/ ? 44' £EP I992 WORLD BANK LATIN AMERICAN AND CARIBBEAN STUDIES w/~~~~~~~~~~~~~P Itet /1',,,,,,, Work in p-ogre.. for publii disssio- ANNUAL WORLD BANK CONFERENCE ON 1997 ISA~~~~~~~~~ F--WA R B S- -EGI O*ML?LSM.. }. _ t- O -t S -M~~~~~~~~~MNTVIE I . . _ 6 . . . U A _~~~~~- - AA/1/,'/(/ I, MONTEVIDEO URUGUAY WORLD BANK LATIN AMERICAN AND CARIBBEAN STUDIES Proceedings ANNUAL WORLD BANK CONFERENCE ON Development in Latin America and the Caribbean 1997 Proceedings of a Conference held in Montevideo, Uruguay Edited by Shahid Javed Burki Guillermo E. Perry Sara Calvo The World Bank Washington, D.C. Copyright © 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing June 1998, revised September 1998 The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denomina- tions, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, with- out asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. Shahid Javed Burki is Vice President and Guillermo E. Perry is Chief Economist of the World Bank's Latin America and the Caribbean Regional Office. Sara Calvo is senior economist of the Poverty Reduction and Economic Management Network Central Unit. The planning and organization of the Third Annual World Bank Conference on Development in Latin America and the Caribbean was a cooperative effort of the Central Bank of Uruguay and the World Bank's Latin America and the Caribbean Regional Office and Economic Development Institute. Particular gratitude is due to Hector Ribeiro, Public Relations Manager of the Central Bank of Uruguay, L. Kumar Arora, Conference Officer, and Lisa Taber of the World Bank, who were responsible for the general coordination of the conference; and Monica Echeverria for organizing external affairs matters. We also thank Dulce Afzal, Liliana Pefia, Jorge Serraino, Angelica Silvero, and Jocelyn Troncoso of the World Bank, as well as staff of the Central Bank of Uruguay, whose collaboration was essential for the event. Cover photo: K. Scholz/H. Armstrong Roberts ISBN 0-8213-4272-X ISSN 1020-6140 Contents Introductiorn .................................................................... . 1 Towards Open Regionalism .............................................................. 3 Shahid Javed Burki and Guillermo E. Perry I. The Challenges of Regional Integration .................................................. 11 The FT11A is Not Free Trade ............................................................. 13 Jagdishl Bhagwati The New Face of Regional Integration .............. ....................................... 20 Enrique V Iglesias Missing Lessons of East Asia: Openness, Education and the Environment ............................. 30 Vinod 1rhomas and Yan Wang II. Assessment of Regional Integration . .................................................... 49 Assessing Regional Integration ........................................................... 51 L. Alan Winters Comment ................................................................... 69 Robert Devlin Comment ................................................................... 72 Ricardo Ffrench-Davis Ill. Andean Pact ................................................................... 77 Trade Flows in the Andean Countries: Unilateral Liberalization or Regional Preferences? .................. 79 Juan J(Dse Echavarria Comment ................................................................... 103 Saratht Rajapatirana IV. NAFTA ................................................................... 107 NAFTA,: An Interim Report ............................................................. 109 Jeffrejy J. Schott Comment ..1.... 125 Nora Lustig iii TRADE TOWARDS OPEN REGIOŽNALISINM V. Mercosur ...... 129 Mercosur: Objectives and Achievements . 131 Sam Laird VI. Roundtable: The Small Economies . 153 Luis Mosca . 155 Fernando Naranjo . 158 VIl. Roundtable: Agricultural Policy and Integration . 163 Bruce L. Gardner . 165 Roberto Junguito . 170 Alberto Valdes . 176 VIII. Infrastructure and Integration . 181 Promoting Latin American Trade: The Role of Infrastructure . 183 Sri-Ram Aiyer and Hans J. Peters Comment . 206 Ernesto Tironi IX. Integration of Financial Services . 211 The Integration of Financial Services in Latin America . 213 Robert Zahler and Carlos Budnevich X. Macroeconomic Policy Coordination . 229 Free Trade and Macroeconomic Policy . 