Preferential trading agreements: adding spices and noodles to a spaghetti bowl Mia Mikic

MASTER CILA Preferential trading agreements: adding spices and noodles to a spaghetti bowl Mia Mikic 2002 U N I V E R S I T É L Y O N 2

Preferential trading agreements: adding spices and noodles to a spaghetti bowl Mia Mikić * Contents 1. Introduction 2. Motivation for Liberalization Through Preferential Trading Agreements 3. Textbook Classification of Preferential Trade Agreements 4. A Non-technical Overview of Approaches to Analyzing the PTAs 5. Adding Spices and Noodles to a Bowl of Plain Spaghetti 6. Preferentialism a Friend And an Enemy to Multilateralism? References Appendix A list of selected PTAs * Department of Economics, The University of Auckland and Graduate School of Economics and Business, University of Zagreb; correspondence to mmikic@efzg.hr 2

Preferential Trading Agreements: adding spices and noodles to a spaghetti bowl 1 INTRODUCTION Economists may agree or disagree on the net effect of the proliferation of the preferential trading agreements in a global economy but none of their analyses will stop the trend of new trade blocs being increasingly built. Many nation states as well as already formed preferential trading agreements (PTAs) work on adding more and more into an already full spaghetti bowl. 1 This paper provides a non-technical overview of the analyses of preferential trading agreements and talks about reasons for their proliferation in the aftermath of the establishment of the WTO. Section two introduces the concept of preferential liberalization (a.k.a. regional trade agreements, regionalism, economic integration) and discusses the motivation for national governments to turn to this approach to liberalization with more zest compared to their efforts for unilateral and multilateral liberalization. The third section gives the textbook classification of the different types of the PTAs. The forth section summarizes the main approaches to analyzing economic impacts of preferential agreements, while section five provides some evidence and discussion on proliferation of PTAs. The last section compares and contrasts the preferential trading agreements with multilateralism in order to examine if and how the preferentialism acts as a building block for multilateralism. This is important because almost 90% of the WTO members are parties to one or more regional arrangements. 1 This term, which describes a mixture of overlapping and intersecting preferential agreements with frequently inconsistent provisions, originally linked to the EU, is credited to Bhagwati (1992). 3

2 MOTIVATION FOR LIBERALIZATION THROUGH PREFERENTIAL TRADING AGREEMENTS * Trade liberalization implies gradual or complete removal of existing barriers to trade in goods and services. If a country decides to lower or remove all of its trade barriers (tariffs and non-tariff barriers) without the expectation of reciprocal actions by other countries, we speak of unilateral liberalization. 2 The case for unilateral liberalization is built along the same reasoning as the case for free trade: The removal of import barriers benefits domestic consumers and producers who gain access to more variety at lower prices. Domestic producers in sectors not directly competing with imports gain as they cease to compete for resources and inputs against previously privileged sectors. Now allocation of resources is in accordance with comparative advantage. Even the sectors which lose privileges (import-substituting sector) benefit in the long run as international competition forces them to become more productive. Needless to say, not all producers would improve sufficiently to remain in the market. But subtracting the losses incurred by those which have to leave the market from the gains other producers and consumers obtain, a society is left with net gains. 3 These net gains should be sufficient reason to wholeheartedly accept unilateral free trade. However due to many, mainly political economy type reasons, governments rarely decide to go along the path of unilateral liberalization. 4 Instead countries seek reciprocal concessions or preferences. In other words, a government is prepared to lower tariffs and * This section borrows from Mikic (2001). 2 Unilateral liberalization should not be confused with unilateralism, which describes a desire to impose one s view of the desirable features of global trade policy or trade in a particular products on other countries and have it accepted by them. 3 The size of the net gains is still a matter of heated debate. While pro-trade economists claim that net gains are underestimated given the imperfections of tools (like so-called Harberger triangles in a partial equilibrium model or estimation of compensating variation in the general equilibrium models), anti-trade activists claim that these net gains are in fact too small to be relevant. 4 Governments typically do not focus on net effects for the whole society. Instead they tend to care more about the net effects for special interest groups, say exporters or producers in the import-substitution sector (or "in import-substitution sectors" or a subgroup of the labour force. 4

other barriers impeding imports in exchange for equivalent concessions from a trading partner, but not otherwise. 5 When liberalization is done on this reciprocal basis, it limits the increase in foreign competition in domestic market and improves market access for domestic producers in foreign markets. Since these markets may be enjoying high protection from the rest of the world competition, there are opportunities to make large profits. This drives producers to lobby with governments for non-unilateral liberalization and therefore reciprocal liberalization might be easier or the only feasible path for liberalization. World Bank publication Trade Blocs cites the former ministry of industry in Morocco Mr. Hasan Abouyoub saying that trade liberalization would have been infeasible without first entering into a free trade arrangement with the EU (p.27). In any case, this practice of quid pro quo or exchanging concessions in the form of lower tariffs has been entrenched into the world trading system and has been the driving force behind the repeated rounds of multilateral trade liberalization. Multilateralism or multilateral trade liberalization is described thus as a simultaneous commitment to the reduction of trade barriers in all member countries. In addition to reciprocity, the principle of non-discrimination has been instrumental in this multilateral approach to trade liberalization. In practice, this principle has taken the form of the socalled most-favoured nation (MFN) principle enunciated in the GATT /WTO Article I. However, there are three alternative provisions allowing the WTO members to discriminate against other members when it comes to their trade policy (Panagariya, 2000, p.289): 1. Generalized System of Preferences (GSP) allows developed countries to give one-way preferences to developing countries. The GSP was designed to promote South North trade, in particular exports from South to North. 2. Enabling Clause allows developing countries to exchange virtually any trade preference to which they mutually agree. This exception to the non-discrimination concept is introduced to promote South-South trade and that is why it s wide 5 In addition to seeking reciprocal tariff concessions (multilateral or regional) instead of unilateral free for trade for terms of trade reasons, countries also have other arguments such as infant industry and specific country arguments, domestic political reasons and occasionally balance of payments arguments. 5

