Comesa and Sadc
Prospects and Challenges for Regional Trade Integration
  • 1, International Monetary Fund

Contributor Notes

Author(s) E-Mail Address:

Regional integration has been seen in Africa as a means of encouraging trade and securing economies of scale. This paper examines in detail the prospects and challenges for trade expansion in the two most prominent arrangements in eastern and southern Africa: the Common Market for Eastern and Southern Africa (COMESA) and the Southern African Development Community (SADC). It finds that possibilities of growth in intraregional trade may be limited, but that the two arrangements offer opportunities for member countries to gain policy credibility for trade reforms and tariff liberalization and to address structural weaknesses. In this regard, the negotiation of the Economic Partnership Agreements with the European Union can also have a significant impact.


Regional integration has been seen in Africa as a means of encouraging trade and securing economies of scale. This paper examines in detail the prospects and challenges for trade expansion in the two most prominent arrangements in eastern and southern Africa: the Common Market for Eastern and Southern Africa (COMESA) and the Southern African Development Community (SADC). It finds that possibilities of growth in intraregional trade may be limited, but that the two arrangements offer opportunities for member countries to gain policy credibility for trade reforms and tariff liberalization and to address structural weaknesses. In this regard, the negotiation of the Economic Partnership Agreements with the European Union can also have a significant impact.

I. Introduction

Regional integration has long been seen in Africa as a means of achieving industrialization and modernization through encouraging trade and securing economies of scale and market access. Consequently, regional arrangements have sprung up all over Africa (Aryeetey, 1997). These have, however, done little to halt the marginalization of Africa in world trade—Africa’s share of world trade in goods and services dropped from more than 5 percent in 1980 to around 2 percent in 2003 (Figure 1). The decline in Africa’s share in world trade has happened simultaneously with an increase in the overall share of developing countries indicating that the growth of African countries’ trade has lagged behind that of developing countries in Asia and Latin America. The weak performance of African trade is a particular concern.

Figure 1.
Figure 1.

Share of Africa in World Trade, 1980–2003 (in percent)

Citation: IMF Working Papers 2004, 227; 10.5089/9781451875447.001.A001

Source: IMF staff calculation based on World Trade Organization (WTO) data.

Several studies (see Iqbal and Khan, eds., 1997) have pointed out that African trade has been hindered by distorted trade regimes and high transaction costs owing to inadequate transport, information, and communications infrastructure. African countries have the most restrictive trade regimes among all groups of countries, with high tariffs, a large number of often specific and seemingly arbitrary exemptions, and significant degrees of tariff escalation (Subramanian and others, 2000). Even though formal nontariff barriers (NTBs) are not very common any more, there continue to be informal barriers such as nonacceptance of rules-of-origin certificates, and cumbersome and inconsistent customs procedures.

Regional integration efforts have also been marked by a lack of political commitment and policy reversals in implementing harmonization provisions, multiple and conflicting objectives of overlapping regional arrangements, and limited administrative resources (Iqbal and Khan, eds., 1997). As a result, even though regional trade arrangements (RTAs) have been in place over the last several years (SADC-1992, COMESA-1982, ECOWAS-1975), they do not appear to have been successful in generating a trade expansion in Africa.

Jebuni (1997) has argued that protectionist trade regimes in Africa have created a strong anti-export bias and have rendered preferential liberalization meaningless. More generally, there is a preponderance of evidence (Baldwin, 2003; Wacziarg and Welch, 2003; and Warner, 2003) linking poor growth performance to trade restrictions, on the one hand, and, on the other hand, linking trade liberalization to improved growth and investment. It is also not clear whether greater integration of regional trade has preceded or followed a formal regional arrangement. For instance, in East Asia (World Bank, 2004), integration with the global economy provided a strong impetus to expansion of intraregional trade and formal regional preferential arrangements were established after this regional integration.

The momentum for RTAs in Africa has picked up in recent times, driven in part by global trends in the same direction and in part by the negotiation of EPAs with the EU. Based on existing knowledge about the links between trade and growth, one clear inference that can be drawn is that African countries should simultaneously pursue nondiscriminatory liberalization and regional integration. As important, RTAs should be used as a mechanism to enhance policy credibility in the region and lock in trade reforms. Regional cooperation can also play an important role in addressing the various infrastructure constraints and limitations on resources in Africa. In this regard, RTAs should be designed to maximize their benefits (Schiff and Winters, 2003; World Bank, 2004; and Lawrence, 1997). These desirable attributes of RTAs may broadly be described as open regionalism and deep integration (see Box 1).

Open Regionalism and Deep Integration

Open regionalism is broadly characterized by a commitment to most-favored-nation (MFN) liberalization; comprehensive product coverage in goods and services with few exclusions; a strong focus on reducing transaction costs at borders; and liberal, clear, and consistent rules of origin (World Bank, 2004).

Open regionalism forms a key ingredient in the success of an RTA and plays an important role in expanding the scope for competition and trade creation and minimizing trade diversion.

Deep integration is the inclusion in regional arrangements of liberalization commitments that go far beyond the multilateral agreements. Commitments in areas such as investments, intellectual property rights, competition policy, standards, and trade facilitation at a regional level help reconcile divergent national practices and can create supranational implementation mechanisms.

When deep integration is tailored to the level of development of a country, it has the potential to improve institutions and impart credibility to trade reforms beyond what would be possible in the multilateral context (World Bank, 2004; and Lawrence, 1997). In addition, regional cooperation has the potential to alleviate trade development and supply constraints.

The purpose of this paper is to examine and discuss prospects and challenges for trade expansion in the two largest RTAs in eastern and southern Africa 2—COMESA and SADC. 3 The paper is structured as follows. Section II describes the features of the two agreements and the extent to which they possess the desirable attributes outlined here. Prospects for trade diversion and trade expansion in the two RTAs are discussed in Section III. Section IV is based on the current plans of COMESA and SADC to implement customs unions in the near future. There are several challenges in simultaneously pursuing the desirable goal of MFN liberalization and agreeing on a common external tariff. These issues are discussed in Section IV. In Section V, we describe and outline some of the issues, challenges, and opportunities in the ongoing EPA negotiations. Section VI concludes.

II. Features of COMESA and SADC

This section describes the features of COMESA and SADC, the two most prominent RTAs in eastern and southern Africa, in terms of membership, and evaluate the extent to which they posses the desirable attributes of open regionalism and deep integration (see Annex I for details).


Features of the Agreement

COMESA originated as a preferential trade area (PTA) in 1982 and has 19 members at present. Eleven of the nineteen member states participate in a free trade area (FTA); others trade on preferential terms.4 The FTA was formed in October 2000 as the result of a long period of tariff reductions and has followed some of the principles of open regionalism.5 This is reflected in a comprehensive product coverage with no exclusions, and a set of liberal and clear rules of origin (see Box 2 for details on rules of origin). These two features of the COMESA FTA agreement, by themselves, have done a great deal in liberalizing intraregional trade.

