4 Realizing the Development Promise of Trade

International Monetary Fund
Published Date:
April 2005
  • ShareShare
Show Summary Details

Removing barriers to trade that discriminate against developing countries is potentially a powerful tool to help achieve the Millennium Development Goals (MDGs). Multilateral, reciprocal, nondiscriminatory trade liberalization offers the best approach for supporting development. Rapid conclusion of an ambitious Doha Round is therefore of great importance. As noted in Global Monitoring Report 2004, the round should spawn a major reduction in trade-distorting policies. Ambitious reference points could be helpful in guiding the negotiations, such as capping the agricultural tariffs of developed countries at 10 percent, eliminating their agricultural export subsidies and manufacturing tariffs, fully decoupling domestic agricultural and rural support policies from production, and pursuing meaningful liberalization of trade and investment in services, including through temporary migration of service providers. Equally important, developing countries need to set ambitious targets to further reduce the average level and dispersion of protection and substantially expand their World Trade Organization (WTO) commitments. To assist in attaining the MDGs, such actions should be fully implemented by 2015, with significant progress before 2010.

A surge in ambition is needed to realize the promise of the Doha Development Agenda. The ambition must focus on transforming agricultural trade policies in OECD countries, devoting greater attention to reducing the trade-restricting effects of nontariff measures, and safeguarding and expanding the scope for developing countries to contest services markets. Given that trade restrictions are much higher in developing than developed countries, significant further liberalization by developing countries will also be required to realize the full potential of trade for development. Indeed, about half of the potential developing country gains from merchandise trade reform will come from reforms by developing economies.

An ambitious Doha Round would benefit the world as a whole and developing countries as a group. But it would not be a panacea, and it would not be costless. Trade is a key driver of growth, but it needs to be complemented by many other policies. In many developing countries “behind-the-border” policies and the investment climate are binding constraints on the ability to realize the benefits from greater trade opportunities. All countries will incur adjustment costs from deep trade reforms, and some developing countries will confront erosion of current tariff preferences. Such changes should be considered within the broader context of development and poverty reduction strategies, with special attention to the concerns of poor households. The standard argument for trade liberalization is that total gains exceed total losses—especially over time—and that gainers can compensate losers while still improving their welfare. But in practice such compensation does not occur. An ambitious Doha Round would generate substantial gains, providing a basis for transferring additional resources to low-income countries to enhance trade capacity. Given that lack of competitiveness due to high operating costs and weak investment climates is the primary factor limiting trade potential in many poor countries, actions to lower such costs are urgently needed, especially in Africa. The agenda includes supporting trade logistics and facilitation, strengthening transport infrastructure, and further reforming policies that create antiexport bias—including import tariffs.

An ambitious Doha outcome—significant liberalization commitments by both developed and developing countries—complemented by a credible commitment to convert part of the resulting gains into increased aid to the poorest countries to enhance their trade capacity—would send a strong signal that political will exists to leverage trade to help achieve the MDGs. This would help countries that have the capacity to export but are now constrained (such as middle-income agricultural exporters) as well as countries that have much less capacity and are unlikely to benefit much from the standard reciprocal exchange of market access commitments in the WTO.

Recent policy in OECD countries has emphasized tariff preferences for small, low-income countries, especially the least developed countries (LDCs) and Sub-Saharan African countries. While actions to make tariff preferences more effective in the short run could be beneficial—for example, through adoption of common, liberal rules of origin—consideration should be given to alternative forms of trade assistance that generate greater benefits for recipients and are less trade-distorting. Tariff preferences have been of limited value to many African countries and have negative effects on the functioning of the global trading system. Alternative measures include sharp reductions in most favored nation (MFN) tariffs and other barriers by high- and middle-income countries on goods and services of export interest to low-income countries, financial assistance, and actions to minimize the incidence of nontariff measures on exports from low-income countries.

“Aid for trade”—assistance to bolster trade capacity and reduce trade costs—can have high returns in terms of increasing growth, reducing poverty, and facilitating the realization of ambitious global trade reforms. The share of aid for trade in total aid commitments increased from 3.6 percent in 2002 to 4.2 percent in 2003.1 A concerted effort to further expand such assistance could do much to improve developing countries’ capacity to benefit from both global and domestic reforms. Numerous trade integration studies undertaken for the LDCs under the Integrated Framework for Trade-Related Technical Assistance, as well as similar analyses done in other developing countries, have identified areas where such aid can be used effectively. The Integrated Framework, a collaborative venture between six multilateral agencies, bilateral donors, and governments of LDCs, offers a mechanism to identify priorities and allocate additional assistance to trade-related investments and policy reforms. To date, dedicated resources to support implementation of the framework have been limited to small-scale technical assistance. Funding for trade priorities is considered in the context of the resource allocation and prioritization process related to country development strategies—poverty reduction strategies (PRSs) or equivalent national development strategies—with donor assistance complemented by loans and other support from international financial institutions. While this should continue to be the case, the Integrated Framework could be used more extensively to both bolster the trade capacity of poor countries and help address the adjustment costs from an ambitious Doha Round.

Regional cooperation is another potentially powerful instrument to leverage trade for development—as long as it does not detract from the pursuit of an ambitious Doha Round and focuses policy attention on domestic (intraregional) policy-related trade constraints. In Sub-Saharan Africa there is a strong regional dimension to growth, absorptive capacity, and the widening of economies of scale needed to lower infrastructure costs and facilitate trade. To fully realize the development promise of both North-South and South-South regional integration arrangements, developing country members of trade agreements should implement significant liberalization on a nondiscriminatory basis in addition to granting preferential access to partner countries. Many Sub-Saharan countries still rely on import duties for a significant portion of government receipts, so revenue concerns and the ability to put in place alternative revenue sources are factors that need to be taken into account in determining the speed of liberalization.

Trade Performance of Low-Income and Least Developed Countries

The past two decades saw a boom in world trade. Non-oil exports from developing countries grew faster than those from developed countries (table 4.1), and their global market share rose from 21 percent in 1980 to 37 percent in 2003 (table 4.2). This increase reflects the growing openness of developing countries to trade and the competitiveness of their exports in global markets. Openness ratios (the sum of exports and imports as a share of GDP) have increased for most developing regions and range from a low of 25 percent in South Asia to 64 percent in Europe and Central Asia.

TABLE 4.1Trade has grown rapidly in recent years, especially in developing countries (average annual growth rate, percent)
Country group1991–951996–20002001–03
Developing countries12.27.77.4
Least developed countries3.810.18.4
African least developed countries0.17.310.2
Low-income countries8.79.68.1
African low-income countries2.812.64.2
Developing countries13.35.38.1
Least developed countries5.83.312.0
African least developed countries3.11.913.3
Low-income countries9.24.412.8
African low-income countries5.20.916.1
Source: IMF, Direction of Trade Statistics.
Source: IMF, Direction of Trade Statistics.
TABLE 4.2Developing countries account for a growing share of non-oil exports (percentage of total)
Country group197519801990199520002003
Developed countries80.378.976.570.665.162.6
Developing countries19.721.123.529.434.937.4
Least developed countries0.
African least developed countries0.
Low-income countries3.
African low-income countries2.
Source: IMF, Direction of Trade Statistics.
Source: IMF, Direction of Trade Statistics.

The foreign exchange earnings of the poorest countries depend on agricultural commodities and labor-intensive manufactures. Although its average annual export growth has increased since the mid-1990s, Sub-Saharan Africa’s share of world trade remains far below what prevailed in the 1970s (see table 4.2). Reasons include the region’s high dependence on agriculture, limited trade opportunities caused by highly distorting policies in OECD countries, and downward global price trends for some commodities (such as coffee) due to expanding supplies.

Problems on the supply side of the market also help explain the below-average trade and foreign direct investment performance of many low-income countries. Civil conflicts have undermined economic performance in a score of African countries, and in others poor macroeconomic policies have often led to price disadvantages as a result of overvalued currencies.2 Moreover, governance problems, corruption, and institutional weaknesses have worsened the local investment climate and inhibited local entrepreneurs from taking advantage of market opportunities. Those same factors have also induced instability in commodity prices and had adverse effects on the demand for commodity exports.

Since the mid-1990s exports from low-income countries and the LDCs have grown faster than the average for the world and for developing countries (see table 4.1). This largely reflects an increase in commodity prices and more recently the decline in the U.S. dollar, as well as an increase in exports of manufactures (textiles). These developments contributed to a minor recovery in these countries’ world market shares, though these remain below those of the 1970s (see table 4.2). Global export shares for low-income African countries and for the LDCs as a group exceed their shares in world GDP.3 Average import growth rates have also picked up recently for low-income countries and the LDCs, outpacing growth in imports for the world and for developing countries during 2001–3.

Exports from the poorest countries remain concentrated in primary commodities. In 2001 the concentration index of Sub-Saharan Africa’s exports was 0.47—similar to that of the Middle East and North Africa, a region heavily dependent on oil exports.4 This is almost twice the figure for East and South Asia, and 50 percent higher than in Latin America. Agricultural products account for at least a quarter of exports from Sub-Saharan Africa and Latin America. As a result countries in both regions are subject to greater commodity price (terms of trade) volatility. The decline in nonfuel commodity prices and the increase in energy prices since 1995 have contributed to a decline in the shares of food and raw materials and an increase in the share of fuels in exports from the LDCs (figure 4.1). Expanding apparel exports from Africa are a noteworthy development of recent years. While reliable data on the composition of exports for 2003 are not yet available, the sharp increase in fuel prices during 2003–4 will likely further increase the share of fuels in LDC exports.

FIGURE 4.1LDC Exports: Less food and raw materials, more energy and apparel

Source: WTO (2004).

Recent Developments in Trade Policy

Global trade policy developments in 2004 provide a mixed picture. After a disappointing outcome at the WTO ministerial conference in Cancun, Mexico, in September 2003, there was considerable gloom about the prospects of the multilateral trading system. But in 2004 WTO negotiations came back on track, with a package of agreements reached in August. The WTO agreed on a framework for negotiations on agriculture and industrial products, a set of recommendations on services, and modalities for negotiating improved customs procedures. Other positive developments included proposals for reforming EU sugar and banana regimes, and an end—after some 40 years—to the regime of managed trade in textiles and clothing.

The proliferation of reciprocal preferential trade arrangements continued in 2004, with activity spread across all regions and countries. Since 1990 the number of preferential trade arrangements in force and notified to the WTO has risen from 50 to nearly 230,5 and another 60 are being negotiated. But there is growing concern that such arrangements create trade diversion and substantial administrative costs due to complicated rules of origin. Although many of the countries most active on this front are strong proponents of the Doha Round and consider the agreements a form of “competitive liberalization” that reinforces WTO-based liberalization efforts, preferential trade may create interests opposed to multilateral extension of market access opening and divert scarce negotiating resources. Moreover, while the agreements can help developing countries by reinforcing internal reforms and lowering trade costs, such benefits do not require preferential access to markets. As discussed later in this chapter, minimizing discrimination and focusing attention on cooperation to lower regulatory trade barriers is an important challenge from a development perspective.