coverage enables developing countries to enter into any type of PTAs among themselves. Therefore all such arrangements formed among the developing countries (e.g. MERCOSUR, AFTA, SAPTA, Bangkok Agreement) are notified to the GATT / WTO under this clause. 3. Article XXIV allows for any two or more members to form a preferential trading area under strictly defined conditions. A key condition is that the exchange of preferences is not partial but extends to substantially all the trade in products originating in participating countries. Another restriction is linked to the height of the protection in trade with the outside world. Depending on the actual type of agreement between countries, the Article stipulates that external tariffs of member countries must not be raised nor must the incidence of the common external tariff exceed that of individual tariffs of participating members prior to agreement. As this is the only avenue for developed countries to grant trade preferences to each other, this Article is behind all preferential agreements formed between developed countries (e.g. EU, EFTA, ANCERTA, NAFTA) and those between developed and developing or transition economies (e.g. former Lome Convention and now Cotonou Agreement, Europe Agreements). These three exceptions allow member countries to discriminate against other members by extending reciprocal or non-reciprocal trade preferences to some but not all the WTO members. Since many of the preferential trading agreements involve geographically neighboring countries, we often refer to them as a regional approach to trade liberalization or regionalism. However it would be better to term this approach as preferentialism since regional aspects of the agreements are less and less important. Obviously preferentialism as an approach to trade liberalization may take place in parallel to multilateral actions or may develop to replace them. 6

3 TEXTBOOK CLASSIFICATION OF PREFERENTIAL TRADING AGREEMENTS We distinguish among four levels of PTA (see Table 1). This classification is not watertight in the sense that most of the real world integrations are mutants across these textbook types featuring elements of more than one form.. The most common type of PTA is known as a free trade area (FTA). A free trade area is the type of regional liberalization where tariffs on goods traded among participating countries are removed but each country retains its own national trade policy and tariff schedule against non-members. It is typically assumed that trade liberalization resulting from such an agreement would cover all tradable products in the participating countries. However, in reality, there are always so-called sensitive activities or sectors that are granted special treatment (including protection from international competition) and this status is kept even when a country forms a regional integration most FTAs apply only to trade in non-sensitive goods (typically excluding agriculture, services, textiles and clothing). Even though in most cases countries forming a free trade area would be at a similar level of development, their trade policies and levels of protection might be quite dissimilar. In such cases, with one country in the free trade area having higher trade barriers than the others, freeing of intra-area trade will result in a phenomenon known as trade deflection. 7

Table 1 Main features of different levels of preferential liberalization Levels of integration Elimination of tariffs on intraregional trade Elimination of tariffs on intratrade and common external tariffs Free movement of capital and labour Free trade area Customs union Common market Economic union Coordination of economic policies and harmonization of standards Trade deflection occurs when imports enter the FTA market through the member country that has the lowest tariff. For example if there was a FTA among A, B and C, and A had the lowest tariffs for widgets, the non-fta exporters of widgets would re-route their trade with the FTA so that widgets would be exported to A and then shipped duty-free from there to B and C. For trade deflection to be effective, transport costs and other costs related to re-routing of imports through the lowest-tariff country must be sufficiently low not to outweigh the difference in tariff rates. Trade deflection affects members of the FTA with higher tariffs in two ways: 1) it shrinks tariff revenue these members would otherwise collect (tariff revenues are in effect redistributed in favor of low-tariff members), and 2) it weakens the protective impact of high-tariff members trade policies. 6 To avoid trade deflection from occurring all FTAs impose rules of origin (ROO). They are used to determine the country of origin of the goods in cases where the production takes place in more than one country (by definition each good can only originate in one country / area). There are basically two types of ROO: non-preferential and preferential. Non-preferential ROO are used for a number of reasons but most often for the application of tariffs and tariff quotas. Preferential ROO prescribe the characteristics of goods 6 In effect, the consequences of trade deflection are equivalent to countries forming a customs union with the common external tariff set to the level of the country within the FTA with the lowest tariffs. 8