Rules of Origin in COMESA and SADC

As has been noted in World Bank 2004, administrative costs to importers due to rules of origin can be fairly high. Therefore, adopting simple and clear rules of origin is an important aspect of regional arrangements.


Goods qualify for preferential treatment if they undergo substantial transformation such that they contain a minimum of 35 percent regional value-added, or include non-COMESA imported materials worth no more than 60 percent of the value of total inputs used, or undergo a single change of tariff heading. There is a list of “goods of economic importance” to member states, according to which the domestic value added requirement is relaxed to 25 percent.


In 1996, SADC trade negotiators initially chose the same simple and transparent rules of origin as COMESA. Goods would qualify for preferential treatment if they underwent a single change of tariff heading, contained a minimum of 35 percent regional value-added, or included non-SADC imported materials worth no more than 60 percent of the value of total inputs used.

By the time of the implementation of the SADC Trade Protocol in 2000, protectionist interests in member states pressed for exceptions to these rules on a sector-and product-specific basis (Flatters, 2002; and Nhara, 2003). The rules were modified to include detailed technical process requirements, lower permitted import content and higher domestic value added requirements.

Wheat flour and related products, coffee, tea and spices, machinery and electrical products, mineral fuels, motor vehicles, and textiles and garments are particular sectors where rules of origin are either not yet agreed or have been made restrictive in order to protect local industry. For instance, in textiles and garments, the rules of origin require double transformation to qualify for preferences—for example, garments must be made from regionally produced fabric and fabric must be made from regionally produced yarn.1 In some sectors such as wheat flour and its products, mineral fuels, motor vehicles, and machinery and electrical products, negotiations to finalize rules of origin are still continuing. With respect to the sugar industry, SADC has agreed on a system of quotas to provide nonreciprocal duty-free market access into the Southern African Currency Union (SACU).

1This is a fairly restrictive provision as pointed out in Flatters (2002), since regional capacity for producing quality fabric and yarn is severely constrained. The double transformation rule has been relaxed to the single transformation rule by SACU for the LDC members of SADC till 2005.

The agreement reaches beyond border measures and provides for the monitoring of NTBs. There is an extensive program of cooperation and technical assistance in trade facilitation and harmonization of standards. A formal dispute settlement mechanism in the form of a COMESA Court of Justice is provided, but disputes in general have been successfully handled through an informal process of diplomatic consultation and the court is not yet operational. The COMESA agreement has also incorporated temporary balance of payments related safeguards. Progress is being made towards a regional framework agreement for investment and a competition policy.

Perhaps due to a lack of political commitment in member countries, the COMESA FTA has been hampered by country level institutional changes and prevailing structural constraints. Eight of the nineteen members have not yet joined the FTA. Among those that have joined, informal NTBs such as nonacceptance of rules of origin certificates, cumbersome bureaucratic procedures, and restrictive standards are some common problems. Some countries have replaced tariffs with discriminatory excises. On the issue of services liberalization, little progress has been made.

Plans for Further Integration

For several years, COMESA has been studying a common external tariff (CET) structure of 0, 5, 15, and 30 percent with plans to implement it by December 2004 (COMESA Secretariat, 2003b). Work ongoing in the areas of customs management and trade facilitation, development of a regional competition policy, etc., has been geared to support the customs union (see Annex I). However, with the setting up of a customs union by the EAC in 2003, two COMESA members, Kenya and Uganda, have agreed on a CET structure of 0, 10, and 25 percent.6 As a result, confusion has ensued, discussions on the CET have been postponed until May 2005, and further delays are likely.7 In addition, a number of alternative CETs are presently under consideration and member states have not yet agreed on the maximum tariff rate or the number of bands. There is stiff resistance to reducing the number of tariff bands and the proposed maximum CET. Going forward, COMESA’s goal to establish a customs union has the potential to unravel the goals of liberalization. Failure to agree on a low and uniform CET will have important repercussions for the region in that several countries could be locked in to fairly restrictive trade regimes.

There are important impediments to the goal of establishing a low and uniform CET. Protectionism has not only manifested itself in resistance to lowering the number of tariff bands and the maximum tariff rate but also in discord over the classification of goods into the four categories of raw materials, capital goods, intermediates, and finished goods.8 One suggestion has been to make an exception from the CET for a limited number of goods for a specified period of time. With 19 members in COMESA, there will be a need for political commitment to ensure that the list of exceptions to the CET does not get extensive. Another significant impediment to agreeing on a low and uniform CET is the dependence on trade taxes as a source of revenue for many countries (Table 5). While a few countries have already aligned their tariff structures with the proposed CET, a number of them have significant adjustment ahead.

Several of the proposed features of the COMESA customs union will make it essential to maintain border controls for intraregional trade. As a result, member countries will be unable to realize efficiency gains from the customs union, making it an exercise in futility. To make things worse, individual countries will have given up a great deal of discretion over their trade policy and future liberalization in the process.

The proposed features of the COMESA customs union that would necessitate border controls in the region are discussed below. First, the proposed exceptions to the CET for a limited number of goods will require member countries to maintain border controls within the customs union. Second, the dependence on trade taxes by member countries has led to an agreement that revenue would be kept by the country of final consumption. This allows countries to keep their import revenues—but will require enormous effort and cooperation on the part of customs administrations to track the country of final consumption. The motivation for countries to avoid a genuine revenue-sharing arrangement comes from the need to maintain fiscal sustainability in the face of difficulties in raising revenue from domestic sources. However, this has the potential to completely overwhelm the regional customs administrations and erode efficiency gains from the customs union. It is worth pointing out that customs administration capacity will be similarly burdened due to overlapping membership between COMESA, SADC, and the EAC and the associated need to track qualifying domestic/regional content for each of the arrangements. These problems will be further accentuated due to the implementation of preferences in African Growth and Opportunity Act (AGOA) or the EU EPAs.

COMESA has plans for eventual formation of a common market and a monetary area, but those plans are in the distant future.


Features of the Agreement

SADC originated as an organization of Frontline states (Angola, Botswana, Lesotho, Malawi, Mozambique, Swaziland, Tanzania, Zambia, and Zimbabwe) to resist the influence of South Africa in the region and was transformed into a development community only in August 1992 with the signing of the Treaty of Windhoek.9 The organization has had a very different approach to regional integration than COMESA—it has concentrated on relaxing the supply side constraints to trade through regional cooperation in various sectors as diverse as infrastructure, agriculture, transportation and human resources (see SADC’s Regional Indicative Strategic Development Plan (RISDP)). Its trade protocol, which was signed in 1996 and led to the launch of an FTA in 2000, is only one of several protocols of cooperation in the organization. Eleven of the 13 SADC member states have joined the FTA.10

The SADC Trade protocol, which is based on negotiations and offers by contracting parties, aims for liberalization of all trade by 2012. Member countries have agreed to liberalize 85 percent of intra-SADC trade by 2008 and liberalize sensitive products by 2012. A small number of sectors are excluded from the liberalization—narcotics, precious and strategic metals such as gold, silver, platinum, etc., second hand goods, and some others for environmental reasons and the like are in this category.