The recent accession of 10 new members to the European Union consolidated the extensive network of intra-European trade agreements and expanded the EU internal market. The European Union is also pursuing association agreements with Common Market for the Southern Core (Mercosur) and Mediterranean countries, and is negotiating reciprocal economic partnership agreements with regional subsets of African, Caribbean, and Pacific (ACP) countries.

Over the past year the United States has also been active in concluding preferential trade agreements, including with Australia, Bahrain, Central America, the Dominican Republic, and Morocco. Several more are under negotiation. In Latin America, Mercosur signed an agreement with Andean countries and is negotiating with India, Mexico, and South Africa (in addition to the European Union). Mexico has entered into a trade agreement with Japan and is negotiating one with Singapore. Similarly, a South Asia free trade area is being negotiated between seven countries, and in Africa the customs union of the East African Community has been launched. The Association of South-East Asian Nations (ASEAN) has agreed to negotiate preferential agreements with China, India, and Japan, among others. Japan, the Republic of Korea, Singapore, South Africa, and Thailand have also been negotiating bilateral accords.6 The rising regional activity in East Asia is particularly noteworthy given these countries’ historical noninvolvement in preferential trade and reliance on the multilateral trading system. An increasing number of developing countries are seeking membership in regional agreements.

Developments also occurred for nonreciprocal preferences. The European Commission introduced a new Generalized System of Preferences scheme offering duty-free access to the EU market for 80 percent of dutiable tariff lines to countries that adhere to a number of international conventions on human rights, labor, the environment, and governance. The new scheme enters into force in April 2005 and provides for better product coverage, more relaxed rules of origin, and greater certainty and predictability in the granting of preferences. In the United States the passage of the African Growth and Opportunity Acceleration (AGOA III) Act extended the general timeframe for AGOA preferences until 2015. It also extended the third-country fabric manufacturing provision for least developed AGOA beneficiary countries until 2007—an important provision since few AGOA beneficiary countries have adequate capacity to manufacture fabric. The legislation provides temporary respite to textile and clothing sectors in several Sub-Saharan countries (including Lesotho, Madagascar, and Swaziland) as competitive pressures intensify following the elimination of all quantitative export restraints in January 2005.

Tariffs, Nontariff Measures, and Contingent Protection

In recent years there has been a decline in tariff barriers worldwide, reflecting autonomous reform, regional agreements, and WTO accession (including by Cambodia, China, Egypt, India, Iran, Pakistan, Tunisia, and Zimbabwe). Globally, average tariffs fell by 2 percentage points between 1997 and 2000 and by 1 point between 2000 and 2003. As measured by MFN tariffs, South Asia is the most restrictive region, followed by Sub-Saharan Africa (table 4.3).

TABLE 4.3Applied most favored nation tariffs are highest in South Asia and Sub-Saharan Africa (simple average, percent)
Region/income groupLate 1980s2004 (preliminary)
East Asia and Pacific18.89.6
Europe and Central Asia10.27.3
Latin America and the Caribbean22.410.4
Middle East and North Africa17.312.4
South Asia68.917.7
Sub-Saharan Africa25.115.3
High-income OECD7.04.1
High-income non-OECD9.06.6
Developing countries25.411.7
Least developed countries28.415.1
Low-income countries31.712.9
Middle-income countries21.810.5
Sources: UNCTAD TRAINS database; World Bank, World Development Indicators; IMF Trade Policy Information Database; WTO Trade Policy Reviews.Note: Classification of regions and income groups based on World Bank 2004b.
Sources: UNCTAD TRAINS database; World Bank, World Development Indicators; IMF Trade Policy Information Database; WTO Trade Policy Reviews.Note: Classification of regions and income groups based on World Bank 2004b.

By contrast, there has been no decline in non-tariff measures affecting trade, such as licensing requirements and antidumping investigations. (Licensing requirements are often associated with health-and safety-related regulatory requirements that apply in principle to both imports and domestic products; see below.) Nontariff measures apply to a large share of trade, with the highest ratios in developing countries (table 4.4). But as discussed below, when converted to ad valorem equivalents, nontariff measures are more important in rich countries.

TABLE 4.4Nontariff measures remain high in several regions, 2002 (percent)
Region/income groupFrequency ratio

(share of tariff lines)
Coverage ratio

(share of imports)
East Asia and Pacific44.032.8
Europe and Central Asia14.623.9
Latin America and the Caribbean26.944.4
Middle East and North Africa49.551.6
South Asia12.128.6
Sub-Saharan Africa24.428.8
High-income OECD22.024.1
High-income non-OECD45.146.9
Developing countries27.234.3
Least developed countries20.036.7
Low-income countries25.536.4
Middle-income countries28.334.0
Sources: UNCTAD TRAINS database; WTO Trade Policy Reviews; EU Product Standards Database; Groupe d’Economie Mondiale, Institut d’Etudes Politiques.Note: Classification of regions and income groups based on World Bank 2004b. Nontariff measures include price controls, quantity restrictions, state trading monopolies, and technical product regulations as classified in the UNCTAD TRAINS database.
Sources: UNCTAD TRAINS database; WTO Trade Policy Reviews; EU Product Standards Database; Groupe d’Economie Mondiale, Institut d’Etudes Politiques.Note: Classification of regions and income groups based on World Bank 2004b. Nontariff measures include price controls, quantity restrictions, state trading monopolies, and technical product regulations as classified in the UNCTAD TRAINS database.

The use of instruments of contingent protection—antidumping and safeguard actions—against imports by WTO members has grown steadily, with more than 2,400 antidumping investigations launched since 1995 (table 4.5). This instrument is increasingly being used by developing countries against other developing countries. Indeed, developing countries have overtaken developed countries in the number of cases launched: between 2002 and June 2004 developed countries initiated 190 cases, compared with 441 by developing countries. India has passed the United States as the top initiator, and Argentina and South Africa have an even higher incidence of antidumping use relative to the value of their imports (table 4.6).

TABLE 4.5Developing countries initiate more antidumping investigations, 1995–2003
Target region
Initiating regionDeveloped countriesDeveloping countriesTransition economiesTotal
Developed countries226574 (129)132932
Developing countries453827 (225)1731,453
Transition economies47 (2)2031
Total6831,408 (356)3252,416
Share of total (percent)
Developed countries24.261.6 (13.8)14.2100.0
Developing countries31.256.9 (15.5)11.9100.0
Transition economies12.922.6 (6.5)64.5100.0
Total/average28.358.3 (14.7)13.5100.0
Source: WTO Antidumping Committee data.Note: Numbers in parentheses are actions against China. Developed countries are Australia, Canada, EU15, Iceland, Japan, New Zealand, Norway, Switzerland, and United States. Transition economies are Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, FYR Macedonia, Georgia, Hungary, Kazakhstan, Kyrgyz Republic, Latvia, Lithuania, Moldova, Poland, Romania, Russian Federation, Slovak Republic, Slovenia, Tajikistan, Turkmenistan, Ukraine, Uzbekistan, and Yugoslavia.
Source: WTO Antidumping Committee data.Note: Numbers in parentheses are actions against China. Developed countries are Australia, Canada, EU15, Iceland, Japan, New Zealand, Norway, Switzerland, and United States. Transition economies are Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, FYR Macedonia, Georgia, Hungary, Kazakhstan, Kyrgyz Republic, Latvia, Lithuania, Moldova, Poland, Romania, Russian Federation, Slovak Republic, Slovenia, Tajikistan, Turkmenistan, Ukraine, Uzbekistan, and Yugoslavia.
TABLE 4.6A few large developing countries have launched the most antidumping investigations, 1995–2003
Initiating country/regionNumber of casesShare of total

Index of initiations per US$ of imports

(United States = 100)
United States32913.6100
European Union (15)27411.3101
South Africa1666.92,634
Sources: WTO Antidumping Committee data; IMF, Direction of Trade Statistics.
Sources: WTO Antidumping Committee data; IMF, Direction of Trade Statistics.

Another important nontariff measure affecting exports and imports are health- and safety-related product standards. In contrast to antidumping efforts, which by definition are intended to protect domestic import-competing industries, health and safety regulations (mandatory product standards) are often applied equally to domestic and foreign (imported) products—so in principle they are not (or should not be) intended to protect domestic industries. Indeed, WTO rules prohibit the protectionist use of technical product regulations. As discussed below, such norms affect trade by increasing the costs of entering a market, with the incidence of the associated costs differing across countries depending on the composition of their exports and the ease of satisfying (documenting compliance with) a specific norm.

Textiles and Clothing, Bananas, and Sugar

A major trade policy change in 2004 was the phase-out of bilateral quotas on exports of textiles and clothing at the end of the year, as required by the Uruguay Round Agreement on Textiles and Clothing. Because the quotas were bilateral and not based on competitive considerations, their removal is likely to lead to export declines and balance of payments swings for several countries. China and India are expected to gain from this liberalization, as they are considered the most competitive producers. But the actual effects will depend on market sourcing decisions by retailers and buyers, as well as the extent to which safeguard measures are invoked and U.S. quotas are maintained under the provisions of China’s WTO membership. In January 2005 China imposed a low MFN tax—ranging from 2–4 percent—on its textiles and clothing exports, to remain in place through the end of 2007.7 Although this will attenuate the expected shift in sourcing to Chinese producers, the elimination of the quotas will still force significant adjustments in less efficient countries (box 4.1).

In July 2004 the European Union announced that it will substantially reduce its sugar production and exports over a four-year period.8 The cut in output is to be accompanied by a reduction in domestic support prices of about one-third, implemented over a three-year period. Import quotas from African, Caribbean, and Pacific countries and India will be maintained at current levels. Farm support payments will be decoupled from production and linked to environmental and food safety standards. While this reform will benefit EU consumers and competitive developing country suppliers, it will reduce benefits from sugar production for several African, Caribbean, and Pacific countries as the value of their quotas declines. The European Union has made provisions for adjustment assistance for such countries.

In October 2004 the European Union announced a proposal to change its import regime for bananas. Under a WTO-backed agreement with Ecuador and the United States, the Union is required to move away from its quota-based system of controlling imports to a tariff-only system by January 2006. Under the proposal the Union would levy a tariff of 230 euros a ton on banana imports from non-African, Caribbean, and Pacific countries, which it argues is equivalent to the level of protection now in place. This move has disappointed producers in several Latin American countries, who are concerned that such a high tariff will make them unable to compete in the EU market, as well as producers in African, Caribbean, and Pacific countries, who fear that the proposed tariff is not high enough to maintain their existing market shares.

BOX 4.1The varying effects of the Agreement on Textiles and Clothing

The Agreement on Textiles and Clothing, agreed to as part of the Uruguay Round, was designed to progressively phase out quotas that limited developing countries’ opportunities to export textiles and clothing. Quotas were to be gradually abolished, and the growth rates of export volumes subject to quotas steadily increased, over a 10-year period ending on 1 January 2005. Importing countries were allowed to choose which products to make quota free at three stages (1995, 1998, and 2002) set out in the agreement. They chose products in which developing countries did not have a comparative advantage—or even face quotas. The acceleration of quota growth did succeed in making quotas nonbinding for some exporters, but for some dynamic exporters—particularly China—the quota growth rates of around 3 percent a year were much lower than their supply capacity. As a result quotas became more restrictive.

