An Analysis of Regional Trading Blocs

Transcription

1 Chulalongkorn Journal of Economics 8(3), September 1996 : An Analysis of Regional Trading Blocs Teerana Bhongmakapat Faculty of Economics Chulalongkom University C P t I 1. Introduction About a decade before ending the millennium, the world economy has witnessed a few major changes that have led to serious debates among many trade economists. Now that the Cold War has ended, thanks to Mr. Gorbachev, one might reasonably take an optimistic view about the future prospects of the world economy. However, international economic rivalry has intensified, and as the Uruguay round has been launched since the mid-1980s, the role of regional economic integration has emerged as a substitute for (rather than a complement to) global trade negotiations under the GATT. Together with bilateralism adopted by the U.S. congress and the executive branch under the Reagan and Bush administrations, regionalism has been seen by many to rise and undermine GATT-wide free trade principles. Completion of the EC's internal market has progressed very rapidly. The result has been considerable worries among many external countries, particularly the U.S. and Western Pacific Rim countries, that the EC may turn into an economic "fortress" and become a closed market. In North America, the U.S. (the GATT's The author is grateful to the Research Institute of International Trade and Industry, MITI. Japan, for research suppon in 1991.

2 404 Chulalongkorn Journal of Economics 8(3), September 1996 founder and long-time strong supporter) has entered into a free trade agreement with Canada and will do so very soon with Mexico. This is partly a strategic response to the EC integration. In East Asia, now the most dynamic region in the world, some nations have called for an economic grouping (though many of them are opposed to bilateralism and regionalism); in fact, ASEAN countries have just agreed to gradually form a free trade area, largely in response to the strong tendency towards rising regionalism and protectionist pressures in Europe and North America. Recent developments in the world trading system indicate "successful" regional trading blocs and the likelihood that a GATT-sponsored multilateral mechanism will not be supported by the two major players as strongly as it was in the past. There is a wide belief that, if these trends continue into the future, the world economy will be fragmented into three major regional trading blocs : (1) the EC with further enlargements to include EFTA and Eastern Europe, (2) the NAFTA, which will probably include some Latin American countries, and (3) East Asia which will be urgently formed. In 1990, the trade between countries belonging to the EC and the NAFTA already accounted for almost 60 percent of world trade (see Table 1 for detailed comparisons). Provided that East Asian countries form trading bloc (which would possibly include Australia and New Zealand who already have made a free trade agreement) the three major regional trading blocs could have their trade shares amounting to approximately 90 percent of world trade. Nonetheless, whether or not the future world trading system will be overwhelmingly dominated by the trading blocs remains an open question, although it may be partially answered by the "success" or "failure" of the Uruguay Round, the conclusion of which will be postponed until early 1993.

3 T. Bhongmakapat : Regional Trading Blocs 405 Table 1 Regional Trading Blocs (1990) Population GDP GDP per Total Trade (million) (bil. US$) capita Exports Imports Trade Share (US$) (bil.$) (bil.$) (bil.$) (%) NAFTA I* NAFTA It** USA Canada Mexico EC (12) W.Gennany EFTA East Asia (excl. China) Japan South Korea Hong Kong Taiwan China ASEAN (6) Singapore Thailand Malaysia Indonesia Philippines Brunei Australia & New Zealand Australia New Zealand World Note : Source * NAFTA I is comprised of the U.S. and Canada. ** NAFTA I1 is comprised of the U.S., Canada, and Mexico. : IMF, International Financial Statistics Yearbook, 1991, and various other sources.

4 406 Chulalongkorn Journal of Economics 8(3), September 1996 Although the formation of any regional economic zone involves political or noneconomic objectives, it is crucial to examine its economic essence, including its role in the world trading system. In this regard, many prominent trade economists, notably Baldwin (1987a and 1990), Bhagwati (1991), Corden (1987), and Krueger (1990) view multilateral free trade as preferable to bilateral and regional approaches, and are alarmed that bilateralism and regionalism may divide the world into three large trade blocs, posing serious dangers to world trade and security (see also Aho and Aronson (1985), Noland (1989b), and Schott (1989). On the contrary, many influential scholars and policy analysts, such as Dornbusch and Krugman, have started to advocate bilateral and regional initiatives (see also Hufbauer (1984), for an earlier view). Dornbusch (1990) praises a U.S. trade bloc, including the U.S.-Mexico deal and free trade arrangements with other Latin American nations, and strongly argues in favor of bilateralism and regionalism. According to Dornbusch, these can be effective bargaining tools for the U.S., and strategic move to handle Japan. More recently, Krugman (1991a and 1991b), whose name is closely connected with the new stream of trade theory, states the importance of the "natural (regional) trading blocs" concept over conventional economic explanations, and prefers that the U.S. forms a free trade zone with neighboring countries, i.e. Canada and Mexico, rather than ones with foreign lands further away, such as Israel and South Korea. Is a geographically discriminatory trading arrangement a good idea? Will its role activate and replace multilateral free trade? How will it affect the future of the global trade system and the world economy? If regional-biased practices are prevalent, how would small countries that tend to be "left on the shelf' respond to such a global movement? The purpose of this paper is to consider these issues and particularly to closely examine the relevance of the natural trading blocs argued for recently and influentially by Krugman (1 991 b).