There is a concern that tariff reductions under SADC have been backloaded and the pace of liberalization has been slow (Kritzinger-van Niekerk and Moreira, 2002). Another important issue is that over time, the rules of origin in SADC have become restrictive and product-specific under pressure from member states (see Box 2). The fact that in some sectors negotiations on the rules of origin are still ongoing is probably indicative of a lack of political commitment to liberalization. The complicated and restrictive rules of origin are likely to increase administrative costs and will make it difficult for exporters to take advantage of SADC preferences. As such, they constitute a serious obstacle to the liberalization of intraregional trade. There is an attempt to ameliorate possible polarization effects on least developed country (LDC) member states through nonreciprocal market access in sugar and more relaxed rules of origin in textiles.11 The agreement also includes provisions for protection of infant industries and antidumping and safeguard measures.

While the agreement calls for elimination of NTBs and the liberalization of services trade, not much progress has been made in either of these areas. There is no institutional mechanism for reporting of NTBs or resolution of disputes, while the liberalization of services is a futuristic provision. Some work has been completed on harmonization of customs procedures. Implementation of the protocol in terms of meeting liberalization commitments on merchandise trade has been uneven with Tanzania, Zambia, and Zimbabwe lagging behind.

While the trade liberalization aspect of SADC appears to be lacking, its approach of addressing structural impediments and supply constraints through sectoral cooperation initiatives is an important one. Active areas of cooperation where there has been progress include the monetary and financial areas—training and capacity building in central banks, development and harmonization of payments, clearing and settlement systems, and some others. Despite the ambitious goals of the SADC sectoral initiatives, progress made thus far has been somewhat limited and it has been widely acknowledged that there is a lack of mechanisms for evaluating and monitoring projects or assessing their effectiveness. Moreover, there is a concern that the organization is still more of a political block than a true instrument for regional economic integration.

Plans for Further Integration

More recently, SADC has focused on macroeconomic stability and convergence in member countries in order to achieve the formation of a common market over the medium term. In this regard, SADC has announced a plan in March 2004 the goals of which include establishment of a SADC customs union and implementation of a common external tariff by 2010, a common market pact by 2012, and establishment of a SADC central bank and preparation for a single SADC currency by 2016.

While the goal of deeper regional integration is desirable, there is a concern that the structure of the proposed SADC customs union will mirror that of SACU. 12 Even though the tariff structure of the latter has been simplified somewhat in recent years, it remains complex, consisting of ad valorem, specific, mixed, compound, formula duties based on reference prices, and other duties and charges. The ad valorem duties cover around three-fourths of tariff lines and comprise 39 bands, ranging between 0 percent and 55 percent. The specific, mixed, compound, formula duties, and other duties and charges, cover an important set of agricultural and agro-industrial products and apparel. While the SACU has harmonized applied customs tariffs, excise duties, customs valuation, rules of origin, and contingency trade remedies, not much else has been harmonized (for instance, customs procedures, technical standards, etc.).13 In addition, at a trade restrictiveness rating of 5, the current trade regime in SACU cannot be characterized as open. 14

III. Prospects of COMESA and SADC

In this section, we evaluate the prospects for expansion of intraregional trade in COMESA and SADC. We discuss concerns with regard to overlapping memberships and product complementarities, as well as trends in intraregional trade. These are useful to shed light on the potential benefits from COMESA and SADC. We find that product complementarities and levels of intraregional trade are low, although intraregional trade has been expanding at a healthy pace in the last few years. Therefore, as we discuss later, the main benefits from COMESA/SADC are likely to come from their ability to impart credibility to trade policies in the region and their success in addressing structural impediments to trade and behind the border measures. There is also an urgent need to rationalize membership in these organizations.

A. Overlapping Memberships in Regional Arrangements

A number of regional arrangements are already in place in the eastern and southern Africa region (COMESA, EAC, IOC, SADC, SACU, among others) and most countries are members of more than one such organization (Kritzinger-van Niekerk and Moreira, 2002). For instance, seven countries are members of both COMESA and SADC. One member of the EAC is a member of SADC while two members belong to COMESA.

Overlapping memberships between the various regional arrangements have costs. Negotiating resources and capacity have been stretched thin across the region. There are administrative costs related to often complex rules of origin. Multiple membership fees are expensive to pay and maintain. Conflicting objectives among rival arrangements have contributed to a lack of progress in many areas (Iqbal and Khan, eds., 1997). As several of these arrangements are in various stages of forming customs unions (COMESA, SADC, and the EAC), it has become clear that conflicts of membership need to be resolved, but this may prove politically difficult.

B. Product Complementarities

Product complementarities between countries are an important indicator of the potential for expansion of intraregional trade. The bilateral product complementarity index between two countries j and k (Cjk) is defined as:


where Xij represents the share of good i in the total exports of country j and Mik represents the share of good i in the total imports of country k (Tsikata, 1999). The index is a measure of similarities between the export basket of one country and the import basket of another country. The value of the complementarity index can range from zero, which represents no complementarity between exports and imports of two countries, to one hundred, which implies a perfect match. The higher the index between two countries, the greater the product complementarity.

We calculate bilateral product complementarity indices for COMESA and SADC member countries using UN-COMTRADE data at the two-digit classification level. These calculations are reported in Tables 1 and 2 below for COMESA and SADC respectively. For example, in Table 1, the complementarity index between Egypt’s exports and Burundi’s imports is 42.1, which is higher than, for instance, that between Malawi’s exports and Burundi’s imports (7.1). Our results indicate that within COMESA, product complementarities between exports of Egypt and imports of the other member countries average to 43.0 while those for Kenya’s exports to the region average to 38.6. For all other countries, the average product complementarity for exports is far lower.15

Table 1.

Bilateral Complementarity Indices in COMESA

article image
Source: IMF staff calculations using UN COMTRADE data.

Indices for Angola, the Democratic Republic of the Congo, Djibouti, and Eritrea were not computed because of lack of data.

Table 2.

Bilateral Complementarity Indices in SADC

article image
Source: IMF staff calculations using UN COMTRADE data.

Indices for Angola, the Democratic Republic of the Congo, and Lesotho were not computed because of lack of data.

