pakistan, Peru, Philippines, Turkey and Venezuela.
286
the political economy of the world trading system
Preferences of the countries associated with each group were by no means always aligned. The G20 included India, which was also a member of the G33. The inclusion of both Brazil—a major exporter with strong interest in liberalization of global agriculture—and India (which sought to maintain significant freedom to intervene in agriculture) meant that the G20 had to balance the offensive and defensive interests of its members. The result was that G20 negotiating positions tended to be 'in the centre'—in effect, already incorporating a deal between members with different preferences and concerns. Constantini and colleagues (2007) use cluster analysis techniques to analyse the internal 'coherence' of the various bargaining coalitions, exploring to what extent they bring together countries with similar structural features (GDP per capita, openness, FDI stocks, human development, food security indicators, share of agriculture in exports, output and employment, level of trade protection). They conclude that the G20 brings together countries that have similar structures, helping to explain the robustness of the group during the negotiations.
Disagreements on the specification of the modalities and the magnitude of allowable exceptions for specific products could not be overcome during 2002-8. Negotiations repeatedly missed deadlines. Despite talks among small subsets of 'representative' WTO members to agree on modalities, progress proved elusive. An example was talks between the so-called five interested parties— Australia and Brazil, representing export interests; the EU and India, representing the 'defensive' interests; and the US (with both offensive and defensive interests)— which did not manage to identify a package of mutually acceptable reforms. The extent of the difficulties were illustrated by the inability to deal with technical issues such as how to go about converting specific tariffs (taxes that are based on volume rather than the value of imported products) into ad valorem equivalents. Although these were important, the fact that ministers were asked to deal with such questions illustrated how deep the differences were in this area (Box 6.3).
The 2005 ministerial meeting in Hong Kong did not lead to any major breakthroughs, with the exception of (conditional) agreement to eliminate export subsidies for agricultural products by 2013. As of mid-2006, the contours of a possible agreement finally began to emerge. It was described as a 20-20-20 package by the Director-General of the WTO. Under this proposal the EU would cut its bound agricultural tariffs by an estimated 54 per cent (following a proposal made by the G20), the US would bind its total trade-distorting agricultural subsidies at a maximum of US$20 billion, and developing countries would agree to a maximum tariff on manufactures of 20 per cent. A deal along the lines of the 20-20-20 package would have implied a significant reduction in not only tariff bindings but also applied policies. Negotiators could not agree on a specific compromise, however. Exporters sought deeper cuts in EU tariffs and both the G20 and the EU wanted the US to accept a lower ceiling on its domestic farm subsidies.
sector-specific multilateral trade agreements 287
Box 6.3. Doha disagreements on ad valorem equivalents
In April 2005 the Chairman of the special session of the Committee on Agriculture (the Doha negotiating group) suspended talks as a result of disagreement on how to convert specific tariffs into ad valorem equivalents (AVEs). This was needed in order to be able to apply a formula for tariff reductions. The problem revolved around what source of data to use to calculate the AVEs in instances where calculations based on the WTO's Integrated Database (IDB) differed significantly from those resulting from use of the UN Comtrade database. Members had agreed that in such cases, more processed products would use the IDB, whereas more basic commodities would be based on UN Comtrade. Rather than just agreeing to split the difference—i.e. use the average of the two estimates-—a more complicated two-stage approach was proposed: (1) a determination if the unit import value for a product in the two databases differed by more than 40 per cent (if not the IDB would used); and, if so (2) a comparison would be made between the AVEs using the two sources, with the IDB used if the difference was less than 20 percentage points. The question was what to do with those products where the difference in AVE estimates was greater than 20 points.
Members had agreed that in such cases the adjustment for commodities should be 25 per cent IDB and 75 per cent Comtrade (as the latter was regarded as having better coverage of processed products), and that the calculation for processed products would be on a 50-50 basis. Switzerland argued that this would favour food exporters, whereas some African countries were of the view that this would put commodities higher up in the band to which the tariff-cutting formula would apply, and thus disadvantage African food exporters.
The issue on which talks broke down was whether the base for the above adjustment should be the unit values or the estimated AVEs. The EU reportedly shifted towards the latter whereas most members thought they had converged to the former. The end result was that this very technical matter was put before ministers at a mini-ministerial meeting in Paris in May 2005. This is not the type of question that ministers should be dealing with. These types of technical issues are matters for negotiators to resolve, working with the WTO Secretariat. The episode illustrated how dysfunctional the talks had become—and that the Secretariat was not able to play the role that it should.
Despite efforts in early 2007 to resuscitate the process, which led to a high-profile meeting between four of the major protagonists (EU, US, Brazil and India) in Potsdam in June, agreement remained elusive. Developing countries remained disappointed by the US offer on agricultural subsidies and the EU unwillingness to open its agricultural markets more. Conversely, the EU and the US were unsatisfied by what they regarded as inadequate offers by Brazil and India on industrial tariffs.
In mid-2008 a draft text issued by the chairperson of the Agriculture negotiating group as a basis for further discussion laid out the most specific and comprehensive blueprint for global liberalization since the negotiations began in 2001, with
288
the political economy of the world trading system
bracketed ranges for key numeric parameters. What follows briefly describes what was on the table, drawing on the excellent summary by Martin and Mattoo (2008).
Export competition. In the course of the 2000s export subsidies became much less of a problem than in preceding years because world prices of agricultural commodities were relatively high. As a result there was much less need for export subsidies to be used to dump surpluses on global markets. The magnitude of export subsidies is determined by the gap between domestic and world prices. Export subsidies are used when high tariffs raise the domestic price of commodities as a result of which domestic output expands. If there are also domestic production support programmes this output expansion will be greater, potentially impacting world prices through an artificially increased global supply.
Unfortunately, the available data on export subsidies are limited. World Trade Organization members do not notify export subsidies in a comprehensive and timely basis—the situation is similar to that pertaining to notification of nona-gricultural subsidies (see Chapter 5).
The July 2004 Framework Agreement had already spelled out in some detail how liberalization was to occur: export subsidies were to be eliminated by a 'credible' date (set at 2013 at Hong Kong), with reductions to be implemented in annual installments during a transition period. Although abolition of all export subsidies would have only a small immediate overall effect given high world prices, a ban would be of significant value in ruling out future recourse to such subsidies in periods with low prices.
The elimination of export subsidies is best seen as a key consequence of reducing the gap between domestic and world prices created by border barriers and domestic support programmes. Eliminating export support without reducing tariffs and domestic support would simply result in putting world agriculture in the situation (no export subsidies, high tariffs and domestic support) faced by manufacturing at the dawn of the GATT in the late 1940s. That in itself would be an achievement, but from an economic perspective is likely to have a limited impact (Hoekman and Messerlin, 2006). What matters is a substantial reduction in border protection (Box 6.4).
As part of the Doha Round discussions on export subsidies, the EU, supported by a number of Cairns Group / G20 members, linked elimination of export subsidies to all existing 'equivalent' forms of export subsidization: specifically the subsidy component of official export credits, the activities of STEs and food aid. One interpretation of this linkage is that it was largely tactical. Although the EU is by far the largest user of export subsidies, a number of traditional export-oriented and pro-liberalization countries make use of these alternative instruments. Although the objective in the case of STEs and food aid is generally not to subsidize exports, they may have that effect. Thus, for example, the US grants both export credits and food aid, and Canada has made long standing use of STEs for specific commodities. Alternatively, the focus on equivalent disciplines for alternative instruments
sector-specific multilateral trade agreements 289
Box 6,4. The three pillars: what matters most?
Of the three pillars, most attention centred on domestic support and export subsidies, with major campaigns by NGOs such as Oxfam lampooning the extent to which EU, US and Japanese farmers were coddled: Creative use was made of the detailed data on agricultural protection that had been compiled by organizations such as the OECD. One oft-cited example of the Oxfam critique of EU protection was that the amount spent on the average cow in the EU was enough to fly it around the world in business class. Another was that more was spent on each cow than the annual per capita income of many households in Africa.
However, economic research suggested that what really mattered most from the perspective of farmers in developing countries was border protection in the EU and other OECD nations. Subsidies are expensive and may be inequitable, but analysis concluded that subsidies had a much smaller impact on world prices than tariffs and TRQs. Domestic support in the OECD involves huge income transfers from OECD taxpayers to OECD farmers, but their impact on (developing country) agriculture exporters is smaller than the impact of tariff barriers in both developed and developing countries. Hoekman, Ng and Olarreaga (2004) and Anderson, Martin and Valenzuela (2006) estimated that border barriers accounted for 80-90 per cent of the impact on world prices and thus welfare. In large part this is because of high tariff peaks in OECD and developing countries in products subject to domestic support or export subsidies: for a given amount of domestic production, tariffs lead to lower levels of domestic consumption than domestic subsidies, and therefore lower world prices (Snape, 1987). Anderson, Martin and van den Mensbrugghe (2006), consistent with other studies done at the time, concluded that reductions in agricultural tariffs would deliver 12 times the gains that would be achieved by abolishing export subsidies and trade-distorting domestic support to agriculture.
that may give rise to export subsidies can be perceived as a necessary step to ensure that governments do not engage in 're-instrumentation' following a WTO ban on export subsidies on farm products.