5 T. Bhongmakapat : Regional Trading Blocs 407 In this paper, we first consider the theory of preferential trading arrangements, sometimes called customs union theory, in order to emphasize and clarify certain economic issues. Since this theory is a large subject and has also been reviewed elsewhere, discussion here is only an attempt at laying down some of the relevant findings and recent theoretical developments in the field. In section 3, we examine the hypothesis of "natural trading blocs" by looking at the resulting trade creation and possible efficiency losses. In this section, an econometric analysis of U.S. bilateral trade flows is also presented. Section 4 discusses the role of regional trading blocs in shaping the future of the post-cold War world economy. Section 4 also glances at possible reactions taken by small countries in East-Southeast Asia or Western Pacific Rim, as well as strategies that these nations may adopt in responding to the geographically discriminatory practices that have emerged in Europe and North America. 2. Trading Blocs in International Trade Theory Theoretical literature on customs unions, or more precisely on preferential trading arrangements, is crucial to understand the gains and losses that may be associated with the formation of trading blocs. ' Forming a free trade area involves substantial reductions in tariffs and nontariff impediments among member nations, therefore one may be led to believe that such a formation is a move fowards 1 free trade that helps raise welfare rather than reducing it. However, trade theory in general asserts that its impact is indeed ambiguous, since it involves both free trade among members, and protectionist 1 discrimination against outsiders. This is fundamentally the question of second-best policy, with one distortion replaced by another, thereby making the question very difficult to answer with certainty. I -- ' The terms, "free trade area," "customs union," "regional economic integration," "preferential trading arrangements," and "trading blocs," are here used interchangeably, although they may not be defined exactly the same.

6 408 Chulalongkorn Journal of Ecotlomics 8(3), September 1996 Although trade theory contains some deficiencies in explaining the existence and effects of preferential trade in which various distortions exist, forming a trading bloc is not a purely political phenomenon. Of course, a good answer should have both empirical and theoretical foundations. In this section then, it is worthwhile to see some major points which trade theorists have asserted on how a bloc-building policy affects trade and welfare. Standard Customs Union Theory One of the early contributions came from Viner's 1950 book. His seminal contributions were the concepts of "trade creation" and "trade diversion" that help identify conditions under which a customs union is welfare improving. Trade creation is the replacement of high cost domestic production by imports from a partner, whereas trade diversion is the replacement of low cost imports from outsiders by imports from a partner. Therefore, for the donor country who gives a preferential tariff reduction, trade creation is beneficial in moving the country toward free trade, but trade diversion is protective and harmful since it results in resource misallocation. Without empirical estimates of these two conflicting effects, the desirability of a customs union is indeterminate. Since these two "Vinerian effects" are very fundamental to analyses of trading blocs, it is probably useful to precisely illustrate them in a more detailed manner. Suppose that we are in an idealized world where we have perfect competition, full employment of resources, no adjustment costs, and perfect mobility of factors within countries, but not across countries. By assumption, there is one good and there are three countries: A, B, and C. Countries A and B are to form a trading bloc, leaving C (or the rest of the world) as an outsider. In Figure 1, D, is A's demand curve, while S,, S,, and S, are supply curves for A, B, and C, respectively. C's supply curve is assumed to be

7 T. Bhongmakapat : Regional Trading Blocs 409 perfectly elastic, so the world price is a constant, P,. Initially before the trading bloc formation, A imposes a non-descriminatory tariff, t, leading to domestic production and imports of OQ, and Q,Q2, respectively. It is simple to see that A pays Q,FGQ2 for imports from C, while domestic consumers pay QIABQ2 for the same amount of imports, giving tariff revenue equal to FABG for the government. With the bloc, imports from B will not be subject to tariff and thus will increase to Q,Q,. Domestic production will decrease to OQ,, while domestic consumption will increase to OQ,. As a result, net gain due to an increase in consumer surplus and a decrease in producer surplus will equal CAI+JBD. Note that, since there will be no tariff imposed on B's imports, tariff revenue equal to IABJ will transferred back to consumers. However, A will have to pay QlIJQ2 for the original amount of imports that used to cost only Q,FGQ2, leading to a loss in government revenue by FIJG. Trade economists show trade creation as Q3Q,+Q2Q4, or in terms of welfare, the areas CAI+JBD, and trade diversion as Q,Q2 or the welfare represented by area FIJG.

8 410 Chulalongkorn Journal of Economics 8(3), September 1996 The above analysis simply shows that, although removing intra-bloc trade barriers improves economic efficiency through trade creation, barriers that remain on extra-bloc trade will lead to an efficiency loss through trade diversion. The latter concept has led many to believe that a unilateral tariff reduction is superior to a customs union. Indeed, Cooper and Massell (1965a) asserted that this was true within the framework represented in Figure 1. Under the small country assumption that international terms of trade are unchanged due to the customs union formation, Cooper and Massell (1965a) showed that tariff reductions indiscriminate from P, to P,, and the subsequent union formation between A and B would yield gains from trade creation (CAI+JBD) without resulting in trade diversion, for imports continue to come from the rest of the world. This led to the proposition made by Johnson (1965) and Cooper and Massell (1965b) that a customs union is not theoretically efficient; therefore, the formation of customs unions and other preferential trading arrangements is driven by non-economic objectives. Parallel to the traditional line of partial equilibrium analysis pioneered by Viner, an alternative theoretical approach was introduced by Vanek (1965), and further expounded and developed by Kemp (1969) and Kemp and Wan (1978). Vanek, in his great book, explored various possibilities within the powerful framework of general equilibrium. He accepted the statement that a customs union may or may not raise welfare, but specified the conditions under which it will be Pareto-superior to the initial situation. The formation of a customs union may make the country granting preferential tariffs worse off and make its partner better off; however, the loser can be compensated, as can the rest of the world which loses from being left out. In compensating the rest of the world, Vanek introduced