231 Barry Eichengreen Comment . 247 Claudio Loser Xi. Roundtable: Is Protectionism Still Alive? . 251 Richard Blackhurst . 253 Alejandra Cox Edwards . 256 David Palmeter . 260 T. N. Srinivasan . 263 XII. Roundtable: The Future of Regional Integration . 267 Richard Feinberg . 269 Roberto Frenkel . 274 Introduction Towards Open Regionalism SH A H I D J AVE D BURK I GU I LLERMO E. PERRY I T IS A GREAT PLEASURE TO WELCOME EVERYONE TO THE THIRD ANNUAL BANK CONFERENCE ON Development in Latin America and the Caribbean (ABCD-LAC). Our first conference, held in Rio de Janeiro, focused on the challenges of reform. Our second conference, held in Bogota, focused on pcverty and inequality. And this, our third conference, here in Montevideo, will focus on trade. Thius, aptly, the title of this conference is "Trade: Towards Open Regionalism." Each subject these ABCD-LAC conferences has covered has been of vital importance to the future stability and success of Latin America and the Caribbean. This year's subject, trade, is of equal importance, and why that is so will become explicitly clear as we hear from our various speakers, discussants and other participants during the next two days. Trade policy can have a significant impact on the health or poverty. Furthermore, trade liberalization ensures that sickness of our economies and, thus, on our future well-being. long-term economic growth in LAC will be more labor Trade policy can make all the difference between stumbling intensive, which helps poverty reduction as greater shares and struggling to achieve gains for our countries, and having of national incomes are received by workers. Consequently, a truly vibrant economy that can lead to vibrant, fulfilled when most of the region was practicing protectionist, lives for our citizens. So let us understand just how important inward-looking trade policies, the World Bank was pro- trade policy is. It is not summarized simply by some mathe- moting a change in outlook. When most of the region dis- matical equation put forth by our gifted economists. At the mantled its protectionist regimes in the late 1980s and end of the day, trade policy affects the lives of our peoples, and early 1990s, the World Bank was there to help with tech- that is something we must always keep in mind. nical assistance and with structural adjustment and trade- reform loans. We also supported the establishment or The World Bank and Trade Policies in LAC strengthening of safety nets to protect the poor from any The World Bank has supported liberal trade policies for a potential short-term consequences of the rapid pace of long time. In addition to strong theoretical arguments in reform. And now we support a policy of "open regional- favor of Dpen trading regimes, there is overwhelming ism" for Latin America and the Caribbean. empirical evidence, both in LAC and worldwide, that such policies promote economic development. In turn, eco- Why Do We Use the Term "Open Regionalism"? nomic growth is a necessary condition for reducing Regional integration in LAC, or regionalism, needs to be poverty, and therefore greater openness will help reduce assessed from a historical perspective. LAC regional inte- Shahid Javed Burki is Vice President of the Latin America and Caribbean (LAC) Regional Office of the World Bank. Guillermo E. Perry is the Chief Economist and head of the Poverty Reduction and Macroeconomic Management Sector of the LAC Regional Office. 3 TRADE TOWARDS OPEN REGIONALISM gration in the 1960s and 1970s had both a positive and a paring the average tariffs imposed by its member countries negative aspect. On the positive side, the regional liberal- in 1986 with those included in the Ouro Preto Treaty ization of trade flows provided at least a limited space for objectives.' These facts leave no doubt that regionalism has greater competition, thus promoting efficiency and export gone hand in hand with unilateral trade opening in the last development in an otherwise closed environment. On the decade, in sharp contrast with what happened in the 1960s negative side, because regionalism was designed to extend and 1970s. That is one reason why we talk of "open" the rapidly declining scope for import substitution, by regionalism in LAC today. providing an enlarged space for a second phase of the pro- The