Quota abolition will lower prices for consumers in markets that had restricted textile and clothing imports. Prices in unrestricted markets can be expected to rise, absent the incentive for efficient exporters to divert products shut out of U.S. and EU markets into these markets. In addition, less competitive suppliers—including small, low-income countries that became exporters of textiles and clothing in part because dynamic exporters were fettered by quotas—may be confronted with lower prices for their products. The impacts on individual exporters will depend on the destination of their exports. One way to identify countries that might lose from this change is to examine what percentage of their exports comes from exports of textiles and clothing to restricted markets. The 18 most vulnerable countries are those for whom exports of textiles and clothing to Canada, the European Union, and the United States exceeded 35 percent of total exports in 2003 (see figure). Most of these are small economies.

Francois and Spinanger (2004) estimate that for China the quotas had the same effect as a tax of about 25 percent on clothing exports to the European Union and United States. Most other exporters faced much lower barriers. The adverse effect on the prices received by vulnerable exporters would be much smaller than 25 percent, partly because textiles and clothing are differentiated products (exports from different countries are very different; see Martin, Manole, and van der Mensbrugghe 2004) and partly because prices in unrestricted markets can be expected to rise following abolition of the quotas.

Exports of textiles and clothing to Canada, the EU, and the United States as a share of total

Source: Manole (2004).

Continuation of quotas against China is a possibility because of the safeguard provisions included when China acceded to the WTO in 2001 (Bhattasali, Li, and Martin 2004). But these apply only against China: if invoked, they would allow other efficient exporters to fill the gap. A safeguard system—or an equivalent export tax on China’s exports—would create global costs of around $13 billion, while transferring some $3 billion to competing exporters (Manole 2004). Most of this benefit would go to countries that are not heavily dependent on exports of textiles and clothing to restricted markets.

Initial indications are that exports of many smaller countries dependent on textiles and clothing are holding up better than had been widely feared. This partly reflects reforms to raise productivity in these countries and partly reflects China’s reluctance to expand its production of textiles and clothing, rather than more sophisticated products. In addition, as noted by Nordas (2004), standard modeling results do not fully incorporate the special features of the supply chain—particularly the advantages of proximity for geographically favored countries in a marketing environment of just-in-time ordering. Still, this initial assessment should not lead to complacency. Countries dependent on textile and clothing exports and their development partners need to focus on how these countries can raise their productivity and improve their trade policies and investment climates to encourage diversification away from dependence on a single sector.

Restrictiveness and Incidence of Trade Policies

As noted in Global Monitoring Report 2004, the trade policy measures reviewed above cannot be readily compared across countries. Such comparison is possible only on the basis of an index that measures the overall trade restrictiveness implied by trade policy. The measure used is the overall trade restrictiveness index (OTRI), which represents the uniform tariff equivalent of the different policy instruments observed for a country that would generate the prevailing level of trade. When the focus is on measuring the trade effects of policies on an importing economy, the trade restrictiveness index can be defined as the equivalent uniform tariff that would keep imports constant. When the focus is on an exporting country’s perspective (market access), the index is the equivalent uniform tariff implied by the set of policies maintained by each country (or all importers) on that economy’s export bundle. The OTRI is calculated as the weighted sum of nominal tariffs and the ad valorem equivalent of nontariff measures at the tariff line level.9

The contribution of nontariff measures to the overall level of trade restrictiveness increases with GDP per capita (figure 4.2). One reason is that average tariffs in rich countries tend to be lower and are bound in the WTO and through regional trade agreements. As a result protectionist pressures, if successful, will by necessity be reflected in other instruments—antidumping being a major example. Most developing countries have tariff bindings (commitments) in the WTO that are well above their applied tariffs, and may not have bound tariffs at all, allowing for the use of tariffs when desired.

FIGURE 4.2Nontariff measures are more important in rich countries, 2002

Source: Kee, Nicita, and Olarreaga 2005b.

Note: AVE stands for ad valorem equivalent.

Trade Restrictiveness, Income, and Poverty

OTRIs are negatively correlated with GDP per capita—the higher is a country’s GDP per capita, the lower is its trade restrictiveness as measured by the OTRI (figure 4.3). Similarly, the higher is a country’s GDP per capita, the lower are the trade barriers imposed by the rest of the world on its exports. Thus there is a negative association between GDP per capita and the trade restrictiveness that countries impose on their imports and that imposed by the rest of the world on their exports.

FIGURE 4.3Trade restrictiveness at home and abroad falls as countries become richer

Regression line shown: coefficient =–0.71; standard error = 0.23; t-statistic =–3.05.

Regression line shown: coefficient =–1.16; standard error = 0.32; t-statistic =–3.52.

Source: Kee, Nicita, and Olarreaga 2005c.

Note: Partial correlation plots obtained by regressing log GDP per capita on constant, log OTRI at home, and log OTRI abroad.

For developing countries for which such data are available, OTRIs are positively correlated with poverty headcounts: the higher is a country’s trade restrictiveness, the poorer it tends to be (figure 4.4). Higher trade restrictiveness on exports is also positively correlated with poverty head-counts—that is, the higher is the OTRI confronting a country’s exports, the more poor households that country is likely to have.10 The implication is that reductions in protection at home and abroad are likely to be pro-poor. Although such correlations of average trade restrictiveness and the number of poor people in a country suffer from aggregation bias and do not imply a causal relationship, more detailed microeconometric studies of trade reforms in developing countries find that liberalization tends to benefit poor people.11

FIGURE 4.4Trade restrictiveness at home and abroad rises with poverty headcount

Regression line shown: coefficient = 0.24; standard error = 0.25; t-statistic = 0.94.

Regression line shown: coefficient = 0.91; standard error = 0.52; t-statistic = 1.74.

Source: Kee, Nicita, and Olarreaga 2005c.

Note: Partial correlation plots obtained by regressing log poverty headcount on constant, log OTRI at home, and log OTRI abroad.

The Incidence of OECD Trade Policies

Estimated OTRIs for high-income OECD countries average 12 percent for all trade. They are much higher for agricultural trade than other merchandise—44 percent compared with 6 percent (figure 4.5). This disparity reflects high levels of support for agricultural production: the OECD Producer Support Estimate rose to 32 percent in 2003, up from 31 percent in 2002.12 A similar sectoral pattern applies to OTRIs for imports from developing countries, although the absolute level of trade restrictiveness is substantially lower for trade with Sub-Saharan Africa as a result of bilateral tariff preference programs (but still high in agriculture).

FIGURE 4.5Agricultural protection is high in OECD countries, and border barriers account for most of it

Source: Kee, Nicita, and Olarreaga 2005c.

Nontariff measures account for a large share of these OTRIs.13 Thus, taking the effects of such measures into consideration is important in accurately assessing the extent to which policies affect trade, both imports and exports. It is important to bear in mind that not all nontariff measures have explicit protectionist objectives. Particularly in the case of health- and safety-related product standards, the intention generally is not (and should not be) to protect domestic producers. To varying degrees, all countries maintain such regulations. WTO rules on such policies require that they be justified by scientific evidence and a process of risk assessment, and based on international norms if any exist. Such standards should also apply to domestic goods, although WTO case law and numerous bilateral disputes over many years regarding their enforcement illustrate that in practice these standards may be used to shield domestic producers from import competition.14 However, it is not possible to discern intent in the data—OTRIs simply show that these types of policies have a major effect on trade, especially in agricultural products.15 No presumptions exist or are implied regarding the intent or scientific basis of any of the nontariff measures captured in OTRIs.16

Because the OTRI does not include services trade and policies, it likely underestimates countries’ actual trade restrictiveness. Trade in services has been growing rapidly for the past two decades and now accounts for some 40 percent of world trade if sales of foreign affiliates are included.17 But statistical agencies do not compile the type of detailed data on bilateral trade in services that exist for merchandise. Nor is there a comprehensive global database on services-related trade policies.

The trade policies of high-income OECD countries imply higher restrictions on imports from low-income countries than from other sources (table 4.7). In part this simply reflects the sectoral structure of protection—most intra-OECD trade is in manufactures that confront low barriers. Barriers are highest for agricultural imports from middle-income countries, which do not benefit from the preferences accorded to poorer countries (table 4.8). Still, even after taking into account the fact that developing countries benefit from tariff preferences under the Generalized System of Preferences and that Sub-Saharan Africa and the LDCs have highly preferred—often duty-and quota-free—access to many OECD markets as a result of programs such as the EU Everything But Arms initiative and the U.S. African Growth and Opportunity Act, OTRIs for low-income countries average 14 percent (figure 4.6).

TABLE 4.7OECD trade restrictiveness is highest toward low-income countries, 2002 (percent)
Importing income group/countryOTRI for all countriesOTRI for low- income countriesOTRI for least developed countriesOTRI for Sub- Saharan AfricaShare of low-income country exportsShare of low-income countries in total imports from developing countries
High-income OECD121412116811
European Union131513142711
United States86542310
Source: Kee, Nicita, and Olarreaga 2005a.Note: OTRI stands for overall trade restrictiveness index.
Source: Kee, Nicita, and Olarreaga 2005a.Note: OTRI stands for overall trade restrictiveness index.
TABLE 4.8Globally, trade restrictiveness is highest for agriculture, 2002 (overall trade restrictiveness index, percent)
Importing region/income group
Exporting region/income groupHigh-income OECDQuadMiddle-incomeLow-incomeLeast developedSub-Saharan AfricaWorld
Total trade
High-income OECD981923212314
Least developed12122425222618
Sub-Saharan Africa11112324212417
High-income OECD36354536283439
Least developed38383932263037
Sub-Saharan Africa34353833263035
High-income OECD551521192110
Least developed332022212412
Sub-Saharan Africa22161918219
Source: Kee, Nicita, and Olarreaga, 2005a.Note: The Quad are Canada, the European Union, Japan, and the United States.
Source: Kee, Nicita, and Olarreaga, 2005a.Note: The Quad are Canada, the European Union, Japan, and the United States.

FIGURE 4.6OECD trade restrictiveness remains high for developing countries

Source: Kee, Nicita, and Olarreaga 2005c.

Note: Quad countries are Canada, the EU, Japan, and the United States.

A reason for this is that the incidence of nontariff measures in a number of OECD countries tends to be disproportionately high on products that developing countries export—especially agricultural products. Thus the higher OTRIs against low-income countries primarily reflect the product composition of imports—developing countries happen to export the goods most affected by nontariff measures—and the fact that not all low-income countries benefit equally from the preferential access regimes offered by the OECD members concerned.18 An implication is that nontariff measures matter, in that they tend to reduce the effective value of the preferential access granted through tariff exemptions. Insofar as the incidence of nontariff measures is higher in developing countries, there is a case for additional assistance to be directed at these countries to help them address the underlying regulatory requirements. The need for this increases the more ambitious the outcome of the Doha Round will be, as deep multilateral liberalization will reduce the value of existing tariff preferences.