Is this a significant issue? Here again there is very limited information. Calculating the subsidy component of an export credit requires data on the amount of the credit, its terms—maturity, interest rate structure, etc.—the credit worthiness / risk profile of the borrower/recipient, etc. The counterfactual will be difficult to determine—would a bank or other financial services provider have provided finance and, if so, at what price? Inherently there will be a subjective element to any assessment of the export subsidy equivalent associated with export credits, the operation of STEs and food aid.
One estimate of export subsidy equivalents of export credits by OECD (2001) for affected products in Australia, Canada, the EU and the US during 1995-8 suggested that these did not exceed 7 per cent for any of the instruments considered. In terms of the overall impact, the share of total agricultural exports
290
the political economy of the world trading system
to which these instruments apply is small—ranging from less than 2 per cent for the EU to around 5 per cent for Canada and the US. It was highest for Australia (15 per cent). Bulk cereals accounted for almost half of the total subsidy element of export credits granted. When used in a simulation model to assess the impact of these programmes on prices, it was found that US export and domestic prices would be only 2 and 1 per cent higher, respectively, following a ban on export credits. Moreover, the bulk of export credits apply to intra-OECD trade—in the case of the US, South Korea and Mexico are the major recipients. It would appear therefore that these are of second-order importance compared to export subsidies proper (which in turn are second order compared to market price support).
Domestic support. Here the focus was on introduction of a number of additional constraints and reductions in bound levels of support. The AMS was to be reduced using a tiered formula. It would cut EU support by 70 per cent; by 60 per cent in members with intermediate levels of support (including the US); and 45 per cent in other members. A new Overall Trade Distorting Support (OTDS) measure—defined as the sum of the AMS, de minimis support, and Blue Box instruments—was to be cut by between 75 and 85 per cent in the EU; 66 to 73 per cent in the US; and 50 to 60 per cent in smaller industrial economies. Permitted Blue Box support would be limited to 2.5 per cent (5 per cent) of the value of agricultural production for developed (developing) members. Product-specific limits would also be introduced on the AMS and the Blue Box, with the allowed level of support to cotton reduced very sharply and under an accelerated timetable (cotton is discussed in more detail below). Projections by Blandford and Josling (2008) suggest that the draft 2008 modalities would not constrain total AMS or OTDS in the US under realistic price projections. However, the product-specific AMS and Blue Box commitments would probably constrain products of particular importance to many developing countries such as sugar, peanuts and cotton. In Europe, the 2008 modalities proposals would lower the OTDS below projected, unconstrained levels of support.
Market access. The method proposed on market access was a 'concertina' approach (see Corden, 1974) under which the highest tariffs rates are reduced more. The proposal was to use a tiered formula that would require proportional cuts in tariffs, with the cut increasing when moving between each of four progressively higher bands (Table 6.2). The proposal would have permitted developed countries to classify 4-6 per cent of tariff lines as sensitive, whereas developing countries could exclude an additional 33 per cent (compared to industrialized economies) and be able to self-designate 'special products' on which they could make smaller-than-formula cuts. Proposals on the number of such special products ranged between 8 and 20 per cent of all agricultural tariff lines. The end result would be to undo much of the harmonization of the basic formula (Martin and Mattoo, 2008).
sector-specific multilateral trade agreements 29I
Table 6.2. Proposed formula for cuts in bound agricultural tariff, July 2008 (%)'
Tariff Band Developed Countries Developing Countries
	Range	Cut	Range	Cut
A	0-20	50	0-30	33,3
B	20-50	57	30-80	38
C	50-75	64	80-130	42.7
D	>75	66-73	>130	44-48.6
Source: Martin and Mattoo (2008).
Application of this proposal would have reduced average bound agricultural tariffs by nearly half, from 40.3 to 20.7. World average applied tariffs would be cut by nearly 40 per cent, from 14.5 to 8.9 per cent. Flexibilities for sensitive and special products would significantly attenuate the impact of the formula: the world average bound rate drops by just over a quarter, and the average applied tariff would fall by one-fifth, from 14.5 to 11.8 per cent. In the high-income country group, the formulae with flexibilities would cut four percentage points off applied tariffs, from 15 to 11 per cent, and reduce average bound tariffs from 40 to 30 per cent. In developing countries, taking into account likely flexibilities for sensitive and special products and the large gap between bound and applied rates (40 percentage points on average), the cuts in developing country applied tariffs would be very small—only 1.9 percentage points. Even more than for OECD countries the main effect therefore would have been to lower average levels of bound tariffs: from 54 per cent on average to 45 per cent.
After the Uruguay Round, OECD countries became intensive users of TRQs. As mentioned previously, under a TRQ, there is an out-of-quota tariff that applies to imports above a specified quota quantity. Volumes below the quota limit pay a lower in-quota tariff. Understanding the impact of TRQs is critical to predicting the outcome of attempts to liberalize trade in agricultural products. For example, reducing out-of-quota tariffs will increase imports only if the current demand for imports exceeds the quota amount such that the out-of-quota tariff is operational. If imports are less than the quota level, reductions in out-of-quota tariffs will be ineffective. On the other hand, marginal expansion of the TRQs will be ineffective if imports are greater than the TRQ—the only effect will be to increase the volume of imports on which scarcity rents are earned. If imports are less than the TRQ, expanding the quota will also be ineffective. Only reductions in in-quota-tariffs will stimulate greater imports in this case. Thus, reductions in out-of-quota tariffs may be the most effective instrument for achieving market liberalization in the majority of cases.
292       the political economy of the world trading system
Special safeguards
Unlike with normal safeguards (discussed in Chapter 9), the special safeguard (SSG) mechanism included in the AoA permits the automatic imposition of higher duties if import volumes rise above or prices fall below a certain level/ It is not necessary to demonstrate that serious injury is being caused to the domestic industry. The SSG can only be used on products that were tariffied in the Uruguay Round by governments that reserved the right to do so in their schedules of commitments. Only 39 countries—17 developed and 22 developing—did so. Safeguards must take the form of temporary duties that may not last more than one year. Tariffs are limited to an additional 33 per cent of the applicable bound rate if the trigger is an import volume surge. Alternatively, if the trigger is a significant reduction in price, the SSG is determined by the difference between the import and the trigger price.
During 1995-2004 only 6 of the 22 developing countries eligible to use the programme actually utilized it, for just 163 tariff lines (Hufbauer and Adler, 2008). Many developing countries that did not use NTMs to distort agricultural trade during the Uruguay Round did not have to engage in the tariffication process and therefore did not have access to the SSG. The right to use the SSG was intended to lapse if there was no agreement in the negotiations to continue the 'reform process' initiated in the Uruguay Round (Articles 5.9 and 20 AOA).
The Doha Round generated numerous proposals regarding the SSG, ranging from keeping it unchanged, revising it to exempt products from developing countries, to abolishing it. Many developing countries proposed that only they be allowed to use special safeguards. This led to proposals to create a new special safeguard mechanism (SSM) that would permit developing countries to raise tariffs temporarily to deal with agricultural import surges. (This rationale for the SSM should be distinguished from protecting farmers in general. The latter objective was addressed by inclusion of provisions allowing developing countries to make smaller or no tariff cuts for 'special products'.) Agreeing on the criteria for a SSM proved to be difficult, with some countries, especially the US, opposing suggestions by the advocates of a flexible SSM—the G33 and its allies—that the SSM should be easy to use, with low thresholds for the import volume trigger, and allow for tariffs above pre-Doha Round bound rates. The US and other agricultural exporters argued that the SSM should not result in tariff increases above pre-Doha Round levels, that it should be limited to a period following liberalization of trade, and not affect the balance of rights and obligations negotiated in the Uruguay Round.
7 The abbreviation SSG comes from the notation used in the tariff schedules of WTO members to indicate that a tariff line could be subject to the special safeguard.
sector-specific multilateral trade agreements
293
The 2008 draft modalities envisaged a special safeguard mechanism with a volume and a price trigger that would be applicable to all agricultural products. The proposal was that import duties of up to 25 percentage points could be imposed if imports exceeded 110 per cent of a three-year moving average. A price-based measure to restrict imports could then be invoked if the price of imports dropped below 85 per cent of a three-year moving average of import prices, with a duty of up to 85 per cent of the gap between prevailing import prices and the three-year moving average. One matter of contention was whether the combination of the special safeguard duty and the applied tariff rate would be permitted to exceed the pre-Doha bound tariff rate levels (Martin and Mattoo, 2008).
No agreement proved possible on the SSM—which became the proximate cause for yet another breakdown of the Doha talks in July 2008. Five main issue areas were in contention in the SSM debate (Hufbauer and Adler, 2008). First, coverage: all agricultural goods or only a subset? Some countries argued that the SSM should not be available for 'special products' that were exempted from the formula cuts. Second, the type of triggers to be used; the thresholds to apply; and whether there should be a single criterion or alternative triggers that would be conditional on the level of commitment that a country had made. Third, whether SSM tariffs could exceed pre-Doha bound rates and, whether this should be subject to limits. The US wanted higher trigger thresholds for SSM tariffs that exceeded pre-Doha bound rates. Fourth, whether SSM tariffs should be conditional on an injury test of some type. A final issue was whether there should be time limits for SSM actions, as for the SSG.
