After decades of continuing progress in multilateral trade negotiations, regionalism is once again being viewed as a solution to the major international economic problems of our times. What's going on?
Slow progress at the GATT has led some economists to conclude that a division of the world into three trading blocs---Europe, the Americas, and East Asia---is the fastest road to multilateral free trade. They argue that negotiations for free trade are far more likely to succeed when conducted among three parties, rather than 154. And for many countries, the proliferation of nontariff barriers in the developed world has made regional integration an attractive policy option. Integration with the United States or European Communities (EC) offers them guaranteed access to a large market.
Not surprisingly, Mexico, Chile, and other Latin American countries are lining up to join a free trade area (FTA) with the United States. And several countries in Western Europe, Eastern Europe, and North Africa are knocking at the EC's door.
Similarly, perceptions of a fortress mentality in Europe and fears of an imminent Western hemispheric FTA have led some East Asian nations to contemplate a defensive regional bloc of their own. And in the former Soviet Union, political disintegration has made regional integration a way of preserving fast-disintegrating trade among the new republics.
This stampede toward regionalism is not new. Four decades ago, regionalism spread across the world like wildfire, ignited by the creation of the European Communities in 1957. The United States, then philosophically opposed to regionalism, saw a united Western Europe as an effective deterrent to the growing Soviet threat and threw its weight behind the EC. This was quickly followed by a proliferation of regional arrangements around the world, especially in the developing countries of Africa and Latin America.
Despite high hopes, this first round of regionalism did not flourish outside Europe. The primary motive behind the movement in the developing world was industrialization through regional import substitution. It was thought that infant industries could first learn to export within a protected regional market---and then face world competition. But for the same reasons that import substitution failed in countries, it also failed in the regions. By the late 1970s, outward-oriented policies had begun to capture the imagination of policymakers. In the years that followed, unilateral, nondiscriminatory trade liberalization became the order of the day, and regionalism was pushed into the background.
Today, regionalism is back with a vengeance. In its current incarnation, regionalism has engulfed all major players in the world economy. Division of the world into three trading blocs---never an issue in the first round of regionalism---is being debated seriously. Confidence in the GATT process is on the decline, and enthusiasm toward the regional approach on the rise. Most notably, the United States---an ardent defender of the multilateral approach and the most formidable opponent of regionalism in the 1960s---has become an active perpetrator of regionalism. The loudest message from these developments is that regionalism this time is here to stay.
In the first round, regionalism was scarcely viewed as a threat to multilateralism. The GATT process was dominated heavily by developed countries, with the United States as the major economic power. Developed countries negotiated tariff reductions on products of interest to each other and extended them to other member countries on the most-favored nation (MFN) basis. So, the integration schemes in developing countries posed no threat to the GATT negotiations. And the integration in Europe helped organize the negotiations better. The United States could deal with the European Community as a single unit, since the EC had a common external tariff.
Today, there is a distinct possibility that regionalism will divide the developed world into three trading blocs: Europe, the Americas, and East Asia. This has brought the global aspect of regionalism to the forefront of the policy debate and divided more sharply than ever the defenders of the multilateral process and the devotees of regionalism. At one extreme, trading blocs are viewed as catastrophic to multilateralism. At the other, they are seen as the building block to successful conclusion of negotiations aimed at free world trade.
The forces of multilateralism have weakened in recent years, and the attraction toward regionalism has strengthened. Why? It is important to recognize that thinking about trade policy has a strong mercantilist bias. In choosing its trade policy, each nation operates on the premise that when it liberalizes trade, it alone bears the cost of liberalization---while all those that obtain freer access to its market share the benefits. So, even though trade liberalization is beneficial in the aggregate, without a coordinated liberalization by all, it is viewed as a losing proposition. Free trade, perceived as a public good, thus suffers from the usual free-rider problem.
In the decades immediately after World War II, the United States accounted for close to half of global GDP. It could generate large benefits for the world economy by freeing up its trade. And it could internalize a large part of those benefits. This gave the United States a reason---and the leverage---to push for freer trade around the world. In the ensuing years, it took on the promotion of the cause of free trade---supplying such international public goods as the Marshall Plan and the various rounds of the GATT negotiations.