9 T. Bhongmakapat : Regional Trading Blocs 411 an important concept: that of the compensating common tariff The compensating common tariff refers to the common external tariff (CET) imposed jointly by customs union members that keep the rest of the world as well off as before. An interpretation of this concept is that the union sets its external tariff such that its volume of trade with the rest of the world remains unchanged. Although the Vanek-Kemp-Wan approach justifies the existence and motives of customs union, under the compensation scheme the union does not necessarily maximize welfare with respect to its joint external tariff; therefore, the compensating common tariff may not be as attractive as unilateral nonpreferential tariff reduction. By using this approach, Berglas (1979) came up with a conclusion which supported that of the Johnson-Cooper- Massell proposition. That is, if world prices are constant for the recipient member who compensates the granting country who loses, the net gain accrued to it can also be achieved through unilateral tariff reduction (while the granting country remains as well off as before the union). Later Contributions In spite of the proposition in favor of unilateral free trade over customs unions, there exist certain conditions under which the latter is superior to the former. The case for unilateral tariff liberalization stated by Johnson (1965) and Cooper and Massell (1965a) is the first-best outcome where there is essentially no change in the terms of trade, no foreign tariff, no economies of scale, etc. In the previous analysis, A and B are assumed small, whereas C is a large country, which means world prices remain unchanged after the union. If the countries that form a union are significant enough in size to have an influence on international terms of trade, there will normally be an additional source of gains from the

10 412 Chulalongkorn Journal of Economics 8(3), September 1996 formatioc; that is, gain due to changes in the terms of trade. In Figure 1, if we allow for increasing costs so that A imports from both B and C, preferential tariff reduction given by A to B will raise B's producer prices, but reduce prices received by C. Changing terms of trade in favor of the bloc and against the rest of the world thus leads to a welfare gain and thereby indicates a trade-off between trade distortion and its offsetting terms-of-trade gain.2 On theoretical grounds, this result is not necessarily valid in all cases. Forming a customs union is usually reciprocal, i.e. both A and B offer preferential tariffs to each other, and the effects on terms of trade between A and B may not be canceled out. If there are more than two goods, and both goods from A and B are imported by C, the union's terms of trade versus C can finally go into any direction, depending on the resulting intra-union terms of trade between A and B. In normal cases, however, outsiders will face worsening terms of trade from trade diversion, while the bloc will improve terms of trade as a result of it. One of the main and obvious reasons why a unilateral tariff reduction is not very attractive to policy makers is that it'is not accompanied by changes in tariffs (or trade barriers) abroad. In the presence of foreign tariffs, there exists an opportunity of mutually provide tariff preferences, as analyzed in Wonnacott and Wonnacott (1981). Multilateral free trade can be seen as explicit cooperation that dominates the option that each country reduces its own tariffs unilaterally. Under some circumstances, reciprocity arranged with certain countries may appear to a given country to be more feasible than reciprocity arranged with an Since under general equilibrium conditions, A's terms of trade is the sum of B's and C's terms-of-trade effects, the terms of trade between A and C can not be predicted theoretically. Nonetheless, as stated by Mundell (1964) in his pioneering work, B's terms of trade unambiguously improve as a result of discriminatory tariff reductions offered by A.

11 T. Bhongmakapa! : Regional Trading Blocs 413 unlimited number of nations. However, the rampant discrimination involved in forming a bloc which is narrowly-based or not open to outsiders is not truly supported by this argument since it ignores possible benefits from mutual cooperation with many other countries. And of course, there is the theoretical question of the optimal choice of partners. In the presence of scale economies, customs union formation was suggested by Corden (1972) to include two additional effects : cost reduction gain and trade suppression loss. With economies of scale for production in A, its average cost curve is sloped downward; therefore, as a result of average cost pricing and preferred tariffs provided by its partner country, B for example, export expansion into B will lead to a cost saving along the average cost curve, thereby yielding the cost reduction effect. A loss from trade suppression for A may occur when the union formation allows production in A to emerge and replace imports into A from the rest of the world, C. If A's production also replaces imports into B from C, there will be trade diversion for B. i Although there have been several theories addressing the question of why a trading bloc or customs union has incentives to from, terms-of-trade changes that could benefit member countries at the expense of outsiders may lead to the undesirable outcome of trade warfare and retaliation. The terms-of-trade gain from the grouping will encourage others to do the same thing. As the analysis of tariff retaliation equilibrium pioneered by Johnson (1954) implies, if only one bloc is large and the others are too small, the large bloc may finally have a net gain, while the others lose. However, if the blocs are large, retaliation may finally leave them worse off than before starting the bloc-building policies. Krugman (1991b) also explains that, in a theoretical sense, since trading blocs can, due to improving terms of trade, raise market power and