same is true with respect to investment flows. tectionist model it may have retarded the process of trade Regionalism in the 1990s has gone hand in hand with a sub- and investment liberalization. stantial liberalization of investment regimes. Nowadays, sev- The revival of regional integration in the early 1990s is eral countries in LAC treat foreign direct investment (FDI) clearly a different process. The new regionalism has on exactly the same footing as domestic investment. New emerged as a by-product of the decisions made by most regional trading arrangements (RTAs) have explicit provi- governments to liberalize their economies. In general, it sions to facilitate investment flows, including national treat- has proceeded hand in hand with unilateral trade liberal- ment provisions.2 This is in sharp contrast to the restrictive ization and the opening to foreign investment. regimes of the 1960s and the 1970s in most LAC countries. Let us first consider the evolution of tariff and non-tar- Table 2 shows that in fact FDI has become an important and iff measures affecting imports from the rest of the world. increasing contributor to domestic investment. This is a sec- Figures 1 and 2 show that protectionist policies have been ond reason why we talk of open regionalism in LAC today. dismantled in the major LAC countries during the last Also, the recent surge of regionalism has not promoted decade, and especially since 1991. Furthermore, Table 1 "closed clubs" as some observers feared. On the contrary, shows the same pattern for Mercosur countries when corn- what we have seen is a trend to enlarge memberships and to FIGURE I Weighted Average Import Charges in LAC, 1984-93 (weighted by a product's share of world trade flows) 80 . . 70 1984-87 75.2 737 61988-90 63.6 60_ 1 1991-93 50 - 43.5 E 40 39- 38 37 30 - 5 20 -1. 10 95 0 Argenrina Bolivia Brazil Chile Colombia Ecuador Mexico Paraguay Venezuela Soam,; UNCTAD, D,actoy of Ioop*. Regme,s. Neo Ytrk. UNCTAD Secretariat, 199'. INTRODUCTION FIGURE 2 Weighted Incidence of Non-Tariff Measures in LAC, 1984-93 (percentages of product categories affected by NTMs, weighted by world trade flows) 90 - ....... 80 _ 1984-87 80.4 70 - 1988-90 60 - [ 1991-93 50 5 52. a - ... .. ......... .......... . . -l E 461 44.1 oP 40-. . .... . 296 32.1 2 rgni Bl.2 Chl Clma2.2 E 22.5 Eo l 20 - . UNCTAD. ....D... of .... . impo ..... N YkU A .19 --l ecti 19.4 3. 3.5>6 l 1 L L.T 1| l 4.6 2l Argentina Bolivia Brazil Chile Colombia Ecuador Mexico Paraguay Venezuela Source: UNCTAID, Dirertor y oflmrport Regirrs. New York: UNCTAD Sec,era,i.t, 1994. superimpose a host of free-trade arrangements. Mercosur These three countries have signed bilateral free-trade agree- has signed trade agreements with Chile and Bolivia, and is ments with Chile, irrespective of their other memberships. busily negotiating new ones wirh the Andean Community Finally, most LAC countries have expressed their willing- and with the European Union. Colombia and Venezuela ness to participate in the construction of a hemispheric free- signed a joint free-trade agreement with Mexico (the so- trade zone and have been active, pro-liberalization members called Group of Three); the fact that Mexico was a member of the World Trade Organization (WTO). This is a third of the North American Free Trade Agreement (NAFTA) reason why we talk of open regionalism in LAC today. Yet and that Colombia and Venezuela were members of the a legitimate question to ask is, Why has regionalism accom- Andean Pact in 1994 did not deter this development.3 panied unilateral trade liberalization? TABLE I TABLE 2 Unilateral and Regional Trade Liberalization in Mercosun Net Foreign Direct Investment as a Share of Gross Domestic Average Tariffs by Country Investment, 1980-95 (percentages) (percentages, annual averages) COUNTRY 1986 OURO PRETO OBJECTIVES 1980-84 1985-89 1990-95 Argenrina 41 12 Argentina 2.4 4.5 6.7 Brazil 80 13 Brazil 4.4 2.1 2.1 Paraguay 20 9 Chile 6.4 6.6 8.2 Uruguay 36 11 Colombia 5.3 8.0 9.9 Mercosur 44 11 Mexico 3.0 6.7 8.8 Peru 0.2 0.6 9.6 Source: Marcelo Olarreaga and Isidoro Soloaga, "Endogenous Tariff Formation: The VenezueLa 0.7 1.0 9.0 Case of Mercosur," mimeo, International Trade Division, Development Economics Department, WArld Bank, Washington, D.C., March 1997. Soarce: World Bank World Development Indicators Database. 