Another striking feature of the OECD OTRI estimates is that agricultural subsidies play only a small role in determining the magnitude of OTRIs for most countries that employ them (see figure 4.5). That is because, from a market access perspective, subsidies are often redundant—because border barriers such as tariffs and tariff rate quotas are the instruments that determine the (world) price impacts. If border barriers were to be reduced significantly, agricultural subsidies would become much more important determinants of trade restrictiveness; thus they are not innocuous. But the implication is that policies (and negotiations) should focus on removing border protection. Lowering agricultural subsidies without lowering border protection will not have much effect.19

Although market access dominates in terms of affecting world prices and distorting domestic markets, subsidies are very important for some developing countries.20 Cotton is an example where global trade barriers are low, but subsidies in several OECD countries are large—with well-documented detrimental effects on West African and other developing country producers. Cotton accounts for about 30 percent of exports from the four West African nations that proposed a cotton initiative in the WTO, and for a significant share of income for millions of poor farming households in the region.21 Recent U.S. cotton subsidies have ranged from $1.5 billion to almost $4.0 billion a year, depending on market conditions.22 Eliminating U.S. cotton programs, while leaving in place other farm programs, would reduce U.S. production by 25–p30 percent, reduce U.S. exports by about 40 percent, and raise world prices by about 10 percent. The annual losses for LDC cotton producers resulting from U.S. and EU support policies are in the range of $120–240 million.23

The Incidence of Developing Country Trade Policies

Developing countries have higher average OTRIs than do high-income OECD countries. Among regions, the Middle East and North Africa have the highest OTRIs, followed by Sub-Saharan Africa, South Asia, and East Asia (table 4.9). Eastern European and Central Asian countries, many of which acceded to the European Union in 2004, have the lowest OTRIs. The average OTRI for all developing countries is 20 percent. As with high-income OECD countries, manufacturing protection tends to be lower than restrictions applied to agricultural trade, though the differences are smaller than in the OECD countries (see table 4.8). On average the world has an OTRI in agriculture that is three times that in manufacturing—35 percent compared with 12 percent—with high-and middle-income countries being more restrictive than low-income countries. The LDCs have the lowest agricultural OTRIs of all income-based country groups. Conversely, developing countries have much higher manufacturing OTRIs, with the highest average levels observed in Sub-Saharan Africa. This generally reflects either a desire to protect local industries or to achieve fiscal objectives; tariffs account for a substantial share of government revenues in many Sub-Saharan countries.

TABLE 4.9Developing countries impose high restrictions on trade with one another, 2002 (overall trade restrictiveness index, percent)
Importing region/income groupOTRI for all countriesOTRI for low-income countriesOTRI for least developed countriesOTRI for Sub-Saharan Africa
East Asia and Pacific21262626
Europe and Central Asia12161516
Latin America and the Caribbean18212020
Middle East and North Africa31484740
South Asia23292728
Sub-Saharan Africa23262624
Developing countries20252523
Least developed countries21222221
Low-income countries22262524
Middle-income countries20252423

In terms of the incidence of OTRIs, a pattern similar to that for OECD countries is observed for developing countries—OTRIs are higher for exports from other developing countries (though the LDCs confront slightly lower average OTRIs than do other low-income countries; see table 4.8). Thus, in addition to facing high barriers in OECD countries, developing countries impose high barriers on trade with one another, and the incidence of these intra-developing country trade restrictions is higher for poorer countries—just as the incidence of OECD trade restrictions is higher for poorer countries.

These estimates show that the protectionist trade policies of high-income OECD countries are not the only external problem undermining the ability of developing countries to use trade for development and poverty reduction. Indeed, middle-income countries maintain much higher levels of protection. Reducing such protection would be in their own interest and that of poorer countries—which often confront higher barriers in global markets than do rich countries, implying that the structure of trade policies in middle-income countries is antipoor. Thus reform of trade policies in developing countries should be a central part of the trade reform agenda. This conclusion also emerges from the research summarized below assessing the potential gains from the Doha Round—which are potentially large only if the round encompasses significant further liberalization of trade by developing countries.

The Doha Round Challenge

The WTO is the primary multilateral instrument through which countries can reduce the trade protection summarized above. In August 2004 the WTO General Council reached several agreements to guide future negotiations under the Doha Round. The agreements were a welcome step forward after the setback at Cancun, but they leave key questions unanswered and extend the timeframe of the round. The agreements include a framework for negotiations on agriculture and industrial products, recommendations on services, and modalities for negotiating improved customs procedures (trade facilitation; table 4.10).

TABLE 4.10Key elements of the August 2004 WTO framework agreement
Agriculture subsidiesAll forms of agricultural export subsidies to be eliminated. Overall bound level of trade-distorting domestic support to be reduced over time, with a 20 percent reduction in the first year. Members with higher levels of support to commit to larger reductions, with product-specific caps on the most trade-distorting forms of agricultural support.
Market access in agricultureNonlinear formula to ensure deeper cuts in higher tariffs. A category of “sensitive” products to be subject to greater flexibility. Possibility of trading off more modest tariff reductions against larger tariff rate quotas.
CottonTo be dealt with “ambitiously and expeditiously” within the agriculture negotiations, by a special subcommittee.
Non-agriculture market accessNo a priori exclusions. Recognition that credit should be given for autonomous liberalization. Nonlinear formula to tackle tariff peaks, high tariffs, and tariff escalation, to be applied on a line-by-line basis.
ServicesRevised offers to be tabled by May 2005, and developing countries to be given technical assistance to help them participate in the negotiations.
Trade facilitationNegotiations to be part of the Doha Round, with a focus on expediting the movement, release, and clearance of goods, with implementation by developing countries linked to provision of technical assistance.

Why Ambition Matters: Likely Impacts of Different Doha Round Outcomes

The extent to which reform commitments under the Doha Round will have economic repercussions depends on whether they affect applied policies. WTO negotiators focus not on applied tariffs and subsidies but on the levels of import tariffs, export subsidies, and domestic support for agriculture that countries commit not to exceed. These so-called bindings are often much higher than applied levels of protection, especially in developing countries. Thus, if cuts to bound tariffs or subsidy rates are not very large, or the gap between bound and applied rates is large, the reforms required by a Doha deal may be minimal. So, much depends on the extent to which changes in bindings translate into real reductions in applied protection levels. Much also depends on the magnitude of exceptions and exemptions for specific instruments or products.

Deep trade reform could generate large global gains. Freeing all merchandise trade and abolishing all trade-distorting agriculture subsidies would boost global welfare by $80–280 billion a year by 2015, depending on whether exogenous population and labor supply growth, savings-driven capital accumulation, and labor-augmenting technological progress are taken into account.24 These estimates ignore possible dynamic productivity gains, exploitation of economies of scale, effects of increased competition on markups and X-inefficiency in production, and liberalization of trade in services—including through greater temporary international mobility of service suppliers. They also assume that beneficiary countries make full use of preferences and that any resulting rents accrue to exporters. Taking such factors into account greatly increases the potential aggregate gains and affects the distribution of net effects across countries.25

Recent research suggests that developing countries would obtain about one-third of the global gain from freeing all merchandise trade, well above their one-fifth share of global GDP. Recent analyses using global general equilibrium models and the latest databases on trade protection and preferences suggest that developing countries would receive a 1.2 percent static gain over initial welfare (real income) levels, compared with just 0.6 percent for developed countries (assuming an overall global gain of $280 billion). The larger gain for developing countries is partly because of their relatively high tariffs—meaning, they would reap substantial efficiency gains from reforming their protection—and partly because their exporters face much higher farm and textile tariffs in developed countries than do exporters from other developed countries, notwithstanding non-reciprocal tariff preferences for many developing countries.26

Agriculture is the most important sector in realizing these potential gains, reflecting the extensive assistance it receives relative to other sectors. Food and agriculture policies are responsible for more than three-fifths of the global gain forgone by merchandise trade distortions (table 4.11)—despite the fact that agriculture and food processing account for less than 10 percent of world trade and less than 4 percent of global GDP. Agriculture is as important for the welfare of developing countries as it is for the world as a whole: their gains from global agricultural liberalization account for almost two-thirds of their total potential gains, compared with one-quarter for textiles and clothing and one-eighth for other merchandise liberalization (see table 4.11).

TABLE 4.11Most economic welfare benefits of full merchandise trade liberalization would come from agriculture, 2015 (percent)
Global welfare benefits
Sector liberalized
Source of benefitsAgriculture and foodTextiles and clothingOther manufacturesAll goods
Developed countries’ policies466355
Developing countries’ policies1592145
Developing countries’ welfare benefits
Sector liberalized
Source of benefitsAgriculture and foodTextiles and clothingOther manufacturesAll goods
Developed countries’ policies2816650
Developing countries’ policies359650
Source: Anderson, Martin, and van der Mensbrugghe 2005, table 12.4.Notes: (1) These data are shares based on a simulated global gain of $280 billion that takes into account exogenous population and labor supply growth, savings-driven capital accumulation, and labor-augmenting technological progress between 2001 (base year) and 2015. Developed countries include the transition economies of Eastern Europe and the former Soviet Union. (2) The share of global welfare gains due to liberalization by middle-income countries is 12 percent for agriculture and food liberalization; 7 percent for textiles and clothing; and 13 percent for other manufactures.
Source: Anderson, Martin, and van der Mensbrugghe 2005, table 12.4.Notes: (1) These data are shares based on a simulated global gain of $280 billion that takes into account exogenous population and labor supply growth, savings-driven capital accumulation, and labor-augmenting technological progress between 2001 (base year) and 2015. Developed countries include the transition economies of Eastern Europe and the former Soviet Union. (2) The share of global welfare gains due to liberalization by middle-income countries is 12 percent for agriculture and food liberalization; 7 percent for textiles and clothing; and 13 percent for other manufactures.

Stronger subsidy disciplines are important, but increased market access in agriculture is crucial. High applied tariffs on agricultural relative to nonfarm products are the main reason food and agriculture policies account for 61 percent of the welfare cost of merchandise trade distortions. Aggregating across all products, subsidies for farm production and exports are only minor contributors: 4 and 1 percentage points, respectively, compared with 56 points due to agricultural tariffs.27 Tariffs are more important in developing than in developed countries, because subsidies are much less prevalent. Given that bound tariff rates are high in agriculture, large cuts are needed to make a difference. The average bound tariff rate in developed countries is almost twice the average applied rate; in developing countries the difference is even greater. To lower the average global agricultural tariff by one-third, bound rates would have to be reduced for developed countries by at least 45 percent and the highest tariffs by up to 75 percent.28

Even large cuts in bound tariffs will do little if exemptions for “sensitive” products are allowed. If WTO members limit tariff cuts for “sensitive” farm products such as rice, sugar, milk, beef, fruits, and vegetables, the prospective gains from Doha could be greatly diminished. Classifying 2 percent of six-digit Harmonized System agricultural tariff lines in developed countries as sensitive (and 4 percent in developing countries, to reflect demands for differential treatment) and subjecting them to just a 15 percent tariff cut would shrink by three-quarters the welfare gains from global agricultural reform.29 These findings illustrate the importance of the call made in Global Monitoring Report 2004 for ambitious targets with respect to lowering agricultural tariffs, for example, to no more than 10 percent.