Measures such as the SSM are essentially motivated by a desire on the part of governments to insulate markets against external shocks—specifically, reductions in world prices that generate import surges. A SSM provides protection and insulation to domestic markets, while reducing market access and increasing the instability of world markets if used by importers accounting for a significant fraction of imports. It was somewhat ironic that while the debates on the design of the SMM were raging in mid-2008 the problem was not low agricultural prices, but the opposite. Governments were taking whatever actions they could to lower consumer prices. In general, the type of SSM that was proposed in the Doha Round might act as a price-insulating measure for developing countries, but as with any trade policy it would generate distributional effects—in this case raising the prices for net consumers of agricultural commodities in the countries using a SSM.
Economists have also pointed to the likely negative spillovers created by SSM-type instruments: any price-insulating effect will likely be diminished by the consequent increase in volatility of world markets for products in which developing countries account for a large share of world production (Ivanic and Martin, 2008). In general, it is unclear why governments should not want to use general safeguards to address import surges if they deem this to be necessary. A major advantage of the general safeguard provision in the GATT is that it requires governments to assess
294
the political economy of the world trading system
the economy-wide implications of taking action. This is very desirable and helpful from a policy perspective. Arguments sometimes heard that the general WTO safeguard provision cannot be used to respond rapidly are incorrect—as discussed in Chapter 9, measures can be imposed provisionally.
Net food-importing countries
Agricultural liberalization, especially moves towards elimination of export subsidies, may increase world prices of food products, and thus have a negative effect on net food-importing developing countries. During both the Uruguay and Doha Rounds a number of countries expressed concerns regarding the impact of liberalization on food security. The main worry was that global liberalization of agricultural trade could give rise to adverse terms of trade effects.
During the Uruguay Round it was noted that institutions such as the IMF and World Bank had instruments available to finance short-term needs should world prices of food increase significantly. Developed country WTO members also committed themselves to continue to provide food aid, as well as technical and financial assistance, all least developed and the 18 developing countries that were classified as net food-importers. Specific recommendations were adopted at the 1996 Singapore ministerial meeting regarding negotiations on international food aid commitment levels and related concessionality guidelines. In response, the Food Aid Committee decided in December 1997 to extend the life of the Food Aid Convention through June 1999 and to open the Convention for re-negotiation.
Research suggests that the impact of global liberalization on net food importers would be limited, because reforms will be spread out over multiple years and price increases will be offset to some degree by an increase in domestic supply that will be stimulated by higher prices. Trade policy is not the appropriate instrument if the objective is food security. Instead, the key need is to have the foreign exchange (and access to credit) to be able to buy food in times of scarcity. Having the domestic ability to produce food is not required—countries should only specialize in food production if they have a comparative advantage in this activity. A necessary condition for this piece of advice to be appropriate is that countries have access to global markets-—that is, can buy supplies. This may not be the case in times of severe global shortage. In 2008, following a steep and rapid rise in food prices, a number of exporting countries imposed export restrictions (taxes or bans). For a brief period, reportedly some countries could not buy staple commodities such as rice at any price. This suggests that disciplining the use of export restrictions should be an important objective of net importers. The WTO only imposes weak disciplines on the use of export restrictions, and efforts in the early part of the Doha Round to put this matter on the negotiating table failed.
sector-specific multilateral trade agreements
295
The reason for the high prices that emerged in 2007 had nothing to do with liberalization—the main fear expressed by net importers in the MTNs—but with global economic developments: the boom in China and other emerging markets up to early 2008, and the resulting rise in oil and other commodity prices. Energy price increases in turn led to greatly increased production of bio-fuels in the US that used maize and other cereals as feedstock, further driving up food prices (Mitchell, 2008). This episode illustrated that liberalization was not a potential problem—it was the lack of global liberalization that had severe implications for net importers. If trade had been free in agricultural products and bio-fuels, Brazil and other major sugar-producing countries could have exported much more efficiently produced ethanol to the US and EU than these countries could supply themselves, without the associated additional negative impacts of diverting food production towards bio-fuel production.
The increase in global food prices had deep roots in decades of trade-distorting policies that encouraged inefficient agricultural production in rich countries (most recently in the 2000s in the form of bio-fuels), led to recurrent dumping of surpluses on global markets, and discouraged efficient production in developing countries (World Bank and IMF, 2008). In turn, as discussed above, developing countries have often taxed their farmers. Overall, the result was declining agricultural prices, overproduction in high-income countries, underproduction in poor countries, thinner global agricultural markets, more volatility, and lower overall reserve supply capacity and food security. Matters were compounded by the use of export controls by Argentina, Ukraine, Russia and Kazakhstan for wheat, and Vietnam, India and China for rice. These restrictions were imposed in an effort to decouple domestic from global markets and rein in domestic food prices. Such export restrictions went beyond food—China, for example, imposed export taxes on fertilizer in an effort to reduce input prices for domestic farmers. Fertilizer costs had increased in line with oil prices; the reduction in supply caused by export taxes increased prices in importing countries further. The result was to put pressure on farmers in low-income countries that are credit-constrained—reducing planting areas and future yields.
Export restrictions tend to: (1) distort prices and the allocation of resources, therefore impeding investment and the supply-side response; (2) prevent local farmers from receiving the higher world market price for their production; (3) displace local production to crops that are not subject to export restrictions, therefore aggravating food security and price concerns; and (4) exacerbate the rise and fluctuations of global food prices, therefore creating a vicious incentive for trading partners to follow suit, curb exports and hoard (Chauffeur, 2008). Moreover, by signalling that global markets cannot be relied upon to function, export controls create incentives for importing countries to subsidize domestic production and emulate the types of policies pursued in many high-income OECD countries.
296      the political economy of the world trading system
Trade policy is therefore part of the problem. Export restrictions can help stabilize domestic prices in the exporting country but at a significant cost in terms of greater world price volatility and higher average prices for net importers. As trade liberalization generally takes a long period of time to be negotiated and implemented, there is, in principle, ample opportunity for governments to develop or strengthen safety-net programmes and complementary policies to maintain real incomes of the poor. Such time does not exist in instances where there are acute shortages that are exacerbated by 'beggar thy neighbour' export restrictions. But in such situations trade policies are useless for net importing countries— governments will want to lower tariffs, not raise them.
The food price increases that occurred in 2007-8—and the response by food exporters—revealed that an exclusive focus on liberalization on the import side and reducing domestic support is too narrow. Export restrictions and export taxes also needed to be on the WTO negotiating agenda. Current disciplines are weak— Article XI GATT is permissive for agriculture export restraints, and export taxes are unconstrained (see Chapter 5). The fact that these measures were not on the table was not because no proposals to this effect had been put forward. Efforts by lapan and the EU to do so in the Doha Round were rebuffed by a number of developing countries such as Argentina. The 2008 draft modalities only required members to notify the WTO of restrictions or bans 90 days after they were imposed and that these measures were not to exceed one year.
Major agricultural subsidy disputes
We conclude this chapter with a discussion of two major agricultural subsidy disputes. One concerned US cotton subsidies, the other the EU policy set affecting sugar. These cases illustrate the complexity of policy in this area, the extent to which policies create distortions that affect negatively developing countries, and the scope that the WTO offers to attack support programmes that exceed what is permitted (has been scheduled).
Cotton. In the mid-2ooos the US was the world's second-largest cotton producer (after China) and the world's leading exporter. The US provided substantial support to its 25,000 cotton producers: averaging some US$3 billion a year in 2005-7 (as subsidies are a function of prices, the amount rises with declines in world prices in any given year). Chinese subsidies were about US$i billion in 2002 (Baffes, 2005). Economists have estimated the effects variously, with the average estimated negative impact on world prices being in the 6-14 per cent range (Alston, Sumner and Buncke, 2007). The negative spillover is particularly detrimental to other cotton producers, including four West African countries: Benin, Burkina Faso, Chad and Mali. In the Doha Round these countries were known as the Cotton 4 or C-4). In Benin, Burkina Faso and Mali, cotton accounts for roughly
sector-specific multilateral trade agreements 297
5 per cent of GDP, with between 200,000 to 300,000 households involved in production. In Benin and Mali, cotton accounts for one-third of total exports; for Burkina Faso the figure is 75 per cent. Sumner (2006) estimated that removing cotton subsidies, as part of freeing all merchandise trade, would expand cotton exports from Sub-Saharan Africa by 75 per cent. The developing countries' share of global cotton exports, meanwhile, would rise from 56 to 85 per cent by 2015. Removal of US cotton subsidies was estimated to increase household incomes of cotton producers by 2.3 to 8.8 per cent, enough to support the expenditure on food for one million people in the four countries concerned (Alston, Sumner and Buncke, 2007).