What these numbers imply is that while there is scope for Egypt and Kenya to export to the region, there is not much scope for other countries to do the same since there are few complementarities for their exports. It is significant that Kenya and Egypt, which can be viewed as the relatively more developed member countries, do not import many of the products exported by other countries in the region. Similarly in SADC, there is complementarity between South Africa’s exports and the imports of the rest of the region, but not vice versa. The asymmetric complementarity essentially implies that the more developed economies of Egypt, Kenya and South Africa are in a much better position to market their exports in COMESA/SADC but the less developed countries are unable to find significant markets in COMESA/SADC. The asymmetric product complementarity in favor of the more developed members raises a concern over possible polarization as investment may be attracted towards the larger and more industrially diversified economies in the region.

Our results are in line with Yeats (1998) which finds that African countries tend to have exports concentrated in a few commodities, reducing their possibilities of intraregional trade. A similar result is obtained in Chauvin and Gaulier (2002) for SADC—they point out that it will be hard for South Africa to function as a “growth pole” or an important outlet for the exports of the rest of the SADC region.16 The authors find that SADC countries have similar disadvantages in manufactured goods (except South Africa) while having similar advantages in primary goods. South Africa appears to be the only source of manufactured goods, and even here, the range of products is limited. The major finding of that study is that potential to expand trade within SADC is small and existing opportunities may already have been exploited.

C. Trade Diversion Versus Trade Expansion

Tables 3 and 4 report growth rates in intraregional and total trade for the entire COMESA and SADC regions—Zimbabwe is excluded because of concerns over data quality. Intraregional trade is computed taking into consideration all present members of COMESA and SADC, even though the membership may have been different over time.17 This is to ensure that growth estimates do not reflect merely additions in membership but genuine trends in trade.

Table 3.

Intraregional Trade

article image
Source: IMF staff calculations based on IMF, Direction of Trade Statistics. Zimbabwe is excluded from all computations.

Data are unavailable for Eritrea and Swaziland.

Table 4.

Total Trade

article image
Source: IMF staff calculations based on IMF, Direction of Trade Statistics. Zimbabwe is excluded from all computations.

Data are unavailable for Eritrea and Swaziland.

Table 5.

Trade Tariffs and Revenues in Eastern and Southern Africa

article image
Source: Tariffs data are from Trade Policy Information Database. Tax data for the tax year 2000 are courtesy of Thomas Baunsgaard.

Within COMESA, it can be seen from Table 3 that intraregional trade has expanded at an average annual rate of 18.8 percent since the creation of the FTA in 2000. Moreover, it appears that growth in intraregional trade slowed in the mid- and late-1990s but recovered after the creation of the FTA. While the time period under consideration is relatively short to arrive at any definitive conclusion about the role of the FTA, the rapid response of intraregional trade raises the question of trade diversion versus expansion. At this time, the literature has not empirically examined the question of trade diversion but the possibilities are widely acknowledged since 7 of the 19 member countries have trade regimes that can be described as at least moderately restrictive (Table 6).18

Table 6.

Restrictiveness Ratings in Eastern and Southern Africa, 2003

article image
Source: Trade Policy Information Database. See footnote 14.

To shed light on the question, it does not appear that an increase in imports from the region has been accompanied by a decline in imports from the rest of the world in recent years. In fact, since total imports have grown at around 8 percent since the mid-1990s (Table 4) and from the very small share of intraregional trade in total imports, it is obvious that imports from the rest of the world must also be growing at close to 8 percent. Moreover, while the proportion of intraregional imports in total imports has gone up, it does not appear to have shifted dramatically. In other words, by this admittedly rather coarse test, we find very limited evidence of trade diversion. Total exports from the COMESA region also have been growing at a healthy pace of almost 10 percent in recent years.

Examining the growth rates of intraregional and total trade for the SADC region in Tables 3 and 4, the period between 1991–95 stands out as the period with significant average annual growth in intraregional trade, around 35 percent. This is far higher than the growth in total imports of the region, which grew at a rate of around 10 percent. This expansion in intraregional exports is mainly due to the trade integration of South Africa and the lifting of sanctions after the abolition of apartheid. During the mid- and late-1990s, total imports into the region stayed flat while intraregional trade grew at a marginal rate of under 2 percent.

In recent years (2000–03), intraregional and total trade in the SADC region appear to have broken out of the stagnation of the mid-1990s. Intraregional imports appear to have grown at a slower annual pace than total imports—clearly, these trends do not support the hypothesis of trade diversion. In addition, growth in total exports from the region looks as though it has picked up as well after being practically flat over the 1990s.

While the time period since 2000 is too short a time frame to draw firm conclusions, the preceding provides some evidence to refute concerns over trade diversion from COMESA and SADC. On the subject of trade expansion, the growth in total exports over the last few years is encouraging and one might cite anecdotal evidence of recent diversification into nontraditional exports in the region. However, it is not clear what impact COMESA and SADC might have had—one cannot say whether growth in exports in the region is due to economies of scale, the competitive effect, or other benefits of COMESA/SADC or due to unilateral liberalization undertaken by several member states in recent years (Subramanian and others, 2000), or other factors such as AGOA.

D. Potential for Trade Expansion in COMESA and SADC

Although intraregional and total trade have been growing at a healthy pace in recent years, opportunities for expansion of trade from greater integration in the COMESA and SADC regions might be somewhat limited since product complementarities and levels of intraregional trade are low and there is a risk of polarization.

As has been noted in the literature (see Iqbal and Khan, eds., 1997; and Oyejide, Elbadawi, and Yeo, 1997), African trade has been hindered by a number of factors including distorted trade regimes, high transaction costs due to inadequate transport, information and communications infrastructure, lack of political commitment and frequent policy reversals, difficulties in implementing harmonization provisions, multiple and conflicting objectives of overlapping regional arrangements, and limited administrative resources.

It has been pointed out that RTAs can play an important role in addressing these hindrances to trade (Oyejide, 1997). They can help address concerns over policy credibility by locking in domestically implemented trade liberalization and functioning as an agency of restraint. Regional cooperation can be effective in addressing weaknesses in infrastructure, and harmonization of standards and customs procedures. A strong RTA can also strengthen Africa’s bargaining power in the multilateral negotiations. In the same vein, Lawrence (1997) argues that regional integration can be a means to achieve deep integration—where deep integration includes integration of national regulatory systems and policies, competition policies and investment rules. Deep integration is viewed as having the potential to create an open and credible policy environment to encourage foreign direct investment and export-oriented growth. According to Chauvin and Gaulier (2002) also, important structural changes may be necessary in order to expand trade potential in the region.

In some ways, the agenda for integration in SADC and COMESA reflects attempts to address these challenges in development of African trade. SADC has taken an approach to regional integration that attempts to address infrastructure constraints, while COMESA has placed more emphasis on harmonization of standards and customs procedures and trade facilitation. However, the lack of political commitment has afflicted both organizations, impeding progress in implementing their vision. The lack of political commitment is also reflected in the fact that the larger economies of the region (South Africa in SADC, Kenya, and Egypt in COMESA) have not taken on active roles as champions of regional integration and liberalization.

