A. Global Background
RTAs have been proliferating exponentially throughout the world. Nearly all countries now participate in at least one RTA, and approximately 300 RTAs, both bilateral and plurilateral, are now in force.3 A sequence of events—the failure to launch a round of multilateral trade talks in Seattle in 1999, their short-lived recovery after the Doha ministerial meeting in 2001, and an impulsive breakdown in Cancún in 2003—has sparked a renewed enthusiasm for preferential arrangements.
More fundamentally, the surge in regionalism reflects changes in the trade strategies of key members of the WTO. The United States, traditionally a champion of the multilateral trade system, has led the current wave of regionalism as it catches up with (and challenges) the EU—the front-runner in regionalism. The U.S. policy shift has resulted in an acceleration of regionalism around the world, including in East Asia, which had, until recently, relied on the multilateral system to support its export-oriented growth. These changes may have set future global trade liberalization on an uncertain path.
Two opposing views describe RTAs as either building blocks for, or stumbling blocks to, further trade liberalization.4 From a global, static point of view, the multilateral approach to trade liberalization is considered superior to the regional approach primarily because RTAs, which are discriminatory, can lead to both trade diversion and trade creation and, thus, reduce welfare when trade diversion is greater than trade creation. However, proponents of regionalism argue that the trade diversion effect of RTAs tends to be smaller than their trade creation effect. They also contend that the multilateral system is unwieldy and that proliferating RTAs can accelerate global trade liberalization. In contrast, opponents of regionalism argue that RTAs produce limited benefits or even losses for their participants and are likely to undermine the multilateral trade system and, ultimately, slow global trade liberalization. They further contend that the overlapping “hub-and-spoke” structure5 of major trade blocs increases the complexity of the world trade system and might suck investment and production into the hub and impose administrative costs on trade, especially for the “spokes.” The debate over how regionalism affects global trade liberalization is unlikely to be resolved in the near future, if at all.
B. Challenges of Regionalism for Africa
Regionalism may distract Africa from its focus on the core objective of its trade policy—creating a liberal trade regime to support its poverty reduction and growth strategies. In a world of proliferating RTAs, African countries, like many in other parts of the world, have responded instinctively to global regionalism by creating more RTAs on the continent and strengthening the existing ones. However, the pursuit of a defensive strategy to counter regionalism in other parts of the world may not be in Africa’s best interest. In particular, as argued below, RTAs aimed at helping local industries through import substitution in Africa are unlikely to expand the continent’s overall trade—even intraregional trade—given the low complementarity of natural endowments, the region’s small markets, and administrative capacity constraints.
In this context, African policymakers need to build political support to make the continent’s existing RTAs more open to the rest of the world through nondiscriminatory liberalization. Opening the RTAs, however, will be a difficult task because of the resistance from highly protected industries—some from pre-independence days—and the ideology of “self-centered development,” which inspired, among others, the Lagos Plan of Action and, generally, import substitution policies. In addition, the RTAs have inevitably created some vested interests that are opposed to broad-based liberalization. Exporters who have benefited from regional preferential market access will want to keep their niche markets. Local producers who have benefited from rules of origin will resist reductions in external trade barriers and efforts to make the rules more transparent and less restrictive. In the case of a customs union, there is a danger that once a country joins it, its trade policy cannot be altered without the consent of its partners. If the initial common external tariff is high and consensus on its reduction is difficult to reach, an RTA member is locked into a restrictive trade regime unless it is willing to withdraw from the union.
Many African policymakers also fear that open trade with the rest of the world would permanently damage Africa’s infant manufacturing base. RTAs are viewed as a training ground to prepare local industries for broad-based liberalization in the future, although few African RTAs intend to liberalize subsequently on an MFN basis. There are, however, some factors that are favorable for further opening up of African RTAs. There is now greater recognition that openness promotes growth—no country has been able to grow fast without opening its economy to international trade.6 The limitations of intra-African RTAs have become more widely acknowledged, leading some to change their views of RTAs. They see them now as a tool to integrate into the world economy by liberalizing intraregional trade first. Nevertheless, it remains to be seen whether the various political forces will result in more broad-based trade liberalization in Africa.