12 414 Chulalongkorn Journal of Economics 8(3), September 1996 thereby tariffs on imports from outside, they may lead to tariflwarfare that quite possibly could leave all of them worse off than they would have been had the trading blocs not been formed. Trading Blocs with Imperfect Competition In the preceding discussion, trading blocs are analyzed by the "conventional" theory of international trade, in which perfect competition is a crucial assumption. Discussion concerning Corden's trade reduction and trade suppression effects is the only exception examined thus far that departs from the traditional perfect competition setup. With a growing belief that imperfectly competitive behavior is of great importance in world trade, much of trade literature has recently been developed by incorporating the theory of imperfect competition into the theory of trade and tariffs. Nevertheless, regarding the topic of trading blocs, very little work has been done. Ethier and Horn (1984) and Krugman (1991a) are probably the only two major theoretical studies that attempt to develop a formal model within the general equilibrium framework of Charnberlinian monopolistic competition. To understand the effects of trade bloc formation under imperfect competition, it is important to look at the "large group" Chamberlinian case and the "small group" oligopolistic competition case separately. In the former case, forming a bloc will have an additional impact on product varieties. Suppose that all three countries, A, B and C, produce manufactured goods that are differentiated and imperfectly substitute, and there are n, m, and m* varieties produced in countries A, B, and C, respectively. Assume that for each country the preference of its representative consumer is that of the Dixit-Stiglitz variant. See Dixit and Stiglitz (1977). This preference is affected by not only

13 T. Bhongmakapat : Regional Trading Blocs 415 I the amounts of varieties consumed, but also the numbers of varieties: n, m, and m*. Theoretically, trade liberalization adopted by A or B can have an ambiguous impact on the total number of varieties, the elasticity of substitution among varieties, and product prices, and thus on elfa are.^ To make things less complicated, let's assume that the substitution elasticity is constant. Now, if A anb B form a trading bloc, leaving C out in the cold, reducing trade barriers by A will force n to fall and m to rise. Extrabloc varieties will be discouraged, so m* will fall. The replacement of n by m illustrates a trade creating effect, and the replacement of m* by m is an example of trade diversion, or more exactly, trade modification. For A, discriminatory tariff reduction may or may not be welfare improving, depending upon the relative gain due to more varieties from B and losses due to less varieties from inside the country and from outside the bloc. If A and B offer reciprocal tariff reductions and both of them are rather symmetric, n and m will increase, and m* will fall in both countries. Moreover, in normal cases, terms of trade will also improve for the bloc and worsen for the rest of the world. As a result, it is quite similar to the concepts introduced by standard theory, forming a trading bloc normally leads to a welfare gain due to trade creating and terms-of-trade effects, and a welfare loss due to trade diversion or trade modification. Countries outside the bloc will unambiguously be worse off. Since in monopolistically competitive industries, firms act as minimonopolist over their varieties, the effects of scale economies or cost reduction effects will be theoretically trivial, unless the substitution elasticity is 3 Changes in varieties play an important role here. As shown in my dissertation (Bhongmakapat, 1988). unilateral tariff reduction will have an ambiguous effect on the total number of varieties; therefore, if the substitution elasticity is not constant, but affected by the total number of varieties, the ambiguous changes in varieties will cause the substitution elasticity, prices, and welfare to go in any direction.

14 416 Chulalongkorn Journal of Economics 8(3), September 1996 allowed to move downward sufficiently as a result of bloc f~rmation.~ A major concern about forming a trading bloc is its impact on world welfare. By assuming symmetric trading blocs, Krugman (1991a) shows that the formation of trading blocs will not only result in trade diversion, but also in an increase in external tariffs due to terms-of-trade changes in favor of the bloc. According to this analysis, without natural trading blocs, world welfare is maximized with free trade and minimized with "three trading blocs." In the case of oligopolistic competition, the effects of trading bloc formation are highly sensitive to what assumptions are made regarding the interaction between firms. This is actually an intrinsic problem of game theory, any correct answer has to rely on empirical work. Due to its partial equilibrium character, this approach has not had much to say about the economy-wide benefits and costs of customs unions, which is an important consideration for economists. Among a few others, Smith and Venables (1988) and Venables (1990) attempted to develop a formal model of intra-industry trade (but not within a general equilibrium setup) to analyze and quantify the effects of a trading bloc under increasing returns to scale and various types of oligopolistic competition. Their economic analysis can be briefly described as follows. Forming a customs union leads to more competition between firms located in different countries, and may influence the form of the competitive game that firms play. More competition should help cut oligopolistic prices, thus benefiting consumers in the union. Another type of gain from competition is the resulting reduction in the number of firms in the industry, causing firms that Theoretically, pricing behavior that is derived from the optimization problem is usually preferred. However, this can be a matter of empirical evidence. If prices set by domestic firms are affected enough by the prices of imported varieties, as in Harris (1984). the effects of scale economies will be significant.