As a result, both organizations have a great deal to do in order for countries to take advantage of possible benefits from regional integration. Going forward, it is important that COMESA and SADC become effective vehicles for nondiscriminatory tariff liberalization as well as addressing weaknesses in infrastructure, harmonization of standards and customs procedures. COMESA and SADC should function as agencies of restraint and provide policy credibility through locking in trade openness.

IV. Challenges in Tariff Liberalization and Harmonization

As discussed in the previous section, regional integration in COMESA and SADC can deliver important benefits to the region. An important component of maximizing potential benefits from COMESA and SADC will be committing to a low and uniform common external tariff in a customs union to provide policy credibility and locking in tariff liberalization and harmonization on a multilateral basis. Below, we discuss some impediments to accomplishing this goal.

A. Dependence on Trade Taxes

The dependence on trade taxes constitutes a major hurdle for tariff liberalization in the region. Some details on the proportion of trade taxes in total revenue and GDP are presented in Table 5. In all of the countries in the region, except South Africa, trade taxes account for over 10 percent of total fiscal revenue. In 8 out of 24 countries, trade taxes are over 20 percent of revenue. In only 4 of the 24 countries, namely Rwanda, South Africa, Tanzania, and Uganda, are trade taxes less than 2 percent of GDP. Lesotho, Namibia, and Swaziland are the most dependent on trade taxes. As shown for SADC in Tsikata (1999), trade liberalization is likely to create a significant fiscal gap through the lowering of import duties in some countries. Moreover, as noted in Chauvin and Gaulier (2002), regional integration may lead to changes in the structure of individual economies that could lead to a contraction of previously import-substituting industries that were important sources of revenue. In the same study, the authors estimate that a SADC FTA may lead to as much as a 6 percent decline in total revenue.

The challenge in maintaining fiscal sustainability in the face of trade liberalization should not be minimized. In particular, as a recent study by Baunsgard and Keen (2004) demonstrates, most low-income countries have not been able to replace lost trade tax revenue from other revenue sources. While the authors do not rule out indirect effects of improved economic efficiency on tax revenue, their results do suggest that concerns over tax revenue are serious. On the other hand, as noted by Mussa (1997), the fiscal situation should not be used as an excuse to inordinately delay trade reform. Distorted trade regimes operate as impediments to improving efficiencies of resource allocation. As reforms are implemented, care should be taken to broaden the effective tax base and seek alternate sources of revenue and if these are limited, better expenditure control should be emphasized. Similar recommendations are made by Ebrill, Stotsky, and Gropp (1999).

According to Tsikata (1999), revisiting the eligibility criteria for tax exemptions in SADC would be an important step in strengthening the revenue effort, particularly since exemptions can account for a large portion of lost revenue in African countries. For instance, exemptions covered over 42 percent of imports in Tanzania in 1999 (WTO, 2000). To the extent that exemptions are for the purpose of offsetting the high tariffs, it should be possible to reduce the tariff rates along with a streamlining of exemptions. These efforts could be taken in conjunction with measures to improve tax administration or broaden the tax base, with the goal of increasing reliance on domestic sources of taxation.

B. Disparities in Restrictiveness of Trade Regime

There are great disparities in the restrictiveness of trade regimes across the region, as can be seen below (Table 6). Several countries such as Djibouti, Malawi, Rwanda, Uganda, and Zambia rank among the most open in the world, while others such as Burundi, Comoros, and Seychelles rank among the most restrictive trade regimes in the world. Only about 9 of the countries have open trade regimes (rated below a 3) and these do not include the relatively developed countries in the region (Egypt, Kenya, and South Africa). Clearly, the more developed countries have not played a leadership role in trade liberalization in the region. In these circumstances, harmonization in the sense of a low and uniform common external tariff will involve significant adjustment from a majority of the countries, including the more developed ones, and agreement will be complicated by the very different starting positions.

At this time, it is worthwhile to re-emphasize the desirability of a low, common external tariff by looking at the potential for trade diversion in the context of a SADC or COMESA free trade area. The more restrictive the trade regime in a country (as in Burundi, Comoros, or Seychelles), the greater the preferential margin granted to partner countries in the FTA. Therefore, the potential for welfare losses due to trade diversion is the greatest for these countries. MFN liberalization can help in reducing these losses. In addition, the preponderance of evidence supports the hypothesis that trade openness contributes greatly to growth (Berg and Krueger, 2003).

In this regard, the RTA should be viewed as an opportunity to lower tariffs on an MFN basis to gain from the policy credibility of the regional commitment. Careful sequencing and appropriate monetary and exchange rate policies can help in addressing any sharp balance of payments swings.

C. Impact of Common External Tariff on Tariff Structures

In this section, we examine additional difficulties in agreeing on a low and uniform common external tariff. We analyze three- and four-band tariff structures, similar to proposed CETs in the EAC and COMESA. Since SADC does not yet have a proposed CET, we assume that they will eventually move to a similar simplified structure. In accordance with the proposals put forward by the two arrangements, tariff lines are separated into raw materials, capital goods, intermediates, and finished goods. An escalating tariff structure is used, with the highest rates reserved for finished goods, and the lowest for capital goods and raw materials. It should be noted that this tariff structure can be a means of providing protection to manufacturing and builds in an anti-export bias. Not only are higher tariffs reserved for finished goods but low tariffs on inputs provide an additional layer of escalation. In thiscontext, agreeing on a low and uniform common external tariff can minimize the degree of protection and help in countering the anti-export bias.

We first examine the impact of the different CETs on the average degree of protection. We have classified all tariff lines into the four categories of goods based on the six-digit level classification. While our classification is not likely to be exactly the same as that which will be finally agreed by the RTAs, it should be indicative. Table 7 displays the impact of 16 possible CET combinations on the simple average tariff—using our own classification of goods into the three or four tariff bands. The maximum tariff is allowed to vary from 20 to 40 percent, the tariff on intermediate goods is allowed to vary between 10 and 15 percent, and the tariff on capital goods is allowed to vary between 0 and 5 percent. The tariff on raw materials is always maintained at 0 percent. Scenarios 1–8 model a three-band CET, and scenarios 9–16 model a four-band CET.

Table 7.

Average Tariffs Under Three- and Four-Band Common External Tariff

article image
1/ Source: IMF staff calculations.

From Table 7, it is clear that raising the maximum tariff, or going from a three- to four-band CET has a tremendous impact on the average degree of protection. Increasing the tariff on capital goods from 0 to 5 percent increases the simple average tariff by 1 percent. Increasing the tariff on intermediate goods by 5 percent increases the simple average tariff by approximately 1.7 percent. Increasing the tariff on finished goods by 5 percent increases the simple average tariff by approximately 1.8 percent.