The global trend toward regionalism tends to disadvantage African countries. Regionalism has created “hub-and-spoke” trade relationships among countries, with large economies being the “hub” and small ones being the “spokes.” Ongoing competitive (or “additive”) regionalism further strengthens such relationships, which tend to benefit the hub disproportionately more than the spokes because of differing rules of origin, product exclusions, nontrade issues, and trade and investment diversion (World Bank, 2004). Because their economies are small, African countries have always been the spokes in their trade relations with industrial countries, especially those in Western Europe. The same World Bank study shows that if all developing countries sign a bilateral FTA (on merchandise trade only) with all the major hubs—the Quad (Canada, the EU, Japan, and the United States) plus Australia and New Zealand, sub-Saharan Africa would lose over $3 billion (in 2001 prices) in 2015, or 0.7 percent of the region’s income (1.2 percent for non-SACU countries). In fact, African countries are likely to lose even more in such a race to bilateralism. First, not all large economies want, at this stage, to establish RTAs with Africa (at least not if agriculture is included and/or antidumping is on the agenda for negotiation). Second, even if Africa can establish RTAs with all the large economies, it is not the same as multilateral free trade. Rules of origin can be and have been used to restrict market access (Krueger, 1999b; Flatters, 2002). Finally, small countries in Africa would face enormous capacity constraints in trying to negotiate and implement multiple RTAs. In contrast, the hubs can essentially replicate rules of origin each time they negotiate a new RTA and are able to access all the spokes’ markets. The hubs would have more access to inputs that are available without duties than any spoke would have (Krueger, 1999a).
The proposed EPAs could increase the benefits of regionalism but need to be carefully designed.7 Key reasons for the failures of intra-African RTAs include the lack of complementarity in trade structure, their narrow focus on intraregional tariff reductions, weak capacity for policy implementation, and a lack of policy credibility. FTAs with the EU as envisaged under the EPAs would, in principle, overcome many of these problems by taking advantage of the EU’s large market size, institutional capacity, and policy credibility. More important, the FTAs would cover more than tariffs and would include liberalization of trade in services and measures designed to foster cooperation on “behind the border” issues (such as trade facilitation and investment climate) and strengthen domestic institutions. To reap the potential benefits of the EPAs, however, these agreements will have to be truly development-oriented, and the EU will have to forgo seeking the “first mover” advantage over other countries outside Africa. This is especially important since the EPAs would essentially treat African countries (and other Africa, Caribbean, and Pacific (ACP) partners) as more equal commercial partners, via reciprocity agreements, rather than as preference beneficiaries, as is the case under the current Cotonou Agreement.
While EPA negotiations will inevitably continue, African countries should consider making MFN reduction of their external tariffs an integral part of the EPA design and implementation. There are, of course, concerns that across-the-board liberalization could entail substantial adjustment costs in the short run. The costs may justify gradual MFN liberalization rather than discriminatory liberalization. Indeed, trade liberalization can be, and often is, phased in over extended periods. If African countries want to use external commitments to facilitate trade reforms and strengthen domestic institutions, the WTO can also serve as a useful multilateral commitment mechanism. In fact, most African countries have yet to commit to more comprehensive tariff bindings under the WTO that are reasonably close to their applied tariff rates.
Preference erosion as a result of proliferating RTAs in other parts of the world reduces Africa’s international competitiveness. The main threat of preference erosion comes from multilateral liberalization and MFN reductions of trade barriers in Africa’s major trading partners, particularly the EU’s reform of its sugar regime and the removal of quotas on textile exports from developing countries.8 However, even without MFN liberalization, continued dependence on preferential schemes—such as the generalized system of preferences (GSP), the African Growth and Opportunity Act (AGOA), and the “Everything But Arms” (EBA) initiative—becomes a risky strategy for Africa as the Quad countries continue to forge RTAs with other developing countries. African countries need to make the necessary structural adjustments to maintain their international competitiveness, a daunting challenge given their heavy dependence on primary commodity exports and weak manufacturing base.