eligible for preferential or duty-free trade within the FTA. The main task of the preferential ROO is to ensure that benefits from preferential trade liberalization (such as a FTA) are restricted to those goods which originate and are traded within the particular preference area (APEC, 1997, p.5). More details on the ROO are contained in Box 1. Box 1 Rules of origin Rules of origin (ROO) is a set of laws and regulations applied by national (or FTA) trade authorities to determine the country of origin of goods. ROO affect different aspects of trade because the origin of goods has a direct bearing on the administration of trade measures such as quota systems, tariff preferences, or anti-dumping and countervailing duties. Another purpose for which rules are applied relates to trade under preferential arrangements, where the ROO is evidence needed for a country to enjoy a preferential status. Furthermore, determination of ROO is necessary in the process of collection of trade statistics and labeling to indicate origin for different reasons including government procurement (ITC, 1999, p.155-6). There are different ways in which the origin of goods produced in more than one country could be determined. Despite a wide variety in the actual methods, in most countries the basic principle behind ROO determination is so-called substantial transformation. This identifies the last country in which substantial transformation of a product has occurred measured by either: - the value added in manufacturing / further processing, or - the changes in tariff classification. 7 Under the system of value added, for a product to be considered to have been produced or manufactured in the country or within a preferential area for which origin is being claimed, a specified percentage (e.g. 40%, 50%, 60%) of the product value has to be added by processing within that country / area. When substantial transformation has occurred so that a good is re-classified into a different tariff heading we talk about applying the principle of the changes in tariff classification to determine the origin of goods. WTO encourages its members to use the tariff classification known as the Harmonized System Nomenclature (HS) for the purposes of collecting trade statistics and assigning custom duties. The HS arranges products according to the degree of processing starting with raw materials and ending with finished products. The country of origin is determined by identifying where the last change in its tariff classification has occurred. GATT did not require parties to adopt any specific set of rules. Most countries typically had two sets of rules: general or non-preferential rules applicable to all countries, and some special rules (often not transparent) for preferential trade and/or regional arrangements. The Generalized 7 There are of course other methods of determining substantial transformation including a specified process of manufacturing. 9

System of Preferences (GSP) is for example one way in which preferential ROO are set. 8 The lack of international agreements on the ROO also allowed for the use of ROO as a protectionist devise. In order to make ROO simpler, more uniform and more stable, the Uruguay Round considered it desirable to prepare a harmonized set of ROO to be adopted by all member countries. The Agreement on Rules of Origin signed in Marrakesh in 1994 (WTO, 1995b, p.31) requires WTO members to ensure that their rules of origin: -are transparent, -have no restricting, distorting or disruptive effects on international trade, -are administered in a consistent, uniform, impartial and reasonable manner, and -are based on a positive standard. The Agreement focuses on the MFN based-trade and does not apply to preferential agreements. It was envisaged that technical work on harmonization of ROO across members would take at least several years and therefore the Agreement carries two sets of provisions: one contains disciplines countries are expected to follow during the transition period (until new harmonized rules enter into force), and another to be applied after the transition period is over and the new harmonized set of ROO are in place. These harmonized rules provide that the origin of goods shall always be the country where the last substantial transformation has been carried out and this should be for any purpose of the origin determination. The adoption of harmonized rules is expected to ease the administrative burden on exporters and reduce exporting costs, as exporters 8 A detailed account of the ROO used in the GSP could be found in Digest of Rules of Origin: Generalized System of Preferences by UNCTAD. 10

should be able to meet ROO requirements much easier. Customs union involves the suppression of all intra-area trade barriers among member countries and the establishment of a common external tariff (CET) on imports from non-members. The European Union began as a customs union (then under the name of the European Economic Community, EEC). Previous to that in 1947, Belgium, the Netherlands, and Luxemburg had formed a customs union called Benelux which became part of the EEC when it was formed in 1958. Because participating countries introduce the CET there is no trade deflection and no need for the rules of origin to police the transshipment by non-members. By the GATT/WTO rules (Article XXIV) the CET must on average be no higher than the pre-union tariff, and the compensation is negotiated by non-member countries when non-members are harmed. Common market is the next level of regional integration in which all tariffs and trade barriers are removed on trade between members, a CET is adopted for outside trade and barriers on movement of labour and capital, as well as enterprises between member countries are removed. The most successful example of a common market in the real world is of course the European Community (which completed the common market stage in 1968). Even deeper integration is achieved in the form of an economic union. It includes all the features of a common market as well as full harmonization of national economic policies As mentioned above, the various regional trade groupings actually established in the real world do not fully correspond to the above classification. For example, a number of real world FTAs also include one or two characteristics of the integration of the higher level: ANCERTA allows for free movement of labour and some movement of capital across the Tasman Sea, while NAFTA in addition to removing most restrictions on foreign investment also deals with intellectual property protection and government procurement. 11

