Regional trading arrangements—bilateral as well as plurilateral—have been a fairly constant feature of the world trading system. The recent trend toward regionalism, however, may be qualitatively different from past efforts and may carry greater risks of becoming a substitute for, rather than a complement to, multilateralism. Apart from the potentially distorting effects of these arrangements on the allocation of world resources (see Chapter II), a number of important elements unique to current regionalism seem to pose a greater threat to multilateralism.

Regional trading arrangements—bilateral as well as plurilateral—have been a fairly constant feature of the world trading system. The recent trend toward regionalism, however, may be qualitatively different from past efforts and may carry greater risks of becoming a substitute for, rather than a complement to, multilateralism. Apart from the potentially distorting effects of these arrangements on the allocation of world resources (see Chapter II), a number of important elements unique to current regionalism seem to pose a greater threat to multilateralism.

Regional arrangements in industrial countries have not appeared—until now—to have been obstacles to efforts to achieve multilateral liberalization; trade barriers against third countries have been low and trade diversion has generally been minimized. Moreover, in some cases it has been easier to negotiate regional, rather than multilateral, liberalization. For the current trend to continue to support multilateralism three conditions are necessary: (1) steady progress with multilateral liberalization; (2) an outward orientation in regional arrangements; and (3) across-the-board intraregional liberalization. These conditions have implications for the form of arrangement chosen, and its membership. Outward orientation, for instance, may be easier to achieve in a free trading area rather than a customs union because it is less restrictive. Similarly, developing countries are likely to exploit gains from trade more readily in arrangements with industrial countries on the basis of differences in resources and productive capacity.

Finally, the globalization of investment and production is likely to lessen the risk of regional arrangements becoming closed blocs. Global investment—carried on by multinational enterprises with global, rather than national, strategies—has exposed the limitations of nationally based policies; it increasingly confronts governments with the need to liberalize and harmonize policies, standards, and regulations with other countries, including in—but not limited to—trade and trade-related areas. Harmonization and liberalization in response to the challenges posed by globalization is in some instances being conducted in the context of regional arrangements. Increasingly, however, globalization calls for multilateral solutions.

Aspects of the New Trend

The first unique aspect of the renewed interest in regionalism is that it has taken place in a weakening multilateral trading system. Substantial difficulties were encountered in launching the Uruguay Round in 1986; it took four years of arduous work—from 1982 to 1986—to break the resistance to a new Round. And progress in completing the Round has been slow. The GATT’s inability to curb rising protectionist pressures since the end of the Tokyo Round has created frustration with the slowness of the GATT process and raised concerns about its effectiveness and its ability to adapt to the emerging trade issues of the 1990s (which include problems posed by high-technology industries and by firms engaged in trade and investment in goods and services on a global scale). Thus, doubts have arisen about the political will of the major players to strengthen the GATT’s credibility through the Uruguay Round. In this context, regionalism has been advocated as an alternative to multilateralism; indeed, regional approaches to liberalization might be seen as a response to the regional and global integration of investment and production with which an unwieldy and slow multilateral approach has been unable to keep pace.

The second novel element is that far-reaching and successful regional initiatives by major industrial countries have created the perception that preferential trading arrangements are here to stay. In part, this reflects the demonstration effects of the extraordinary, albeit fortuitous, conjunction of EC 1992 and the United States-Canada FTA. Some have interpreted this conjunction of events as a confirmation of a trend toward a fragmentation of the multilateral trading order into a tripolar system of trade blocs—centered around the United States, the EC, and Japan (although the latter does not form part of a trade bloc). The accord between the EC and EFTA to form a European Economic Area (EEA) and the prospect of a North American Free Trade Area (NAFTA)—involving Canada, Mexico, and the United States—have heightened this perception.

The above concerns have been reinforced by the shift in the trade policy stance of the United States—the main architect of the GATT system—from a virtually one-track approach emphasizing nondiscrimination to a multitrack approach encompassing unilateralism, regionalism, and multilateralism.36 To be sure, the United States has sought to use regional initiatives as complements to multilateralism, both substantively (for example, to build useful precedents for broader liberalization in areas not yet covered by the GATT) and tactically (for example, to goad other countries into participating more seriously in the Uruguay Round (Schott, 1989)). However, this has been accompanied by the perception that, if GATT negotiations fail, the United States would be prepared to use bilateral or plurilateral arrangements as an alternative course of action.37 In all, there is an increased concern that the big players in the multilateral trading system may be giving greater priority to regionalism in their trade policy agendas at the expense of multilateralism.