C. The Web of Africa’s Trade Arrangements
Africa has been a pioneer of regional integration. From the SACU established in 1910, through the historic Pan African Congresses, to the first regional federations, African trade integration has included numerous arrangements at regional and subregional levels. The Lagos Plan of Action, followed by the 1991 Abuja Treaty, established the AEC with subregional economic communities envisaged as the building blocks for the AEC.9
African RTAs have largely been motivated by the continent’s desire to promote growth through regional cooperation. Many African countries are landlocked small economies with inadequate infrastructure.10 Of the 53 African countries, 39 have fewer than 15 million people, and 21 have fewer than 5 million (ECA, 2004). Although Africa has 12 percent of the world’s population, it produces just 2 percent of the world’s output because its productivity is low. In 2003, sub-Saharan Africa’s GDP was 17 percent lower than Australia’s; when South Africa is excluded, Africa’s GDP is about the same as Austria’s. RTAs, by creating larger markets, are thought to enable African countries to exploit economies of scale and enhance domestic competition as well as to raise returns on investment and, hence, attract more foreign direct investment (FDI).11 Africa could realize most, if not all, of such benefits by liberalizing unilaterally and/or by participating in multilateral liberalization sponsored by the WTO (Oyejide, 1997). African leaders also believe that RTAs would increase their bargaining power in international trade negotiations and that trade integration would help reduce regional conflict.
RTAs can be viewed as particularly beneficial to Africa’s landlocked countries. A CU with coastal countries, for example, may effectively connect landlocked countries to the ocean. Most of Africa’s landlocked countries and their coastal neighbors are members of RTAs, sharing long-standing trade routes and port facilities. The share of regional trade in total trade, especially on the import side, also tends to be higher for landlocked countries than for other countries. Nevertheless, these countries stand to gain most from reductions in the region’s trade barriers vis-à-vis the rest of the world, because the marginal cost of imports from the region is generally higher. In addition, past experience suggests that RTAs have not generated the expected benefits for landlocked African countries because coastal countries have more often than not created obstacles, administrative (customs procedures) or physical (roadblocks), resulting in excessively high costs of transit or even double taxation at entry.
Africa is now a dense web of RTAs and a classic example of a variable geometry in integration (Figure 1). Most African countries have multiple RTA memberships. There are four major RTAs, with 24 member countries, at different stages of development toward a CU: the West African Economic and Monetary Union (WAEMU), the Central African Economic and Monetary Community (CEMAC), SACU, and the East African Cooperation (EAC, successor to the defunct East African Community). The first two are simultaneously monetary unions with a common currency—the CFA franc. Within SACU there is a smaller common monetary area. Four additional large groups of countries are FTAs with long-term goals of becoming CUs, monetary/economic unions, or common markets: the Economic Community for Central African States (ECCAS), which includes all CEMAC members; the Economic Community of Western African States (ECOWAS), which includes all WAEMU members; the Common Market for Eastern and Southern Africa (COMESA), and the Southern African Development Community (SADC), which includes all SACU members and overlaps with COMESA. Most other arrangements are cooperation agreements and have a limited economic impact.
The African Galaxy
Source: Authors’ compilation.In this enormous diversity, some common features of African RTAs stand out.
-
The goals are usually ambitious. These usually include promoting economic prosperity through regional integration and enhancing growth by exploring economies of scale; enlarging local markets and encouraging investment; facilitating trade liberalization; harmonizing economic policies; enhancing economic convergence; and accelerating economic diversification. As previously noted, many African RTAs are part of deeper regional integration schemes. These goals are legitimate and reflect the aspirations of the African people. Nevertheless, they are long-term development objectives and can be achieved only through sustained effort at both the regional and national levels. Moreover, they require a strong political commitment, which does not seem to have been sustained in the past, judging from the long delays and reversals of policy changes.