Comparing the results in Table 7 to the existing levels of the simple average tariff in Table 5, it is apparent that a maximum tariff of 25 percent and a three-band CET will yield, at the minimum, a simple average tariff of 12 percent (Scenario 2). This results in little liberalization or a reversal for several countries in the region (including Djibouti, Eritrea, Mozambique, Rwanda, Uganda, and Zambia, and members of SACU) and could potentially imply welfare losses to these countries from the arrangement. On the other hand, the maximum tariff of 25 percent reduces significantly the degree of protection available to industries in several other countries. Countries like Angola, Burundi, Comoros, Ethiopia, Kenya, Mauritius, Madagascar, Seychelles, Sudan, and Zimbabwe have significantly higher protection at present and may face temporary adjustment costs from liberalization. As emphasized earlier, implementing a CET in this situation will require careful sequencing of reforms and complementary policies. In addition, for all member countries to benefit from liberalization under the RTA, it is important that no country suffer a reversal in liberalization. A three-band tariff structure with a maximum tariff of 20 percent would be the preferred option in this regard.19

In Table 8, we analyze the impact of the five most liberal CETs studied above on the tariff structures in the region. Information is provided on the number of tariff lines that would remain unchanged, or experience an increase or decrease in tariffs. Acronyms describing the countries are BDI (Burundi); COM (Comoros); MOZ (Mozambique); RWA (Rwanda); TZA (Tanzania); UGA (Uganda); ZMB (Zambia); ZWE (Zimbabwe); MUS (Mauritius); MWI (Malawi); MDG (Madagascar); and KEN (Kenya).20 As can be seen, the implementation of the CET based on the classification of raw materials, intermediates, capital and finished goods will imply substantial changes in the tariff structure in all countries. Tariffs on a number of lines will go down, and on a number of lines will go up.

Table 8.

Impact of Common External Tariff on Selected Countries 1/

article image
Source: IMF staff calculations.

Country acronyms are as follows: BDI (Burundi); COM (Comoros); MOZ (Mozambique); RWA (Rwanda); TZA (Tanzania); UGA (Uganda); ZMB (Zambia); ZWE (Zimbabwe); MUS (Mauritius); MWI (Malawi); MDG (Madagascar); and KEN (Kenya).

Clearly, this has the potential to hurt some sectors and benefit others in each country. In fact, both COMESA and EAC are facing tremendous difficulty in agreeing on the classification of goods since an input for one country may be a finished good for another.

While there may be a great deal of temptation to protect domestic industries by negotiating for a suitable classification of goods, countries in the region should bear in mind that this will only serve to highlight their lack of commitment to true liberalization and will cause delays in agreement and implementation problems down the road. To the extent that their goal is to liberalize and gain policy credibility, protracted negotiations in this area are likely to be counterproductive.

Another issue with agreeing on a CET is that there may be a conflict with tariff bindings at the WTO. When countries bind their tariffs at the WTO, they put a ceiling on their applied tariff rates. Therefore, tariff bindings may get violated if the bound rates are lower than the corresponding rates under the CET. In Table 9, we present some estimates on the number of bound tariff lines that may be in conflict with the CET for COMESA and SADC member countries, depending on the agreed CET.

Table 9.

Number of Tariff Bindings to Be Renegotiated at World Trade Organization (WTO)

article image
Source: IMF staff calculations.

From Table 9, it appears that over half of the countries except Mozambique, Tanzania, Uganda, and Zambia will exceed their tariff bindings in order to implement the suggested CET. The SACU countries will have the largest number of tariff lines in excess of bindings, reflecting the fact that they have bound a greater number of lines in the WTO. For other countries—Burundi, Rwanda, Zimbabwe, Mauritius, Malawi, Madagascar, and Kenya—the number of lines is not very large. It should be pointed out that the enforcement of tariff bindings in the WTO depends on whether a country’s trading partners raise a complaint at the WTO.

V. European Union—Economic Partnership Agreement (EPA) Negotiations

In this section, we examine and discuss developments and issues relating to the ongoing negotiations with the EU on EPAs. The EU is an important trade partner in respect of both imports and exports for countries in the region (Table 10), and the outcome of these negotiations will have a significant impact. At this point in time, not much progress has been made in the negotiations as individual countries are in the process of formulating their trade negotiating positions.

Table 10.

European Union (EU)—Share in Trade

article image
Source: IMF staff calculations.

Historically, a group of 77 African-Caribbean-Pacific (ACP) countries, which includes all member countries of COMESA and SADC except Egypt and South Africa, have been recipients of unilateral preferences into the EU market under the Lomé conventions. These preferences have provided them with important market access for agricultural and other exports. However, these unilateral preferences are incompatible with WTO rules—since the “Enabling Clause” in the GATT does not allow unilateral preferences that discriminate between groups of developing countries, except in favor of LDCs. Since preferences granted to the ACP countries are neither available to all developing countries, nor restricted to just LDCs, the Cotonou agreement, concluded in 2000, requires all ACP countries to negotiate WTO-compatible EPAs with the EU to replace unilateral preferential arrangements by end-2007. The EPAs will involve reciprocal market access into the ACP countries for the EU with a possible transition period of 10–12 years for the phasing out of trade barriers between the parties (in accordance with GATT Article XXIV).

The EPAs have a development focus—they are to assist ACP countries in enlarging their markets by improving the predictability and transparency of the regulatory framework for trade and creating conditions for increased investment. In this context, the EU has placed strong emphasis on South-South integration through reinforcing existing regional integration initiatives, harmonization of rules and creation of customs unions. Negotiations are being conducted with regional economic groupings instead of individual countries.

In eastern and southern Africa, the two regional economic groupings that are negotiating with the EU are de facto COMESA and SADC. At present, the so-called ESA (Eastern and Southern Africa) negotiating group comprises 16 of COMESA’s 19 member countries.21 The SADC negotiating group, on the other hand, includes 7 of the 13 member countries (Angola, Mozambique, Tanzania, and the SACU countries) of SADC with South Africa participating as an observer.

A. Features of EPA

The EPA is expected to take the form of an FTA.22 It is expected that the EU will provide market access similar to that under the “Everything But Arms” (EBA) initiative.23 The provision of such market access under EPAs, it should be noted, will not be an improvement for LDC members of the negotiating group and by itself, will provide them with few incentives to participate. The rules of origin under EBA are more restrictive than those under the Cotonou Agreement. Therefore, LDCs are likely to seek an improvement through simpler and less restrictive rules of origin and full cumulation across all ACP countries, the EU, and other EBA countries. Non-LDC members of ESA and the SADC negotiating groups, on the other hand, have greater incentive to participate in order to preserve market access to the EU.