Likewise, the EU still does not show full harmonization of fiscal or financial systems even though it has reached the level of economic union. One of the reasons for the mismatch between theoretical or textbook and practical types of integration is that often after starting to form a particular level of integration, countries find themselves pressured by initial success or failure to either proceed / evolve into a more complex type of integration or revert to a more primitive type of integration. 4 A NON-TECHNICAL OVERVIEW OF APPROACHES TO ANALYZING THE PTAS 9 The basic ideas on economic effects of PTA could be obtained by using an example of a customs union between two countries (as in such way we do not have to worry about differential external tariffs nor of movements of factors of production or harmonization of other economic policies). Take then two countries A and B, which form a customs union. The analysis we use is of a partial equilibrium nature as we also apply other standard assumptions of trade theory. 10 Suppose that country A has a tariff t A against country B and the rest of the world (W) and suppose that country B is more efficient than the rest of the world (p B <p W ). Also assume that supplies from country B and the world are perfectly elastic at prices p B and p W (i.e. country A is a small country). In that case prior to any agreements the country A will be importing from country B (0M 0 in part a of Figure 1) and not from the world. When Country A and B form a union agreeing on a common tariff t A against the rest of the world the effects will be as shown in part a of Figure 1. After the customs union all imports still come from the country B but trade expands (to 0M 1 ) because now without a tariff country A pays less for imports - there's a surplus gain (note that this is an excess demand curve) to country A and when lost tariff revenues are 9 This section leans heavily on Mikić, 1998, pp.445-59 and Richardson (2002). 10 That is perfect competition, no distortions other than the tariffs and the CU, no externalities, no international movements of factors, and no other trade costs. 12

subtracted, a country is left with a net gain, as shown. This is called trade creation and is unambiguously a good thing for country A. 13

Figure 1 Part a Part b p A p A p 0 A p 1 A gain n p W +t A P B +t A pw W P B p 0 A p 1 A loss gain P B +t A p W +t A P B pw W ED A ED A M 0 M 1 M A M 0 M 1 M A Trade creation describes trade that occurs between members of a regional integration that replaces what would have been produced in the importing country were it not for the preferential liberalization. It is associated with welfare improvement for the importing country since it represents a substitution of higher-cost domestic production by efficient regional partner country production. As this is a shift in the direction of what would have been a free-trade allocation of resources, it has a beneficial welfare effect. Another net effects is however possible given that our preferential partner might not be the most efficient of all countries we trade with. This situation is presented in part b of Figure 1. Before the customs union country A buys only from the rest of the world at a price p W +t A. After the union with country B, it does not pay to import from the world as the imports from a partner country B is now duty-free. With p B <p W +t A country A switches to purchasing imports from the partner country. Again there is a surplus to be reclaimed, but there is also a loss in tariff revenues (see Figure 1, part b). A net effects 14

may be negative if country B is much less efficient than the rest of the world. This effect is trade diversion. 11 Trade diversion is trade that occurs between members of a preferential agreement that replaces what would have been imported from a country or countries outside of the bloc. It is a consequence of discrimination introduced by a preferential agreement. Because it represents a shift away from a free-trade allocation of resources it is associated with welfare reduction for the importing country it increases the cost of imports. Both trade creation and trade diversion are possible effects and the net effect of the preferentialism depends on the relative strengths of these two. The following are some circumstances where trade diversion will be less of a problem (World Bank, 2000, pp.40-1): 1) as trade diversion can only occur if the country has a tariff on imports from the rest of the world, it is clear that lowering those tariffs would reduce the cost of trade diversion 2) as trade diversion arises only if partner country costs are out of line with costs and prices in the rest of the world, if partner country itself had lower tariffs against the rest of the world there will be a downward pressure on it s costs 3) with transport and transaction costs, it is likely that costs of trade diversion are smaller if partner country is close (geographically) and more similar (in terms of production structure). To summarize, the net effect of forming a customs union (or some other form of PTA) could be positive or negative. However, this result depends on the external tariff of the PTA against the rest of the world to remain the same as it was before the PTA. This is of course very restrictive assumption and there are different reasons why this might not 11 This is one reason why the GATT provisions on bilateral deals are inadequate to protect other countries: even with domestic tariff unchanged the external volume of trade falls and when the third countries are not large they might very well suffer a loss (cf. Richardson, 2002). 15

happen with real PTAs. Namely the external tariff of the PTA may be influenced by the parallel unilateral or multilateral liberalization of a partner or partners; by the particular set of political-economy factors; by the delegation of tariff-setting authority within the customs union, or by the free-riding within the union. Another dimension lacking in the above analysis is the economic impact on the countries outside the PTA. Article XXIV of the GATT of course attempts to minimize the adverse effects on these countries by requiring that the level of external tariffs of the PTA countries is not higher than before the PTA. It seems however that this does not prevent the harm PTAs could cause to the third countries; it is the PTAs effects on the terms of trade and the imports of the countries left outside the PTAs that determines the changes in their economic welfare (see more in Winters, 1996). There are also so-called dynamic effects arising from preferential trading agreements, which help integrated larger integrated markets. They refer to: better utilization of economies of scale increased competition within the integrated market greater investment and technology transfers in countries within the regional integration benefits from increased labour and capital mobility. There are of course political impacts as well. Often the purpose of integration is political while the economic consequences are only side-effects of the political payoff (World Bank, 2000, p.11). The most important of these political payoffs are: intra-regional and extra-regional security (like in case of MERCOSUR, Association Agreements with Central Europe, and in Africa, etc.) bargaining power (like in the case of Cairns Group, the EU, and OPEC) project cooperation (like in areas of fishing -EU and EFTA, energy SADC, etc.) 16