-
Intraregional tariff reduction is generally the primary focus. A noticeable exception is the SADC, which has devoted considerable resources to regional sectoral projects (e.g., energy and infrastructure) since its inception. Starting with an FTA, most functioning RTAs are gradually reducing intraregional tariffs while preserving their individual tariffs vis-à-vis third countries. Tariff reduction targets have been implemented to various degrees but, in most cases, are behind schedule. Tariffs are not usually reduced on sensitive commodities. With the exception of SACU, none of the RTAs has reached the point of making all regional trade tariff-free. Some RTAs also stipulate the elimination of nontariff barriers (NTBs), but have made only limited progress toward that goal. Quantitative restrictions, import bans, roadblocks, and administrative charges are still in place in COMESA. There is also evidence that new NTBs are being erected (ECA, 2004); in Western Africa, for example, they include unofficial fees at border crossings, administrative delays at ports, cumbersome customs formalities, and multiple interstate checkpoints and roadblocks (Table 1). The narrow focus on intraregional tariff reduction and failures to eliminate NTBs have limited the economic benefits of RTAs.
Official Checkpoints on Selected Routes of Western African Highways, December 2000
Official Checkpoints on Selected Routes of Western African Highways, December 2000
| Highway | Distance (Kilometers) | Number of Checkpoints | Number of Checkpoints per 100 Kilometers |
|---|---|---|---|
| Lagos, Nigeria, to Abidjan, Côte d’Ivoire | 992 | 69 | 7 |
| Lome, Togo, to Ouagadougou, Burkina Faso | 989 | 34 | 4 |
| Niamey, Niger, to Ouagadougou, Burkina Faso | 529 | 20 | 4 |
| Abidjan, Côte d’Ivoire, to Ouagadougou, Burkina Faso | 1,122 | 37 | 3 |
| Cotonou, Benin, to Niamey, Niger | 1,036 | 34 | 3 |
| Accra, Ghana, to Ouagadougou, Burkina Faso | 972 | 15 | 2 |
Official Checkpoints on Selected Routes of Western African Highways, December 2000
| Highway | Distance (Kilometers) | Number of Checkpoints | Number of Checkpoints per 100 Kilometers |
|---|---|---|---|
| Lagos, Nigeria, to Abidjan, Côte d’Ivoire | 992 | 69 | 7 |
| Lome, Togo, to Ouagadougou, Burkina Faso | 989 | 34 | 4 |
| Niamey, Niger, to Ouagadougou, Burkina Faso | 529 | 20 | 4 |
| Abidjan, Côte d’Ivoire, to Ouagadougou, Burkina Faso | 1,122 | 37 | 3 |
| Cotonou, Benin, to Niamey, Niger | 1,036 | 34 | 3 |
| Accra, Ghana, to Ouagadougou, Burkina Faso | 972 | 15 | 2 |
-
The rules of origin (ROOs) differ among the RTAs, and some of them are restrictive. SACU is the only African RTA that has achieved relatively free intraregional trade. Internal trade within all other RTAs, including FTAs and CUs that are expected to become effective in later years, is subject to ROOs. In the case of SADC, the initially proposed ROOs were relatively simple and uniform across products: they involved a change of tariff heading, a minimum of 35 percent of the value added within the region, or a maximum import content of 60 percent of the value of total inputs. However, these rules were subsequently revised and became more restrictive—not only did they become product-specific, but they also required detailed technical processes (Flatters, 2002; Flatters and Kirk, 2004). ROOs can be costly to enforce because they require documents to prove origin and verification procedures at border crossings. Estimates of such costs for African RTAs are not available, but evidence elsewhere suggests that they can be considerable.12
-
External trade barriers remain relatively high. Much progress has been made to reduce external tariffs in Africa, especially in the context of IMF- and World Bank–supported programs. Nevertheless, the simple average of applied MFN tariffs in Africa is higher than that in other developing regions of the world (Table 2). As in other regions, the average tariffs in Africa often mask large variations between countries and products. While some African countries have reasonably low average tariffs (e.g., 9 percent in Eritrea and 11 percent in Mozambique), others have high tariffs (e.g., 39 percent in the Comoros and 37 percent in Nigeria) that could lead to large trade diversion under the RTAs. NTBs against non-RTA members remain significant, even though information documenting them is lacking (ECA, 2004).