Conditions have changed. The United States is no longer the leader it was in the 1950s and 1960s. The EC has grown large---and driven by its farm lobbies and the fortress mentality promoted by Europe 1992, become increasingly inward-looking in the view of many observers. Japan and the East Asian newly industrializing countries (NICs), hardly significant players in the world economy immediately after World War II, have emerged as major economic forces. The United States is now what Jagdish Bhagwati calls a "diminished giant," suffering from large current account deficits and a sense that it can no longer dominate the world markets the way it did a few decades ago.
These developments---along with foot-dragging by the EC at the Uruguay Round and a perception that Japanese markets are closed to U.S. suppliers---have fueled the (already existing) anti-multilatera-list ethos in the United States and moved it to consider forming a bloc of its own. The results: the Canada-U.S. Free Trade Agreement (CUSTA), the negotiations for a North American Free Trade Agreement with Mexico (NAFTA), and the Enterprise of Americas Initiative.
The conversion of the United States to regionalism has encouraged similar thinking in East Asia. The CUSTA and NAFTA negotiations have prompted some countries in East Asia to reassess the possible gains from regional integration. A move is under way to turn the Association of Southeast Asian Nations (ASEAN) into the ASEAN Free Trade Area. There is also concern that Asia's future access to the U.S. market will be curtailed substantially if the Western hemisphere's FTA becomes a reality. This concern has encouraged the view that East Asia, too, should form a defensive bloc of its own, centered on Japan. Given its stake in the U.S. market, Japan has resisted the idea so far---but that could change rapidly if the Americas turn into an inward-looking bloc.
Whether trading blocs are good---or bad---is the most controversial issue. There is general agreement that complete free trade in the world is the most desirable goal and that today's trade regime is not the best we can do. But in getting to there from here, there are many possibilities, and the disagreements among economists over these possibilities are pronounced.
The issue has two broad aspects. First, the static impact question focuses on whether the immediate effect of regional integration on world welfare is positive or negative. Second, a dynamic time-path question asks whether, regardless of the impact, regionalism will lead to multilateral free trade by merging regional blocs into a single world bloc.
The questions are illustrated in figure 1, borrowed from Bhagwati. The static impact question is whether the regional integration today will take the world economy to a higher welfare (path 1) or keep it at a lower welfare (path 2). The dynamic time-path question is whether blocs will eventually converge (path 3) or remain fragmented, leaving world welfare lower (path 4). Moreover, will the multilateral process take the world economy along path 5 or leave it at a lower level of welfare along path 6? How do the various time paths compare?
A partial answer to the first question, based on trade creation and trade diversion, was provided earlier. This answer was adequate for the first round of regional arrangements, which could be reasonably evaluated in isolation, taking the rest of the world as a passive entity. But in the new regionalism, the issue is more complex. Today, large blocs are forming almost simultaneously in different parts of the world. So, a proper analysis must take into account the interdependence of blocs, including strategic interactions among them.
The theoretical literature on the economics of trading blocs is still in its infancy. Some innovative models have been developed, but their applicability to the real world is unfortunately limited. Indeed, even the authors of these models hesitate to draw policy implications from them.
At the forefront of the literature is Paul Krugman's model, which T. N. Srinivasan has already declared "theory without relevance." To give some flavor to the model, assume that the world consists of a large number of identical countries and that there is one differentiated product with many potential varieties. Each country specializes in one of these varieties and imports the varieties produced by other nations. Suppose we divide this world into several identical blocs, such that each bloc chooses the tariff on the other blocs optimally. A trivial but important point is that world welfare is maximized when the number of blocs is either one or very large. In the first case, we have free trade by definition. In the second, the optimal tariff is near zero, so that once again we obtain free trade after the fact. Clearly, starting from one bloc, as the number of blocs increases, welfare must first decline, reach a minimum, and then rise. The critical question is: at what number of blocs is welfare minimized? For a variety of simulations, Krugman finds that this number is approximately 3!
Taken seriously, this result suggests that the static impact of a division of the world into three blocs is negative and will move the world along path 2. But it is easy to construct alternative models that yield different conclusions. In particular, after allowing for asymmetric blocs, there is no clear relationship between the number of blocs and welfare. Srinivasan shows this in a two-good, multicountry Ricardian model. Alas, that returns us to the Vinerian conclusion that the static impact of trading blocs is ambiguous.
