Regional integration schemes have featured highly in the policy agenda of developing country governments during the past three decades, although the strength of the commitment to such arrangements has fluctuated considerably. Developing countries have typically considered regional integration efforts to be, at a minimum, conducive to economic development and political stability in the region as well as to the bargaining position of developing countries in multilateral forums. At times and in some cases, regional integration has been considered essential for self-sustaining growth and development. Views on the role that regional integration might play in development have evolved in line with changing historical circumstances and economic and political philosophies.

Regional integration schemes have featured highly in the policy agenda of developing country governments during the past three decades, although the strength of the commitment to such arrangements has fluctuated considerably. Developing countries have typically considered regional integration efforts to be, at a minimum, conducive to economic development and political stability in the region as well as to the bargaining position of developing countries in multilateral forums. At times and in some cases, regional integration has been considered essential for self-sustaining growth and development. Views on the role that regional integration might play in development have evolved in line with changing historical circumstances and economic and political philosophies.

The great variety of regional integration arrangements among developing countries defies easy generalization. Box 2 provides general information and Table 7 more detail on the specific objectives of existing regional groupings, particularly those emphasizing preferential (reciprocal) trading arrangements, in the form of free trade areas or customs unions. (See also Appendix for the membership of these groups.)

Table 7.

Selected Regional Trade Arrangements in Developing Countries: Main Objectives

article image

A movement toward regional integration schemes in Latin America began in the 1950s, leading to the establishment of the Latin American Free Trade Association (LAFTA) and the Central American Common Market (CACM), both in 1960. Although stimulated in part by regional arrangements formed in Europe during the 1950s, the Latin American cases were based on an entirely different rationale. Reflecting the predominant economic philosophy of the times—promoted largely by the Economic Commission for Latin America (ECLA)—intraregional trade liberalization was seen mainly as a means to overcome the limits imposed by small domestic markets on an import-substitution development strategy. The expectation was that larger regional markets would engender economies of scale and provide firms with a “training ground” to gain the competitiveness needed for later success in world markets (Morawetz, 1974). This view permeated the design and choice of integration instruments in Latin American arrangements, most notably the CACM and the Andean Pact but also the Caribbean Community (CARICOM), where explicit means were devised to allocate import-substituting industries among members while attempting to ensure “balanced” industrial development. Interest in regional integration declined somewhat from the mid-1970s to the mid-1980s, against a background of disappointing results in the major Latin American group (the LAFTA) and the emergence of the debt crises in the 1980s. Under the current revival of regionalism on the continent (see Chapter III), the previous emphasis on import substitution appears to be giving way to a more outward-oriented and market-based approach, both in new regional arrangements and in old arrangements being revived.

The historical roots of regional integration schemes in sub-Saharan Africa may be traced to decolonization. Regional integration has been seen as a key means for cooperation in the midst of the often centrifugal forces stemming from the complex process of nation building. Against the background of fragile economic units, all regional integration arrangements in sub-Saharan Africa have emphasized collective economic self-reliance and security, with intraregional trade liberalization considered a key mechanism for achieving shared aspirations for growth and development. Sub-Saharan Africa has the largest number of regional groupings in the world, with considerable overlapping membership. It also has the largest number of ineffective or dormant arrangements. The Lagos Plan of Action, launched by the Organization of African Unity in 1980, provided a unifying framework for these arrangements. This plan envisages separate but convergent efforts toward integration in the three sub-Saharan African subregions: West Africa, East and Southern Africa, and Central Africa. The main arrangement in the first subregion is the Economic Community of West African States (ECOWAS)—established in 1975—which encompasses and broadens the memberships of the Economic Community of West Africa (CEAO) and the Manu River Union (MRU). The Preferential Trade Area (PTA), in operation since 1984, was established for Eastern and Southern Africa. With regard to Central Africa, the Economic Community of Central African States (CEEAC) has been under negotiation for some time; it would encompass and broaden the memberships of existing arrangements in the subregion, namely, the Customs and Economic Union of Central Africa (UDEAC) and the Economic Community of the Great Lakes Countries (CEPGL). Approaches to regional economic cooperation not emphasizing preferential trading agreements have also emerged, illustrated in particular by the Southern African Development Coordination Conference (SADCC), established in 1980. The SADCC is excluded from this analysis, since it did not aim to increase intraregional trade; however, it made significant progress in other areas (see Box 3).

Selected Regional Trade Arrangements in Developing Countries1

article image
article image

See Appendix for membership.

Data from World Bank, World Development Report, 1991; and IMF, Direction of Trade Statistics.

Broadly speaking, the enthusiasm for regional integration arrangements in Asia and the Middle East has not been to date as pronounced as in sub-Saharan Africa and Latin America. Sharp ideological differences, political instability, frequent military hostilities, and substantial economic heterogeneity (in terms of market size, resource endowments, and per capita income level) among countries explain why regional integration schemes involving intraregional trade liberalization have not been numerous in these areas (Langhammer and Hiemenz, 1990). The same reasons, however, seem to have favored a variety of political and economic cooperation agreements. For example, the success of the Association for South East Asian Nations (ASEAN), established in 1964, stems not from intraregional trade but mainly from it having become an effective interlocutor for cooperation in economic matters and foreign affairs with OECD countries. Ambitious plans for integration contained in the design of the Arab Common Market (ACM) and the Maghreb Customs Union, both established in the 1960s, have remained largely unimplemented. By contrast, cooperation efforts have achieved considerable success in a number of cases, including, among other things, an earlier agreement on Regional Cooperation for Development among the Islamic Republic of Iran, Pakistan, and Turkey—replaced in 1985 by the Economic Cooperation Organization (ECO)—and the Gulf Cooperation Council (GCC), established in 1981.