locking-in domestic reforms (like in the case of Mexico with NAFTA, central European countries with the Association Agreements with the EU, or Greece and Portugal with the EU, etc.). As this overview indicates, the estimation of economic impacts of PTA is a complex task. Even if the net result is positive we need to ask if there is a better alternative to reach such an outcome. In other words, if a country liberalizes unilaterally would it be worse off compared with liberalization through a PTA? The answer based on economic implications is simple: unilateral liberalization is at least as good as preferential (if partners are internationally efficient) or superior to preferential liberalization (if partners are inefficient by world standards). There must be then the reasons belonging to the area of political control of liberalization (monitoring and exchange of concessions) that tip the scale in favour of preferential agreements compared to unilateral liberalization since we witness only an upward trend in the number of the PTA under negotiation. 5 ADDING SPICES AND NOODLES TO A BOWL OF PLAIN SPAGHETTI In the 20th century we have witnessed two large waves of regionalism: one during the 1950s and 1960s (Bhagwati, 1992), and the other during the 1990s. As Panagariya (2000, p.287) claims the effective regional liberalization during the first wave did not spread beyond Western Europe (EC and the EFTA). The second wave of regionalism however was much more successful driving virtually all countries developed and developing - to seek some form of regional trading arrangement. By 1999 more regional agreements had been notified to the WTO than it had countries as members (World Bank, 2000, p.123; see Box 2). Even Japan, Hong Kong and Korea, countries that refused repeatedly to enter into any contractual trading arrangements, are presently working on forming free trade areas. Due to this continuing proliferation of regionalism real pressure has arisen on the multilateral trading system (see also section 6). 17

According to Hoekman and Kostecki (2001, p.348-9) there are a number of factors behind the steady growth of preferential trade agreements: 1) Former socialist economies using regional liberalization as a tool of re-integration into the global economy. 2) Changed attitude of the US towards regionalism. 3) Domino effects putting pressure on countries remaining out of the block to get in. 4) Globalization i.e. increased internalization of all markets putting pressure on firms to seek efficiency through larger markets. 5) Using regional agreements to lock in domestic reforms and grow credibility. 6) Regionalism is often driven by foreign policy and national security considerations. Box 2 Preferential trading agreements in the world today Similar to GATT previously, the WTO requires its members to notify it of the preferential (regional) trade arrangements in which they participate. Most of the current members have notified participation in one or more of the arrangements. However it seems the number of notifications has increased since the establishment of the WTO. During the GATT era in the period 1948-94, there were about 125 notifications (relating to trade in goods), and since 1995, another 125 more arrangements related to trade in goods and services have been notified to the WTO (or are to be notified). Many of the notified arrangements are not in force any longer. Some of them have been superseded by redesigned arrangements among the same members. Out of total 250 agreements notified, more than 170 are deemed to be active. However a large number of those are just components of the two major regional integration agreements/arrangements among developed countries (the EU and NAFTA). It seems that less than 40 arrangements presently exist with a wide variety of rules and characteristics. Historically, most preferential agreements were concentrated on the European-Mediterranean region (Chart 2) which accounts for more than 50% of all agreements in force. However, as number of agreements currently under negotiation in the Americas is quickly increasing, the European region might lose its traditional supremacy. Another region which was quite dormant in proliferating regionalism Asia-Pacific- has lately also become surprisingly active in adding to the total tally of agreements. Another interesting phenomenon is a rise in cross-regional negotiations. As can be seen from Chart 2, 14 such agreements are being negotiated with 15 already in force. 18

Number of RTAs No. of RTAs CAHIERS DE COMMERCE INTERNATIONAL No 3, 2002 Chart 1 The number of preferential trade agreements notified each year (1948-2000) 300 250 Establishment of the WTO Source: WTO (2000) 200 150 100 50 0 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 Year Chart 2: Geographical Distribution of RTAs, both in force and under negotiation 120 100 80 60 40 FTAs under negotiation CUs under negotiation FTAs in force CUs in force 20 0 Americas Asia Pacific Eastern Europe & Central Asia Euro- M editerranean Sub-Saharan Cross Regional Region Source: WTO (2000) 19