An important issue to be negotiated is the level of market access offered to the EU in reciprocity. Since the EU has committed to an asymmetric approach in terms of product coverage and transition periods, it is likely that African countries will not be required to liberalize all sectors and will be allowed a reasonable transition time. In order to be compatible with WTO rules (GATT, Article XXIV), the agreement will require the reciprocal liberalization of substantially all trade by African countries—the EU has expressed its ambition that over 90 percent of trade will be covered over the long term. However, the interpretation of this provision leaves considerable room for uncertainty and countries in the region are likely to push for exclusion of a greater number of products from the agreement. It has been pointed out that even if the agreement were to liberalize 90 percent of trade, African countries could potentially end up without liberalizing most of their important domestic sectors because with current high tariffs, imports in the important sectors might be small (Hinkle and Schiff, 2004). This would be disappointing and another missed opportunity to reap efficiency gains from trade. A suggestion that has been made is that the agreement should aim for liberalization of 90 percent of tariff lines and that there should be limits on the number of tariff lines within a chapter or heading that can be excluded from liberalization. This will ensure that entire sectors do not escape liberalization. Moreover, the agreements should phase in additional tariff reductions over a period of time to cover over 95 percent of tariff lines.

Nontariff barriers such as sanitary and phytosanitary standards are also to be addressed in the agreement. Here, the development aspect of the EPA is expected to come into play through technical assistance and capacity building measures. For instance, the EU is expected to support regional and national capacity building to develop regional testing and certification institutions in Africa (COMESA Secretariat, 2003a). With regard to other features of the agreement: safeguard measures for both industrial and agricultural products are expected to be included. The ACP countries are to provide the EU with a list of products for the priority removal of export subsidies.24

A significant part of the EPA is that ACP countries and the EU have agreed to liberalize trade in services on the basis of a positive list approach. To address concerns over weak regulatory and supervisory systems in African countries, the EU is to provide support in the development of the services sector and in appropriate sequencing of liberalization commitments. There is also provision for a special safeguard in the area of services. The coverage of services in the agreement along with appropriate sequencing and regulation could be very beneficial for African countries provided they take advantage of the opportunity to liberalize and make substantial commitments. Services provision in areas such as accounting, telecommunications, transport, finance, etc., is fairly inefficient in the region and drives up the transaction costs of doing business.

The negotiations will also cover the issue of improved market access for mode 4 supply of services (movement of natural persons). This is important since worker remittances from temporary movement of workers can be a significant source of income for developing countries and can stimulate growth through dissemination of technological and business skills. It is unclear how generous the improvement in market access for movement of natural person will be because of Europe’s concerns over absorbing labor flows from the accession countries.

B. Overlapping Membership Issues

As discussed above in Section II, overlapping memberships between RTAs in Africa are a significant obstacle to achieving the goals of regional integration. In this regard, the EU’s plan to push the formation of customs unions is a particularly beneficial aspect of the EPAs in that they will require a long-overdue rationalization in the membership of COMESA and SADC, since it is technically impossible for a country to be a member of more than one customs union. The problems of overlapping membership are:

  1. Seven countries (Angola, the Democratic Republic of Congo, Malawi, Mauritius, Swaziland, Zambia, and Zimbabwe) are members of both COMESA and SADC.

  2. Kenya, Tanzania, and Uganda are members of the EAC customs union. While Kenya and Uganda are members of COMESA, Tanzania is a member of SADC. Even if COMESA and the EAC harmonize (there is some potential here), there is still the issue that Tanzania is a member of SADC and not COMESA.

  3. Egypt is not an ACP country but is a member of COMESA.

  4. SACU is already a customs union and includes South Africa, which is not an ACP country.

One option, albeit unlikely, may be that COMESA and SADC decide to harmonize their tariff structures and form one large customs union, which would also be in line with the goals of the African Union. Alternatively, countries in the region would have to choose the RTA they wish to be a member of. Based on the membership in EPA negotiating groups, it is possible that Angola, and Swaziland will drop out of COMESA while the Democratic Republic of Congo, Malawi, Mauritius, Zambia, and Zimbabwe will drop out of SADC over the medium term. Given that the EAC customs union protocol has already been adopted, the member states—Kenya, Uganda and Tanzania—can decide to harmonize with either COMESA or SADC. Tanzania might elect to leave SADC or Kenya and Uganda might leave COMESA. This might be ultimately determined by which group is viewed as having negotiated the more favorable EPA or regional agreement.

It remains to be seen how the issues of Egypt and South Africa will be addressed. Given that Egypt’s exports to the rest of COMESA are approximately 1 percent of its total exports, and that it does not provide great scope for exports from other countries in the region, its membership of COMESA is likely to come under greater scrutiny.25 With regard to South Africa, it will be extremely difficult administratively to separate goods originating in South Africa versus the rest of SACU. Given that South Africa has an FTA with the EU, it may be much simpler and possible for them to participate in the agreement.

C. Revenue Losses Owing to EPA

One of the major concerns is over possible losses in fiscal revenue given that the EU is the biggest trade partner for the region. In Table 11, we present estimates of the revenue loss from the removal of all tariffs on all EU imports. These estimates have been computed using the UN COMTRADE data on imports classified at the six-digit tariff line level. The revenue loss, as a percent of GDP, is computed by assuming that the statutory rates are actually applied to each of the tariff lines so that this is an upper bound on the possible revenue loss. These estimates do not take into account the impact of exemptions, which are reported to be widespread in sub-Saharan Africa. In order to estimate the possible impact of exemptions, we assume that they are applied equally to imports from the EU and the rest of world. We accordingly present an adjusted estimate of loss in trade revenue as a percentage of total revenue.

Table 11.

Trade Revenue Loss Owing to EPA

article image
Source: IMF staff calculations.

Adjusted for tariff exemptions.

As can be seen from Table 11, Burundi, Comoros, Kenya, Mauritius, and Rwanda are likely to face the most significant losses in fiscal revenue (over 6 percent of total revenue) from the EPA. While these results are not unexpected for the other countries, it is surprising that Rwanda is likely to face a significant loss in revenue from the EPA. Rwanda does not appear very vulnerable if one examines the proportion of its imports from the EU and tfrom its dependence on trade taxes (Tables 5 and 10). However, our analysis indicates that Rwanda’s imports from the EU are concentrated in the tariff lines with the highest tariffs so that removal of tariff barriers vis-à-vis the EU can lead to substantial revenue losses. One shortcoming of our analysis is that since we lack information on exemptions, we assume that they apply equally across all tariff lines.

Within SACU, we have been unable to separate the customs revenue because of lack of data, but the loss of revenue will be an important concern for Lesotho, Namibia, and Swaziland. In other countries also that we have been unable to examine because of lack of data, there may be some revenue loss and alternative revenue sources will need to be identified.