The West African countries formed a coalition in the Doha Round. They pushed for the abolition of export and other trade-distorting subsidies granted to cotton producers in the US, EU and China, and that their cotton farmers be compensated during the transition period in which subsidies were to be phased out. This was a first for the multilateral trading system: LDCs coming forward with a specific demand. The West African proposal attracted much support from other developing countries, as well as the donor community in several OECD nations (Lee, 2007).
The US strongly resisted, arguing that agricultural policies needed to be addressed horizontally as part of an overall agreement on agriculture and not on a product-specific basis. It also argued that African countries should focus on diversifying their economies away from a reliance on cotton towards producing textiles, which could then be granted preferential market access to the US under the African Growth and Opportunity Act. At the Cancun ministerial the second draft of the ministerial text called on the WTO Director-General to consult with the international agencies, including the Bretton Woods institutions, to redirect their programmes and resources to assist these countries to diversify out of cotton; the same draft was deafeningly silent on price-depressing subsidies in the US, China and elsewhere. The lack of willingness to address the African demands on cotton was a factor that strengthened the resolve of African delegations more generally to hold firm on their opposition to accept to launch negotiations on the Singapore issues.
From a visibility standpoint, the Cotton 4 gained considerably from tabling their initiative—the direct link between policies in the most powerful trading economies and their anti-development consequences in depressing incomes of the world's poorest emerged with worldwide notoriety. However, the proposal for 'special treatment' or 'early harvest' met resistance from developed and developing countries alike, and gained only limited support within the WTO membership. Moreover, the proposal for compensation was difficult for trade ministers to address.
In 2002 Brazil had initiated a WTO dispute settlement case against the US cotton programme (US—Upland Cotton, WT/DS267). In September 2004, a WTO panel
298       the political economy of the world trading system
ruled against the United States, as did the Appellate Body in March 2005. The panel/AB ruled that: US cotton subsidies had exceeded the 1992 benchmark year level of subsidy commitments; the direct payments made under US farm programmes were not covered by the Green Box as they were not fully decoupled income support (because payments were accompanied with planting restrictions); so-called Step 2 programme payments that compensated US exporters and cotton mills for the difference between domestic and world prices were prohibited subsidies; US export credit guarantees were prohibited export subsidies; and US domestic support measures that were 'contingent on market price levels' had resulted in excess cotton production and exports that, in turn, reduced world prices (led to price suppression) and as a result caused 'serious prejudice' to Brazil. The AB recommended removal of the 'prohibited subsidies' by July 2005 and the serious prejudice resulting from 'actionable subsidies' by September 2005.
Following nonimplementation by the US, Brazil sought authorization to retaliate against US$4 billion of US exports based on the magnitude of the subsidies granted by the US. It also requested authorization to cross-retaliate in other areas (i.e. TRIPS). The US in turn requested WTO arbitration, which was suspended following a mid-2005 agreement between the parties. In February 2006, the US Congress approved a bill that repealed the Step 2 subsidy programme for upland cotton. As these export subsidies accounted for about 10 per cent of total subsidies to the US cotton industry, Brazil requested a compliance panel to determine whether the US cotton programme continued to violate WTO rules. In December 2007, the panel ruled that the US was still not in compliance; the AB agreed in its June 2008 report (see Schnepf, 2008).
This was an important case on a number of dimensions. Together with the sugar case discussed below, it signalled that developing countries could and would use the dispute settlement mechanism to contest agricultural support policies if they violated the WTO. It also illustrated that the SCM agreement has a broad reach and that arguments can be brought to the WTO that subsidy programmes cause price suppression on world markets. The case was also important in revealing that the distinction between Green and Amber categories of subsidies may not be very useful from a 'legal certainty' perspective—a number of programmes that had been assumed to be in the Green Box were found to be contestable. More importantly, the case illustrates that it is only when there is intense and focused scrutiny of a set of policies that it may become clear whether a programme satisfies the Green Box legal criteria (economists will argue that the separation is simply not possible to make, as in 'general equilibrium' any policy can have an indirect effect on output). Although certain aspects of the reasoning and approach used by the panel and AB can be criticized—e.g. whether and to what extent US policies suppressed world prices, a question on which economists disagree (Sapir and Trachtman, 2008)—much of the information that emerged as a result of the case was certainly not common knowledge. The case, as do others,
sector-specific multilateral trade agreements
299
reveal the limits of the Trade Policy Review Mechanism and other transparency and notification requirements of the WTO: what is needed is analysis of the effects of policies—-not just to determine economic impacts but also whether they violate WTO disciplines.
Sugar. Trade protection for sugar production has been a longstanding feature of the international economy, dating back to at least the 1800s (Mitchell, 2005). It has been greatest in countries of the northern hemisphere that produce sugar beets, which is twice as expensive to produce as sugar produced from cane. Over the years, high protection lowered consumption, reduced imports and led to surplus production that was dumped on the world market supported by export subsidies. As world market prices fell and producers confronted subsidized competition from beet sugar, all governments of sugar-producing countries confronted calls for protection.
The EU, Japan and the US all impose high levels of support. Since the early 1970s, US sugar imports declined from more than five million tons per year to about one million tons per year. Japan's sugar imports fell from 2.5 to 1.5 million tons between 1980 and 2000. The EU was a net importer of about 2.5 million tons of sugar in the early 1970s, compared to net exports of about five million tons in the early 2000s. In 1999-2001, the value of gross receipts of sugar producers in the EU was more than double the value of their output measured at world prices. In this period, total OECD support for sugar was equivalent to about half of global exports (US$6.35 billion compared to US$11.6 billion), similar in value to the total exports of sugar of all developing countries (US$6.5 billion). The EU accounted for 43 per cent of the US$6.35 billion in OECD support for sugar. Much of this support was provided through very high border protection—around 90 per cent for the EU. Its support policies resulted in the EU becoming the second largest exporter in the world (after Brazil), accounting for 12 per cent of world exports. At the same time, the EU was also the world's fourth-largest importer, an idiosyncrasy that reflected the preferential access granted to African, Caribbean and Pacific (ACP) countries under the Sugar Protocol of the Cotonon Convention between the EU and ACP states, the successor to the earlier Lome Convention. The end result of the interventions in the major countries was that they became self-sufficient and effectively closed to competition.
The decline in import demand for sugar by the EU and the US depressed world prices and adversely affected more efficient producers in developing countries such as Brazil and a number of African economies. Estimates indicate that world prices could have been 40 per cent higher in the absence of the protection of sugar in OECD nations (Mitchell, 2005). High protection led to the emergence of high fructose corn syrup as a substitute for sugar in the US and Japan, which came to account for 40-50 per cent of sweetener use in these countries. Developed in the 1960s, corn syrups were profitable because of high sugar prices, and over time became cheaper than beet sugar. As is often the case, the protectionist policy led to
300
the political economy of the world trading system
a market reaction that undermined the original objective—leading to pressure in the EU for controls to be imposed on corn syrup production.
In 2002, Australia, Brazil and Thailand launched a dispute against EU sugar.8 In EC—Sugar, the complainants argued that the EU violated its WTO export subsidy commitment levels, in part through de facto cross-subsidization of exports as a result of guaranteeing high annual intervention (support) prices for a given quantity of EU sugar, and in part as the result of re-exporting an amount of sugar equivalent to what it imported from ACP countries on a preferential basis.
The EU policy regime for sugar complemented high intervention or support prices for sugar with production quotas. These were of two types: so-called quota A and quota B. The sum of A and B quotas determined the maximum amount of sugar that could be sold in the EU market in a given year. All excess production had to be exported. Production of A and B sugar benefitted from the high intervention price in the EU; excess production—so-called C sugar in EU jargon—did not. There is no physical difference between these various categories: there is one world price for sugar, be it A, B or C sugar. The producer price for A sugar was greater than the producer price received for B sugar. Both were less than the basic intervention or support price as a result of a levy that was used to finance the export subsidies needed to sell excess production on the world market.
In addition to (part of) the B quota, the EU also exported an amount of sugar equal to what it imported from ACP countries under its preferential access programme (the Cotonou Convention). This equalled some 1.3 million tons, rising to up to 1.6 million tons in some years. Given that EU production exceeded consumption at the intervention price, in effect all the ACP sugar was 're-sold' on the world market. That is, the effects of the ACP sugar protocol imports on the EU market were 'sterilized' by exporting the amount imported. As the ACP sugar was bought at the intervention price, the export sales incurred a significant loss, which was absorbed by the EU budget (taxpayers). These costs are clearly export subsidies, and were recognized as such by the panel and the EU.
A key question in the dispute revolved around what producers do with the rents, in particular whether they use them to cross-subsidize production and exports of C sugar. De Gorter, Just and Kropp (2008) show that cross-subsidization is possible for a variety of permutations of production costs, world price levels and support prices implied by a quota level B. Article 9.1(c) AoA requires that payments on the export of an agricultural product be 'financed by virtue of governmental action', a condition that was met in this case. The panel argued that the EU policy was 'a governmental action' that allowed the cross-subsidization to be 'financed' by EU exporters. This is consistent with economic analysis. Note that the panel finding does not expand the scope to argue more generally that
8 What follows draws on Hoekman and Howse (2008).
sector-specific multilateral trade agreements
301
policies support cross-subsidization. A pre-condition for such a case is specific subsidy disciplines. These have only been negotiated for agriculture in the WTO.