TABLE 2 MERCHANDISE TRADE OF SELECTED REGIONAL INTEGRATION ARRANGEMENTS, 1999 (BILL DOLLARS AND %) Share in total Value exports/imports Annual percentage change 1999 1990 1998 1999 1990-99 1998 1999 APEC (21) Total exports 2497 100.0 100.0 100.0 7-4 6 Intra-exports 1774 67.5 69.3 71.0 8-7 9 Extra-exports 723 32.5 30.7 29.0 6 4 0 Total imports 2625 100.0 100.0 100.0 7-7 10 Intra-imports 1883 65.4 71.3 71.8 8-6 11 Extra-imports 741 34.6 28.7 28.2 5-10 8 EU (15) Total exports 2180 100.0 100.0 100.0 4 4-1 Intra-exports 1385 64.9 63.1 63.5 4 7 0 Extra-exports 796 35.1 36.9 36.5 5 0-2 Total imports 2232 100.0 100.0 100.0 4 6 1 Intra-imports 1389 63.0 62.8 62.2 4 7 0 Extra-imports 843 37.0 37.2 37.8 4 5 3 NAFTA (3) Total exports 1070 100.0 100.0 100.0 7 0 6 Intra-exports 579 42.6 51.2 54.1 10 5 11 Extra-exports 491 57.4 48.8 45.9 5-5 -1 Total imports a 1420 100.0 100.0 100.0 9 5 12 Intra-imports 575 34.4 40.3 40.5 11 6 12 Extra-imports 846 65.6 59.7 59.5 7 5 11 ASEAN (10) Total exports 359 100.0 100.0 100.0 11-7 9 Intra-exports 79 20.1 21.9 22.1 12-18 10 Extra-exports 280 79.9 78.1 77.9 10-3 9 Total imports 299 100.0 100.0 100.0 7-25 7 Intra-imports 69 16.2 22.6 22.9 11-16 9 Extra-imports 231 83.8 77.4 77.1 6-27 7 CEFTA (6) Total exports 107-100.0 100.0-12 1 Intra-exports 13-13.0 12.0-6 -7 Extra-exports 94-87.0 88.0-13 2 Total imports 134-100.0 100.0-11 -2 Intra-imports 13-9.7 9.5-8 -4 Extra-imports 122-90.3 90.5-11 -1 MERCOSUR (4) Total exports 74 100.0 100.0 100.0 5-2 -9 20

Intra-exports 15 8.9 25.0 20.3 15-1 -26 Extra-exports 59 91.1 75.0 79.7 4-3 -3 Total imports 83 100.0 100.0 100.0 12-3 -16 Intra-imports 16 14.5 21.0 19.0 16-1 -24 Extra-imports 67 85.5 79.0 81.0 12-4 -14 ANDEAN (5) Total exports 43 100.0 100.0 100.0 4-16 10 Intra-exports 4 4.3 13.9 8.9 13-3 -29 Extra-exports 39 95.7 86.1 91.1 3-18 16 Total imports b 36 100.0 100.0 100.0 8 2-20 Intra-imports 4 7.7 11.7 11.7 13-10 -20 Extra-imports 32 92.3 88.3 88.3 8 4-20 a Imports of Canada and Mexico are valued f.o.b. b Imports of Peru and Venezuela are valued f.o.b. Note: The figures are not fully adjusted for differences in the way members of the arrangements in this table record their merchandise trade. Source: WTO, Annual Report 2000, Table 1.9, p.23 21

There are two processes taking place on the two sides of the Pacific Ocean, which deserve commenting on. One is related to the negotiation of the Free Trade Area of Americas and another one relates to increased activities among the Asian nations that normally were not involved in preferentialism (e.g. Japan, Hong Kong, Singapore, Korea). In both of these processes there is a strong involvement of the US seeking to advance the trading agenda that is still not embraced by the WTO (e.g. trade linkages to environmental and labour standards) or which they could better monitor and/or enforce in preferential agreements (related to investments, intellectual property protection and similar). These activities add new flavour to the old spaghetti bowl as it is not only the number of preferential agreements (density in the bowl) that is increasing, but they also change texture and taste as some cover areas that normally would not enter an FTA (as agreements with Singapore and Chile do) and also for the first time we find possible agreements with countries that so far had none (Japan, South Korea). 22

6 PREFERENTIALISM A FRIEND AND AN ENEMY TO MULTILATERALISM? Following Bhagwati (1992) the implications of PTAs for multilateralism can be examined along two separate lines: 1) the immediate (static) impact of the PTAs on world welfare, and 2) the impact on a dynamic liberalization path since PTAs might expand to encompass the whole world (regionalism as a stepping stone to global free trade) but might also fragment the world economy (regionalism as a stumbling block to global free trade). 12 The static impact effect is of course the question related to the net welfare effect of the regionalism: are they on balance more trade creating or trade diverting? The answer is really an empirical one. There are some conditions which if satisfied may increase the likelihood for a regional integration to be trade creating, but they cannot be generalized (see Mikić, 1998, p. 450-52). Thus regional trading arrangements can raise welfare but there can be no presumption that they will. When looking at the impact on the dynamic liberalization path, it is helpful to consider elements of speed, efficiency and certainty. The speed of trade liberalization on a global level has become an issue during the 8 th GATT Round the Uruguay Round where the major players could not agree on commitments in the newly added areas of agriculture and services trade. Many commentators claim that the most important reason for that round s movement away from the stale position was a credible threat by the US of going on a separate (e.g. bilateral and regional) path away from multilateral talks. The situation has not improved since the Uruguay Round closed and with the establishment of the WTO. An increased 12 There are also impacts of multilateralism on regionalism that are worth mentioning. Following Panagariya (2000, pp.324-25), claims could be made that: a) multilateralism makes regional liberalization more sustainable, and that b) multilateral liberalization among the developed countries leads to more preferential trading arrangements between developed and developing countries. 23