Finally, the recent trend toward regionalism involves preferential trading arrangements among industrial and developing countries involving reciprocal trade concessions. (Reciprocity in concessions has, of course, been a standard feature of the numerous regional arrangements among developing countries.) This is illustrated by the ongoing United States’ Enterprise for the Americas Initiative (EAI) and the negotiations between Canada, Mexico, and the United States on a NAFTA, as well as the association agreements between the EC and Czechoslovakia, Hungary, and Poland. Previous arrangements with developing countries generally took the form of one-way trade preferences granted by industrial countries in favor of selected developing countries (examples include the U.S. Caribbean Basin Initiative and EC preferences extended to certain developing countries under the Lomé Convention).38

Clearly, the main intention of industrial countries in offering opportunities for (reciprocal) preferential trading arrangements to developing countries has been to encourage further economic and political reform, particularly at a time when budgetary constraints in industrial countries have placed limits on foreign aid. Nonetheless, there is a risk that developing countries may interpret such moves as a further sign that the major players have doubts about the viability of the GATT and are offering preferential trading arrangements as an alternative. Such an interpretation may in turn induce developing countries to divert scarce resources away from multilateral negotiations so as to concentrate efforts on securing a place in the queue for preferential trading arrangements with the major players.

While the above concerns may be overblown, they raise the important question of whether the recent revival of regionalism is likely to retard, rather than promote, multilateralism. This issue is addressed below by examining the compatibility of regional arrangements with the GATT and by focusing on some of the risks and limitations of regionalism.

Compatibility of Regional Arrangements with the GATT

The principle of nondiscrimination is a cornerstone of the GATT. It is embedded in Article I, which commits the contracting parties to grant most-favored-nation (MFN) treatment with respect to trade concessions to all other GATT members. Regional trading arrangements (bilateral or plurilateral) are, by definition, discriminatory: trade concessions are granted only to members whose preferential access to other members’ markets may enable them to displace exporters from nonmember countries.

Although discrimination in regional arrangements runs counter to the principle of non-discrimination, the GATT permits regional arrangements under certain conditions. The architects of the GATT recognized that such arrangements could provide a complementary, practical route to universal free trade—where politics permitted a faster and deeper degree of liberalization. Thus, Article XXIV of the GATT permits departures from the MFN obligation for preferential trading arrangements (free trade areas, customs unions, or interim arrangements leading to either of the former) provided that: (1) other GATT members are notified of the details; (2) such arrangements do not raise trade barriers on the whole against other Contracting Parties to the GATT; and (3) such arrangements cover “substantially all trade” between partners and commit them to reduce barriers to intraregional trade, possibly by means of a schedule and within a “reasonable length of time.” (As indicated below, less strict conditions apply to developing countries.) The first condition was intended to enable other GATT members to scrutinize such arrangements to ensure their consistency with Article XXIV. The second condition sought to minimize the adverse effects of trade diversion on third countries and on members of preferential trading arrangements. The third condition, as further explained below, aimed to limit the number of preferential arrangements.

At first glance, the third condition appears some-what paradoxical, especially considering Meade’s (1955) conclusion that, other things being equal, a movement in the direction of 100 percent tariff cuts within a union would be associated with an increase in the probability of losses from trade diversion. This condition was central in the view of the GATT architects, however, in that it aimed at limiting the number of preferential trading arrangements to those where the intensity of the political commitment was commensurate with the liberalization and structural adjustment effort required (Bhagwati, 1990a). The underlying intent was to minimize the risk of the world trading system fragmenting into the discriminatory and sectoral bilateral arrangements that dominated the 1930s. At the same time, the “substantially-all-trade” clause recognized that full integration of trade among members of a preferential trading arrangement would create “an important element of single-national [sic] characteristics among [these members] . . . the resulting quasi-national status following from such integration in trade [would legitimate] the exception to MFN obligation towards other GATT members” (Bhagwati, 1990a, p. 1308). The latter argument takes on more force with higher degrees of economic and political integration among members: after all, each nation state is a customs union, and no one would consider the removal of barriers among states and provinces of a nation state as a threat to multilateralism.

The implementation of the provisions of Article XXIV was further weakened by the 1979 Enabling Clause. Under this clause, arrangements that are exclusively among developing countries can bypass Article XXIV altogether, unless the arrangements contemplate the selective removal of nontariff barriers, in which case approval by other GATT members is required. The Enabling Clause is thus substantially less demanding than Article XXIV in terms of notification and consultation, substantially-all-trade requirements, ex ante and ex post protection vis-à-vis third countries, and the timeframe for implementation of intraregional trade liberalization.