A second key issue concerned the total volume of sugar exports. For a WTO member to be able to grant an agricultural export subsidy it must be scheduled and be subject to reduction commitments to ensure that the percentage reductions in budgetary outlays and quantities specified in Article 9.2(b) (iv) AoA are achieved by the end of the implementation period. The EU scheduled: (1) a 'base quantity level' of 1,612,000 tons, to be progressively reduced to 1,273,500 tons in 2000 as the 'final quantity commitment level' for sugar; and (2) a 'base outlay level' of €779.9 million, to be progressively reduced to €499.1 million in 2000.
In 2001 the EU exported 4.1 million tons. The excess over what was permitted comprised 1.3 million metric tons of ACP sugar and 1.5 million tons of C sugar. The EU claimed that, by virtue of Footnote 1 to its Schedule, the total ceiling it had bound itself to achieve was not 1,273,500 tons but this amount plus a maximum of an additional 1.6 million tons. Footnote 1 reads as follows: 'Does not include exports of sugar of ACP and Indian origin on which the Community is not making any reduction commitments. The average of export in the period 1986 to 1990 amounted to 1,6 mio t.' The EU essentially argued that the purpose of the footnote was to allow it to meet its commitments in the body of its schedule while continuing to subsidize exports of sugar of ACP and Indian origin up to 1.6 million tons. The panel and AB rejected this argument and ruled against the EU. As discussed further in Chapter 12, this case had major implications for the ACP countries that had benefitted from preferential access to the EU.
How important is agriculture?
The deadlock over agriculture in the Doha Round raises the question of what the opportunity cost was of putting agriculture so much at the centre of the Doha Round. As noted in Chapter 1, agriculture accounts for only a small share of global trade and an even smaller share of the GDP of the rich countries—less than 5 per cent. The political economy factors discussed above imply that in rich countries there is simply not a significant constituency that feels strongly about the transfers that are made to farmers. At the end of the day food is too small a share of the consumption basket and expenditures of most households. Moreover, many groups in OECD countries actively support agricultural support programmes on the basis of equity or re-distributional grounds, or see it as a matter of national (food) security. The reaction of net exporters in 2008 to the rapid escalation of world food prices bolstered the views of those who argued that countries should have domestic food production capacity. Many developing countries went into the Doha Round ambivalent about global agricultural liberalization because they feared they might lose as a result of rising prices as the global price suppressing effects of OECD
302
the political economy of the world trading system
protection was removed. Although this would benefit their farmers, the historical pattern of taxation of the farm sector in many low-income countries indicates that farmers weigh less heavily than do urban consumers of food. (Although simulation models generally concluded that the overall impact on prices from liberalization would be limited on an annual basis-—given that the results of the MTN would be implemented over many years—this never appeared to have much of an impact on those arguing that a Doha Round would be bad for net importers.)
Agriculture is very important for countries with a comparative advantage in the sector, many of which are developing countries (Hertel and Keeney, 2006). Many of the poorest people in developing countries depend on agriculture and higher prices can therefore have major implications for poverty reduction (Hertel and Winters, 2006). The problem is that many of the groups concerned are in countries that are not major markets and which therefore have little to offer in a MTN. Incentives were skewed further by the fact that a large number of the poorest countries were not going to make any market access concessions (see Chapter 12): they essentially removed themselves from the quid pro quo bargaining game altogether. They did not need to bargain because they had duty-free, quota-free access to major markets such as the EU (under its Everything But Arms, EBA, initiative) and a number of other OECD countries. But most important from a development and poverty reduction perspective is that what matters most is to enhance productivity and reduce trade and transactions costs for the countries concerned. This is a domestic policy reform and investment agenda, and only indirectly a function of the policies of OECD countries (Chapter 12).
The Doha Round experience raises questions as to whether MTNs are capable of generating significant additional liberalization or whether they may not be better used as a mechanism to lock in national reforms that have resulted from a domestic political process. The issue linkage literature discussed in Chapter 4 suggests that it makes sense—-and indeed, will be necessary—to link agriculture to a broader agenda. The empirical question is whether the resulting negotiating set has enough in it to induce movement on agriculture.
Paarlberg (1997) has argued against issue linkages when it comes to agriculture. In his view the Uruguay Round neither facilitated nor motivated agricultural liberalization beyond what had been decided in the EU through the MacSharry reforms of 1992 and the domestic US reforms of 1990 and 1995-6. Before the Uruguay Round, Runge and von Witzke (1990) predicted that EU expansion to include Eastern European countries of Poland and Slovenia would be the source of demand for liberalizing the CAP. Budgetary pressures, linkages with new issues such as the environmental consequences of the CAP and the emergence of interest groups around these issues, and the long-term decline in the power of agricultural lobbies through attrition in the number of people working the land did force a rethinking of institutions such as unanimity in voting for policy changes (that existed in the EU before the 1992 reforms).
sector-specific multilateral trade agreements 303
These analysts suggest that the basic driver of reform in this sector must be national (and, increasingly, as countries pursue regional integration of markets, regional). Doha suggests that aggressive attempts to force multilateral liberalization may not have much success—in the end, what was on the table was a significant package of additional binding of past and ongoing national reforms, but little in the way of additionality once flexibilities and sensitivities are taken into account. International pressure through MTNs may play a useful role in helping to push along domestic reforms that are already being considered, but the experience to date in the WTO suggests that issue linkage is very difficult to operationalize in practice. From this perspective the Uruguay Round outcome may have been an outlier in generating an agricultural deal because the formation of the WTO created a take-it-or-leave-it situation. In the Doha Round the threat of exclusion did not exist.
Matters were compounded by what was arguably an excessive focus on reduction of applied levels of protection. If this is not achieved and the MTN breaks down, this has a high opportunity cost: it comes at the expense of not achieving greater lock-in of national/regional policy reforms. History illustrates that a focus on locking-in policy reforms can be very valuable when the economic situation deteriorates and pressures for protection rise as was the case in 2008-9. The emphasis that was put in the Doha Round on actual liberalization of agricultural trade may therefore have been an example of letting the best become the enemy of the good—as the opportunity costs of nonagreement were significant, including absence of progress on NAMA and services, which together account for more than 95 per cent of the trade of most countries.
6.2. Textiles and Clothing
Starting in the late 1950s trade policies towards textiles and clothing imports were gradually exempted from many GATT 1947 disciplines. Being labour-intensive and requiring relatively low technology inputs, the production of textiles and clothing is an activity in which many developing countries have a comparative advantage. Indeed, for a large number of countries this sector is the entry point into the production of manufactures. As domestic industries in high-income nations came under pressure from cheaper imports, initially from lapan, and subsequently from other Asian countries, they successfully lobbied for trade restrictions. Bilateral, discriminatory trade restrictions steadily expanded in terms of product and country coverage, and by the early 1990s a global web of QRs existed.
304      the political economy of the world trading system
Protectionism was driven by a desire to maintain employment of unskilled or semi-skilled workers. Textile and clothing industries were often regionally concentrated, and accounted for a substantial share of total manufacturing employment in many OECD countries in the 1960s. Trade protection slowed down the adjustment process in OECD countries, but did not stop it. Total employment in the sector declined steadily over time. Trade policy therefore can be seen as attenuating pressure from imports, giving industries more time to adjust (downsize, improve productivity). The policy came at a high economic cost, however, and one that was inequitably distributed. The price-increasing effect of protection impacted especially hard on lower income groups. For example, estimates for Canada revealed that in relative terms the burden of protection was four times higher for low-income consumers than for higher income groups (UNCTAD, 1994).
It was on the occasion of Japan's accession to GATT in 1955, at that time still a developing economy and a major exporter of textiles and clothing, that the concept of market disruption was first extensively discussed in the GATT. The first step towards formalization of a system of managed trade in this sector was the Short-Term Arrangement on Cotton Textiles, introduced during the Dillon Round (1961). This rapidly evolved into a Long-Term Arrangement (1962), which in turn led to four successive Multifibre Arrangements (1974-94) (Table 6.3). The discriminatory character of the MFA was progressively intensified and country and product coverage considerably extended. Initially limited to cotton fabrics, over time wool, man-made fibres, vegetable fibres and silk blends were added. By 1994 MFA-IV had 45 signatories, including 31 developing and Central and Eastern European countries that exported textiles and clothing, and eight importers. Among these, Austria, Canada, the EU, Finland, Norway and the United States applied restrictions, whereas Japan and Switzerland did not.9 Exporters were subject to bilaterally agreed quantitative export restrictions or unilaterally imposed import restraints. As textiles and clothing accounted for about 45 per cent of total OECD imports from developing countries in the early 1980s, it was the MFA and not MFN that was the cornerstone of the institutional framework for North-South trade.