While agricultural market access liberalization is by far the most critical merchandise trade issue on the negotiating table from a developing country perspective, disciplining domestic production subsidies and phasing out export subsidies is also important. In part these steps are needed to prevent “re-instrumentation” of assistance from tariffs to domestic subsidies. Moreover, some OECD subsidies are very important for developing countries. Removing cotton subsidies would raise the price of cotton in international markets and benefit developing country exporters. For Sub-Saharan Africa, recent research suggests that the increase in export prices relative to manufactures overall would be small. But cotton exports from Sub-Saharan Africa would increase 75 percent. The share of all developing countries in global cotton exports would expand to 85 percent, instead of the 56 percent projected for 2015.30

The foregoing analyses of possible outcomes of negotiations consider policy instruments such as tariffs, preferences, subsidies, and tariff rate quotas in agriculture. But they do not consider the effects of nontariff measures and mostly ignore the implications of removing barriers to trade and investment in services. Given that nontariff measures play an important role in determining the overall trade restrictiveness of countries, reducing the incidence of such policies could generate additional gains for developing countries and the world as a whole. Using a partial equilibrium framework that allows much greater disaggregation across products and countries than is possible in global computable general equilibrium models, a recent study compared the likely effects of an unambitious Doha Round and an ambitious one that addresses nontariff measures.31 The low-ambition round is defined as causing a 40 percent cut in bound tariffs, a reduction in tariff peaks to a maximum of 50 percent, a 40 percent reduction in domestic agriculture subsidies, elimination of export subsidies, and an improvement in trade facilitation that increases world trade by an average of 2 percent.32 Conversely, an ambitious Doha Round would result in free trade (elimination of all applied tariffs), a 50 percent cut in the restrictiveness of nontariff measures, elimination of trade-distorting agriculture production and export subsidies, and the same improvement in trade facilitation.

An unambitious Doha Round would not do much to improve developing country welfare or help achieve the MDGs. Static global welfare gains would be about $59 billion. In contrast, an ambitious Doha agenda could nearly quintuple these gains, to $269 billion. Full tariff and subsidy reform without any changes in nontariff measures would generate “only” $111 billion.33 Sub-Saharan Africa would see its welfare increase by an estimated $2.8 billion under the ambitious scenario, compared with only $140 million under the type of agreement that current trends suggest is likely to emerge. The low-ambition scenario would raise Sub-Saharan GDP by 0.1 percent. In contrast, an ambitious Doha Round outcome could boost GDP by 1.3 percent. Around 40 percent of the gains from the round would accrue to developing countries under both scenarios, with the greatest absolute amount going to middle-income nations. Low-income countries would obtain 14 percent of the developing world welfare gains. Under both scenarios welfare gains relative to GDP are largest for low- and middle-income countries (figure 4.7).

FIGURE 4.7A low-ambition round vs. deep WTO reforms

This analysis suggests that efforts to reduce the incidence of nontariff measures are important. But as noted, not all these measures are discriminatory. In the case of regulatory product requirements—which account for a large share of nontariff measures in many high-income countries—attenuating the effects of enforcement is unlikely to be costless. Thus the additional $150 billion that could be attained by halving the ad valorem equivalent of nontariff measures is an upper bound estimate of potential net gains. Determining how to attenuate the impacts of nontariff measures at the lowest cost requires both country-specific analysis and a concerted effort to improve the global information base on applied nontariff policies. But it is probable that much of the appropriate response revolves around “aid for trade” to improve quality, testing, conformity assessment capacity, and so on.

Prospects for realizing the potential of trade to support development are also conditional on further liberalization by developing countries. As revealed by the OTRIs, developing countries maintain much higher trade barriers against nonagricultural products than do OECD countries and impose high trade barriers on each other. Combined with the fact that many (especially middle-income) developing countries have been growing more rapidly, prospects for greater trade flows are greatest between developing countries. Moreover, most of the benefits of trade reform derive from what countries do themselves—as discussed in Global Monitoring Report 2004 and numerous recent reports, including the United Nations MDG taskforce report on trade, maintaining high barriers to trade imposes costs on an economy.

If developing countries as a group were not to participate in liberalizing trade under the Doha Round, not only would global gains fall significantly—by close to one-half—but potential gains to developing countries would fall by 50 percent.34 Sub-Saharan Africa and the LDCs could even end up losing from the Doha Round if they do not pursue reforms. The reason is that the efficiency gains from own liberalization are needed to help offset terms of trade losses that may arise because food prices rise (in the case of net food importers), tariff preferences are eroded, or market share is lost to other developing countries that pursue reforms (or have already done so).


The foregoing discussion largely ignores domestic reform and liberalization of service sector policies, an area where research suggests there is great potential for gains, in terms of both efficiency and equity. Countries such as Brazil and India have seen exports of business process services such as call centers, software development, and back-office services expand rapidly. African exports of commercial services grew by more than 20 percent in 2003, in line with merchandise export growth, and account for 18 percent of total exports—the largest share of all developing regions.35

Developing countries have called for specific commitments on improving access to OECD markets for service suppliers, especially through temporary migration (mode 4 in the General Agreement on Trade in Services, or GATS). This is the main area where virtually no commitments have been made (figure 4.8). As noted in Global Monitoring Report 2004, it is also the area with the greatest potential for increased exports from developing countries (figure 4.9).

FIGURE 4.8WTO Market access commitments for services by mode of supply

Source: Marchetti (2004), updating WTO document S/C/W/99.

Note: The above information is as of August 2004 and is based on a sample of 37 representative sectors. DC (developed countries); Tran (transition economies); Dev (developing countries). Full implies a binding commitment to market access and national treatment; unbound implies no commitments were made; and partial implies specific policies limiting access or involving discrimination were bound.

FIGURE 4.9Foreign direct investment and cross-border exchange account for most trade in services

Source: Maurer (2004)

To date, little progress has been made on services negotiations. Although numerous requests have been put on the table, only 47 first-round offers have been put forward by WTO members.36 Most of these are of limited value as they do not go beyond status quo policies. The LDCs have made the fewest commitments on services in the WTO—in contrast to recently acceded countries, which have been asked to make far more commitments than most existing members (table 4.12). There is great scope for more advanced countries to lock in liberal trade and investment policies in areas that matter to developing countries—especially on cross-border trade (mode 1)—and to increase the extent to which service markets can be contested through temporary migration of service providers. Conversely, developing countries can do much more to use the WTO as an instrument to commit to greater liberalization of their services. While considerations of “balance” create incentives not to do so until greater clarity has emerged on what is likely to be feasible in other areas (such as agriculture), services are an area where a quid pro quo can be offered that is likely to benefit the committing country. Much of the agenda here revolves around making commitments targeted at attracting greater foreign direct investment (FDI) in services (mode 3 of the GATS). Encouraging such investment in both goods and services production is a direct way to improve the investment climate.

TABLE 4.12Developing countries have made fewer market access and national treatment commitments for services under the WTO
Country groupAverage number of subsectors committed per countryRange

(lowest/highest number of committed subsectors)
Developing countries541-154
Least developed countries201-110
Transition economiesa10658-154
Developed countries10887-117
Countries acceded since 199510637-154
Source: Marchetti 2004.Note: Data as of August 2004. The total number of subsectors in which commitments can be made is 160. These commitments may be full or partial bindings; see note to figure 4.8. a. Includes only European and Central Asian transition economies.
Source: Marchetti 2004.Note: Data as of August 2004. The total number of subsectors in which commitments can be made is 160. These commitments may be full or partial bindings; see note to figure 4.8. a. Includes only European and Central Asian transition economies.

FDI can do much to enhance an economy’s productivity, because foreign service suppliers offer new technologies, new products, and more differentiated and higher-quality services. The increased competition associated with open access to service markets also puts pressure on the prices and performance of incumbent firms, reducing input costs for firms that buy the services. Moreover, the importance of an efficient services sector goes beyond the sector’s contribution to a country’s balance of payments; such efficiency is also a key determinant of domestic firms’ competitiveness. Key services that influence firms’ ability to participate in world trade include telecommunications, transportation, financial, and other business services such as accounting and legal services.

In many of the LDCs, inefficient business services are an important barrier to effective integration with world markets. High-cost trade support services act as a tax on exporters and ultimately on growth and poverty reduction. For a small economy confronting given world prices of traded goods, higher service costs reduce export prices and increase import prices. As a result exporting industries must pay lower wages or accept lower returns on capital. For example, freight rates for goods originating in Sub-Saharan African countries are often considerably higher than for similar goods originating in other countries, and thus have contributed to the region’s weak trade performance and high poverty in recent decades. High transportation costs are partly due to shortcomings in infrastructure, geography, and the economic size of markets, but they may also reflect policies that restrict competition—such as cargo reservation schemes or entry barriers into air and maritime transport.

The Doha Round, Low-Income Countries, and Sub-Saharan Africa

Market access is a necessary but insufficient condition for harnessing trade for development. As noted in Global Monitoring Report 2004, domestic supply constraints are the main reason for the lack of trade growth and diversification in many developing countries. Without action to improve supply capacity and reduce transactions costs, trade opportunities cannot be fully exploited and the potential gains discussed previously will not be realized. The agenda spans many areas, many of them “behind-the-border.” Dealing with these issues will require action in developing countries and assistance from developed countries, including both finance and efforts to reduce the costs of nontariff measures.

Recent country studies that analyze the potential impacts of global reforms on African poverty, using information from household surveys, show that trade policy actions by themselves will have limited effects—positive or negative. Complementary reforms and investments are needed to stimulate the desired supply response to changed incentives. In Ethiopia, one of the world’s poorest countries (with more than two-thirds of the population living on less than $2 a day), the Doha Round alone will not have much significant effect on poverty.37 That is because any impact on the prices and quantities of Ethiopia’s main exports—coffee, qat (a mild stimulant exported mostly to neighboring countries), and livestock—will not be large. Moreover, exports represent a small part of GDP: annual exports are about $600 million, or $10 per capita, equal to 15 percent of GDP—one of the lowest shares in Sub-Saharan Africa. Thus even a hypothetical doubling of international demand for Ethiopian products would have limited impact on poverty. Domestic markets are poorly connected, limiting the price and quantity effects of any trade policy changes in rural areas, where most poor people live. If price signals fail to reach households, supply responses to international shocks will be low. In addition, low agricultural productivity implies that supply responses will be associated with higher production costs, reducing net benefits. Finally, Ethiopian households, especially the poorest, are insulated from market shocks (and thus effects of trade policies) because they are widely engaged in subsistence activity. More than half the average budget of poor households is from subsistence.