Intraregional Trade

The following analysis is based on regional arrangements in which intraregional trade liberalization had been expected to play a central role. (The assessment does not focus on noneconomic objectives of regional integration arrangements. Nor does it apply to agreements geared exclusively toward economic cooperation or to the economic cooperation aspects of integration arrangements.) In general, regional arrangements among developing countries have not led to a large or sustained rise in intraregional trade (Tables 8 and 9, and Chart 2), but they have been more successful in other areas of economic cooperation. Further scope may there-fore exist for enhancing such cooperation in the joint production of “software” public goods, such as education and training; research and development, particularly in agriculture; joint management and control facilities; and coordination of investment incentives.

Table 8.

Selected Developing Country Regional Arrangements: Intraregional Exports

(In percent of region’s total exports)

article image
Sources: IMF, Direction of Trade; International Financial Statistics; World Economic Outlook; and IMF staff estimates.
Table 9.

Selected Developing Country Regional Arrangements: Changes in Trade to GDP Ratios

article image
Sources: IMF, Direction of Trade; International Financial Statistics; World Economic Outlook; and IMF staff estimates.

Southern African Development Coordination Conference (SADCC)

This paper focuses on regional arrangements aimed mainly at trade liberalization. Trade liberalization is not a major goal of the Southern African Development Coordination Conference (SADCC), and its strong performance has, therefore, not been discussed. Its effectiveness as a regional organization may, however, have lessons for others, while its focus on establishing cooperation in other areas has been important in facilitating trade (Chart 2).

The SADCC, founded in 1980, comprises ten member states.1 Its main objectives are to reduce economic dependence, particularly on South Africa, foster cooperation among members on projects to promote a balanced regional development, and coordinate and secure support from “cooperating partners” (that is, foreign donors).2 The SADCC’s pragmatic approach emphasizes flexible and feasible cooperation on specific projects rather than detailed and highly formalized multicountry agreements. The SADCC does envisage expanding trade among members, but mainly along the narrow lines of production covered by SADCC-sponsored industrial programs. Although more recently, the SADCC has shown interest in liberalizing internal trade, it continues to adhere to the motto: “let production push trade rather than trade pull production.”

Consistent with its pragmatic approach, the SADCC has a small secretariat (in Gaborone, Botswana), which performs limited coordination, conducts research on regional issues, and takes the lead in certain regional policy discussions. The main responsibility for coordinating sectoral programs is assigned to individual members (for example, food security to Zimbabwe, and energy conservation and development to Angola). SADCC policy is set annually at a summit meeting of heads of state, where priorities are established and decisions on which projects to include as official SADCC projects are made. Decisions on individual sectors are made by the relevant ministries. Eligible projects must have substantial regional content. The annual Consultative Conference, which includes open meetings with “cooperating partners” (that is, foreign donors), is equally important because it provides funding to SADCC projects.

The SADCC as a whole is relatively small in economic size. Its GDP of about $25 billion compares to that of Chile or Morocco, while its population of 79 million is comparable to that of Mexico, yielding a low per capita income (about $315). A number of SADCC countries, however, are relatively rich in such mineral resources as copper, cobalt, chrome, diamonds, gold, silver, and uranium. Nearly all members are heavily dependent on South Africa, which supplies about one third of SADCC imports. Intra-Conference trade has remained below 5 percent of total SADCC trade since the 1970s (Chart 2), and nearly half involves semi-manufactured and manufactured goods, compared to 10 percent for total SADCC trade, which suggests some potential for regional trade creation. However, about 80 percent of intra-Conference trade is accounted for by Zimbabwe, which is the most industrialized member of the SADCC, deriving about 30 percent of its GDP from manufacturing. Six of the ten member countries are landlocked. Transport and communication networks—particularly rail links to Mozambique ports—have been often and severely disrupted by guerilla attacks, causing certain landlocked members to shift to substantially longer and costlier routes to South African ports.

The SADCC has been rather successful in achieving some of its objectives, although its economic dependence on South Africa has not been appreciably reduced. The SADCC’s focus on specific projects has helped mobilize foreign donor support and the delegation of responsibilities has enhanced efficiency in project implementation. Attention has appropriately centered on projects with substantial positive externalities for the region, with transport and communications at the top of the SADCC priority list during most of the 1980s. Despite military activity around transport corridors, by 1990 60 percent of transit traffic from the six landlocked states moved through SADCC ports, compared with only 20 percent in 1980, and national power grids in six member states had been interconnected. Progress was also made in the 1980s in other priority areas, particularly food security. By the late 1980s, the SADCC could point to visible achievements: firms were increasingly shipping and receiving cargo through ports and railways improved under SADCC auspices (which was particularly beneficial for landlocked countries); air links had also improved, including cooperation between national carriers; the intra-Conference share in the SADCC total international telecommunication traffic had risen to about 15 percent (compared to 8 percent in 1980), and new crop strains from the SADCC research center were being used. In terms of value, transport and communications accounted for over 75 percent of SADCC-sponsored projects followed by food and agriculture (12 percent), and energy (8 percent). Mining, industry and trade, manpower, and tourism accounted for the remainder (Hanlon, 1989).

This has entailed a shift in favor of what the SADCC calls the “enterprise sector” (private firms and parastatals), including through greater efforts at cooperating with the private sector and improving the environment for foreign and domestic investment in regional industrial enterprises. The SADCC has also stepped up efforts to improve the investment climate and to address the problem of human resource development, especially the shortage of well trained manpower and low productivity.

1 Angola, Botswana, Lesotho, Malawi, Mozambique, Namibia (joined in 1990), Swaziland, Tanzania, Zambia, and Zimbabwe.2 All SADCC members are eligible for membership in the Preferential Trade Area for Eastern and Southern Africa (PTA) and eight SADCC members (Angola, Lesotho, Malawi, Mozambique, Swaziland, Tanzania, Zambia, and Zimbabwe) have joined it. Four SADCC members (Botswana, Lesotho, Namibia, and Swaziland) belong with South Africa to the Southern African Customs Union (SACU).