Determining the impact of the MFA is quite complex. Although it was clearly very detrimental to the most efficient suppliers (such as China), to some extent the losses imposed on developing country exporters were reduced because the quotas were generally enforced by the exporters themselves. Insofar as the quota was
9 On the export side, MFA-IV covered Argentina, Bangladesh, Brazil, China, Colombia, Costa Rica, Czech Republic, Dominican Republic, Egypt, El Salvador, Fiji, Guatemala, Honduras, Hong Kong, Hungary, India, Indonesia, Jamaica, Kenya, Macao, Malaysia, Mexico, Oman, Pakistan, Panama, Peru, Philippines, Poland, Republic of Korea, Romania, Singapore, Slovakia, Slovenia, Sri Lanka, Thailand, Turkey and Uruguay.
sector-specific multilateral trade agreements
305
Table 6.3. A chronology of managed trade in textiles and clothing Date Event
1955 Japan introduces 'voluntary' export restraints (VERs) on cotton textiles shipped to
the US. Restraints are continued in 1956. 1956-60       The UK imposes VERs on cotton textiles from Hong Kong, India and Pakistan.
1961 The US textile and clothing industry makes its support for the 1962 Trade Act and
the Kennedy Round conditional on interim restrictions to deal with 'market disruption' caused by surges of imports from low-cost countries. The Short-Term Arrangement on Cotton Textiles is negotiated in July 1961.
1962 . The Long-Term Arrangement regarding International Trade in Cotton Textiles (LTA)
imposes a 5 per cent growth limit on imports of cotton products and places an important portion of the North-South trade in textiles under a managed trade regime.
1967 The LTA is extended for three years.
1970 The LTA is extended for another three years.
1973 To gain the support of the textile industry for the 1974 Trade Act (granting
negotiating authority to participate in the Tokyo Round). The US Administration persuades major developing-country garment exporters to accept a Multifibre Arrangement (MFA). The MFA limits the growth of textile and clothing imports to 6 per cent per annum.
1974 A Textile Surveillance Body is created to supervise the implementation of the MFA
under the auspices of the GATT textile committee, which is composed of the parties to the arrangement 1977 An extension is agreed for a five-year period (MFA-II), including a provision for
'jointly agreed reasonable departures' from MFA rules under special circumstances.
1982 MFA-III is negotiated, extending the arrangement for five more years. The
'reasonable departure' clause is dropped.
1985 Developing countries covered by the MFA establish an International Textile and
Clothing Bureau to promote the elimination of the arrangement and the return of trade in textiles and clothing to the GATT.
1986 The MFA is extended until 1991 (MFA-IV). 1991 The MFA is extended again until 1994.
1995 The Uruguay Round Agreement on Textiles and Clothing (ATC) sets out the rules for
a transition process, which is expected to result in 2005 in the full integration of textiles and clothing, into the GATT system.
2005 The ATC provides for its own termination on 1 January 2005.
binding, this implies that rents were being transferred to the exporters that had obtained licences to export (see Annex 2). Estimates of the magnitude of these quota rents are difficult to obtain as few countries auctioned off the quota licences or established markets in which quota allocations could be traded. An exception was Hong Kong, where quota prices for constrained items such as dresses, woven parkas, knitted pullovers and cotton sweaters ranged from US$6 to US$40 per dozen in 1996-7 (Spinanger, 1999).
306      the political economy of the world trading system
The MFA created strong incentives for geographic diversification of textile and clothing production. For example, as Hong Kong became more constrained by QRs, Chinese investors established production facilities in other countries such as Mauritius, which then became significant exporters. During the MFA years, a pattern of 'quota-hopping' FDI emerged as newly constrained firms set up shop in markets that were not (yet) constrained. A number of developing countries therefore benefitted from the quota regime by obtaining a 'guaranteed' market in the US or the EU. Such countries were often higher cost suppliers than large producers such as China, and confronted the prospect of increasing competition if the MFA were to be abolished.
Another effect of the MFA was that it created incentives for quality upgrading. Given that the VERs constrained quantities (number of shits, etc.), suppliers that were restricted could earn more if they could increase the unit values of the products they shipped. Harrigan and Barrows (2006), in a study of the effects of the removal of MFA restrictions as a result of the ATC (see below), estimate that the average price of textile and clothing imports by the US from all exporters fell significantly in product categories that had been subject to restrictions.
Bringing textiles and clothing into the fold
As in the case of agriculture, it was only in the Uruguay Round that textiles and clothing were seriously discussed in a MTN. The reasons were not the same, however. In agriculture, important factors were the financial burden of agricultural support programmes and the trade tensions that these programmes had caused. In textiles and clothing there was no pressure from OECD Finance Ministries. Although consumer organizations in high-income countries undoubtedly did not welcome the cost-increasing effect of the MFA, their voice was barely heard. The main common element was pressure from exporters, in particular those countries that perceived they would do better under a more competitive (less managed) trade regime. An implicit link was established between the demands by the US and the EU to address issues such as services and TRIPS in the Uruguay Round, and the desire of many developing countries to see an improvement in the market access conditions for their manufactured exports, in particular clothing.
Not surprisingly, negotiations were quite difficult. Major areas of disagreement concerned the application of general GATT rules, the modalities of phasing out of MFA restrictions, the duration of the transitional period and its product coverage, and the need for special safeguards. However, these areas were all addressed without the type of brinkmanship that characterized the agricultural negotiations. The ATC stipulates that the MFA was to be phased out over a ten-year period (1995-2004) and that standard GATT rules prohibiting the use of QRs and VERs (see Chapter 9) would apply. Products covered by the ATC were to be integrated
sector-specific multilateral trade agreements 307
into GATT in four stages. In 1995, at least 16 per cent of HS categories that were subject to MFA restrictions in 1990 were to be 'integrated'—i.e. no longer be subject to QRs. In 1998 (stage two) another 17 per cent of tariff lines would be integrated, followed by a further 18 per cent in 2002 (stage three) and the remaining 49 per cent by the end of 2004.
The ATC implementation strategy followed by the US and the EU complied with the letter, if not the spirit of the agreement. Very few textile or clothing categories that are important for developing countries were liberalized in the first stages of the MFA abolition (Spinanger, 1999). The EU and the US carefully chose to liberalize categories where imports were either already unrestricted or were relatively capital-intensive. Virtually all of the liberalization of the politically sensitive items was left for the final stage—the end of 2004. Not surprisingly, this gave rise to concerns on the part of developing country exporters regarding the implementation of the agreement.
Supervision of the implementation of the agreement was in the hands of a Textiles Monitoring Body (TMB), comprising an independent chairperson and 10 individuals who were broadly representative of the WTO membership, balancing export and import interests. Textiles Monitoring Body members rotated periodically and were expected to act on a personal basis. The TMB had a conciliatory and semi-judicial role. It examined all measures taken under the ATC, and their conformity with the agreement's rules and programmes for integration and liberalization. Matters on which agreement could not be reached could be brought to WTO dispute settlement. As noted in Chapter 3, a number of textile-related disputes were brought to the WTO after 1995. Indeed, one of the first cases to be brought by a developing country (Costa Rica) concerned US restrictions on this sector.
Implementation of the ATC resulted not only in the abolition of QRs, but also in the demise of the special, bilateral safeguard measures permitted under the agreement. The ATC contained a special safeguard clause in Article 6, which could be invoked during the implementation period of the ATC (that is, up to 2004) for products being integrated into the WTO. Under the ATC, safeguard actions could be discriminatory, were subject to a less stringent injury criterion and did not require compensation of affected exporters. Actions could be taken if imports of a product increased so much as to cause serious damage, or threat thereof, to the domestic industry producing like (or directly competitive) products. Damage indicators included standard economic variables such as output, productivity, capacity utilization, inventories, market share, exports, wages, employment, domestic prices, profits and investment (Article 6.3 ATC). Transitional ATC safeguard actions could be applied on a discriminatory basis, in contrast to measures taken under the Agreement on Safeguards. They required demonstration of a sharp and substantial increase in imports, actual or imminent, from the targeted countries. Measures were not to exceed three years duration or until the product is integrated into GATT 1994, whichever came first.
308      the political economy of the world trading system
Over 30 safeguard actions under the ATC were taken between 1995 and 2002, mostly by the US against developing country textile exporters but also by developing countries against each other (for example, Brazil was an active user). The measures were reviewed by the TMB, and many were rescinded. Three ATC safeguard actions led to WTO disputes, all involving the US. In all three cases, the panels, supported by the AB, concluded that the US had violated the provisions of the ATC.10 The panels signalled that the transitional safeguards in ATC were to be regarded as exceptional instruments and that members invoking this provision of the ATC had to be in full compliance with the various criteria laid out in the agreement. As from the end of 2004, safeguard measures on trade in textiles must be compatible with WTO rules—that is, be applied on a nondiscriminatory basis and conform to other WTO rules.
The back-loaded nature of ATC implementation created the possibility that importing countries might not remove all QRs by 2004. To reduce the probability of this occurring, the ATC required that quotas grow substantially over the ten-year transition. This ensured that import-competing industries would gradually be subjected to more competition. Quotas were to grow by 16 per cent in stage one, 25 per cent in stage two and 27 per cent in stage three. Thus, a 6 per cent permitted growth rate in 1994, became 7 per cent per year during 1995-7; 8.7 per cent during 1998-2001; and 11 per cent per year during 2002-4.