15 T. Bhongmakapar : Regional Trading Blocs 417 I continue to stay in the industry to become larger and further?, rationalize or extend the lengths of production runs to exploit more 1 economies of scale. If markets are fully integrated, as in the single i market model, a lower degree of market segmentation will reduce the domestic firms' market power as their ability to price discriminate ) between markets goes down. The changing form of the game being i played is supposed to be pro-competitive at home. In this type of model, 1 the effects from economies of scale and terms of trade will benefit t the bloc, but as in Corden's analysis, there will be trade diversion, 4 5 trade suppression, and worsening of outsiders' terms of trade. '! 3 3. The Irrelevance of Natural lkading Blocs As trade theory generally suggests, a trading bloc is a distortion i comprised of both free trade and protectionist components. After ; the bloc formation, member countries may or may not experience 1) a net gain. This depends on the relative magnitudes of trade I creation, terms-of-trade, and the effects of scale economies, versus trade diverting and trade modification effects. Outsiders are normally worse off due to trade diversion, trade modification, and terms-oftrade changes. If the protectionist component is significant and encourages trade warfare or retaliation, world welfare will ultimately diminish and every one will be worse off. I r Since theoretical predictions suffer from the second-best nature of the problem being analyzed, any good answer cannot be overly generalized and empirical evidence should be seriously considered. A number of studies have assessed the effects of discriminatory trading arrangements on trade and welfare, mostly in the case of European integration. According to most of these studies, trade creation and terms-of-trade benefits would outweigh trade diversion losses. Under imperfectly competitive markets, the net welfare effects of EC integration and the U.S.-Canada Free Trade Agreement for member countries are positive and larger than under

16 418 Chulalongkorn Journal of Economics 8(3), September 1996 perfect competition, see Smith and Venables (1998) and Venables (1991) for European integration, and Brown and Stem (1989) for the U.S.-Canada pact.5 These empirical results might be embraced by European federalists and NAlTA advocates, but in fact they do not indicate impressively large gains. Furthermore, in general, as illustrated by Richardson (1989), empirical research on trade policy under imperfect competition has shown that trade liberalization leads to gains much larger than those estimated under perfect competition. This implies that under imperfect competition with increasing returns, multilateral free trade may be much more beneficial than customs unions that result in protectionist trade diversion and trade modification effects. What is clearly suggested by empirical evidence currently available is that trade bloc formation could make outsiders significantly worse off. Sampson and Snape (1980) estimated substantial welfare diverted from nonmembers to EC members. Hamilton and Whalley (1985) simulated general equilibrium welfare effects of various free trade areas involving the U.S., and found that most excluded regions generally would experience sizeable.welfare losses. Considering that the welfare gains from a trading bloc, either the EC or NAFTA, are not very large, but lead to adverse effects on outsiders and the world at large, one is reasonably convinced that regionalism is probably a bad idea, particularly from long-term and global perspectives. Nonetheless, Krugman (1991b) views recent popular movements towards free trade areas as being With increasing returns and oligopolistic competition, one often accepts the idea that government interventionist measures, such as tariffs and subsidies, are beneficial to "national interests" to a greater extent than what is actually suggested by standard trade theory, see an influential paper of Krugman (1987). However, as indicated by Venables (1991), removing intra-ec trade barriers would yield relatively larger welfare gains with oligopolistic competition than with perfect competition,

17 T Bhongmakapar : Regional Trading Blocs 419 "natural" since the countries that are engaging in free trade agreements are actually natural trading partner^.^ According to Krugman, if the formation of preferential trading agreements follows the line of natural regionalism, they will have "a much better chance of improving welfare than trade arrangements between "unnatural" trading partners." (Krugman, 1991 a : 122),' In supporting the hypothesis of natural trading blocs, Krugman (1991b) estimated a gravity equation with data from G7 countries.' He regressed the value of trade, in logarithm, on the products of the incomes between countries, in logarithms and dummy variables for countries that belong to trade groupings, i.e. the U.S.-Canada, and Europe (4 European countries), and found that the estimated coefficients of the dummies for U.S.-Canada and Europe were statistically significant and equal to and , respectively. Therefore, as interpreted by Kmgman, the regression result indicated "very strongly" the regional bias of trade (Krugman, 1991b : 20)8. Of course, Krugman's analysis advocating natural regionalism is subject to criticism. One of the few major drawbacks r An earlier idea of natural trading blocs was given by Balassa (1961). who conjectured that neighboring countries would have an additional stimulus to trade, but relying on reasons quite different from Kmgman's. Whereas Balassa's idea was based on similarity of tastes and an awareness of common interests, Krugman's explanation aims at economic geography, particularly the role of transportation costs. ' The gravity model relates the value of trade between two countries positively to their incomes using a functional form similar to the gravity equation in physics. Applying it to econometric analyses of international trade flows has been very successful. It has also been standard for estimating trade effects of regional integration. For example. see Aitken (1973). Pelzman (1977), and Brada and Mendez (1985). Based on the above-mentioned coefficients for the dummy variables, the U.S. and Canada did 13 times as much trade as they would if they were not neighbors, while the European countries did 7 times as much.