D. Other Challenges and Opportunities in EPA Implementation

Here we discuss additional opportunities, challenges, and concerns that are associated with implementation of the EPAs. In terms of opportunities presented by EPAs—these agreements can act as a mechanism to lock in trade reforms and gain credibility for liberalization policies, and trade-related governance and institutions (Lawrence, 1997; and Lewis and others, 2003). The negotiations envisage technical assistance in the area of standards and liberalization of services, both of which are critical areas in Africa.

A major concern on the other hand, in EPAs is that they might impede or slow down the process of liberalization in many African countries and create incentives against multilateral reform. There are several reasons to view EPAs with suspicion in this regard. First, given the high tariff barriers and dependence on trade taxes in a number of ESA and SADC countries, and the EU’s stated desire to push for the formation of customs unions, there is a distinct possibility that the agreed common external tariff will be high and increase the average degree of protection in the more liberal countries in the region. Thus, not only would past liberalization efforts be reversed but once the common external tariff is set, it would not be possible for individual countries to pursue unilateral tariff reductions. Future liberalization could thus be difficult and slow—not the least because of difficulties in coordinating across many countries but also because of resistance from domestic entrenched interests. Second, as with any other preferential arrangement, there is the concern that ACP countries and/or the EU will have secured market access and will have incentives to oppose and undermine future multilateral liberalization. The concern is magnified in this instance by the size of the preferential arrangements being envisaged—with 77 ACP countries and the EU participating, these account for around half of the total WTO membership. Interests in the ACP countries that would have preferential market access to the EU might oppose future multilateral liberalization, including in agricultural trade. Moreover, with the EU obtaining preferential market access in a large number of developing countries, they might have little incentive to push for further MFN liberalization in these countries.

On balance, it is not clear if EPAs will be welfare-enhancing or not. In large part, this will be determined by the actual scope of the liberalization commitments undertaken and their implementation. Therefore, in order to maximize benefits and minimize concerns, it is critical that the EPAs result in the adoption of a low and uniform common external tariff. The agreements should also aim for comprehensive product coverage in both goods and services and appropriate sequencing, including in trade and tax policies. In both goods and services, multilateral liberalization should be pursued concomitantly with preferential liberalization. There is also a need to ensure that rationalization of membership does indeed take place and these agreements are not a force for further fragmentation in the RTAs.

VI. Conclusion

Regional integration has gained momentum in recent years in eastern and southern Africa, driven in part by global trends and in part by the negotiation of the Economic Partnership Agreements with the EU. This paper examines prospects and challenges facing two prominent regional arrangements in eastern and southern Africa, COMESA and SADC, with regard to their role in promoting trade.

Analyzing the features of the agreements in detail, this paper finds that the COMESA FTA has taken a market-liberalization approach to regional integration but has been hampered by country-level implementation issues. On the other hand, SADC has taken the approach of addressing infrastructure and supply constraints but has also had implementation problems. Restrictive rules of origin are also a concern in SADC. The problems faced in either arrangement point to a lack of political will and commitment. Going forward, both COMESA and SADC have plans to form customs unions, although the plans of COMESA are more advanced. It will be important to ensure that the customs unions are implemented in a way that furthers the goals of liberalization.

This paper’s empirical analysis confirms the results of other studies that suggest prospects for expansion of intraregional trade might be limited—it bases its inference on low levels of intraregional trade and product complementarities. It has been previously documented that exports of eastern and southern African countries are concentrated in a few primary commodities. In this regard, our estimates confirm existing results indicating that the relatively developed economies of South Africa, Egypt, and Kenya might not be able to function effectively as markets for the products of other economies in SADC and COMESA. Furthermore, there is a risk that investment might be attracted to the relatively more developed economies resulting in polarization.

In this environment, benefits of COMESA/SADC are likely to flow from other sources. They can be a useful avenue for harmonization and capacity building in various trade-related areas (standards, customs procedures, other NTBs, etc.) and help to alleviate infrastructure constraints through regional cooperation. In addition, they can provide policy credibility to trade reforms through commitment to MFN liberalization, potentially demonstrated by a low and uniform common external tariff.

There are several challenges in implementing a low and uniform common external tariff. The most prominent of these relates to the dependence of a large number of countries in the region on trade taxes as a source of revenue. In almost all of these countries, trade taxes are over 10 percent of total revenue. Moreover, past experience indicates that it has been difficult for low-income countries to replace lost tariff revenue with revenue from other sources. Clearly, serious efforts will need to be made to broaden the tax base and develop alternate sources of revenue. Another challenge relates to the wide disparities among the trade regimes of member countries in COMESA and SADC, which will make it difficult for them to harmonize and agree on a CET that is liberalizing and beneficial for all countries.

An interesting feature that comes out in our discussion is that the initial tariff structures of the countries in the region are very different. Agreeing on a common external tariff based on categorizing goods into raw materials, capital goods, intermediate goods and finished goods (as in COMESA/EAC) will require substantial changes in the tariffs on individual lines. This has also been an area of discord in the COMESA and EAC negotiations. We point out that protracted negotiations have a cost in terms of delaying the agreement and causing implementation problems down the road. In addition, they can diminish the potential gains from policy credibility, since some countries may be perceived as having protectionist intent. An additional result of our analysis is that the CET might result in an apparent conflict with WTO bindings for some of the countries.

We also examine several issues regarding the EPA negotiations with ACP countries in Africa. One of the upcoming challenges, once again, relates to the potential loss of fiscal revenue since the EU is an important trading partner for the region. As discussed previously, significant adjustment efforts will need to be made in order to compensate for the loss of revenue. Another significant concern is the impact of the EPAs on the multilateral liberalization efforts of individual African countries and on the multilateral trading system. Given the emphasis on reinforcing RTAs and the push for the formation of customs unions, it will be important to take the opportunity and implement a low and uniform CET—not doing so will slow down future liberalization and has the potential to fragment the multilateral trading system. The tendencies opposing multilateral liberalization are likely to be aggravated with the creation of vested interests that would like to preserve their special market access privileges granted under the EPAs.

At this time, it is not clear whether the EPAs will prove beneficial for African countries over the long term. The EPAs’ emphasis on reinforcing RTAs can be beneficial because countries will be required to rationalize their overlapping memberships and because it counteracts the formation of “hub-and-spoke” agreements. If EPAs succeed in strengthening the commitment to nondiscriminatory trade liberalization and harmonization in various areas, African countries can gain from more policy credibility and better-integrated regional markets. Technical assistance provided in standards etc. will also be very welcome.

In conclusion, it should be emphasized that in the past regional integration has not yielded results for Africa in terms of economic performance. Going forward, it is important to heed the lessons and pursue nondiscriminatory liberalization concomitantly with regional integration. Deep integration and open regionalism are considered desirable attributes for RTAs; there is a need for political will and commitment to ensure that COMESA and SADC can effectively pursue these objectives in the greater interest of promoting trade. They must meet several challenges to achieve these goals, but the cost to them of not pursuing reforms may ultimately be greater.