The ATC was implemented as scheduled at the end of 2004, albeit in a rather messy way, with continued restrictions being imposed on China in particular. China's protocol of accession allows WTO members to take product-specific safeguard actions on the basis of 'market disruption' rather than the more constraining 'serious injury' criterion required by the WTO for regular safeguards for a 12-year period (until 2013). In principle, the impact of the elimination of QRs in this sector depends on how constrained the most efficient exporters were. Because of the 'voluntary' nature of the restrictions, the restrictiveness of the MFA (and later ATC) has been measured in the literature as an export tax equivalent: the implicit tax on exports that is associated with the quantitative limit imposed by the importing country. These taxes were on the order of 20-50 per cent for China, and, in the case of the US, were estimated to have increased during the ATC implementation period (Francois and Worz, 2007). The reason was the huge increase in export potential of China as a result of sustained high growth in that country. The export potential greatly surpassed ATC quota growth rates, resulting in policy becoming more restrictive over time,
Under the ATC, policy was more restrictive towards China than other exporters. China's QRs were more likely to be binding; grew at a slower rate; and were subject
10 United States: Restrictions on Imports of Cotton and Man-Made Fibre Underwear from Costa Rica, WT/DS24/AB/R (10 February 1997); United States: Measures Affecting Imports of Shirts and Blouses from India, WT/DS33/AB/R (25 April 1997) and US-Cotton Yarn from Pakistan (WT/DS/192).
sector-specific multilateral trade agreements
309
to greater constraints in terms of the ability to shift QRs across product categories and time (Brambilla, Kandelwal and Schott, 2007). Francois and Worz (2007) conclude that as measured by export tax equivalents, the US did not implement the ATC as envisaged by negotiators, because protection increased between 1996 and 2004 for 15 of the exporting countries that were subject to restrictions. In contrast, Canada implemented the ATC according to plan—export tax equivalents had dropped to zero by 2004. In the EU, the average tax equivalent for clothing fell from 13 to 3.6 per cent between 1996 and 2004, except for China, which still confronted an equivalent tax of 19.4 per cent in 2004.
The implication of removing what were still high export tax equivalents at the end of 2004 was clear: large export surges. Chinese exports to the US increased by 39 per cent in 2005, with exports of formerly quota constrained items rising by 270 per cent (Brambilla, Kandelwal and Schott, 2007). In response the EU and the US re-imposed restrictions in the form of negotiated export growth quotas with China. Dayaratna-Banda and Whalley (2007) argue that the result of implementation of the ATC was to shift from a general quota regime affecting all (competitive) developing countries to one where the major markets targeted restrictions on the major supplier—China. Once the provisions of China's accession protocol have expired, WTO members will have to limit themselves to standard instruments of contingent protection to protect domestic producers. As discussed in Chapter 9, China is already the primary target of antidumping actions.
The demise of the MFA was a major achievement, not least because the agreement to reintegrate this sector into the GATT reflected a major change in the negotiating strategy of developing countries. They insisted that progress in this area was a quid pro quo for the TRIPS agreement and the GATS. Full liberalization of trade was not achieved, of course. Tariffs remain much higher on textile and clothing products than most other manufactures. In the Uruguay Round the trade-weighted tariff average in developed countries for these products fell to 12.1 per cent, down from 15 per cent. Access conditions for developing countries continue to differ as a result of PTAs with, and unilateral preference programmes of, OECD countries. These create incentives for so-called outward processing trade and related investments in this sector by providing duty- and quota-free access for products that satisfy the applicable rules of origin (see also Chapters 10 and 12).
A major motivation for these preferential access regimes on the part of developing countries is to attenuate the competitive impacts of China and other more efficient exporters becoming less constrained by QRs. The MFA and the ATC created rents for suppliers that were not constrained. As the ATC was implemented, less efficient suppliers saw market shares erode as a result of greater competitive pressure. This was particularly marked for Sub-Saharan African countries, which confront especially high trade costs. Countries such as Lesotho that had increased exports of clothing to the US fourfold between 2000 and 2004, driven by liberal rules of origin and duty-free access under the African Growth and Opportunity
310       the political economy of the world trading system
Act, saw exports decline by 43 per cent in items for which China was constrained in 2005 (Brambilla, Kandelwal and Schott, 2007).
6.3. The Information Technology
Agreement
Both the agreements on agriculture and on textiles and clothing are multilateral agreements—they apply to all WTO members. Both are sector-specific and the objective in both cases is to (re)integrate these sectors into the WTO. They differ in that the ATC was a time-bound agreement that expired in 2005. As mentioned in Chapter 4, another type of agreement that may be negotiated under the WTO involves the elimination of barriers for subsets of products. These so-called zero-for-zero agreements became prominent in the Uruguay Round, and continue to be strongly supported by industry groups. Examples of sectoral zero-for-zero agreements concluded in the Uruguay Round—under which subsets of (mostly) OECD countries agreed to eliminate tariffs, either immediately or following a transition path—included deals on agricultural, construction and medical equipment, beer, furniture, paper, pharmaceuticals and toys (Mann and Liu, 2009).
The most prominent example of a zero-for-zero deal that was incorporated into the WTO is the Ministerial Declaration on Trade in Information Technology Products, generally called the Information Technology Agreement (ITA), concluded in 1997 by 39 countries accounting for 90 per cent of world information technology (IT) trade. Participants agreed to eliminate tariffs over a three-year period on almost all IT products on a MFN basis. The major product categories covered by the agreement include computers, parts and accessories, telecommunication equipment (including modems, pagers and fax machines), semi-conductors, semi-conductor manufacturing equipment, and certain software and scientific instruments. Consumer electronics are excluded. Tariffs were cut in four equal installments, with developing country signatories having until 2005 to eliminate tariffs on certain items. Other duties and charges were to be abolished upon the entry into force of the agreement.
The ITA was driven by a coalition of IT firms and industry associations that sought to eliminate barriers to trade in their products and used a variety of international non-WTO mechanisms to build a constituency for liberalization. The ITA was the first liberalization agreement concluded after the Uruguay Round (the agreements on finance and telecoms concluded in 1997 were not stand-alone but continuations of Uruguay Round negotiations that could not be brought to closure during the round). At the time the agreement was concluded, some
sector-specific multilateral trade agreements
311
observers noted that such issue-specific, targeted agreements illustrated that the WTO, in contrast to the GATT, could make progress on liberalization without launching a round. It also suggested that 'privileged groups' of the kind discussed in Chapter 4 could be constructed, that is, deals involving enough players with an interest on an issue to allow free riding by nonparticipants to be discounted.
A closer look at the ITA negotiating history suggests that there is little reason to believe that it will (or should) be a model for future liberalization initiatives under WTO auspices. Discussions on the coverage of the agreement were contentious, and numerous linkage strategies were employed by participants in efforts to ensure that a 'balance of concessions' would be attained. The IT 'sector' spans many different products, and much of the negotiation involved discussions regarding the coverage of the agreement. The EU insisted that trading partners offer concessions on market access for alcoholic beverages as a condition for signing the ITA. Developing countries attempted to obtain concessions on textiles, although at the end of the day their negotiating leverage proved insufficient.11
The product coverage of the ITA ended up reflecting primarily the interests of the Quad. It deals only with a subset of the policies affecting trade in the IT products selected: the ITA is a tariff-only agreement. It proved impossible to address any nontariff policies affecting market access. Consumer electronics-— products of greater interest to developing countries as suppliers than to the Quad—were not included under the ITA. The ITA is therefore a rather unbalanced agreement from a developing country perspective. Elimination of tariffs on the products included under the ITA will be of benefit to consumers in developing country signatories (including foreign investors), but no quid pro quo was obtained. Standard reciprocity and 'internalization' considerations of the type discussed in earlier chapters were important in the ITA. However, given that any deal would have to be applied on a MFN basis, the US insisted that the ITA signatories must cover at least 90 per cent of total production of the IT products included under the agreement.
The ITA was to a large extent the brainchild of major IT companies in the EU and US. These firms and their industry associations—the US Information Technology Industry Council, the European Association of Manufacturers of Business Machines and Information Technology Industry (EUROBIT) and the Japanese Electronic Development Association—were prime movers behind the initiative to eliminate tariffs on their products. They jointly developed recommendations for the February 1995 G7 ministerial conference on a Global Information Society, proposing that tariffs on the building blocks of the infrastructure of such a society be abolished by 2000. Industry groups continued to push the idea in the context of other fora, especially the Transatlantic Business Dialogue (TABD) and APEC. This
11 At the time, some were calling the ITA the Information, Textiles and Alcohol Agreement (Fliess and Sauve, 1998: 62).
312
the political economy of the world trading system
constellation of interests was successful at the end of the day in abolishing tariffs on much of their output, but this success came at a cost. It essentially involved taking care of the concerns of a set of large and powerful enterprises in mostly OECD countries. These firms will in future have less of an incentive to support more general liberalization of their home markets.