The performance of regional integration arrangements may be assessed by comparing actual results to initial objectives. Two measures of performance are useful in this regard. The first examines whether the arrangements have succeeded in achieving their fundamental goal of raising intraregional trade. The second looks at the extent to which the arrangements have succeeded in implementing initially agreed liberalization timetables. Given that an increase in the share of intraregional trade is not, in itself, an indication of welfare gains for member countries (trade-diversion effects may predominate), an alternative approach would be to assess, ex post, the net gains associated with the arrangements, in terms of static trade-creation versus trade-diversion effects and in terms of dynamic effects on growth and development. The latter assessment is inherently difficult because it presupposes an ability to guess correctly what would have happened in the absence of the regional integration and under alternative policy scenarios.

The main exception to the general rule that the formation of these arrangements did not increase trade among members is the CACM during its first decade of existence; its share of intraregional exports rose sharply between 1960 and 1970 but stagnated in the 1970s and declined in the 1980s (Table 8). The LAFTA also experienced an appreciable increase in intraregional exports during 1960–80. In the case of the ASEAN, while the share of intraregional trade is not insignificant (18.6 percent in 1990), it declined in the 1970s and has remained fairly stable since then. Moreover, the ASEAN’s relatively high share of intraregional trade can be explained by transshipment trade. The remaining arrangements that have operated for over a decade have recorded only small increases, at best, in the share of intraregional trade, which in many cases has been subject to fluctuations and reversals.

The relative importance of intraregional trade remains low for most groupings, although the variance is significant. In 1990, the share of intraregional exports ranged from a high of 18.6 percent in the case of the ASEAN, to regional groups in Africa with shares typically below 5 percent. The share of intraregional trade in developing country regional groupings is not any greater—and, in most cases, considerably less—than the share of South-South trade in the total trade of the South, even after excluding trade of the oil exporting developing countries and of the dynamic Asian economies; South-South trade excluding the major oil exporters and the four most dynamic Asian economies—Hong Kong, Korea, Singapore, and Taiwan Province of China—was about 17 percent in 1988–90. (Including the oil exporting developing countries and the four most dynamic Asian economies, the 1988–90 share of South-South trade in the total trade of the South was about 34 percent.)

Total trade shares may not, however, always be the best measure of the importance of intraregional trade. For example, in the Andean Pact, where large oil exports heavily influence aggregate statistics, non-oil intraregional trade more than tripled to 10 percent during the 1970s—the first decade of the Andean Pact—with most of the increase accounted for by manufactured exports (although a narrow range). Similarly, the bulk of the higher share of intraregional trade in the LAFTA during 1960–80 was due to increased manufactured exports. Despite its weak implementation, it appears that the LAFTA did provide a wider market for the expansion of manufactured exports within the region and contributed to general export diversification, particularly for Argentina, Brazil, and Mexico. In a number of sub-Saharan African groups, the low shares of intraregional trade conceal the fact that such trade is important (ranging between 10–50 percent) for some of the individual countries in regional groups, particularly those that are landlocked. Such considerations do not appreciably alter the basic conclusion, however, that regional arrangements have fallen short of their goal of fostering substantial intraregional trade.

Implementation Record

The failure of regional integration schemes to increase trade among members was due partly to structural elements limiting the scope for potential gains, and partly to the low degree of implementation. Implementation delays generally reflected a basic incompatibility between the inward-oriented development strategies of most members and regional liberalization. While there are differences among developing country arrangements, significant difficulties have generally been encountered in implementing trade and other integration measures on schedule. This has been particularly evident in four areas: the removal of barriers to intraregional trade, an important target in all arrangements; the establishment of a common external tariff; the implementation of nonmarket mechanisms to allocate new industries among participating countries; and the removal of constraints on intra-area factor mobility (particularly on labor mobility). (The last three were operational targets in only a subset of arrangements.)

In most cases, initial deadlines for the removal of barriers to intraregional trade were postponed, often several times. Normally a reduction in quantitative restrictions and import tariffs was contemplated. In a number of cases (for example, the LAFTA/LAIA and PTA), delays in liberalizing intraregional trade reflected a lack of automaticity in implementation timetables, with reciprocal tariff reductions made subject to periodic negotiating rounds and consensus, either on a product-byproduct basis or on a request-and-offer basis. In addition, in many cases (for example, the ASEAN, Andean Pact, CEAO, LAFTA/LAIA, and PTA), reductions in trade barriers were not based on across-the-board reductions in tariff and nontariff barriers but on a system of positive “lists,” that gave participating countries considerable latitude to exclude sensitive products from the set of items subject to reciprocal trade concessions, and biased the selection of products in favor of those with limited potential for intraregional trade. In a number of cases (for example, the ASEAN, ECOWAS, and PTA) that aimed to create a free trade area—permanently or as a precursor to a customs union—the coverage of intra-area trade liberalization was also limited by fairly restrictive rules of origin; these narrowed the range of goods eligible for tariff preferences on the basis of value-added criteria, and in some cases (for example, the ECOWAS and PTA) also on the basis of enterprise ownership criteria.

The degree of implementation of intraregional liberalization programs also varied widely. Implementation was fairly successful in the CACM, which broadly adhered to the original liberalization timetable, and maintained a substantially unrestricted intraregional trading environment throughout most of the 1960s and 1970s; this record was subsequently undermined, however, particularly by nontariff barriers, in the face of recurrent balance of payments crises during the 1980s. The GCC also appears to have removed tariff barriers to intra-area trade broadly on schedule, although with special reservations granted to Oman. At the other extreme are some sub-Saharan African groupings (for example, the ECOWAS and UDEAC), that are characterized by an almost complete nonimplementation of intra-area liberalization. In most other cases, implementation has been partial, and initial liberalization has often been followed by impasses, delays, and sometimes reversals—the latter frequently taking the form of erosion of previously granted trade preferences because of exchange and trade restrictions introduced in response to balance of payments difficulties.