18 420 Chulalongkorn Journal of Economics 8(3), September 1996 is his regression formulation and interpretation. Since his regression tactic was to "pool" trade flows between the seven countries and include only two dummy variables for the U.S.-Canada, and Europe into the equation, the coefficients he obtained simply indicate that the U.S. bloc and the "EC" would be creating much more trade than an "average pair" of G7 countries would do, and that the U.S. bloc would stimulate much more trade than would Europe. By my interpretation, it simply means that Kmgman has not done enough to show that forming a bloc among neighboring nations will be much more trade creating than forming a bloc between distant countries. Maybe, in terms of trade creation, the U.S. trade bloc would gain more by including economically efficient trading partners located far away than by including those which have a sound geographical advantage, i.e. Canada and Mexico. Therefore, it is very important to test the hypothesis of natural trading blocs by examining the trade-stimulating effects that could possibly occur with some other trading partners as well. In considering the relevance or irrelevance of natural regionalism here, my regression tactic is to employ a crosssectional gravity model of bilateral trade between the U.S. and her trading pa~tners.~ It is then asked whether for the U.S., (i) trading arrangements within NAFTA are more trade stimulating than with any other regions, and (ii) in selecting a country to form a bloc, Canada and Mexico are superior in terms of trade creation with In this paper, the same approach as used in Krugman (1991b) is employed. Although gravity models are quite standard in the empirical analysis of trade policy and regional economic integration, it is by no means an ideal approach in analyzing the economic impacts of trading bloc formation, particularly when the economy-wide effects are concerned. However, considering the questions at hand, this approach should be sufficient.

19 T. Bhongmakapat : Regional Trading Blocs 421 the U.S.1 The model here specifies the value of trade between the U.S. and 84 trading partners as a positive function of the incomes of the trading partners, distance between economic centers, and dummy variables for trade preference areas. The functional form being estimated is as follows: where Tu, is the value of merchandise trade between the U.S. and country j, GDPj is the nominal GDP of country j, measured in millions of US dollars, YN, is the GNP per capita of country j, measured in US dollars, DIST~, denotes the shortest navigable distance between the commercial centers in the U.S. and country j, measured in nautical miles, and DAZ represents the dummy variable for trade between the U.S. and the region or area z (it can be a country).'' As normally specified in gravity models, the GDPj coefficient should have a positive expected sign. NYj measures demand structure or tastes, and is included according to Linder's hypothesis, as suggested by Bergstrand (1989) and Thursby and Thursby (1987). Its expected sign is positive. DISTu, serves as a lo Although in a number of gravity models, data on bilateral trade flows across more than one pair of countries are used and pooled together, it is not uncommon to employ a bilateral trade flow model for a single country, see Thursby and Thursby (1987) and Summary (1989). From an econometric standpoint, Thursby and Thursby (1987) pointed out that a bilateral trade model for each country had an advantage over a bilateral trade model which pooled data from all countries. However, a bilateral trade model for the US is performed here essentially due to the nature of the questions considered, not merely due to statistical criteria. 'I All data used are those for 1989, except for navigable distances. Trade flows data are obtained from the IMF, Directions of Trade Statistics Yearbook, Data on GDP and per capita income are from the World Bank, WorU Development Report, 1991, except for a few countries, data which are from the IMF, Internarional Financial Sratisrics Yearbook, Distances are from the Japan Shipping Exchange, Disrance Tables for World Shipping, Tokyo, 1983.

20 422 Chulalongkortz Journal of Economics 8(3), September 1996 proxy for natural trade resistance which in turn refers to a composite of transportation cost and transport time, and thus its expected sign is negative. The dummy variable, DAz, is actually used to approximate the trade-stimulating effect of trading arrangements between the U.S. and the region (or country) z. Therefore, for any z, its dummy coefficient helps estimate the trade-stimulating effect, in a very approximate fashion of course. A high estimated coefficient value for any group z implies its strong trade creation with the U.S., either with or without existing special trade preferences. It is common for a cross-section analysis using the ordinary least square method to have the heteroscedasticity problem since, in the presence of heteroscedasticity, the variances of the OLS estimators are biased. Therefore, the model is estimated by using the heteroscedasticity-consistent covariance-matrix estimator of White (1980). Table 2 shows the estimated results based on two gravity equations, one with regional dummy variables and the other with country dummies as independent variables with regard to trade stimulation. The former and latter equations are respectively related to questions (i) and (ii) mentioned above. In the first regression, North America is considered versus other major regions, where DNAFTA is the dummy variable for U.S. bilateral trade with the rest of North America (including both Canda and Mexico), DANIE with Asian NIEs (including Taiwan, Hong Kong, and Korea), DSEA with Southeast Asia (including Singapore, Thailand, Malaysia, Indonesia, and the Philippines), DEC with European Community, and DLATIN with Latin America (excluding Mexico). In the second regression, the trade-stimulating effects of trading with Canada and Mexico are compared with other prospective countries, especially those highly competitive in world markets, such as Asian NIEs. DCND denotes the dummy veriable for U.S.