5 TRADE TOWARDS OPEN REGrONALJISM Why Has Regionalism Accompanied Unilateral goes for investors in Colombia and Venezuela, who benefit Trade Liberalization? from greater certainty about the future access to their Although it is a legitimate intellectual exercise to ask respective markets that is preserved by the Andean Com- whether it would not have been more convenient to have munity agreements. This consideration may provide an had only unilateral liberalization without the new RTAs, important motivation for many LAC countries to sign free- we think it is more interesting to ask why unilateral open- trade treaties with the United States. Even if the United ing was accompanied by the revival of regionalism. Let me States is generally open, once in a while there are protec- refer to a similar question posed by Professor Paul Krug- tionist impulses affecting imports of particular products, man to free-trade advocates. He asks, Why is it that free- some of which are precisely those in which LAC countries traders support WTO and GATT negotiations when, on are more competitive. This reciprocal "lock-in" effect purely theoretical grounds, they should oppose them?4 In reduces uncertainty and risks, and may thus increase fact, trade theory suggests that countries should unilater- investment levels. ally liberalize to reap the benefits of freer trade, even if oth- It also makes sense to "lock yourself in." Making liber- ers do not do so. Then why doesn't it happen? Why don't alization credible is often a necessary condition to ensure we see unilateral trade liberalization taking place through- that sound economic policies will yield their benefits. If out the world? Why do we need an international organiza- investors fear that liberal policies may be reversed, they tion like the WTO? may invest less or nothing at all. "Locking in" through Of course, political economy considerations come into RTAs increases the credibility of trade and investment play. There are interests that oppose unilateral trade liber- opening, because RTAs raise the "costs" of policy reversals. alization; there are those who think in mercantilistic terms "Entry costs" also enhance credibility, because the negotia- ("exports are good, imports are bad") even if economic the- tions leading to RTAs and their implementation require ory would prove them wrong. When free-traders support the use of public resources and political capital, which sig- WTO negotiations, they are implicitly accepting these nal a government's commitment to trade and investment constraints. liberalization. Why would a government go through all Let me suggest that there may be similar motives dri- these lengthy negotiations if it is not really committed to ving the process of unilateral liberalization accompanied keeping the economy open? Notice that these arguments by RTAs. It is easier to have public opinion on your side, that emphasize "locking-in" effects imply that the eco- and to counteract opposing interests, in favor of "integra- nomic motivation behind RTAs is more related to their tion" with your neighbors and/or for securing access to for- effects on investment than on trade itself. 5 eign markets for exporters through reciprocity, than it is It should not come as a surprise to anybody that the new for unilateral liberalization. A policy "package" that con- RTAs have come at the same time as the new surge in FDI tains both unilateral liberalization and RTAs with your and other capital flows in a world of higher financial inte- trading partners is likely to receive greater political sup- gration. Countries want to share the bonanza by giving port than one that is limited to unilateral liberalization. more credibility to their policies and more certainty to But this is only one reason why policymakers may have their access to foreign markets. Of course, the credibility of chosen to engage in both processes simultaneously. Let me macroeconomic policies may rank higher in investor con- also suggest that securing access to crucial markets may cerns than credibility of trade policies, but the latter may indeed be worth some costs. Even if your trading partners be a useful complement to credible stabilization and struc- liberalize today, they may change their minds in the future. tural reforms. To "lock-in" access