This does not mean that an ambitious Doha Round would not bring opportunities to Ethiopia. It would increase international demand for Ethiopian products. But for that to translate into quantifiable benefits, the increased demand would have to be accompanied by improved productivity (rather than a simple increase in production), to also spur nonexported agricultural production (especially staple crops). And given that most poor Ethiopians live in remote areas, increased productivity alone may not have a large effect on poverty in the absence of complementary policies. Better infrastructure and access to credit are two preconditions for integrating the rural poor with domestic markets and reducing subsistence.

A similar conclusion arises from a recent study on Zambia.38 In 1998 more than 70 percent of the country’s population lived in poverty. Using an approach similar to the analysis of Ethiopia, in terms of projecting the likely effects of a Doha Round on world prices, the authors investigate the impact of trade reforms on household welfare through effects on consumption and income. They conclude that only small impacts can be expected from a Doha Round set of trade reforms. This reflects the fact that the round will likely generate only a small change in prices, so gains and losses for producers and consumers will be small as well. Moreover, for Zambian households large shares of spending and income are related to home-produced goods—which are unlikely to be affected by trade liberalization. Larger effects are predicted on the income side, particularly in terms of higher wages, employment opportunities, and a move from subsistence to market-oriented agricultural activities. Again, a key finding is that trade policy by itself is not enough. Complementary policies are important in allowing households to take full advantage of trade liberalization and global market opportunities. This result is illustrated by two recent case studies: one on the role of extension services in boosting productivity among poor Zambian farmers, and another on the role of job programs in providing employment for heads of households.39

An analysis for Madagascar similarly concludes that global trade reforms by themselves may have limited impact.40 Many of the country’s exports do not confront high protection (vanilla, cloves, crustaceans, semiprecious stones), most poor people live in rural areas not connected to markets (preventing the pass-through effects of border price changes on supply), and productivity is low. Moreover, limited competition in domestic markets is likely to concentrate gains from trade in the hands of traders and exporters, as opposed to poor farmers.41 Global liberalization would increase international demand for Madagascar’s exports, but to take full advantage of this it would be necessary to raise productivity and reduce the transportation costs of getting products to urban areas and export hubs. Past unilateral reforms and improved access to export markets show that trade can reduce poverty—although given their low initial base, the impact of greater exports can be limited only in the short to medium term (box 4.2).42

In sum, country-based analysis of the poverty impacts of global trade reforms in very poor countries suggests that the first-order (or static) effects are small. Price effects, whatever the magnitude, will be passed through incompletely, little will be done to attenuate the effects of nontariff measures, and many LDCs have concentrated export bundles involving products that often are not heavily protected in world markets (such as vanilla in Madagascar and coffee in Ethiopia). Second-order effects—allowing for adjustments on the supply side through movements from subsistence into market participation, changes in crops, and moves out of unemployment—can be larger. But for any positive effects to be significant, global trade reforms must be accompanied by complementary actions targeted at increasing the supply response. Such actions could center on reducing transportation costs from remote areas, increasing farm productivity through extension services, and so on.

BOX 4.2.Why has rapid export growth failed to significantly reduce poverty in Madagascar?

Fueled by low labor costs, relocation of textile and apparel firms from neighboring Mauritius, preferences granted by the U.S. African Growth and Opportunity Act, and the creation of export processing zones, Madagascar’s textiles sector experienced dramatic growth in the late 1990s. Textile and apparel exports increased from $100 million in 1995 to almost $500 million in 2001. In the same period employment in the sector grew from 30,000 to 200,000 individuals. Expansion of the textiles and apparel industry has enabled many individuals and households to increase income and in some cases escape poverty. But the welfare effects of this expansion have been geographically confined: most of the employment has been created in a few urban areas with relatively low poverty.

Most of the gains, in both absolute and relative terms, have accrued to nonpoor households. As a result this export-led growth has increased inequality between poor and nonpoor people, between urban and rural areas, and between skilled and unskilled workers. Skilled workers have experienced rapidly increasing wages, while rural areas have largely been cut off from the effects of the growing textiles and apparel sector. Unskilled workers have benefited only marginally, because a large pool of unemployed and underemployed workers has kept their wages from rising. But poor households located in the capital, where most of the employment was created, have benefited. It is estimated that five years of expansion of the textiles sector reduced poverty by almost 1 percentage point, or about 150,000 individuals.

These results suggest that two factors are required if export-led economic growth is to reduce poverty significantly. First, growth and job creation must not be restricted to a few geographic areas. Dispersion could be achieved by creating geographically diverse export processing zones to generate job opportunities for a wider range of the population. Second, to facilitate their absorption into the formal labor force, poor people must be assisted in obtaining the skills sought by expanding industries. This can be achieved by providing government-sponsored training or incentives for firms to provide training.

Source: Nicita 2005.

This conclusion is supported by the many Diagnostic Trade Integration Studies conducted under Integrated Framework auspices for the LDCs, as well as similar assessments for other developing countries. These find that trade cannot play much of a positive role if the macroeconomic climate is unstable and the business environment is hostile, and that for many countries the key challenges are to reduce constraints on competitiveness and supply responses—especially lack of transport, energy, and water infrastructure, lack of access to finance, and weak worker skills. Improving trade policies and taking action to facilitate trade by strengthening and streamlining customs can have a significant payoff, but a major impact on poverty requires reducing domestic supply constraints.

Thus trade may make only a limited contribution to reducing poverty in the short to medium run. And though its potential is substantial, especially in the long run, trade is just part of the solution.43 Even in Kenya and Lesotho, African countries with dynamic export sectors (cut flowers and apparel, respectively), booming exports have not had major benefits for employment, and development of supply industries (backward linkages) has been slow.44

Trade and Development—Toward a Concerted Response

The traditional international policy response to the challenge of expanding trade in developing countries has been a mix of preferential access and development assistance. In recent years preferential access has been deepened and made more meaningful through the EU Everything But Arms initiative and the U.S. African Growth and Opportunity Act. Efforts have also been strengthened to provide more aid for trade.

All these programs have been beneficial. But preferences are an eroding asset: they have become less valuable as importing countries have liberalized their trade—both on a nondiscriminatory MFN basis and as a result of regional trade agreements. The value of preferences will be further eroded the more the Doha Round results in significant further liberalization. The answer is not to stop global liberalization to maintain the value of preferences. Instead, action is required on complementary instruments to help countries address competitiveness challenges and exploit market access opportunities. This includes measures to reduce the incidence of nontariff measures.

Market Access: Beyond Tariff Preferences

To some extent tariff preferences offset the explicit discrimination against developing country trade reflected in higher tariffs on key exports and the higher incidence of nontariff measures revealed by the OTRIs discussed previously. Given that developing countries generally do not give significant tariff preferences to other poor countries, preferences only significantly affect the incidence of trade policies in developed countries. A seemingly logical implication of this is that advanced developing countries should also offer preferential access to their markets to the poorest countries. At the 2004 meeting of the United Nations Conference on Trade and Development in Brazil, it was agreed that countries would seek to do so in the framework of the Global System of Trade Preferences.45

Much of the literature on this subject concludes that preferences have not resulted in the desired export expansion and diversification in many countries. There have been several notable exceptions, but these are associated with quantitative restrictions being imposed on competitive suppliers. Reasons why benefits have often proven limited include restrictive rules of origin—reducing utilization of preferences below 100 percent, especially for key manufactures such as clothing,46 and excluding “sensitive” products from the coverage of arrangements. Recent preference schemes for the poorest countries—such as the EU Everything But Arms initiative for the LDCs, the U.S. African Growth and Opportunity Act for eligible Sub-Saharan countries, and Canada’s duty-and quota-free access program for the LDCs—provide more comprehensive coverage than do traditional General System of Preferences-type programs. In the case of the African Growth and Opportunity Act, they also apply more liberal rules of origin and have allowed exports of textiles to expand rapidly from eligible countries.

While preferential access may generate rents and stimulate export-oriented production in countries that are granted liberal origin rules, any rents are shared with importers (box 4.3).47

BOX 4.3.Many of the rents created by trade preferences accrue to importers

For exporters, benefits from trade preferences depend on the size of the rents associated with being able to sell into protected markets and on their capacity to capture those rents. Recent research has found that importers may be able to capture some, if not most, of such rents. For example, one study found that the average export price increase for products benefiting from preferences under the African Growth and Opportunity Act was about 6 percent, whereas the average MFN tariff for these products was some 20 percent (see figure). Thus, on average exporters received around one-third of the tariff rent. Moreover, poorer and smaller countries tended to obtain smaller shares. The estimates ranged from 14 percent in Malawi to 46 percent in South Africa. Reasons why importers and retailers capture a large share of the rents include the monopsony power of buyers and limited bargaining power of African exporters—perhaps reflecting limited or asymmetric information.

More important, such rents are highest where there are also quantitative restrictions on competitive suppliers (as under the nowexpired Agreement on Textiles and Clothing and the EU sugar regime). These regimes are being (or have been) dismantled, implying erosion of rents even before the conclusion of the Doha Round. One response is to deepen and extend preferences to additional countries. The experience under AGOA or EBA suggests that existing tariff preference programs could be made more effective by adopting common, liberal rules of origin. Several OECD countries have also proposed that middle-income countries apply access schemes like those under the act.

While a decision to adopt common, liberal rules of origin would be desirable if put in place rapidly, and preferential access to large middle-income markets would be of value, it must be recognized that such “affirmative action” is not a panacea for poor countries to expand and diversify their trade. One reason is the earlier observation that preferences that have generated rents have largely been accompanied by quotas. With the implementation of the Agreement on Textiles and Clothing, these have largely been removed. A second reason is that many countries have not been able to benefit from existing deep preferences, so extending these may not have much effect. Third, there is a global negative externality associated with preference programs—they create incentives not to pursue global liberalization, especially in sectors where distortions are highest. Preferences for some countries by definition imply greater discrimination against other, less preferred developing countries, and thus trade diversion—because the set of goods that beneficiary developing countries produce overlaps with those of other developing countries that are not beneficiaries. This trade diversion and discrimination has created significant tension between developing countries and has recently given rise to several formal WTO disputes. Finally, in terms of numbers, most of the world’s poor people do not live in the poorest countries, and preferences for some come at the cost of equally poor households in other, less preferred developing countries.

Given the systemic downsides, limited benefits, and historical inability of many poor countries in Africa and elsewhere to use preferences, a decision to shift away from preferential “trade as aid” toward more efficient, effective instruments to support poor countries could both improve development outcomes and help strengthen the multilateral trading system. More effective integration of the poorest countries with the trading system requires instruments aimed at improving the productivity and competitiveness of firms and farmers in these countries. Supply constraints are the primary factors that have constrained the ability of many African countries to benefit from preferences.48 This suggests that the main need is to improve trade capacity and facilitate diversification. In part this can be pursued through a shift to more (and more effective) development assistance that targets domestic supply constraints as well as measures to reduce the costs of entering foreign markets. Such aid for trade is particularly urgent for countries that confront preference erosion as a result of multilateral reforms. This is a significant economic issue for small countries dependent on sugar, bananas, and to a lesser extent clothing exports—products where protection and thus preference margins are high. Importantly, most of these are not the poorest countries, but middle-income economies.49

Aid for Trade

Bolstering trade capacity by linking farmers to markets and improving their productivity, lowering transaction costs by identifying and removing red tape and improving customs clearance, or putting in place the regulatory preconditions for benefiting from and managing liberalization—especially in the area of services—are all institution-intensive. The needs are not just for trade infrastructure and trade facilitation, but also involve complementary actions to implement safety nets, enhance access to credit, improve health and education services, and pursue effective regulation to attain efficiency and equity objectives. This agenda extends well beyond trade integration. It is largely “additive” in that actions on these fronts can be pursued in parallel.50 On all these fronts development assistance can play an important role in enhancing the overall gains from trade and managing costs and downside risks. A specific area where assistance may have high payoffs is in reducing the costs of satisfying nontariff regulatory policies in export markets.