In practice, the provisions of Article XXIV have not been strictly observed partly because of ambiguities in the language that have permitted different interpretations. The notification and examination of the European Community was an important watershed for the subsequent fate of Article XXIV. For political reasons, GATT Contracting Parties decided not to issue a formal ruling on whether the Treaty of Rome was compatible with GATT provisions (Patterson, 1966; and Patterson, 1989). A precedent for inaction was thus set and has not subsequently been reversed (Schott, 1989). Of the 70 regional trading arrangements (including amendments to existing arrangements) notified during 1948–90 and examined by GATT working parties, only 4 were deemed by consensus to be compatible with Article XXIV.39 No arrangement, however, has been formally declared to be incompatible with Article XXIV.

With most GATT members involved in some form of preferential trading arrangement, members have tended to refrain from forcefully criticizing preferential trading arrangements involving other GATT members, particularly since the United States—previously the major critic of discriminatory arrangements—has become a user of Article XXIV.

The concerns with the lax implementation of Article XXIV are nevertheless not trivial. As stated in a report commissioned by the GATT (Leutwiler, and others, 1985; p. 41): “. . . exceptions and ambiguities have been permitted . . . [which] have set a dangerous precedent for further special deals, fragmentation of the trading system, and damage to the trade interest of non-participants.” And a former Deputy Director General of the GATT has lamented that “of all GATT articles, [Article XXIV] is one of the most abused, and those abuses are among the least noted” (Patterson, 1989, p. 36).

Notwithstanding the lax implementation of Article XXIV, there is broad agreement that these arrangements have not, in themselves, hindered multilateral liberalization, at least until now. Part of the reason is the maintenance of low trade barriers visà-vis third countries in most areas of trade (see Chapter IV). Through successive rounds of GATT negotiations, industrial countries have eliminated formal quantitative restrictions on most manufactured goods and drastically reduced import tariffs. Moreover, unilateral liberalization by some countries participating in regional arrangements have helped minimize trade diversion that might otherwise have occurred. The obvious exception to these developments is the EC’s adoption of the Common Agricultural Policy (CAP), which has sheltered the agricultural sector and stifled the views of the more liberal EC members that could have, in the absence of the CAP, joined forces with the advocates of liberalization. Regional arrangements cannot justifiably be blamed for the proliferation over the past fifteen years of nontariff measures (such as voluntary export restraint arrangements, countervailing and antidumping measures, and orderly market arrangements). These reflect a resistance in industrial countries to structural adjustment required by changes in comparative advantage and have not been confined to members of regional arrangements.

In addition, liberalization under regional arrangements can serve as a useful precedent that could be subsequently multilateralized through GATT negotiations. In effect, regional negotiations may be less hampered than GATT negotiations by problems of “free riding” (members enjoying MFN benefits while escaping obligations), “foot dragging” (members taking advantage of consensus rules to block progress until their demands are met), and “convoy effects” (the least willing participant determining the pace of negotiations). This, together with similarity in economic philosophies and a high degree of pre-existing integration, may allow certain countries to liberalize faster and more deeply within regional groups than multilaterally—including in areas previously not subject to GATT rules and disciplines. This has been illustrated, for example, by the liberalization of services and investment under the United States-Canada FTA, and by the liberalization of capital and financial services in the context of EC 1992. As noted below, however, there are limits to regional liberalization in those areas where protection is most ingrained.

The multilateralization of liberalization gains under regional arrangements presupposes, of course, that a credible and well functioning multilateral system is in place. Thus, steady progress with multilateral liberalization is essential to subsume preferential trading arrangements into a broader and more open trading system; to hold in check—and indeed to erode—the inherent discrimination of such arrangements; to convert their regionally circumscribed liberalization into building blocks for freer trade on a global basis; and to prevent regionalism from fragmenting the world trade system.

Risks and Limitations of Regionalism

As suggested above, concerns that the recent revival of regionalism may undermine the GATT-based trading system are related to a perception that the major players in the system may have shifted their priorities away from multilateralism. At one extreme, pessimistic scenarios based on a faltering of the Uruguay Round envisage a fragmentation of the trading system into large blocs with a “fortress mentality” and a proliferation of bilateral and minilateral preferential trading arrangements. Absent the restraint of multilateral disciplines, regionalism could, in effect, become divisive. Nevertheless, a more widely held view in recent years has been that a vigorous pursuit of regionalism, including by the major players, is consistent with a strengthened multilateral trading system. Undue emphasis on regionalism, however, would necessarily undercut multilateralism and render it inoperative; proponents of this view tend to overestimate the virtues of regionalism and underestimate its risks.40

The risks of regionalism include the possible diversion of scarce skills from multilateral to regional negotiations, the increased potential for friction among regional groups, and possible adverse effects on countries excluded from such arrangements.