number of members (at the moment 144 countries are full members) and widened agenda with new issues still emerging, the multilateral negotiations are becoming complex and definitely slower that they used to be in the early years of GATT. The failure of the Ministerial in Seattle 1999 and the stalling Doha negotiations are clear evidence of that. But we must be careful not to overestimate the speed of preferential trade liberalization either. The most successful preferential integration in terms of achieving free trade for its members is the European Union. Yet it took a long time for free trade of goods and services within the EU to be in fact implemented. There is a perception of regionalism being quicker in liberalization because perhaps the negotiating time on the agreement might be relatively short. However, the time involved from the moment an agreement is announced until free trade is in fact materialized is very long, and in many of the agreements it does not happen ever. With respect to regionalism producing more efficient results, one has to be careful how to define efficient. If efficient is to mean more liberalization per unit of time or resources spent to get the result, then we cannot really say much about the efficiency of either path, as it would be difficult to measure it. The issue of certainty is linked to stability and irreversibility of trade liberalization commitments made under one of the two approaches to liberalization. This is where multilateralism is definitely superior as proven by experience. All of the trade negotiation rounds have been mostly implemented, even though it took longer for some than others. With respect to commitments to reductions in tariffs, these are bound and it is very difficult for a member to renege on the agreement. 13 On the other hand, it is easier for regional agreements to fail (e.g. LAFTA) or to adjust (ASEAN) the speed of 13 It is not to say that some members might occasionally try to introduce some additional import surcharges without previously clearing it up with other members. But sooner or later they will have to step back in line because the large number of members in the WTO also means that it would be more difficult for all to agree to a change. 24

liberalization to meet the specific interests of some of the members. Many transition economies have sought membership of the WTO for the reason of obtaining a means of locking in their economic reforms. In conclusion, it is unclear whether regionalism can be considered a stepping-stone to global free trade or a stumbling block. What is certain however is that regionalism is here to stay and that the world economy in the future will rely upon both regional trading blocs and the WTO. This two-way commitment of individual countries to trade liberalization may push the world economy further down the path towards free trade than would sole reliance upon multilateral negotiations. On the other hand, as Panagariya, (2000) argues, the proliferation of the regional trading agreements with differentiating tariff and other barriers will result in a chaotic and non-transparent trading regime - spaghetti bowl spiced with agreements from the Americas and with added twists from bilateral agreements in Asia. The only way out is to speed up liberalization based on the MFN principle. Once external tariffs drop to zero, tariff preferences and the spaghetti bowl created by them will automatically disappear (p.328). Despite this being the best solution, it is not the most likely one (at least not in a short run). From the point of view of developing countries and their need to secure access to developed countries markets, if the WTO-led path is not feasible, the only available avenue is through proliferation of association agreements with the other established integrations such as the EU, NAFTA, and individual large traders like Japan and consequently developing countries agreements such as AFTA or MERCOSUR. In fact what the World Bank (2000, p.117) proposes is modification of the WTO rules regarding the regional integrations to create a presumptive right of association. Analogous to the MFN clause, if association is granted to one country, there should be a presumption that similar terms should be available to others: if Iceland is offered reciprocal freedom from antidumping suits from the EU, then the same option should be available to Ghana. By opening the clubs to the membership by developing countries the advantages of regionalism are combined with the advantages of multilateralism. In terms of the effects 25

discussed above, if this type of regionalism was pursued outsiders would be able to enjoy the positive trade creation effects while the negative effects of trade diversion would be minimal. 26

References APEC, (1997) Compendium On Rules of Origin, Singapore: APEC. Bhagwati, J. (1992) Regionalism versus Multilateralism World Economy, 535-55. Bhagwati, J., P. Krishna and A. Panagariya (1999) Trading Blocs, Cambridge, Mass.: The MIT Press. Bhagwati, J., D. Greenaway and A. Panagariya (1998) Trading Preferentially: Theory and Policy, The Economic Journal, 108 (July), 1128-48. ESCAP (2000) The Future WTO Agenda and Developing Countries Studies in Trade and Investment 41, New York: United Nations. Hoekman, B. and M. Kostecki (2001) The Political economy of the World Trading System: The WTO and Beyond Oxford: Oxford University Press ITC (1999) Business Guide to the World Trading System, 2 nd ed, Geneva and London: ITC and Commonwealth Secretariat. Krueger, A. (1999) Are Preferential Trading Arrangements Trade-Liberalizing or Protectionist? Journal of Economic Perspectives, 13, 4, pp. 105-24. Mikić, M. (2001) Training manual on Increasing Capacities in Trade and Investment promotion, UN/ESCAP, New York Mikić, M. (1998) International Trade, Basingstoke: Macmillan. Panagariya, A. (2000) Preferential Trade Liberalization: The Traditional Theory and New Developments, Journal of Economic literature, XXXVIII (June), pp. 287-331. Richardson, M. (2002) PTAs an overview, presented at the University of Auckland, September, mimeo Winters, A.(1996) Regionalism versus multilateralism World Bank Policy Research, International Economics Working Paper No. 1687 retrieved http://monarch.worldbank.org/pub/decweb/workingpapers/wps1600series/wps1687/ World Bank (2000) Trade Blocs, Washington, D.C: World Bank available at http://www.worldbank.org/research/trade/pdf/trade%20blocs.pdf WTO, (2000) Mapping of Regional Trade Agreements, Geneva: WTO, WT/REG/W/41 27