In recent years progress has been made in expanding trade-related technical assistance and capacity building. The WTO and OECD database that tracks such assistance distinguishes between trade policy and regulations (technical assistance for product standards, integration of trade with development plans, trade facilitation), trade development (trade promotion, market development activities, and so on), and infrastructure. Infrastructure is the largest of these categories but is not limited to trade-related projects. The share of total aid commitments that went to support trade policy and regulations and trade development rose from 3.6 percent in 2002 to 4.2 percent in 2003.51 About one-third of this support went to the LDCs. Aid has been expanding most rapidly for trade policy and regulations (figure 4.10), reaching almost $1 billion in 2003, with the amount allocated to Africa tripling between 2001 and 2003. Commitments for trade development activities, which amounted to some $1.35 billion a year in 2001 and 2002, reached $1.8 billion in 2003.52 Infrastructure assistance was stable at around $8 to $9 billion a year between 2001 and 2003, with Asia by far the largest recipient region. In 2002 Africa’s share of infrastructure aid diminished, reflecting an emphasis on social sectors.

FIGURE 4.10Distribution of ODA for trade-related activities and infrastructure by region and main category

Source: WTO and OECD, 2004.

Note: These regions were defined using U.N. classification.

The World Bank has continued to expand its activities and support for trade, focusing on trade integration broadly defined to encompass trade policy, infrastructure, and customs and trade facilitation. Trade diagnostic work is being carried out in more than 35 countries and is complemented by regional studies that examine issues for several countries at a time. There is a growing lending program for trade-related activities (figure 4.11). Projected commitments for new trade operations between fiscal 2004 and 2006 are significantly larger than in previous years, with trade facilitation lending being a major growth area.

FIGURE 4.11Bank trade-related lending

Source: The data comes from the Bank’s Business Warehouse, which consolidates information recorded in SAP and other Bank Systems.

Trade policy has been a frequent component of programs supported by the International Monetary Fund (IMF). In April 2004 concerns about the balance of payments impacts of trade policy changes motivated the introduction of the Trade Integration Mechanism. Although the core of the IMF’s mandate involves support for orderly external adjustment in the face of shocks, the mechanism provides added assurances and represents the first explicit attempt to help members adjust to shocks that emanate from multilateral trade liberalization. The mechanism is designed to help mitigate the stated concerns (by many developing countries) that implementation of WTO agreements might give rise to temporary balance of payments shortfalls—through the erosion of tariff preferences, adverse changes in the terms of trade of net food importers, or expiration of quotas following full implementation of the WTO Agreement on Textiles and Clothing. In July 2004 Bangladesh became the first member to obtain support from the Trade Integration Mechanism, followed in January 2005 by the Dominican Republic.

These activities reflect the growing recognition that trade-related reforms and infra-structure—both hard and soft—can have high rates of return in increasing economic growth rates. Realizing an ambitious Doha Round and addressing trade capacity priorities in poor countries calls for more to be done. Not only will the associated trade and related reforms give rise to adjustment costs, but to enhance the benefits for poor households in poor countries, as argued previously, investments are needed to facilitate trade, lower transport costs, link farmers and producers to markets, and improve productivity. Strengthening “aid for trade” mechanisms to pursue this agenda would help improve the development relevance of both the WTO and regional integration agreements. An expanded allocation of aid for trade would also help build support for the gradual elimination of discriminatory tariff preferences. Indeed, such support is likely to be a precondition for attaining an ambitious Doha outcome that benefits all developing countries.

In considering options for expanding financial support for trade adjustment and integration, two issues are particularly pertinent: the “additionality” of aid resources and the operational framework through which additional funds are made available. Additionality may be achieved by recognizing that the Doha Round, especially an ambitious one, will generate substantial gains for the world as a whole, and that much of these gains will accrue to developed countries. The potential gains create an opportunity for donor countries to consider transferring an increment of the net gain from trade liberalization to help address adjustment costs and improve trade capacity in developing countries. Various options could be considered in this connection, in addition to more official development assistance and stronger public-private partnerships: a commitment to transfer a proportion of the tariff revenue currently collected on imports from developing countries, capturing some of the consumer gains resulting from lower agricultural support for specific products, or partial real-location of current subsidies and income support to development assistance.53 In all such cases explicit earmarking of additional funding for trade is not desirable—instead the goal would be to establish a framework to commit to redistributing some of the gains from global trade reforms.

The operational framework for allocating aid for trade must ensure harmonization of donor efforts (bilateral and multilateral) and place assistance for trade reform, adjustment, and competitiveness within the broad context of a country’s development program. The attention given to the trade and investment agenda in a country’s national development strategy depends on many factors. Whether trade capacity and trade policy reforms should be given greater priority is a decision for governments. To determine this requires that trade be considered when designing national development and poverty reduction strategies. Progress has been made in this direction and continues to be pursued, but much more can be done to better integrate trade considerations into this process. Equally important, demands for funding to address trade and supply capacity constraints need to be met by the donor community in instances where trade is considered a priority and requests for assistance have been put on the table.

The institutional mechanism to identify priorities and allocate additional aid for trade could build on the Integrated Framework for Trade-Related Technical Assistance. This mechanism has already been established, piloted, and subjected to external evaluation, and has broadly based donor and recipient support. The framework is a unique collaborative venture between 6 multilateral agencies—the IMF, International Trade Centre, United Nations Conference on Trade and Development, United Nations Development Programme, WTO, and World Bank—17 bilateral donors, and LDC governments. The primary objective is to integrate assistance in alignment with national development strategies and priorities (including poverty reduction strategies), based on diagnostic country studies and national consultations. Demand for the framework’s trade assessments is high—14 LDCs have completed the assessments and another 14 are under way or planned.54

Progress in subsuming the Integrated Framework’s recommendations into country growth strategies has been slower than anticipated. Without additional assistance, over time, legitimate questions can be raised about the efficacy of the program in providing a more effective enabling process of integration with the global trading system. The dedicated resources associated with the Integrated Framework have been limited to small-scale technical assistance. Funding for trade priorities is considered in the context of the poverty reduction strategy resource allocation and prioritization process, with donor assistance (grants) complemented by loans and other support from international financial institutions and regional development banks. While that should continue to be the case, the Integrated Framework could be used more extensively to both bolster the trade capacity of poor countries and help address the adjustment costs from an ambitious Doha Round. To do so, consideration should be given to expanding the framework’s reach beyond LDCs to include other low-income countries and consider regional trade cooperation priorities. (In Africa, for example, members of regional integration initiatives span both LDCs and non-LDCs.) Given the strong case for additional funding to meet trade adjustment and integration costs, building on the Integrated Framework could improve aid effectiveness in the trade area and ensure that identified trade priority projects are implemented.

Leveraging Regional Integration as a Tool for Development

Regional agreements among neighboring developing countries can help lower trade transaction costs and remove policy barriers to intraregional trade. Achieving this is particularly important for landlocked economies that must transport goods through neighboring countries. The agenda extends beyond tariffs and nontariff barriers imposed at borders—it includes regulatory cooperation and joint provision of infrastructure to facilitate trade. South-South integration can also play a beneficial role in helping to create the preconditions for agglomeration, clustering, and the linkages that are critical in diversifying productive activities and sustaining growth.

In Africa in particular, there is a strong regional dimension to growth, absorptive capacity, and the widening of the economic space needed to lower infrastructure costs and facilitate trade. This goes well beyond the trade agenda narrowly defined and includes improving natural resource management and increasing energy supplies and reliability. In terms of absorptive capacity, regional integration provides a means to increase aid flows in order to open markets and enhance growth prospects in developing countries. Some poor African countries are too small to raise savings for investment on their own, improve growth given limited national markets, or effectively absorb large aid flows for infrastructure. A regional strategy—such as that foreseen by the New Partnership for Africa’s Development (NEPAD), revolving around improving policies and governance and developing regional programs for groups of countries to cooperate on infrastructure (power, roads, transportation)—can do much to increase the capacity of the countries concerned to effectively use greater aid flows. As discussed in a number of recent studies,55 narrow trade agreements between developing countries have limited potential to foster economic growth, because preferential liberalization of trade policies offers substantial scope for costly trade diversion. For regional integration to be beneficial, external barriers must be low.

North-South agreements can help provide both a focal point for behind-the-border trade reforms in developing countries and mechanisms through which to engage stakeholders and policymakers in the design and implementation of reforms. Because they require trade liberalization by their developing country members, such agreements can also help reduce trade diversion created by South-South agreements among the countries concerned. They also create more opportunities for economic benefits to be realized given larger Northern markets and higher income levels.

Preferential trade agreements revolve around trade discrimination: this is an explicit objective. Because global development prospects are best served by nondiscrimination,56 from a development perspective the best outcome would be if regional agreements promoted the principle of nondiscrimination. Achieving this is a major challenge given that the driving force of trade agreements is in part mercantilist—improving (preferential) access to markets. This specific challenge is an illustration of a more general one: to change the modus operandi of designing and implementing trade agreements to place development (economic efficiency and equity) considerations more at the center of deliberations.

Agreements that the European Union and United States are negotiating with developing countries can do much good if designed in a way that puts development first. The ongoing Economic Partnership Agreement (EPA) negotiations are the most important in terms of scale and country coverage, but Japan, the United States, and other large traders are also negotiating agreements with other developing countries.57 Taking development seriously in such negotiations has a number of implications, including identifying the most appropriate form of, and membership in, counterpart regional arrangements, addressing trade barriers with neighboring countries, and identifying actions to reduce trade costs—something of great importance for landlocked countries.58 Perhaps most important in the context of EPAs and Africa is to avoid trade diversion costs and attenuate tariff revenue losses associated with a move to reciprocal free trade with the European Union.59 If African, Caribbean, and Pacific partner countries maintain current levels of protection against the rest of the world, the effect of moving to free trade with the European Union will be to transfer part of what is now collected in revenue on imports to EU producers in the form of higher prices. This points to the importance of complementing reciprocal liberalization with reductions in external barriers to trade, as well as effective assistance to facilitate trade and improve services and related behind-the-border regulation. Pursuit of nondiscriminatory tariff liberalization by developing countries where such barriers are still significant would be better than full preferential liberalization relative to the developed partner only. Many Sub-Saharan African countries still rely on import duties for a significant portion of government receipts, so revenue concerns and the ability to implement alternative revenue sources are a factor influencing the speed of liberalization.