Simple Average of Applied MFN Tariffs, 1997 and 2004
(In percent)
Simple Average of Applied MFN Tariffs, 1997 and 2004
(In percent)
| 1997 | 2004 | ||
|---|---|---|---|
| Africa | 21.6 | 17.2 | |
| CEMAC | 19.9 | 19.4 | |
| COMESA | 23.6 | 18.5 | |
| ECOWAS | 20.0 | 16.8 | |
| WAEMU | 22.6 | 17.8 | |
| SADC | 20.0 | 15.9 | |
| Other developing countries | 14.4 | 11.6 | |
| Asia Pacific | 16.1 | 12.1 | |
| Europe | 11.2 | 9.7 | |
| Middle East and Central Asia | 16.9 | 12.2 | |
| Western Hemisphere | 13.2 | 12.2 | |
| Industrial countries | 8.7 | 5.7 | |
Simple Average of Applied MFN Tariffs, 1997 and 2004
(In percent)
| 1997 | 2004 | ||
|---|---|---|---|
| Africa | 21.6 | 17.2 | |
| CEMAC | 19.9 | 19.4 | |
| COMESA | 23.6 | 18.5 | |
| ECOWAS | 20.0 | 16.8 | |
| WAEMU | 22.6 | 17.8 | |
| SADC | 20.0 | 15.9 | |
| Other developing countries | 14.4 | 11.6 | |
| Asia Pacific | 16.1 | 12.1 | |
| Europe | 11.2 | 9.7 | |
| Middle East and Central Asia | 16.9 | 12.2 | |
| Western Hemisphere | 13.2 | 12.2 | |
| Industrial countries | 8.7 | 5.7 | |
-
Among the three formal customs unions, SACU has a common revenue pool and redistribution system among members, but it has maintained high levels of protection against the rest of the world. WAEMU has made headway in reducing tariffs vis-à-vis the rest of the world, but it still has to deal with official and unofficial internal trade barriers (including ROOs). CEMAC has functioned poorly, making little progress in reducing its tariffs against the rest of the world. Similarly, the record of the African FTAs on reducing trade barriers against nonpartner countries is mixed.
-
Revenue losses are important concerns in RTA design and implementation. Trade taxes13 generate almost one-third of all government revenues in African countries (Agbeyegbe, Stotsky, and WoldeMariam, 2004). It is not surprising, then, that a discussion on trade liberalization—whether in the context of multilateral, regional, or bilateral arrangements—also provokes a discussion of its potential consequences for government revenues.14 Since the 1980s, the revenue-to-GDP ratio in sub-Saharan Africa has remained virtually stagnant, while the resource needs for the provision of public services and infrastructure have increased sharply. Potential revenue losses from RTAs are generally small because intraregional trade in most RTAs typically accounts for about 10 percent of total trade. In cases where intraregional trade is important and the CET is also reduced when a customs union is formed, revenue losses can be significant.15 Because members of most RTAs trade mainly with the rest of the world, concerns over revenue losses have been a key obstacle to broader trade liberalization in Africa.
-
The desired level of regional integration is high. The five major RTAs (ECOWAS, WAEMU, COMESA, CEMAC, and SADC) aim to establish a customs/economic union and a common market. CEMAC is nominally already a customs union, while the other four are at various stages of forming an FTA—an interim step for deeper integration. While the WAEMU member states of ECOWAS have established a customs union with a CET, similar progress has not yet been achieved for the ECOWAS countries as a group. By aiming for deep integration, African countries are seeking to maximize the benefits of regional integration, but they are also imposing on themselves a substantial demand for administrative capacity and a high degree of economic convergence among themselves. The prospect of surrendering trade policy to CU authorities and accepting the resulting redistribution of customs revenue has often aroused concerns and tensions. Although the push for a high degree of integration is not entirely responsible for the poor record of RTA implementation, it may have exceeded the limit of regional capacity and political will, leading to delays in implementation. Diverse national interests have led to the establishment of a large number of RTAs on the African continent, often with overlapping membership and conflicting commitments, which further complicate implementation.