Regional and multilateral negotiations are time-consuming and skill-intensive; they both compete for scarce—and often the same—human and administrative resources. The obvious concern is that the current proliferation of regional trading arrangements that have high priority in governments’ agendas may be diverting these scarce resources away from multilateral negotiations, precisely when concentrated negotiating energy and political commitment are required to complete the Uruguay Round successfully. Diversion of resources may occur even where governments would rather focus on multilateral talks, if they feel a need for defensive actions in response to moves toward regionalism by other trading partners.

An excessive number of preferential trading arrangements would lead to significant frictions, political pressures, and practical problems. Each new arrangement would inevitably undermine the value of the preferential access granted to some countries under existing arrangements, and lead to increased tension.41 The trade-diversion effects of regional arrangements may also lead nonparticipants to retaliate with preferential arrangements of their own, possibly triggering a damaging chain reaction. A proliferation of preferential trading arrangements would give rise to a host of technical problems, including mismatches in the phasing of tariff reductions under overlapping arrangements, inconsistency of rulings under different dispute settlement mechanisms, and confusion and conflicts in implementing and enforcing different rules of origin under separate FTAs. As Schott (1989) suggests, negotiation and maintenance of the GATT system appears “simple and straightforward compared with the maze of problems” that would result from an expanding network of preferential trading arrangements. These problems would be reduced, however, if preferential trading arrangements were open to new members, as this would help to convert these arrangements into building—rather than stumbling—blocks to universal liberalization (Bhagwati, 1990b).42

Smaller countries would be the most hurt by a drift toward regionalism that undermines the principle of MFN nondiscrimination. As Bhagwati (1990a, p. 1304) puts it, the GATT system is intended, among other things, to shield “the weak against the strong who would otherwise, in bilateral one-to-one confrontations, have an advantage constrained only by altruism and conscience.” An open multilateral trading system is particularly important to developing countries undertaking structural adjustment programs including the liberalization of their exchange, trade, and foreign direct investment regimes. Partly reflecting the high stakes they have in a strong system of multilateral disciplines, developing countries have been engaged in the Uruguay Round much more intensively than in previous GATT rounds, including through their participation in coalitions (for example, the Cairns Group of agricultural exporting countries and the de la Paix Group of middle-sized countries) that have served as catalysts for compromise proposals.

In view of the importance of the multilateral system for developing countries, it is not clear why they are interested in forming preferential and reciprocal trading arrangements with industrial countries. Several reasons have been advanced to explain this interest. First, these types of arrangements may have important elements of economic and political cooperation and may extend to areas (for example, foreign direct investment) hitherto not covered by the GATT. Second, for developing countries that have already liberalized trade substantially on a unilateral basis, these arrangements might serve to “lock in” liberalization and protect it from domestic protectionist pressures at home. (Of course, this objective can also be achieved by “binding” such liberalization in the GATT.) Third, developing countries might enter these arrangements defensively to hedge against the risk of rising protectionism in industrial countries. In this context, the reciprocity of trade preferences is puzzling. Industrial countries with whom developing countries might wish to enter into preferential trading arrangements are likely to account for a major share of the latter’s trade, which implies that an FTA that is comprehensive (as required by Article XXIV of GATT) would be tantamount, from the developing country point of view, to broadly based unilateral liberalization. In these circumstances, there would seem no reason for a developing country to circumscribe its reciprocal concessions by making them available only to the industrial partner in the FTA, rather than extending them on an MFN basis. As Fritsch (1989, p. 343) succinctly puts it:

  • … on the one hand, if [developing countries] tend to favor the maintenance of a restrictive trade regime, the proposal of an FTA [with a major industrialized trading partner] is by definition a non-starter, whereas on the other hand, if they are willing to liberalize, they should in fact do so on an MFN basis to avoid the negative trade diversion effects of the formation of an FTA.