Where the establishment of a customs union has been an operational target (and not only a statement of intent), the implementation of a common external tariff has proven elusive. The initial dead-lines were in most cases not met. The extent of partial implementation has varied, however, among cases. For example, the CACM agreed in principle to a common external tariff but the common tariff was rendered largely ineffective—mainly because of exemptions from external tariffs granted by some member countries for “necessary” imports from outside the region—over most of the period since the CACM was formed. The GCC adopted a minimum common external tariff in principle—albeit with some delays—but exceptions have been applied for certain members, significant variations among countries remain, and noncompliance has persisted in some areas—particularly concerning luxury items. In most other cases, nonimplementation of the envisaged common external tariff has been the rule.

The implementation of nonmarket mechanisms to allocate new manufacturing industries among members30 has also fallen substantially short of initial targets. For instance, the Andean Pact had originally intended to include in its Sectoral Programs of Industrial Development (PSDIs) about one third of all manufactured items in the tariff code; in the event, only 4 PSDIs were approved and most were only partially implemented. In the case of the ASEAN, only 2 out of the originally envisaged 5 ASEAN Industrial Projects (AIPs)—and only 1 out of the originally proposed 32 ASEAN industrial complementation projects (AICs)—were implemented. In the CACM, only 2 firms were created under the Regime for Central American Integration Industries, and the scheme became inoperative in the late 1960s. In the MRU, no “union industry” has gone beyond feasibility status.

The ECOWAS, CEAO, and the GCC are among the few regional groups that have included operational commitments to allow freer intra-union factor (labor) mobility. The implementation record in these cases is mixed. The ECOWAS signed a protocol on labor mobility in 1979, but it has not been implemented. The GCC took action to permit skilled professionals in member countries to register and practice in any other member country of their choice. The CEAO adopted and implemented a community convention recognizing the right of residence of migrant workers within the region.

Factors Influencing Low Implementation

Implementation problems do not, however, fully explain the failure of developing country regional integration schemes to raise intraregional trade in line with initial expectations. For one thing, the share of intraregional trade increased in some cases, as has been noted, even as the implementation of the regional liberalization program came to a standstill. For another, the implementation problems themselves need to be explained. The explanation seems to lie in a number of political and structural features—an incompatibility between members” national pursuit of inward-oriented development policies and the goal of intraregional trade liberalization, combined with strong vested interests in import-competing industries and a weakening external environment.

The deterioration in the external environment in the 1970s—and particularly in the early 1980s—exposed the underlying weaknesses of import-substituting development strategies and further reduced the scope for trade liberalization. Fundamental questions about the role of integration were underscored by other considerations, including differences in economic philosophies and strategies among members of some regional arrangements (such as the ASEAN); the role of powerful vested interests in highly protected sectors; constraints imposed on liberalization in many cases by inappropriate macroeconomic policies; and political problems that sometimes led to military conflict. At the operational level, these economic and political elements led to implementation delays because the underlying lack of commitment was not offset by a strict timetable for reforms. The timetables lacked automaticity, included “positive” lists rather than across-the-board liberalization of tariff and non-tariff barriers, and entailed restrictive rules of origin. This made it easier for members to backtrack in the face of balance of payments difficulties or not to implement reforms owing to the lack of supporting policies. In the case of arrangements that aimed to allocate industries among members, their failure to do so reflected administrative problems and conflicts of interest arising from the difficulty of determining industrial location in the absence of market criteria and clear comparative advantages. In addition, while many arrangements were ostensibly modeled on the EC, most lacked the necessary institutional mechanisms to achieve their objectives; this was evident in inconsistencies between national legislation and integration commitments, the absence of strong enforcement mechanisms and ineffective dispute settlement procedures, and lack of compensation mechanisms to address distributional concerns. (The latter was important in schemes that grouped members with widely differing stages of development.)

An inward orientation tended to lead to high levels of effective protection in the form of tariff structures characterized by high and highly dispersed rates and complex systems of nontariff barriers. Firms established behind such high protective walls have often enjoyed strong lobbying power. Most developing country governments signing regional integration pacts during the 1960s, 1970s, and 1980s typically did not intend to alter fundamentally their inward-oriented policies; on the contrary, they often saw regional integration as an extension of import-substituting industrialization, as a means to enlarge the markets for their protected firms. In the absence of a wholehearted commitment to shift toward outward-oriented policies, governments were unwilling to sacrifice the demands of their influential national import-competing sectors to the goal of free trade within regional groupings. The political economy of vested interests in import-competing sectors, as noted above, in turn limited the scope for governments to shift the orientation of their policies.

Under these circumstances, governments engaged in regional integration schemes often preferred to avoid automaticity in the implementation of intraregional liberalization programs and time-tables. The tendency in many regional arrangements was rather to choose an implementation approach based on negotiation and consensus, often on a product-by-product basis, which in most cases led to negotiating fatigue and excessive selectivity of offers. The “lists” of products eligible for trade preferences within the region were relatively easy to construct as long as the products in question were subject to trade within the region prior to the arrangement or were not produced in the region at all. However, heavy resistance and negotiating impasses tended to be encountered when it came to the “hard core” products of highly protected firms. The significant differences in efficiency of import-competing industries across member countries only worsened this problem. Even in the case of the ASEAN, differences in economic philosophies and strategies among members constituted a major stumbling block to intraregional trade liberalization—with Singapore, at one extreme, pursuing a decidedly outward-oriented strategy, and Indonesia, at the other, espousing import-substituting industrialization.