to export markets by imposing costs to In sum, there are both political economy and purely "exiting" RTAs makes some sense. For example, investors economic arguments that support the political decision to in Mexico appreciate the more secure access that NAFTA move forward with unilateral trade liberalization and provides to the U.S. market. Investors in Argentina, regionalism. In any event, there is little doubt that they Uruguay or Paraguay similarly appreciate the insurance were part and parcel of the same political decision to inte- that the Mercosur treaty provides regarding their gained grate LAC economies into the global economy, and we access to the Brazilian and each other's markets. The same could even argue that one process would not have hap- 6 INTRODUCTION pened without the other in many LAC countries. There is, TABLE 3 thus, a case to judge jointly the outcome of both processes Trade Growth in LAC and not to attempt an artificial separation in the analysis. (growth rates of trade value in US$, exports FOB, imports CIF) There are other non-economic reasons for joining RTAs. A. EXPORTS FROM LAC A well-known argument is that integration enhances the TO 1986-90 1991-95 capacity to influence worldwide outcomes. Although it is WORLD 6.4 11.8 doubtful that RTAs significantly enhance the voice of 1. LAC 11.3 16.0 developing countries in the WTO, it is a fact that a strong 2. ROW 5.7 11.0 Mercosur has changed the landscape for future negotiations INDEX 5.0 4.0 of a free-trade zone in the Western Hemisphere. B. IMPORTS TO LAC A new argument is that RTAs may contribute to demo- FROM 1986-90 1991-95 cratic consolidation if maintaining a democratic regime is WORLD 9.3 17.7 a prerequisite for continued membership. This type of 1. LAC 8.6 16.8 "political lock-in" consideration weighs heavily in today's 2. ROW 9.5 18.0 efforts of Central European countries to join the European INDEX -1.0 -1.0 Union. While it may be less important in LAC, last year's events in Paraguay suggest that this may indeed be an INDEX ={ £(l+LAc)/(lROW)u-ls*100; if INDEX > 0, intra-LAC trade growth important consideration for LAC countries. A less-publi- was faster than trade with ROW cized political consideration is that RTAs, by strengthen- Notes: LAC includes the developing countries of the Western Hemisphere (i.e., intra- LAC trade does not include trade berween LAC countries and Canada and the United ing business links across countries, reduce the likelihood of States) conflicts between countries that have border disputes. A In principle, intra-LAC exports and imports should be equivalent, but the differences similar "peace lock-in" effect has been a prime mover in this data reflect the incidence of transportation costs; hence the difference between the rate of growth of exports from LAC to LAC and the growth rate of imports from behind the EEC initiative since the 1950s. LAC to LAC. Soarce: International Monetary Fund, Diecrtion of Trade Statistics. What about the Results? Coming back to hard economic facts, overall trade trends duced evidence of this sort of effect for specific sectors in in the 1990s provide justifications for optimism. It is true some regional groupings. Such cases are usually the out- that LAC intra-regional imports and exports grew rapidly come of exceptions to the general policy: They are rem- in the first half of the 1990s, but imports from the rest of nants of trade distortions of the past that were designed to the world grew even faster than intra-regional imports. protect specific sectors. Such distortions do not promote Moreover, exports to the rest of the world have grown economic efficiency, nor do they help the poor, and thus, as somewhat less rapidly than intra-regional trade, but still at I have said before, we are all obliged to point them out, faster rates than during the late 1980s (see Table 3). When hoping that they will be corrected as soon as possible. Such one looks at the data for each subregional grouping, it may be the case with those sectors under special regimes or becomes evident that intra-regional imports have grown higher common external tariffs in Mercosur. We are well more rapidly than imports from the world in all cases aware, of course, that in some instances there may be some except CARICOM. Howcver, trade with rhe rest of the room for