As noted earlier, regional arrangements can forge—and have forged—liberalization in areas not yet covered by the GATT. It is doubtful, however, that these arrangements can deliver, on their own, substantial trade reform in areas where protectionism has proven to be most resilient. The latter include agriculture—which is subject to heavy protection and subsidization in most industrial countries, including in the EC, EFTA, Japan, and the United States—and textiles and apparel, where a highly restrictive regime of bilateral quotas under the Multifibre Arrangement (MFA) has been in effect for more than twenty five years. Regional arrangements involving industrial countries have done nothing to dislodge protection in these fields.

By contrast, it is widely believed that the Uruguay Round offers the best avenue—perhaps the only one—for achieving meaningful liberalization in these areas. This reflects the fact that the GATT forum—unlike negotiations toward regional arrangements—allows for “overall reciprocity” (Patterson, 1989), where the give-and-take in negotiations and trade-offs are not limited to country-by-country or sector-by-sector concessions. The success of the Uruguay Round hinges, to a significant extent, on whether a bargain can be struck in which industrial countries accept substantial liberalization in such key fields as agriculture and textiles in return for liberalization by developing countries in the “new” areas of services, investment, and intellectual property (Schott, 1990; Bhagwati, 1990b).

Finally, the ongoing process of globalization of investment and production is likely to lessen the risks of regionalism. The globalization process—carried on by multinational enterprises with global, rather than national, strategies—has exposed the limitations of nationally based policies and increasingly confronts governments with the need to liberalize and harmonize policies, standards, and regulations with other countries, including in—but not limited to—trade and trade-related areas. Harmonization and liberalization in response to the challenges posed by globalization is in some instances being conducted in the context of regional arrangements. Increasingly, however, the process of globalization calls for multilateral solutions; this fact will likely limit the risk that regional arrangements will turn into closed blocs.


During most of the post-World War Hera the United States took on the leadership role in building the GATT-based trading order, typically remaining suspicious of regionalism (Bhagwati, 1990a). A key exception was its support for the EC in 1958—largely because of political and military considerations (Patterson, 1966). Until the mid-1980s, the United States had not made use of Article XXIV of the GATT—which sets out provisions under which preferential trading arrangements involving GATT members may qualify for a derogation of the most-favored-nation (MFN) obligation (see below). Since 1985, beginning with its free trade agreement with Israel, the United States has been participating in reciprocal and preferential trading arrangements. A discussion of the relevance of, and reasons for, the U.S. “conversion to Article XXIV” is found, among other things, in Bhagwati (1990a) and Schott (1989).


For instance, George Schultz, former U.S. Secretary of State, noted in a speech in 1985 that “. . . bilateral arrangements . . . can strengthen the multilateral system. Our hope … is that the example of greater liberalization [under these arrangements]—and the recognition that the United States can pursue another course—will motivate a larger group of nations to tackle the job of expanding trade on a global basis.” (Quoted in Patterson, 1989; emphasis added.) A similar point has been made by James Baker III, former U.S. Treasury Secretary (Baker, 1988): “If possible, we hope . . . liberalization will occur in the Uruguay Round. If not, we might be willing to explore a ‘market liberalization club’approach, through minilateral arrangements or a series of bilateral arrangements” (emphasis added).


The proliferation of preferential arrangements built by the EC after 1958 with its former colonies and Mediterranean neighbors was subject to strong criticism, led by the United States, in the GATT, and resulted in a standoff in the early 1970s. The reciprocal nature of such arrangements was a key point of contention because the EC was perceived as using these arrangements to secure preferential access to developing country markets. According to Luyten (1989), the impasse was broken through an informal understanding (the “Casey-Soames understanding”) through which the EC obtained “armistice” on the legal issue in the GATT by committing itself not to insist on reverse trade preferences from developing countries.


The four cases are: (1) the South Africa-Rhodesia Customs Union (1948); (2) the Nicaragua-El Salvador Free Trade Agreement (1951); (3) Nicaragua’s participation in the Central American Common Market (1958); and (4) the Caribbean Common Market (1973).


Schott (1989) discusses the weaknesses in arguments advanced by the proponents of regionalism.


Schott (1989) notes that this type of problem has already occurred in the case of the United States-Israel FTA, with Israel considering that Canada obtained better terms in its FTA with the United States, particularly regarding the provisions on trade in services and the dispute settlement mechanism.


Schott (1989, p. 23) notes in this connection that the U.S. Congress has not favored open-ended FTAs, “because of the uncertainty about which countries would join and thus what the anticipated adjustment pressures and trade effects would be. In fact, under U.S. law only self-contained FTAs may qualify for ‘fast-track’implementing provisions.”