Implementation difficulties in intraregional liberalization did not, therefore, stem from inward orientation in economic policies alone, but rather from the combination of these policies with strong vested interests in import-competing industries. Thus, as Edwards and Savastano (1989) note, the relative success of the regional liberalization process in the CACM was made possible by the initially weak position of import-competing sectors in member countries, even if the CACM itself was seen as—and in fact became—a means to foster import-substitution industrialization within the region. By contrast, powerful vested interests in highly protected import-competing sectors existed before the establishment of the LAFTA; these explain to a significant extent the major impasses in the liberalization program that relied on a product-by-product negotiation of reciprocal trade concessions, rather than on a more automatic process.

Developing country regional groups typically did not, therefore, intend to implement a “classical” model of integration involving an across-the-board removal of barriers to intraregional trade followed by market-based resource reallocation. Instead, many regional pacts tended to engage in limited and highly selective intraregional trade concessions, so as to maintain protection for many existing industries. Moreover, in a number of cases, regional pacts also set up mechanisms to allocate—on a nonmarket basis and often through negotiation—new industries among member countries. The failure to implement these latter mechanisms reflected not only huge administrative problems (as in the case of the Andean Pact, which also faced severe geographical limits on integration) but also, and perhaps more fundamentally, sharp conflicts of interests arising from the inherent indeterminacy of industrial location in the absence of market criteria and clear comparative advantage.

The scope for liberalization within regional groups was in many developing country cases constrained by inappropriate macroeconomic policies. Against a background of external and internal imbalances—often reflecting expansionary fiscal and monetary policies—country authorities unwilling to allow a substantial devaluation of their over-valued currencies also tended to resist meaningful reductions in trade barriers. This resistance reflected an unwillingness to release foreign exchange reserves and fiscal revenues from import tariffs. In part, the adoption of realistic exchange rates was resisted because of concerns regarding the potentially contractionary short-term effects of devaluation, particularly where import-substitution policies had led to a heavy reliance of manufacturing output on artificially cheap imports of intermediate and capital goods. Thus, import substitution tended not only to be inconsistent with far-reaching intraregional liberalization, it also introduced structural rigidities in macroeconomic management. These conditions could be sustained only to the extent that commodity export prices remained buoyant and no major constraints arose in terms of access to foreign capital inflows.

The CACM in the early 1980s provides an example of a successfully implemented regional arrangement with a relatively high level of intraregional trade (in a developing country context) that was undermined by unsupportive national policies. While political and military conflicts in the area contributed to the decline of intraregional trade in the 1980s, the main reasons for this development were economic. With the onset of the debt crisis in the 1980s, external imbalances gave rise to severe foreign exchange shortages; these were initially addressed through the imposition of trade and exchange restrictions, including on intraregional trade. Inadequate macroeconomic policies contributed to higher inflation rates in all member countries, while official exchange rates remained pegged or closely linked to the U.S. dollar. The resulting overvaluation of the CACM currencies was worsened by the concurrent appreciation of the dollar relative to other major currencies. The competitiveness of CACM exports consequently declined, which reduced recorded intra-area trade. At the same time, cross-country differences in the degree of overvaluation gave rise to informal border trade as imports from third countries were channeled through the countries with the most overvalued currencies.

Commodity price shocks in the 1970s and the broader and deeper deterioration of the external environment in the early 1980s (including those owing to terms-of-trade losses, sharp increases in real international interest rates, and a severe curtailment of access to foreign credit), exposed the underlying weakness of macroeconomic policies under import-substituting development strategies and further reduced the scope for trade liberalization within regional arrangements. Although developing countries pursuing export-oriented industrialization were generally better able to adjust to the difficult external environment, those emphasizing import substitution encountered relatively more severe constraints on their ability to implement appropriate adjustment policies. As a result, liberalization in many regional groups tended to be reversed, particularly in the early 1980s, as country authorities, faced with mounting macroeconomic imbalances, reintroduced or increased nontariff barriers and exchange restrictions, including on intraregional trade.

With chronic shortages of foreign exchange, the functioning of some clearing and payments systems attached to regional integration arrangements also tended to falter. This often reflected an accumulation of arrears and the growing unwillingness of net creditor partners to maintain large amounts of non-convertible currencies in their asset portfolios. The existence of monetary unions in some of the African regional groups, although ensuring the convertibility of the common currency (through the willingness of a major industrial country to defend the peg), could not prevent inappropriate fiscal policies from causing inflation. The resulting over-valuation of these unions” common currency, in turn, reduced the scope for intraregional trade.

Political and institutional factors also constrained integration. Implementation of integration schemes requires a degree of shared sovereignty—reflected in binding commitments and possibly in a body of community law and supranational institutions—commensurate with the degree of integration sought. Many of the regional integration arrangements among developing countries, although ostensibly modeled on the EC, were unable to endow their integration agreements with the institutional effectiveness needed to achieve their often ambitious objectives. This has been evidenced by, among other things, inconsistencies between national legislation and integration commitments, the absence of strong enforcement mechanisms, and ineffective dispute settlement procedures. The difficulties encountered in subordinating nationalistic goals to supranational structures were in a number of cases compounded by contradicting political philosophies among members of the same regional group; political instability in national systems (for example, frequent changes of government and military coups); and at times also bilateral (military) border conflicts within regions.

In many cases, important distributional questions were not addressed. Issues of uneven distribution of the benefits and costs of integration have been high on the agendas of regional integration arrangements among developing countries. Such concerns have focused not only on the potentially uneven distribution of short-run costs of economic dislocation (transitional unemployment and idle capacity) and losses of fiscal revenues, but also on economic polarization over the medium term—that is, the possibility that the dynamics of integration would favor development in the more developed member countries. Regional arrangements among developing countries have attempted to address these concerns in several ways, including special treatment for lesser developed member countries (for example, through relatively less demanding liberalization schedules and special access to regional credit facilities); “balanced growth” criteria in the allocation of new industries among members; and explicit compensation funds. However, compensation funds intended under several arrangements were never established—owing to design problems and budgetary constraints in member countries. And other mechanisms tended to be perceived as insufficient by the smaller, less developed members. As a result, distributional concerns added to the resistance against deeper and broader liberalization within regional groups.