gaining cconomies of scale and dynamic bencfits world has in all cases grown at very high rates, certainly that partly compensate for some of those hopefully short- higher than in the late 1980s (see Table 4). term inefficiencies. In general, though, we would urge the These aggregate figures cannot be taken as definite countries to accelerate convergence toward lower and more proof that healthy trade creation has outweighed any inef- uniform tariff levels. ficient trade diversion, but they certainly do not give much There are also other problems created by recent devel- support to the opposing argument. opments. The array of overlapping trading arrangements is It is true, however, that we are concerned about ineffi- creating distortions and difficulties, especially in the cient trade diversion in some instances. Several studies, enforcement of rules of origin and other complex techni- conducted at the World Bank and elsewhere, have pro- calities. As Enrique Iglesias, one of the champions of 7 TRADE TOWARDS OPEN REGIONALISM TABLE 4 U.S.-based FDI reacted more vigorously to the deregula- lntra-Regional Import Growth in LAC, 1986-90 and 1991-95 tion of FDI in Mexico in the second half of the 1 980s than (growth rates of imports value in US$, CIF) to NAFTA.' ln fact, data published by the U.S. govern- 19s6-90 1991.95 ment shows that U.S. direct investment in Mexico grew at Intra-MERCOSUR 18.5 24.6 a much slower pace than that in other LAC countries dur- From ROW 7.9 19.7 ing the early 1990s (see Table 5). INDEX 9.9 4.0 There has been a distinctive jump of total FDI flows to 1986-90 1991-95 the Andean Community. Their share of FDI flows to the Intra-ANDEAN 9.8 34.6 LAC Region increased substantially in the 1990s, which From ROW 2.7 16.0 INDEX 6.9 16.0 may be related to the vigorous market integration between Colombia and Venezuela. In contrast, there is no such evi- iQIs6-p)o 1991-95 dence in the case of Mercosur (see Table 6), though last Intra-CACM 3.0 20.8 year's figures looked promising. Changes in the composi- From ROW 5.7 t5.6 tion of FDI in Mercosur countries seem more related to pri- INDEX -2.5 4.5 vatization efforts than to Mercosur itself. However, highly 1986-90 1991-95 protected sectors in Mercosur countries have been attract- Intra-CARICOM 6.7 -8.0 ing a higher proportion of that investment, a development From ROW 0.4 5.3 INDEX 6.2 -12.6 that may prove to be not so desirable in the long run. In ROW - Rest of World other words, FDI flows motivated by the market expansion INDEX = [YI +LAC)!iI+ROW)1-11*100 When INDEX 0, n-ra-LAC imports effects of RTAs tend to be more welfare enhancing than growing faster than imports nto, ROW. FDI flows driven by "tariff-jumping" motivations. MERCOSUR = Argentina, Brazl, Paraguay and Uruguay That said, we must acknowledge that the jury is still ANDEAN = Bolivia. Colombia., Ecuador, Peru and Venezuela CACM = Costa Rica. El Salvador, Guatemala, Honduras and Nicaragua out evaluating the outcomes of the recent revitalization of CARICOM = Bahamas, Barbados, Belize, Guyana, Jamasca, Sutoarne and Trinidad & Tobago regionalism in LAC. Progress has been uneven across Soare, Internarional Monetary Fond, Direction of TYrad Stzostza. TABLE 5 regional integration in LAC, stated in his speech at the United States: Average Annual Growth Rate of Direct Investment recent Inter-American Development Bank annual meeting, in LAC, 1991-95 it is time to begin to tackle some of these problems in a (rate of change of dollar value of direct investment, measured systematic way. Also, it is particularly worrisome that at cost value) some of the small economies, particularly in Central Amer- GROWTH RATE 1995 SHARES ica and the Caribbean, have been left out, temporarily, of To the World 11.1 some of the action and have consequently suffered some Latin America & the Caribbean 12.2 17.3 erosion of their previous preferential access to some mar- (of total) kets without gaining new terrain. South America 17.6 38.3 What is the evidence on FDI flows? In the case of Mex- (of total to LAC) ico, there appears to have been a strong response to Argentina & Brazil 28 25.7 Chile 28.8 4.5 NAFTA from non-NAFTA countries, but not from U.S. Andean Group firms. The combination of locational advantage and secure (Col, Ecu, Per. Ven) 24.3 7.2 and preferential access to the U.S. market was a powerful Mexico and Central America 7.0 25.6 incentive for non-U.S. firms to establish themselves in Mexico 3.3 11.4 Mexico. Such access was, of course, less important for U.S.