The departure of RTAs from the General Agreement on Tariffs and Trade (GATT)/World Trade Organization (WTO) principle of nondiscrimination (the most-favored-nation—MFN—clause) is sanctioned by GATT Article XXIV and the Enabling Clause. However, Article XXIV requires that an RTA must (1) not “on the whole” raise protection against nonmembers; (2) reduce internal tariffs to zero and remove “other restrictive regulations of commerce” other than those justified by other GATT articles; and (3) cover “substantially all trade.” The Enabling Clause relaxes the conditions for RTAs among developing countries. For example, it drops the conditions on the coverage of trade and allows developing countries to reduce tariffs on mutual trade in any way they wish. See World Bank (2000) for a more detailed description of the WTO rules on RTAs.
See Bhagwati and Panagariya (1996), Baldwin and Venables (1996), and Panagariya (2000) for comprehensive surveys of the theoretical literature on preferential trade agreements. World Bank (2000, 2004) and Schiff and Winters (2003) provide extensive surveys of the empirical literature from a development perspective.
A “hub-and-spoke” structure refers to a set of trade relationships in which a dominant (hub) country simultaneously has separate RTAs with individual smaller countries (spokes), which do not normally form RTAs between themselves—analogous to a hub-and-spoke system in air transportation.
For the growing body of empirical literature on trade, growth, and poverty, see Sachs and Warner (1995), Dollar and Kraay (2001), Berg and Krueger (2003), Baldwin (2003), and Panagariya (2004). For a critique of the literature, see Rodriguez and Rodrik (2001). The thrust of this literature is that growth is essential for reducing poverty and that trade and growth are positively related. However, the role of trade policy in the trade-growth-poverty link is still a subject of debate.
See Panagariya (2002) and Hinkle and Schiff (2004) for a detailed discussion of key issues concerning African EPAs with the EU.
Under the Uruguay Round Agreement on Textiles and Clothing (ATC), all quotas on textile and clothing exports from developing countries were phased out at the beginning of 2005. However, pressure has been mounting in the United States and EU to curb Chinese exports.
Most African RTAs have yet to be recognized by the WTO (United Nations Economic Commission for Africa (ECA), 2004). For those that have been notified to the WTO, they were mostly reported under the Enabling Clause. No WTO members have requested examination of any of these RTAs for consistency with WTO rules.
There are 15 landlocked countries in Africa: Botswana, Burkina Faso, Burundi, Central African Republic, Chad, Ethiopia, Lesotho, Malawi, Mali, Niger, Rwanda, Swaziland, Uganda, Zambia, and Zimbabwe.
However, there is little evidence that RTAs among very small countries promote growth.
A study undertaken by the European Free Trade Association (EFTA) estimated that the cost to EFTA members of documenting origin to receive duty-free entry into the EU averaged 3–5 percent of the export price (Herin, 1986, cited in Krueger, 1999a, p. 112). On the other hand, it is worth noting that ROOs may help reduce the potentially negative welfare effect of an FTA when these rules prevent or reduce costly trade diversion. See Krishna and Krueger (1995) for details.
Trade taxes include border taxes on trade only (excluding the VAT and excise on imports).
Ebrill, Stotsky, and Gropp (1999) find that trade tax revenues tend to fall with tariff levels whenever the latter—measured as the ratio of trade tax revenue to import value—are below 20 percent, whereas Khattry and Rao (2002) estimate this threshold to be around 40 percent. Ancharaz (2003) finds that fiscal dependence on trade taxes makes trade reform less likely to happen.
The World Bank (2003c) estimates that the planned CU among members of the EAC could result in a revenue loss of 5.2 percent for the region (5.9 percent for Kenya, 3 percent for Uganda, and 2 percent for Tanzania). The prospective FTAs with the EU would result in more significant revenue losses, given that the EU accounts for a larger share of total African imports (about 40 percent).