The ECOWAS’ failure to enhance intra-Community trade, for example, partly reflected the economic diversity of members. (The ECOWAS includes Côte d’Ivoire, with a per capita income of $770, and Burkina Faso, with a per capita income of $210.) Less developed partners tended to resist a faster reduction of intra-Community tariff and nontariff barriers. They feared that large revenue losses and the transitional costs of economic restructuring would be insufficiently compensated by access to the Community’s compensation fund. Initial deadlines for even minor liberalization measures—such as a freeze on all intra-Community tariffs and initial moves toward selective, modest tariff cuts—were therefore repeatedly postponed. This contrasts with the experience of regional groups in industrial countries—particularly the EC—where “structural funds,” although limited, have been carefully targeted to assist depressed regions and remove structural rigidities in labor markets, thus facilitating adjustment following liberalization (Gordon, 1991).

Even if intraregional trade liberalization programs had been successfully implemented, some structural elements would still have limited the scope for potential gains. A first constraint has been the point of departure of regional trading arrangements among developing countries. Typically the trade systems of member countries were characterized by high and highly dispersed effective protection, with little intraregional trade occurring prior to integration. This contrasts with the starting position of regional arrangements among industrial countries, where the countries involved tended to be each other’s best customers. A second—and fundamental—constraint has been that regional pacts among developing countries with similar factor endowments have offered, by their nature, limited scope for trade creation along the lines of comparative advantage (for which different factor proportions in technology are key); at the same time, their small markets and low per capita incomes have not been conducive to significant trade based on scale economies and product differentiation. Finally, underdeveloped capital markets, barriers to entry, differential tax and regulatory environments among member countries, and restrictions on intraregional factor mobility have seriously constrained the capacity to reallocate resources and rationalize production in response to the removal of barriers to intraregional trade.

Quantifying Net Gains

A number of studies have attempted to estimate the net static gains and losses of developing country regional arrangements. These produce mixed results for Latin American arrangements, while studies of sub-Saharan African regional arrangements generally indicate that they resulted in negligible trade diversion. The results have been highly sensitive to, among other things, the choice of time period and the technique used to determine the path for trade and other economic variables that would have obtained in the absence of the arrangement. The results of a sample of these studies are given below.31

Studies of the CACM—the most intensively examined developing country regional group—have shown inconclusive results. Wilford (1970) and Aitken and Lowry (1973), focusing on the eleven to twelve-year period ending in 1967, concluded that trade creation predominated in the years following the formation of the union. Willmore (1976), on the other hand, using 1958–68 data and a more disaggregated approach, found that the CACM had predominantly trade-diversion effects in nondurable consumer goods but led to significant external trade creation in intermediate goods, while results for durable goods were inconclusive. Cline (1978) expanded the measurement of net static gains for the CACM beyond trade creation and diversion to include the effects of extraregional trade creation, trade suppression (replacement of imports by domestic production), increases in domestic production at wage rates above the social opportunity cost of labor, economies of scale, foreign exchange savings, and industry intermediate-input effects (reflecting output and import requirements via input-output relationships). According to Cline, although trade diversion exceeded trade creation in the CACM, the effects of the other factors more than compensated for this, leading to substantial net static gains—albeit unequally distributed among member countries—on the order of 2 percent of the region’s 1972 GDP.32

With regard to the LAFTA, George and others (1977), using 1952–69 data, and Langhammer and Spinanger (1984) using 1962–79 data, conclude that trade diversion accounted for most of the expansion of intraregional trade in the LAFTA. By contrast, Aitken and Lowry (1973), focusing on the 1955–67 period, find no evidence of significant trade diversion.

In the case of the Andean Pact, Khazeh and Clark (1990), using 1968–77 data, estimate static gains to be sizable—about 9 percent of the region’s 1977 GDP—in part because of substantial extra-regional trade creation. (Khazeh and Clark’s results are, however, difficult to accept at face value. The measured gains appear disproportionate, exceeding most estimates of static gains for the EC; they also incorporate a substantial upward bias by using the counterfactual, which assumed that the trade of Andean Pact countries would have unfolded, in the absence of the arrangement, in step with the trade of certain comparator countries.)

As regards sub-Saharan African groups, Lang-hammer (1978) found that the CEAO led to negligible trade creation and that extraregional imports were replaced more by domestic production (trade suppression) than by intraregional imports (trade diversion). A World Bank (1989a) study suggests that negligible trade creation has been the rule, rather than the exception, in the case of sub-Saharan Africa, as indicated by only marginal increases in the ratio of intraregional trade to GDP during 1965–83. Most of these groups displayed a significant increase in the ratio of extraregional trade to GDP over the same period, but this can hardly be attributed to the regional integration schemes themselves, especially since most have not been implemented to an appreciable extent. These conclusions are corroborated in Table 9, which decomposes the changes in the trade openness ratio (trade to GDP) over relevant periods into intraregional and extraregional trade components—that is, crude measures of trade creation and trade diversion, respectively.

Few studies have attempted to quantify the trade impact of the trade liberalization measures undertaken by the GCC. Cain and Al-Badri (1989) have identified the industries responsible for the relative increase in intra-area trade as those of textiles, nonmetallic products, and other manufacturing. Metwally and Daghistani (1987) attribute the weakness of the industrial base in these countries, and therefore marginal intraregional trade, to the oil boom experienced by GCC members since 1973, which has transformed these economies into major importers.

Reflecting the inherent methodological difficulties involved, few studies have attempted to quantify, ex post, the dynamic gains or losses of regional integration arrangements among developing countries. One of the few studies available is of the CACM (Cline, 1978); it attempted to gauge the impact of integration on structural change—assuming that this would raise welfare of risk-averse nations—and on foreign capital inflows. The study concludes that these types of dynamic effects in the CACM led to additional gains, equivalent to about 1 percent of the region’s 1972 GDP.