- ' ~~~~~~Caribbean 11.5 36.2 based FDI. Some U.S.-based firms now have fewer incen- tives to invest in Mexico to have access to its market. In Saoer.: U.S. Department of Commerce, Surrey ofCaentz Basess, June 1996. other words, "tariff-jumping" incentives to invest in Mex- N5tOre< Based on data on the direct investment position of the United States in LAC, ,huch measuLres che value of the net accumulated stock of capital that U.S. parent ico disappeared for U.S.-based firms. As a matter of fact, companies provide to these foreign affiliates. 8 I N TROD L CT ION TABLE 6 process of building an open regionalism in the Western Growth of Nei FDI to LAC Groups Hemisphere. We want to help promote a healthy, open (average annual growth rates of US$ values) regionalism and freer trade for the well-being of all LAC SHARE OF populations. TOTAL We have invited a very distinguished group of policy- FDI TO LAC makers academics and members of the private sector and REGION OR COUNTRY 1986-90 1991-95 1995 m MERCOSUR 37 21 25 multilateral institutions. Keynote speakers include Presi- ANDEAN 10 55 22 dent Enrique Iglesias of the IDB, Professor Jagdish Bhag- CACM 19 21 2 wati of Columbia University and Mr. Joseph Stiglitz- CARICOM 122 13 3 until recently Chairman of the Council of Economic MEXICO 12 38 27 Advisors for the U.S. government, and currently Chief CHILE 58 31 7 Economist of the World Bank. We thank them all for their To all LDCs 32 cooperation and participation. And we are especially grate- MERCOSUR = Argentina, Brazil, Paraguay and Uruguay ful to our host, Minister of Finance Luis Mosca, for the effi- ANDEAN = Bolivia, Colombia, Ecuador, Peru and Venezuela cient organization of the conference and the warm welcome CACM = Costa Rica, El Salvador, Guatemala, Honduras and Nicaragua CARICOM = An,igua & Barbuda, Bahamas, Barabdos, Belize, Dominica, Grenada, we have received. Jamaica, St. Kitts, St. Lucia, St. Vincent and Trinidad & Tobago We look forward to a productive and enjoyable Sourres International Monetary Fund, International Financial Statistics, meeting! groups of countries and sectors. Action in trade in services Notes and agricultural products has lagged behind other lati- 1. The presidents of Mercosur countries met in Ouro Preto, tudes globally. Improving the infrastructure for integra- Brazil, on December 17, 1994. The "objectives" were the tariff lev- . . ~~~els that will be in place by 2006. tion remains a challenge. The case for harmonizing some 2. Such provisions are included in the agreements or treaties of trade-related policies, such as subsidies, indirect taxation, tefloigACrangragmns:Mcsu,ndnCo- > > v r~~~~~~~~he following LAC trading arrangements: Mercosur, Andean Com- and competition policies is becoming more compelling. munity, Central American Common Market (CACM), Caribbean There is also the difficult questiori of how much macroeco- Community (CARICOM), NAFTA, Group of Three (Colombia, nomic policy coordination will be needed as integration Mexico, Venezuela) and several bilateral agreements. deepens. Finally, at some point a dilemma may arise 3. The Andean Pact became the Andean Community of Nations between "deepening" integration under current schemes, in 1997. l4. Paul Krugman (1997). "What Should Trade Negotiators expanding memberships, and building a new free-trade Negotiate About?"JosurnalofEconomicLiterature35 (March): 113-20. area in the Americas. 5. These arguments are further elaborated in Raquel Fernandez We expect to cover all these rich topics in this confer- (1997). "Returns to Regionalism: An Evaluation of Non-Traditional ence. The purpose of the World Bank in choosing this sub- Gains from RTAs," mimeo, International Trade Division, Develop- ject matter for its Third Annual Bank Conference on ment Economics Department, World Bank, Washington, D.C. 6. See Magnus Blomsrrom and Ani Kokkc (1997). "~Regional Development in LAC is to promote a wide discussion of all Integration and Foreigna Direct Iivestmeiit," mimeo, Iriternational issues that will be crucial for the success of both regional- Trade Division, Development Economics Department, World Bank, ism and