WTO, (1995b) Trading Into the Future, Geneva: WTO. 28

Appendix The list of selected PTAs As Chart A1 shows PTAs are concentrated in Europe. However developments in the last two years since this mapping was done by the WTO indicate that while Europe will retain its supremacy for the number of PTAs other regions, e.g. Americas and Asia/Pacific will decrease the current gap. The following list of PTAs is far from comprehensive; it includes only some of the agreements which are in force. The detailed description of all preferential agreements see in Frankel, J. (1997) Regional Trading Blocs in the World Economic System (IIE, Washington). Europe 14 EU European Union [including its subset Economic and Monetary Union (EMU)] Established as the customs union in 1957 Current level of integration economic union [and monetary union for a subset] Current members: Austria, Belgium, Denmark*, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweeden*, United Kingdom* (countries labeled by * are not in the EMU) Potential members by 2005: Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovak Republic, Slovenia EFTA European Free Trade Area Established as a free trade area in 1960 Current level of integration free trade area Current members: Iceland, Liechtenstein, Norway, Switzerland 14 Various bilateral agreements including partnership agreements between the EU and a number of other countries in Europe and in the world at present are mostly non-reciprocal, non-symmetrical agreements and are not considered here. 29

EEA European Economic Area Established as free trade area between EU and EFTA in 1994 Current level of integration free trade area Current members: members of the EU and the EFTA (except Switzerland) CEFTA Central European Free Trade Area Established as free trade area in 1992 Current level of integration free trade area Current members: Bulgaria, Croatia, Czech Republic, Hungary, Poland, Rumania, Slovak Republic, Slovenia Americas NAFTA North American Free Trade Area Established as free trade area in 1994 Current level of integration free trade area (including agriculture, intellectual property protection, investment, etc) Current members: Canada, Mexico, United States Potential members: Chile and CARICOM countries MERCOSUR Mercado Comun del Sulde Established as free trade area in 1991 Current level of integration free trade area Current members: Argentina, Brazil, Paraguay and Uruguay Andean Community Established as free trade area in 1989 after an attempt to create a customs union as the Andean Pact has been failing since 1969 (another failed attempt at a customs union in 1992) 30

Current level of integration free trade area Current members: Bolivia, Colombia, Ecuador, Peru, Venezuela CACOM Central American Common Market Established as free trade area in 1959 Current level of integration free trade area Current members: Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua CARICOM Caribbean Community Established as free trade area (CARIFTA) in 1968 Current level of integration customs union (1973) Current members: Bahamas, Jamaica, Belize, Montserrat, St. Kitts and Nevis, Antigua and Barbuda, Dominica, Saint Lucia, Barbados, St Vincent and the Grenadines, Trinidad and Tobago, Grenada, Guyana, Suriname Under negotiation: FTAA Free Trade Area of Americas Proposed free trade area among 34 countries Asia/Pacific APEC Asia Pacific Economic Cooperation Established as a nondiscriminatory association of countries in 1989 Current level of integration negotiation in process, still not formally a trading bloc (vision to become a free trade region by 2020) Current members: Australia, Brunei, Canada, Chile, China, Chinese Taipei (Taiwan), Hong Kong, Indonesia, Japan, Korea, Malaysia, Mexico, New Zealand, Papua New Guinea, Peru, Philippines, Russia, Singapore, Thailand, Vietnam, US 31

ANZCERTA Australia and New Zealand Closer Economic Relations and Trade Agreement (aka CER) Established as a free trade area in 1983 Current level of integration free trade area (covering lots of extras including free movement of labor, extensive trade facilitation and harmonized competition law provisions) Current members: Australia, New Zealand AFTA ASEAN Free Trade Area Established first in 1967 as ASEAN to foster economic, cultural and social cooperation among the original five members Current level of integration free trade area Current members: Brunei, Cambodia, Indonesia, Malaysia, Myanmar, Philippines, Singapore, Thailand, Vietnam SAARC South Asian Association for Regional Cooperation Established in 1985 Current level of integration free trade area (SAPTA) since 1992 Current members: Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, Sri Lanka, Under negotiation: FTA between ANZCERTA and AFTA ASEAN China FTA ASEAN Japan FTA CEP Hong Kong and New Zealand Africa Southern African Customs Union Established in 1910 32

Current level of integration customs union and an integrated labour market and all members except Botswana integrated in the Common Monetary Ara with SA rand as a legal tender Current members: Botswana, Lesotho, Namibia, South Africa, Swaziland African Economic Community Established as a supranational trade body in 1994 Current level of integration bloc striving for an EU type of integration but nothing much achieved yet Current members: 51 countries Multilateral agreements World Trade Organization Established as preferential trading bloc in 1947 (GATT) Current level of integration preferential trading bloc Current members: 144 full members, plus 2 to be ratified 33

Chart A1 Geographical distribution of the regional trade agreements in the world 34