Contrary to the expectations of many developing countries, empirical studies typically have found that regional arrangements have not yielded significant economies of scale (see Carnoy, 1970, for Latin America; and Pearson and Ingram, 1980, for Ghana and Cote d’Ivoire). In the case of the CACM, while Cline (1978) estimates positive gains from economies of scale, Willmore (1979) finds that the common market has tended to support oligopolistic behavior (for example, market sharing arrangements) more than specialization and economies of scale.

Views on the Role of Regional Integration

The past experience with regional integration arrangements among developing countries raises questions about the future role of these in developing country trade policies. These questions are all the more relevant in view of the recent resurgence of interest in regionalism.

Views among economists on this matter vary widely. Some authors focusing on Latin America—for example, Dornbusch (1989)—argue that, given the often prohibitive protection systems and the political resistance to a substantial dismantling of these systems, intraregional liberalization behind (presumably transitory) high external barriers offers little risk of trade diversion and potentially substantial gains; these gains include those derived from industrial complementarity, economies of scale, rents associated with preferential market access, greater product variety for consumers, and competition-induced improvements in efficiency.

The above reasoning is based, in part, on considerations of political feasibility and seems to assume that political resistance to far-reaching unilateral liberalization would be significantly greater than to more limited liberalization within regional groups. Intraregional liberalization is, furthermore, assumed to be easier to “sell” domestically in view of the reciprocal nature of regional trade concessions. Experience suggests, however, that the political resistance to liberalization is not necessarily a continuous, upward sloping function of the degree of liberalization; rather, there appears to be a “threshold” to resistance, associated with shifts in economic philosophies. Thus, as long as policies have been firmly committed to protecting import-substituting industries, little progress has been made toward meaningful intraregional liberalization, reciprocity in regional trade concessions not-withstanding. By contrast, once sufficient political consensus has been mobilized in favor of a shift in policies toward outward orientation and greater reliance on market forces, developing countries have tended to opt for comprehensive trade reforms on a unilateral basis, regardless of the actions of regional and nonregional trading partners.

Intraregional liberalization behind (temporarily) high external barriers is often also proposed as a means to provide a “training ground” for firms to gain experience and reap economies of scale, before they are exposed to import competition or are ready to move toward export markets outside the region. Experience suggests, however, that this approach has inherent dangers. First, most regional groups among developing countries are not likely to offer sufficient market size for exploiting economies of scale. High protection against third countries in these cases would tend to foster excess capacity and inefficient and monopolistic market structures, in which rents would not constitute a net income gain for society—but mainly a transfer of income from consumers to firms and a net loss of welfare relative to free trade. This would tend to create vested interests that would resist subsequent liberalization, thereby perpetuating high external barriers initially envisaged as transitory.

Second, even where economies of scale might be reaped through regional pacts—and, thus, where rents might represent a net increase in member countries’income that could not be obtained without temporary protection of regional markets—the prolonged maintenance of high external barriers would tend to lead to excessive entry, as new firms try to capture monopoly rents. This would in turn drive incumbent firms up their average cost curves, with excess profits eventually absorbed by reduced efficiency owing to operation at below the optimal scale (Eastman and Stykolt, 1962; and Horstmann and Markusen, 1986). A frequently cited example in this connection is the automobile industry in Latin America, where—perhaps with the exception of Brazil—most plants have average production runs well below what is generally considered to be the minimum efficient scale (Rodrik, 1988).33 Thus, to realize the gains from economies of scale, external protection would have to be lifted relatively soon—not an easy thing to do. In sum, as Krugman (1989, p. 359) notes, protectionism has extra costs when domestic market structures are concentrated and “import substituting industrialization looks even worse in the new theory [which models trade under imperfect competition and economies of scale] than in standard theory [based on models that assume perfect competition and constant returns to scale].”34

In contrast to Dornbusch, other authors contend that the case for preferential trading agreements among developing countries is not convincing. This view is based partly on their disappointing past experience, including the fact that no developing country case can be cited where a regional integration scheme was, in itself, a significant contributor to development (Langhammer and Hiemenz, 1990). More fundamentally, it is argued that there is a “fallacy of transposition” in the expectation that the success of regional integration among industrial countries could be replicated in arrangements among developing countries. In particular, many developing country regional groupings include members with small markets, low per capita incomes, and similar factor endowments—leading to similar productive structures. The similarity of economies in these cases does not serve as a basis for trade expansion based on intra-industry specialization and product differentiation—which contrasts with industrial country cases, where intra-industry trade expansion is made possible by large markets, product differentiation, and high per capita incomes. Developing countries would thus be better off exploiting gains from trade based on differences in resource endowment and productive structure—seizing comparative advantages dynamically at every point in time as they build their human, capital, and technological stocks. This strategy would be best pursued in the context of unilateral and multilateral liberalization (Greenaway and Milner, 1990).

While the above conclusion is compelling, it tends to discount excessively the potential for intraregional trade and thus needs qualification. As Greenaway and Milner (1990) note, similarities among developing countries should not be exaggerated. Some regional groups display differences in per capita incomes—although typically within relatively low levels—and in productive structures among members, suggesting the existence of complementarities and the potential for a degree of intraregional trade expansion based on inter-industry specialization (and perhaps even intra-industry specialization in the larger groups).35 Even in sub-Saharan Africa, where per capita incomes and market sizes are lowest among developing countries, the World Bank (1989b) estimates that some $4–5 billion of this region’s imports from the rest of the world could be potentially supplied by other sub-Saharan African countries already exporting similar products to the rest of the world. Should this potential be realized, the share of intraregional trade in sub-Saharan Africa would treble to 18 percent. Although subject to qualification, this result could be seen as evidence of the possibility of efficient trade switching, provided that countries are willing to remove barriers to trade on a fairly comprehensive basis—which has not been the case to date. The thrust of this argument would apply, a fortiori, to regional trading arrangements among middle-income developing countries.