The policy agenda confronting developing countries at the regional level is similar to that at the multilateral level: in both cases much of the challenge is to use these instruments to help address specific national and foreign trade constraints. A major difference is that in the case of regional agreements, reducing external barriers is important to reduce trade diversion. Another difference is that in the case of North-South agreements there may be significant development assistance commitments associated with the implementation of agreements. Allocating and channeling such aid for trade through a multilateral mechanism—such as the Integrated Framework—would ensure explicit recognition that the trade agenda is (and should be) largely a national, country-specific one, and that technical and financial assistance should be directed toward trade-related areas identified as priorities by countries.


3.In 2003 shares in global GDP for all LDCs, African LDCs, all low-income countries, and African low-income countries were 0.5, 0.3, 3.4, and 0.6 percent, respectively.
4.The measure of concentration reported here is the Herfindahl-Hirschmann index (the sum of squared shares).
6.World Bank (2004a) provides a comprehensive discussion of the ongoing proliferation of regional trade agreements.
7.These are reportedly specific taxes, not ad valorem, and thus imply an incentive to upgrade quality (produce higher-value products). These taxes may seem low, but because they apply to all exports (not just exports to restricted markets), they are significant.
8.In a case brought by Australia, Brazil, and Thailand against the EU sugar regime, a WTO dispute settlement panel ruled that a number of aspects of EU sugar policy violated WTO rules, including on export subsidies. African, Caribbean, and Pacific countries opposed the case on the basis that reforms in the EU regime would erode existing preferential access arrangements.
9.The OTRI belongs to the family of trade restrictiveness indexes developed by James Anderson (Boston College) and Peter Neary (University College, Dublin) for the World Bank in the early 1990s. A welfare-based measure was originally developed in Anderson and Neary (1994, 1996); an import volume-based measure was first defined in Anderson and Neary (2003). The OTRI is a theoretically well-grounded measure of protection. It does not suffer from the well-known drawbacks of simple or import-weighted tariff averages, and allows the effects of both tariffs and nontariff measures to be estimated. The methodology used for the measure comprises four steps. First, import demand elasticities by country and by product are estimated at the 6-digit level of the Harmonized System (HS) of commodity classification (some 4,200 product categories). Second, an estimate is made of the impact on imports, again at the 6-digit HS level, of core nontariff barriers (quotas, nonautomatic licensing, minimum prices, and similar policies), measures of a regulatory nature (particularly technical, product-specific regulations, and sanitary and phytosanitary measures), and domestic support granted to agriculture. Third, using these demand elasticity and impact estimates, the product-level ad valorem equivalent of the nontariff measures and agricultural subsidies is calculated for each country in the sample. Finally, tariffs and ad valorem equivalents of nontariff policies at the product level are aggregated to produce an overall measure of trade restrictiveness. The index is calculated bilaterally for every country and its partners. The OTRI calculations use the most recent measures of applied preferential tariffs—both reciprocal and nonreciprocal—as well as ad valorem equivalents of specific tariffs. In both instances the source of these data is the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII, France). It is assumed that tariff preferences are fully utilized by eligible countries and that any associated rents accrue to the exporting countries. As discussed below, however, in practice a significant share of such rents often accrues to the countries providing preferential access.
10.The statistical robustness of these relationships is not strong, especially for own trade restrictiveness imposed on imports. Removing just a small number of (outlier) countries results in the relationship either disappearing or becoming a negative association—that is, more protection is associated with less poverty. The relationship between OTRIs on exports and poverty is stronger (both the coefficient and its statistical significance).
11.See Nicita (2004c) on Mexico and Porto (2003) on Argentina, as well as the Africa-specific work summarized below. Hertel and Winters (2005) contains a review of recent evidence.
12.That is, 32 percent of the gross receipts of farmers were generated by transfers from consumers and taxpayers (OECD 2004).
13.Relative to the Global Monitoring Report 2004, the OTRIs reported here are higher on average due to the inclusion of information on mandatory product standards. These were not included in last year’s report due to an absence of data for EU members in the UNCTAD TRAINS database. Such data were collected in preparation for this report. Detailed information, including OTRIs by country that use tariff data only, and welfare-based measures of trade restrictiveness, can be found at; click on Data & statistics.
15.Whether the intent of these measures is protectionism or based on sound scientific evidence is an important question, but one that is difficult to address. Does the ban on hormone beef imports or restrictions on genetically modified food products in Europe have a protectionist element? Is the ban on Iberico ham “pata negra” or on some types of French epoisse cheese in the United States protectionist? Are they legally and scientifically defensible on health grounds? The answers to these questions will clearly vary depending on which side of the Atlantic they are asked, influenced by risk perceptions and attitudes of those who answer, regulatory requirements, and so on. The OTRI simply measures the impact that these policies have on imported quantities and prices.
16.It is not possible to make clear-cut distinctions in the data between discriminatory and regulatory measures. Even so-called core nontariff barriers identified in the UNCTAD database—such as nonautomatic licensing—may be used as a measure to enforce regulatory requirements.
17.This comprises trade as defined by the WTO General Agreement on Trade in Services (GATS).
18.These preference regimes generally target the LDCs, which are a subset of the broader low-income group. Note also that many African countries are not LDCs.
19.As discussed below, this conclusion regarding the relative impact of subsidies on global prices is supported by both partial and general equilibrium analysis. See Hoekman, Ng, and Olarreaga (2004); Anderson, Martin, and van der Mensbrugghe (2005); and Hertel and Keeney (2005).
20.Argentina is an example of a country that is significantly affected by OECD agricultural subsidies—they account for about half of the OTRI it confronts in the EU market.
21.Minot and Daniels (2002).
25.For example, Walmsley and Winters (2003) conclude that allowing inflows of service suppliers (natural persons) equivalent to 3 percent of the labor force would generate global (static) gains of more than $300 billion. Estimates of global gains from deep liberalization that include induced productivity effects and services liberalization exceed $1 trillion (Anderson 2004a; Anderson, Martin, and van der Mensbrugghe 2005) and range to up to $2.1 trillion (Brown, Deardorff, and Stern 2003). The extent to which such gains can be realized depends on complementary actions being taken to improve the investment climate and business environment so as to allow factors of production to be allocated to their most productive uses, markets to clear, and so on. This “behind-the-border” complementary agenda will require investment and is an area where development assistance can play an important role in allowing countries to capture the potential dynamic gains from trade.
29.The potential loss can be reduced by requiring any product with a bound tariff in excess of 200 percent to be reduced to the cap rate. See Anderson, Martin, and van der Mensbrugghe (2005).
32.Based on estimates reported in Francois, van Meijl, and van Tongeren (2005).
33.This is larger than the $80 billion found by Hertel and Keeney (2005) using a static computable general equilibrium (CGE) model and the latest GTAP database (which also includes preferences) because of the greater disaggregation of products.
36.These are mostly confidential; only some countries have derestricted them.
41.Regarding cereal and particularly rice, which are largely produced and consumed in Madagascar, predicted price increases following an ambitious Doha Round would have mostly neutral effects from a poverty perspective because the poorest households are both producers and consumers of rice. An increase in the price of rice would likely produce a redistribution of income between geographic areas, with rural areas benefiting at the expense of urban areas. Given that Madagascar is a net importer of rice (although only 7 percent of consumption is imported), an increase in the price of rice would have a small negative effect on overall welfare.
42.Similar conclusions emerge from other LDC country case studies undertaken as background research for the MDG taskforce report on trade (Soloaga 2004), as well as a number of recent country cases collected in Hertel and Winters (2005).
43.Previous reports, including the Global Monitoring Report 2004, have argued that trade has significant potential to help achieve poverty reduction and support higher growth rates. As noted in those reports, however, trade offers an opportunity, not a guarantee. Realization of significant reductions in poverty rates is dependent on functioning markets, linking farmers to markets, and so forth. Dealing with the complementary “behind-the-border” and investment climate agenda is critical.
45.The Global System of Trade Preferences (GSTP) was established in 1988 as a framework for negotiations aimed at the exchange of trade preferences. It is based on (differentiated) reciprocity—that is, is not unilateral.
46.Recent research suggests that rules of origin generate costs on the order of 3–5 percent of the value of the goods (Brenton 2003; Anson and others 2003). Given that many MFN tariffs in OECD countries are less than these levels, this has obvious implications for the effective value of tariff preferences. Candau, Fontagne, and Jean (2004) find that underutilization is highest in textiles and clothing (for EU imports under both the Global System of Preferences and Everything But Arms programs). In the case of the Everything But Arms initiative, exporters in principle benefit from 100 percent duty-free access, but are found to pay up to 6.5 percent average tariffs. The authors also find that the utilization of trade preferences is lower when the preferential margin is small, suggesting that compliance costs (rules of origin) are a factor.
47.See Olarreaga and Özden (2004) on the Africa Growth and Opportunity Act, Özden and Sharma (2004) on the U.S. Caribbean Basin preference program, and Tangermann (2002) on agricultural preferences in general.
49.IMF (2003) estimates the potential loss from preference erosion resulting from a 40 percent cut in protection in the Quad at 1.7 percent of total LDC exports. However, individual LDCs may suffer bigger losses due to the concentration of their exports in products that enjoy deep preferences. Of these, Cape Verde, Haiti, Malawi, Mauritania, and São Tomé and Príncipe are the most vulnerable to preference erosion. The total (aggregate) value of export revenue that would be lost by LDCs as a whole is estimated at some $530 million—similar to the $600 million figure found by Limão and Olarreaga (2005). This assumes that preferences are fully utilized and that developing countries get all the associated rents, so this is an upper bound. Alexandraki and Lankes (2004), focusing on middle-income countries, conclude that potential erosion impacts are less than 2 percent of total exports for countries that are most “preference dependent.” Six countries—Belize, Fiji, Guyana, Mauritius, St. Kitts and Nevis, and St. Lucia—would be significantly affected, with predicted export declines ranging from 7.8 percent for Fiji to 11.5 percent for Mauritius.
50.Winters (2004).
52.Examples mentioned in WTO and OECD (2004) include a regional trade facilitation program of the United Kingdom that aims to increase trading opportunities for small-scale farmers and traders through the development of common standards across goods and services and the streamlining of customs procedures in Southern Africa. Another example is France’s regional program that aims to develop “fair trade” in Africa. Increases in trade development activities have centered on regional programs, such as an International Development Association (IDA) credit to assist the development of power exports between Southern African countries and an EU multisector small and medium-size enterprise development project in Latin America.
54.For information on the Integrated Framework and the status of its activities, go to
58.These issues are discussed at greater length in Hinkle and Newfarmer (2005) in the specific case of the EPAs. For example, the form of African, Caribbean, and Pacific agreements is an issue because a customs union—the preference of the European Union—may be difficult to implement given overlapping regional agreements in Africa and the need to develop revenue sharing mechanisms.

    Other Resources Citing This Publication