21 T. Bhongmakapar : Regional Trading Blocs 423 bilateral trade with Canada, DMEX with Mexico, DTWAN with Taiwan, DKREA with Korea, DHK with Hong Kong, DSPORE with Singapore, DJPAN with Japan, and DISREL with Israel. Table 2 Regression Results Right-Hand-Side Estimated Variables Coefficients t-statistic (I) Estimated Gravity Equation with Regional Dummies In GDP In YN In DIST DNAFTA DANIE DSEA DEC DLATIN Adjusted R2 = (2) Estimated Gravity Equation with Country Dummies In GDP In YN In DIST DCND DMEX DTWAN DKREA DHK DSPORE D JPAN DISREL Adjusted R2 =

22 424 Chulalongkorn Journal of Economics 8(3), September 1996 As shown in Table 2, the regression results of gravity equation (1) support the idea that U.S. bilateral trade within NAFTA is highly trade creating, but do not indicate that natural trading arrangements do strengthen trade significantly more than trade with "unnatural," remote regions. Although intra-nafta trade has already been promoted by various trade preferences made by all the members, according to the results, U.S. trade within NAFTA is not more trade creating than is bilateral trade between the U.S. and Asian NIEs, and only slightly more trade creating than is U.S. trade with Southeast Asia.12 Of course, it is not surprising that it does stimulate trade relative to U.S. trade with the Latin American region, while the trade creating effect between the U.S. and the EC is unfounded. Furthermore, the importance of natural trading arrangements is not supported by the estimated results in equation (2). The results more or less help indicate trade creating gains from choosing one trading partner over another. As discussed by various authors, all countries considered here are possible candidates or in possible regions for a free trade agreement with the U.S., see Schott (1989). According to the regression, the trade preference effects of NAFTA are significant. U.S. trade with Canada is trade stimulating relative to those with Mexico and some other nations, particularly Japan and Israel. However, trade arrangements with Canada only result in slightly more trade than those with Hong Kong and Taiwan, and less trade than those with Singapore. More importantly, Mexico is quite different from Canada. The U.S.-Mexico grouping is not as beneficial, in terms of trade stimulus, as a US.-"Tiger" grouping, except for Korea, though it is superior to arrangements with Japan and Israel. This is tantamount to saying that building a " During the past decade, the U.S. trade policy has turned to aggressively reduce imports from East Asia, as pointed out by Bhagwati and Patrick (1990). Otherwise, one may see stronger trade creation with Asian NIEs and Southeast Asia.

23 T. Bhongmakapat : Regional Trading Blocs 425 trade bloc with a country or region based on the concept of a natural or geographic alliance is probably not a good idea. Natural regionalism suffers from its lack of solid economic reasoning, although it has political and cultural advantages to sell to short-sighted politicians, special interests, and the public who may not be aware of the long-term, economy-wide costs of the inward protectionist elements embodied in such discriminatory trading arrangements. If we examine trade beyond its aggregate flows and highlight conventional trade economists' view in more detail, we will clearly see the opportunity costs of creating a trading bloc, and that the best trading bloc is by and large multilateral free trade under which a country can broadly acquire potential efficiency gains from specialization. Tables 3-8 show the "revealed" comparative advantage, RCA, of exporters to the U.S., the EC, and Japan, as well as that of the three major trading nations of the world.13 For the U.S., gains from trade can simply occur by doing more trade with countries which have high RCA values for commodities groups in which the U.S. lacks comparative advantage. As shown in Tables 3 and 6, the U.S. has strong revealed comparative advantages in mining or crude materials (except fuels) and machinery and equipment, and the U.S. can find various trading partners that are, in the revealed sense, more efficient than Canada and Mixico in commodity groups in which the U.S. has strong comparative disadvantage, especially energy and non-mechanical manufacturing. If the machinery and equipment group is producing under increasing returns to scale, it " Revealed" comparative advantage is a crude indicator for comparative advantage based on patterns of trade. and it is usually defined as the ratio of the share of a commodity group in total exports for a country or region to that commodity group's share of world exports. However, in Tables 3-5, the indicator is applied to exports to the three markets to allow for market segmentation in world markets. A value greater than 1 indicates a relative comparative advantage for the commodity group.

24 ,426 Chulalongkorn Journal of Economics 8(3), September 1996 will imply that the U.S. can find trading partners with a strong advantage in exporting the product group to the U.S. By exporting more of the commodity groups in which the U.S. had a strong comparative advantage to various countries that experience exceptionally low comparative advantage, the U.S. will clearly gain. In fact, they will gain much more than if they entered a narrowly conceived restrictive trade bloc with a few countries that do not offer economic gains widely and substantially. Multilateral free trade is simply a route which allows every one, who is an importing country for varieties of products, to benefit from trade with least (comparative) cost suppliers. This is indeed a very major reason why trade diversion and trade modification is costly and can be avoided via multilateral free trade. By the same token, for the EC, and much more apparently for Japan, multilateral free trade can be roughly illustrated to be better than regional free trade in terms of gains from specialization. One may argue that in practice multilateral free trade allows the presence of free riders who avoid giving concessions, but that a regional free trade is relatively effective, and can be a force for liberalization. Of course, gains from either regionalism or multilateralism do not depend only on the specialization pattern, but are also crucially dependent on how much trade barriers are to be removed reciprocally by trading partners who join the negotiating tables. However, as illustrated by Baldwin (1987a), if. there are countries other than those engaging in a bilateral free trade agreement and they indeed would like to trail liberal trade policies, multilateral negotiations will lead more to trade liberalization, while still being able to meet other requirements as much as the bilateral deal can do, and while multilateral negotiations do not eliminate the free-rider problem, they help reduce it.i4 Therefore, l4 For a similar line of arguments, see an explanation made by Lawrence (Lawrence and Schultz, 1990 : 23).