Regardless of the economic case for regional trading arrangements among developing countries, these arrangements can be expected to continue to be a feature of the international trading system—indeed, they may become increasingly important given the perceived noneconomic benefits and other factors discussed earlier. Hence, it is important to identify conditions under which the potential gains of such arrangements can be maximized while the losses—to members and nonmembers—are minimized. The review of experience suggests that the potential gains of regional arrangements can be maximized if at least two key conditions are met; (1) maintenance of an outward orientation; and (2) provision for across-the-board intraregional liberalization with automatic timetabling. These two conditions broadly correspond to the provisos in Article XXIV of the GATT (see Chapter IV).

Outward orientation in development strategies would be reflected, among other things, in relatively low and uniform levels of protection vis-à-vis third countries and in open and transparent foreign direct investment regimes. These conditions are crucial not only for minimizing the trade-diversion effects of regional trade preferences, but also for ensuring that developing countries are part of the process of globalization of investment and production, the main channel of access to technology transfer. The process of dynamically seizing comparative advantages in a multilateral context would allow developing countries to achieve stronger growth while diversifying their productive structures. As per capita incomes and capital stocks rise, so would the opportunities for trade expansion based on similarities among developing countries. South-South trade is more likely to grow if developing countries are integrated into the multilateral system of trade and investment than if they are not (Greenaway and Milner, 1990).

The need to achieve outward orientation in developing country regional arrangements would appear to favor free trade areas rather than customs unions, particularly where little shared sovereignty through supranational structures for common decision making is envisaged. A free trade area allows importing member countries to avoid undue costs from trade diversion by unilaterally reducing trade barriers, thus shifting the source of supply back to lower-cost outside producers. Also, member countries already maintaining more liberal trading regimes or wishing to liberalize their trade policies would not be constrained by free trade areas. Under a customs union, protectionist pressures could end up forcing the more liberal members of the union to adopt higher levels of external protection vis-à-vis nonmembers. Free trade agreements would, by contrast, allow the more liberal members to proceed unilaterally with the ongoing non-discriminatory removal of trade barriers; this would, in turn, impose a salutary—and noncoercive—liberalizing pressure on external trade policies of the rest of the members.

Experiences with regional trading arrangements among developing countries clearly illustrate the drawbacks of undue selectivity in trade concessions and non automatic timetables. These features have generally reflected deep-seated unwillingness to liberalize intraregional trade in a meaningful way. They have also tended to bias the negotiations in favor of trade concessions on products not yet produced in the region—thereby widening the scope for trade diversion or suppression, especially where external tariffs are high—or on products already traded relatively freely within the region—thus adding little trade-creation potential. An approach to regional integration based on automatic time-tabling and—as required by Article XXIV of the GATT—an across-the-board reduction of trade barriers would help avoid these drawbacks. It would ensure participation in regional trading agreements only where the liberalization commitment is serious. It would therefore also enhance the probability that such arrangements would be outward oriented, thus reducing the risk that they might be used as a substitute for—or an excuse for not engaging in—nondiscriminatory liberalization.

The full benefits from trade liberalization in developing countries will not be reaped, however, unless industrial countries remove their high protective barriers on products where developing countries possess clear comparative advantages—for example, agricultural products, textiles and clothing, and leather products. As emphasized earlier, the Uruguay Round of GATT negotiations offers the best forum for developing countries to secure from industrial countries the removal of these barriers.

Finally, economic cooperation is also relevant. While the experience with the trade liberalization aspects of regional integration arrangements has been disappointing, considerable success with economic cooperation has been achieved in a number of such arrangements. Potentially vast mutual gains may be secured through regional economic cooperation among developing countries in areas where significant externalities and public goods exist and, therefore, where market mechanisms tend to fail. As Langhammer and Hiemenz (1990) note, experience indicates that success in economic cooperation among developing countries is likely to be greater where the joint production of “software” public goods (for example, education and training; research and development, particularly on agriculture; joint management and control facilities; coordination of investment incentives; and so on) is given sufficient weight in comparison to the joint production of “hardware” public goods (for example, infrastructure). This type of economic cooperation has the advantage of minimizing the risk of “white elephants” and their associated—often unsurmountable—management difficulties.


Nonmarket mechanisms to allocate manufacturing industries must be distinguished from regional cooperation on infrastructure projects of mutual advantage. In the latter case, successes in regional cooperation are not difficult to find, as illustrated, for instance, by the experience of SADCC (see Box 3).


El-Agraa (1989, Chapter 15) discusses some of the approaches used in measuring the effects of economic integration among developing countries.


Cline’s conclusion, however, hinges largely on the sizable measured “gains” from foreign exchange savings, which are based on a rather questionable method (Bulmer-Thomas, 1988). According to this method, trade suppression and diversion evaluated at the shadow exchange rate are deemed to constitute a gross saving of foreign exchange. Net savings are then arrived at by subtracting the domestic resources used, but these are valued at the official or market (rather than shadow) exchange rate.


In the Andean Pact, excessive entry in the automobile industry was administratively decided from the outset, as various types of models were allocated via political negotiation among member countries, in the context of implementing the Pact’s joint industrial planning program.


Rodrik (1988) similarly argues that import-substitution policies in developing countries “look doubly bad” in the presence of imperfect competition and increasing returns to scale. He concludes (p. 132) that “the levels of protection observed in the manufacturing sectors of most developing countries vastly exceed any that could be justified by the presence of imperfect competition.”


Ex post econometric evidence of a possible link between the CACM and/or LAFTA and intra-industry trade is found in Balassa (1979), Havrylyshyn and Civan (1983), Balassa (1986), and Balassa and Bauwens (1987 and 1988). As Greenaway (1989, p. 35) notes, however, “all of this documentary work suffers from the classic antimonde problem. Would the growth in [intra-industry trade] have been as rapid in the absence of integration?”