The implications of theoretical models notwithstanding, Krugman and Larry Summers have argued separately that since blocs are likely to form among geographically proximate countries that trade intensely among themselves to begin with, trade-creating effects will almost certainly dominate trade-diverting effects. The obvious example is NAFTA. Mexico exports more than 80 percent of its goods to the United States even without a free trade area. So, it is unlikely that NAFTA will be trade-diverting on balance.
Although there is some truth in this, the general applicability is in doubt. Geographical proximity need not be accompanied by intense trade: Africa and South Asia export 95 percent of their goods outside their regions. In the Western hemisphere, trade of most Latin American countries is very diversified. Excluding Mexico, roughly 55 percent of Latin America's exports go to regions other than the Western hemisphere. Even East Asia exports more than 65 percent of its goods outside East Asia.
More important than the static impact question is the dynamic time-path question. Very little is known about this issue, and Bhagwati addresses it systematically for the first time. The ultimate issue here is whether, over time, a few large trading blocs will lead to free world trade faster---and with greater certainty---than the multilateral process. As with the static impact question, the answer is not clear-cut. A small number of blocs can alleviate the free-rider problem that afflicts the GATT process and make the approach to a cooperative solution faster and more certain. But the larger the blocs, the greater the market power they enjoy, and the greater the temptation to impose restrictions on trade outside the bloc. There is no clear presumption on whether blocs will lead to free trade faster and with greater certainty than the multilateral process.
Experience also gives mixed guidance on this issue. The EC stands in sharp contrast to other parts of the world. Once the Treaty of Rome was signed, intra-union barriers to trade came down rapidly and with certainty. Elsewhere in the world, regionalism has been piecemeal. In Africa and Latin America, regionalism did not go very far, and there were many reversals. In some cases, regionalism even became a barrier to liberalization. In agriculture, the regional approach was no more successful than the multilateral one. Indeed, the EC-wide existence of the Common Agricultural Policy (CAP) was perhaps a more effective barrier to freeing up trade in agriculture than would have been the case without the EC.
To some extent, the question of whether trading blocs are good or bad is moot. Regionalism is here---and likely to stay. So, a more constructive approach is to ask whether mechanisms can be devised to turn the stumbling blocs into the building blocs of a free world trade system. In particular, how can we reform the GATT rules and design regional arrangements in such a way that regionalism complements rather than substitutes for multilateralism?
Three suggestions have been made. First, Article XXIV could be modified to rule out FTAs and allow only customs unions. A customs union requires a common external tariff and, given GATT bindings on most tariffs, ensures that all tariffs come down to the lowest level prevailing in the union at the time of its formation. This happened recently when Greece, Portugal, and Spain joined the European Community. Because the European Community is a customs union, these countries had to substantially liberalize their trade regimes to reach the EC's common external tariff.
Looking to the future, trading blocs are likely to form around countries with liberal trade regimes. So, a common external tariff can serve the cause of liberalization in countries with inward-looking regimes that want to join a bloc. A common external tariff can also generate some important side benefits. For one thing, it eliminates the need for rules of origin, which often become instruments of protection in the hands of a weak authority. Moreover, to the extent that lobbies are active, their effectiveness may be diluted in a customs union. A union requires unionwide lobbying, while tariffs in an FTA are responsive to lobbying at the national level.
Second, Article VI on antidump- ing and Article XIX on voluntary export restraints (VERs) must be reformed. The liberalization of tariffs by the GATT over the years has been accompanied by the erection of alternative trade barriers in the form of antidumping actions and VERs. The persistence of these barriers can make regionalism a vehicle for extending national protection to regions. A customs union will intensify competition among partners, which can, in turn, increase the pressure for anti- dumping actions and VERs against countries outside the union. To reduce this possibility, Articles VI and XIX must be strengthened.