25 T. Bhongrnakapar : Regional Trading Blocs 427 Table 3 "Revealed" Comparative Advantage of Exporters lo the U.S. ( ) Agriculture Non-Mechanical Mach~nery & Partners\SITC and Food Mining Energy Manufacturing Equipment NAITA II* Canada Mexico EC (12) Germany EITA Latin America Middle East Far East Far East excl. Japan Japan South Korea Taiwan Hong Kong China ASEAN (6) Singapore Thailand Malaysia Indonesia Philippines Brunei Others Australia & New Zealand Austrialia New Zealand Rest of the world (0+1) (2+4) (3) (5+6+8) (7) Note : *NAFTA I1 is comprised of the U.S., Canada, and Mexico. Source : Data compiled from OECD, Foreign Trade by Commodity.

26 428 Chulalongkorn Journal of Economics 8(3), September 1996 Table 4 "Revealed" Comparative Advantage of Exporters to the U.S. ( ) Agriculture Non-Mechanical Machinery & Partners\SITC and Food Mining Energy Manufacturing Equipment (0+1) (2+4) (3) (5+6+8) (7) NAFTA I* NAFTA II* Canada U.S.A Mexico EC (12) Germany EFTA Latin America Middle East Far East Far East excl. Japan Japan South Korea Taiwan Hong Kong China ASEAN (6) Singapore Thailand Malaysia Indonesia Philippines Brunei Others Australia & New Zealand Austrialia New Zealand Rest of the world Note : * NAFTA I is comprised of the US and Canada. ** NAFTA I1 is comprised of the US, canada, and Mexico Source : Data compiled from OECD, Foreign Trade by Commodity.

27 T. Bhongmakapat : Regional Trading Blocs 429 Table 5 "Revealed" Compara~ive Advantage of Exporters to Japan ( ) Agriculture Non-Mechanical Machinery & Partners\SITC and Food Mining Energy Manufacturing Equipment NAFTA I* NAFTA 11* Canada USA Mexico EC (12) Germany EFTA Latin America Middle East Far East South Korea Taiwan Hong Kong China ASEAN (6) Singapore Thailand Malaysia Indonesia Philippines Brunei Others Australia & New Zealand Australia New Zealand Rest of the world (0+1) (2+4) (3) (5+6+8) (7) - Note : * NAFTA I is comprised of the US and Canada. ** NAFTA I1 is comprised of the US, canada, and Mexico Source : Data compiled from OECD, Foreign Trade by Commodity.

28 430 Chulalongkorn Journal of Economics 8(3), September 1996 Table 6- "Revealed" Comparative Advantage of the U.S. ( ) Average Agriculture and Food Mining Energy Non-Mechanical Manufactures Machinery & Equipment Source : Data compiled from United Nations, International Trade Statistics Yearbook, Table 7 "Revealed" Comparative Advantage of the EC ( ) Average Agriculture and Food Mining Energy Non-Mechanical Manufactures Machinery &Equipment Source : Data compiled from United Nations, International Trade Statistics Yearbook, Table 8 "Revealed" Comparative Advantage of the Japan ( ) Average -- - Agriculture and Food Mining Energy Non-Mechanical Manufactures Machinery & Equipment Source : Data compiled from United Nations, International Trade Statistics Yearbook, 1991.

29 T. Bhongmakapar : Regional Trading Blocs 431 even in the presence of asymmetric willingness among a large number of countries to liberalize trade, the broad-based multilateralism can potentially benefit each country in the process of liberalization by bringing all interested parties to the table simultaneously. Furthermore, over a period of time, as a number of trading partners experience changes in comparative advantage (which may not only be due to possible discriminatory trading arrangements, but also to the countries' supply-side dynamism), forming a trading bloc (which usually requires long-term commitments with "preferred" trading partners) may not be consistently desirable. This is especially true when the bloc is relatively inefficient, or leading to a relative decline in its international competitiveness. 4. Regional Trading Blocs and Future of the World Economy As we are approaching a new millennium, the world economy is increasingly being transformed for the new coming era. It is an era in which butter will be far more vital than guns, and international competition among major powers, defined mainly by economic criteria, will be intensified. While the Cold War ended with the collapse of the former Soviet Empire, the relative power of the U.S. has dramatically declined. In the 1990s, it is expected that the hegemonic role of the U.S. will decline further, whereas the EC will grow relative to the U.S., or more or less maintain its world GDP share as it did during the 1980s. The rise of Japanese will continue (e.g. Bergsten, 1990 and Lo et. al., 1989). See also Table 9. This will create a state of US.-Europe-Japan "Tripolarity" which will replace the US.-dominated world order of the postwar period. Such a global transformation will result in strong international economic rivalry among the three major world powers. If the international rivalry is viewed by the major players as a zero-sum