The Declaration calls for periodic review of the product coverage of the ITA. The first such review took place very soon after the ITA was agreed. The associated talks and subsequent reviews are often described as working towards an TTA-2'. To date, no agreement has proved possible on extending the product coverage and going beyond tariffs to cover disciplines on NTMs. Extension of product coverage has been contentious because of disagreements on where to draw the boundary between an TT product5 and other electronic products. Problems have also arisen on how to classify products that use new technologies and that were not explicitly listed in the original ITA schedules of signatories.
Disagreements between ITA members on such classification issues led to a formal dispute in the WTO in May 2008. The US and Japan contested decisions by the EU to impose tariffs on products such as cable boxes that can access the internet, certain flat panel LCD monitors, and computer printers that can also scan and fax documents. These tariffs created an inducement for export companies to assemble the final product inside the EU. The tariffs averaged 10 per cent, and were supported by new EU member countries such as Poland that had attracted substantial FDI in the affected high-tech sectors.
6.4. Conclusion
If GATT was perceived to lack teeth, it was in part due to the de facto exclusion of trade in agriculture and textiles and clothing from the reach of its disciplines. The agreements reached in the Uruguay Round therefore constituted a significant step forward in the process of reasserting the relevance of the general principles of nondiscrimination and open markets. Without these agreements the WTO would have been much less credible as an organization.
The examples of both agriculture and clothing have much to teach about the political economy of multilateral liberalization and negotiation. The agriculture case illustrates that if domestic lobbies are strong and can mobilize the support of other groups (who may be primarily driven by quite different objectives, including noneconomic reasons), multilateral cooperation can break down. Standard reciprocity does not work in the sense that intrasectoral tradeoffs are not feasible. The domestic interests seeking better access to foreign markets could be and were
sector-specific multilateral trade agreements 313
satisfied through negotiations that were limited to manufactures. The potential gains from trade in policies affecting market access for manufactures were more than large enough to allow significant progress to be made in reducing barriers to trade in manufactures. No linkage was required with agriculture, and US attempts to impose such linkages in recurrent MTNs failed because they were not credible. The cost of total breakdown of a MTN because of lack of agreement on agriculture was simply too great.
Progress was made in the late 1980s on agriculture because new interest groups appeared that sought to control agricultural support programmes. In the case of the EU, these included finance ministries. The adoption of the Maastricht treaty, which set targets for government deficits and public debt in the run-up towards European Monetary Union and the prospect of future enlargement of the EU maintained serious pressure on agricultural expenditure. The emergence of environmental lobbies also played a role. An increasing awareness of the environmental downside of intensive and polluting farming encouraged by existing production support policies helped to undercut support for production-increasing policies. Last but not least, the emergence of the Cairns Group was an important factor. It was less inclined than the US to compromise, as the issue was vital to export interests of the group, and therefore could act as both a proposal making and a blocking coalition.
Although internal political dynamics played an important role in reintegrating agricultural trade policies into the GATT/WTO, the power of the agricultural lobby remains very strong. Tariff protection remains formidable in many WTO members. Noneconomic considerations continue to play a major role in domestic and international discussions on agriculture. The AoA states that nontrade concerns' must be taken into account in future efforts to liberalize trade in this sector. The preamble to the agreement identifies food security, protection of the environment and ensuring the viability of rural areas as examples of objectives that may be realized through agricultural policies. The 'multifunctionality' of agriculture is frequently used by the EU and other WTO members as a justification for agricultural intervention. Where to draw the line regarding the 'legitimacy' of agricultural policies is an ongoing source of debate. However, there is no compelling rationale for permitting the use of trade measures to attain noneconomic objectives.
The Doha Round made clear how hard it is to use a MTN to generate significant additional liberalization of agriculture. The differences in interests across and within countries are great, and putting together a package of tradeoffs within the sector that was acceptable to the major players proved very difficult. The issue linkage literature discussed in Chapter 4 suggests that in such situations it is necessary to link agriculture to a broader agenda. With the removal of the Singapore issues in 2004, such linkage was largely limited to concessions by the major agricultural exporters on manufactures and/or services. In principle a good case can be made that there should have been enough on the table. In mid-2008
3H
the political economy of the world trading system
negotiators had come close to a deal on agriculture and NAMA modalities, but at the end of the day agricultural sensitivities led to yet another breakdown of talks— specifically the design of a special safeguard mechanism for agricultural imports.
The focus of the WTO is on policy bindings—the maximum level of protection that can be accorded to goods or services. Although what matters for exporters are applied levels of protection, this is not the focus of WTO negotiations. For many developing countries applied trade policies are much more liberal than is implied by their commitments in the WTO. As a result, deep 'liberalization' commitments associated with a specific formula to cut tariffs may not do much, if anything, to lower applied rates of protection. The extent of the 'binding overhang' is significant. For Brazil, for example, the import-weighted average tariff for agricultural merchandise is 40 per cent, compared to an applied average MFN rate of 10 per cent. Similar ratios prevail for other countries.
An important determinant of the perceived value of such bindings is what traders expect to happen in the future—is it likely that governments will raise tariffs above applied levels? Absent tariff bindings that are at, or close to, applied rates, under the WTO rules they are free to do so. If the expected probability of 'backsliding' is low, negotiators will be either (1) pressured to seek very deep cuts in bindings so as to reduce actual levels of protection; or (2) not pressured much insofar as exporters confront relatively low applied tariff barriers and see only a limited net benefit associated with further reductions (given the costs they need to expend to lobby for the cuts to be achieved) (Hoekman and Vines, 2007). Deep cuts in bindings may be resisted by negotiators as 'giving up too much'—in the Brazil case just mentioned to get close to the current applied average tariff, the cut in bound tariffs would have to be some 75 per cent. This is hard to sell politically at home—it appears to be a lot, even though such a cut would not do much to reduce actual levels of protection. The insistence by developing countries that they make less deep liberalization commitments than high-income countries makes it more difficult to agree on a level of reduction that would be meaningful in terms of actual liberalization. In principle there should be a feasible deal that links agriculture, manufactures and services, but the experience to date suggests that such a deal may need to put much greater emphasis on the value of binding unilateral (and regional) liberalization.
A different story applies in textiles and clothing. Here there were also powerful lobbies in OECD countries that were successful in obtaining protection. But there were no direct budgetary implications that created pressure to abolish such protection. Although regressive in income distribution terms, protection of textiles and clothing was not subject to strong opposition from consumer groups, in part because competition was not choked off completely as exporters diversified across developing country locations and firms in the domestic industry improved their productivity or exited. The explanation for the agreement to integrate textiles and clothing into the GATT in this case is more in line with standard reciprocal
sector-specific multilateral trade agreements
315
negotiating dynamics. Developing countries insisted on liberalization as a quid pro quo for agreeing to accept the TRIPS and GATS agreements. This gave the lobbies in the US and the EU> who sought disciplines on services policies and stronger enforcement of intellectual property law, an incentive to confront the domestic clothing industry. This industry had in any event become smaller, more specialized and itself increasingly engaged in international production. The lower quality garments industry had declined substantially in size in both the US and the EU in the 1980s, reducing its political clout. Here there is also much less scope to raise 'multifunctionality' concerns and argue that trade policies are required to meet noneconomic objectives. Nonetheless, liberalization will occur slowly. Textile and clothing tariffs continue to offer domestic industries levels of protection that greatly exceed those applying to other manufacturing industries. Moreover, import-competing industries are very well aware of the existence of antidumping and safeguard instruments.
6.5. Further Reading
T. Warley, 'Western Trade in Agricultural Products', in International Economic Relations in the Western World 1959-71 (London: Royal Institute of International Affairs, 1976) provides a historical overview of agricultural policies and trade of OECD countries. L. Alan Winters, 'The Political Economy of the Agricultural Policy of Industrialized Countries', European Review of Agricultural Economics, 14 (1987): 285-304, discusses the question of why farmers have been able to obtain high levels of protection. The same author, in 'The So-called Noneconomic Objectives of Agricultural Support', OECD Economic Studies, 13 (1989): 238-66, critically addresses the rationales that have been offered for such policies. Anne Krueger, Maurice Schiff and Alberto Valdes, 'Agricultural Incentives in Developing Countries: Measuring the Effect of Sectoral and Economy-wide Policies', World Bank Economic Review, 2 (3) (1988): 255-72, analyse the effects of agricultural and nonagricultural policies on farmer's incentives in developing countries.
Tim Josling, Stefan Tangermann and T. Warley, Agriculture in the GATT (London: Macmillan, 1996) provide a comprehensive treatment of 40 years of discussions and negotiations on agriculture in the GATT, and include a summary evaluation of the Uruguay Round Agreement on Agriculture. Richard Higgott and Andrew Cooper, 'Middle Power Leadership and Coalition Building: Australia, the Cairns Group and the Uruguay Round', International Organization, 49 (1990): 589-32, discuss the formation and operation of the Cairns Group. Tim losling, 'Agriculture in the Next WTO Round', in Jeffrey Schott (ed.), The WTO After Seattle