Third, countries should be encouraged to adopt liberal rules of entry when designing regional arrangements. In essence, this amounts to a conditional MFN, where an existing regional arrangement is open to any nation willing to abide by its rules and responsibilities. This is the type of suggestion also made for liberalizing trade in services---and it was tried to some extent in designing the government procurement code in the Tokyo Round. More to the point, the conditional MFN was behind the liberal trade regime that emerged through the bilateral route in the second half of the nineteenth century.
Lest this discussion make the problem of turning trading blocs into an instrument of multilateralism appear too simple, it is important to add a word of caution. Mike Finger reminds us emphatically, and painfully, that rule-writing in the GATT game is less than half the battle. It is the enforcement of rules that is most difficult. Repeatedly, regional arrangements have violated the articles of the GATT: to date, the GATT has formally declared only four arrangements to be compatible with its articles. Yet, it has never censured a single agreement as being incompatible with its standards.
Despite this distressing record on implementation, there is something to be said in favor of why GATT rules on regional arrangements should nevertheless be strengthened. Robert Hudec calls attention to the important fact that the GATT rules have had a definite impact on the discussions that preceded the conclusion of various arrangements. So, even though the GATT failed to censure any GATT-incompatible agreements, the presence of rules had an unmistakable influence on regional arrangements concluded.
Quite apart from how regional arrangements affect the world trading system is the issue of what individual developing countries can hope to achieve from them. As with the multilateral dimension of regionalism, the developing country dimension today is different from that in the 1960s.
Two differences are fundamental. First, the initial conditions are different from those prevailing three decades ago. Trade reforms of the 1980s have led to the creation of a liberal trading environment in a large number of Latin American countries and an increasing number of African countries. The additional gains from FTAs through trade creation are thus negligible. So, the attraction of regionalism must lie outside the trade creation effects.
Second, in the previous round, developing countries attempted integration almost exclusively with other developing countries. In the current round, FTAs are also being proposed between developing and developed countries. This is clearly a new dimension.
The experience with South-South integration has been discouraging, and any temptation to promote such schemes in the future should be resisted. Max Corden reminds us: "The plain fact is that such free trade areas (or preferential areas) would not make a great deal of difference, and have not in the past... It is far better for Argentina to go for the world market---i.e., to liberalize unilaterally and in a non-discriminatory fashion, as she has been doing---than just to go for the Brazilian market. Brazil has the largest economy in the Third World, and yet it is smaller than Canada's (as measured by the dollar value of GDP). And this applies even more to Brazil."
As for North-North integration, despite the reservations of Alan Winters, it has been widely successful in Europe. Intraregional trade expanded greatly, but not at the expense of trade with nonpartners, which also grew rapidly. And European integration has greatly diminished, if not eliminated, the possibility of future internal conflicts.
Looking to the future, North-South integration holds much promise for developing countries. Regional arrangements of this type can solidify past reforms, guarantee future access to a large market, and stimulate growth via increased direct foreign investment, more intense competition, and faster technological diffusion. On the Mexico-U.S. FTA, Richard Cooper writes: "As I understand it, Mexico has two principal reasons for wanting a free trade area. The first is to foster and to lock in through international commitment a set of liberal trade policies... The second is to lock the United States into a set of procedures that are more predictable for Mexican exporters with respect to the whole range of 'safeguards'---antidumping and anti-subsidy procedures in particular."
The implications of regionalism for the global trading system nevertheless remain contentious. Rudi Dornbusch argues that regional and multilateral approaches are complementary. He applauds GATT: "GATT is wonderful---has been, is, and will be." To those who declare that GATT is dead, his response is: "GATT is not dying---it is resting." While GATT rests, he feels we need the regional approach to complement it. Bhagwati is skeptical: "We are only walking on two legs is the popular argument. That we may wind up walking on all four is ignored."
In the end, the ambiguity of the economics of regionalism and trading blocs divides economists on this issue. The dilemma in the economists' thinking on the issue is captured well by Vinod Thomas: "One can make an argument, depending on initial conditions, either that: regional blocs may be a bad idea in principle, but good in practice, something Professor Krugman seems to argue. Or that even if they are a good idea in principle, they are likely to be bad in practice, as Professor Bhagwati seems to conclude."
Drawn from Jaime de Melo and Arvind Panagariya, The New Regionalism in Trade Policy, World Bank (forthcoming in September).