development in the region. This choice under- Washington, D.C. scores the importance the World Bank attaches to the 9 I I. The Challenges of Regional Integration The FTAA is Not Free Trade J A GD I S H B HA GWATI *t s HEN THE CLINTON ADMINISTRATION'S SPOKESPERSONS, OR PERHAPS WE should call them spin-masters, recently released the results of the commissioned study by the consulting firm DRI of NAFTA's economic effects, the public debate was so flawed analytically that it might as well have been taking place in a coun- try populated by illiterates. Even the Lehrer News Hour, our justly celebrated f T ~~intellectual news show, had Ambassador Mickey Kantor and Ms. Thea Lee debat- ing the employment effect of NAFTA, when trade policy can generally affect employment only in specific sectors, not economy-wide. The public debate over NAFTA, then and now, has suf- between FTAs and FT, and because of the legacy of Ross Perot fered from a worse, even crippling confusion. Few in the during the NAFTA debate, when to be anti-NAFTA came to media, and fewer still in the administration and Congress, be popularly regarded as being protectionist. understand that there is a world of difference between a This fall, as we move in the direction of the FTAA under free-trade area (FTA), which moves to free trade only the false banner of free trade, there is still a window of among its members, and free trade (FT), which lowers opportunity to introduce Washington to the real issues trade barriers for all. Indeed, there are now many well- about FTAs and even to stop the train in its tracks: Ideas can known free traders among economists who oppose FTAs be potent. I intend, therefore, to argue here why FTAs, and consider them a pox on the world trading system, so including the FTAA, are a bad idea, and to outline the alter- that the widespread assumption in the policy debate over native trade policy that we should embrace. NAFTA arLd Free-Trade Area of the Americas (FTAA)- that free traders support them and protectionists oppose Why Are FTAs Bad News? them-is rank nonsense. An FTA, because of the inherent discrimination that it As it happens, the idea of the expansion of NAFTA to implies in freeing trade, is different from free trade. Chile, and eventually of an FTAA, which the Clinton admin- Indeed, its flip side is protectionist-protectionist against istration is wedded to, is a mistake. There is an alternative non-members, against whom the relative protection is trade policy that would better fit free trade as well as our increased because barriers fall in favor of members, while regional objectives. But it has failed to be debated because the ones against non-members remain in place. Recalling the administration has consistently fudged the distinction Orwell, therefore, serious economists have now abandoned Jagdish Bhagwati is Arthur Lehman Professor of Economics and Professor of Political Science at Columbia University. He is author of Protec- tionism (MIT Press, 1988) and The World Trading System at Risk (Princeton, 1991) and was Economic Policy Adviser to the Director General of GATT. 13 TRADE: TOWARDS OPEN REGIONALISM the phrase Free-Trade Agreements for one that reminds us envelope estimate by University of Maryland economist of what they truly are: Preferential-Trade Agreements Arvind Panagariya suggests that Mexico's annual loss from (PTAs). Their preferential nature, and the huge growth in this trade diversion could be as high as $3 billion. their number in the last decade, has raised alarms. The growing empirical evidence of trade diversion in Ti-ade Diversion: An important reason, first noted by these preferential-trade agreements is, in fact, not surprising Jacob Viner in 1950 in a pathbreaking study for the because there has been abundant evidence that many busi- Carnegie Commission, is that the preferences create incen- ness lobbies prefer them to genuine, multilateral free trade tives to divert, not just create, trade. Thus, member coun- because they give them preferred access to foreign markets at tries, enjoying the enhanced protection vis-a-vis more effi- others' expense. Just recall President Clinton's resort to cient non-members, may replace them as producers and Japan-bashing during the NAFTA debate: that it would exporters within the PTA, distorting world trade and give our firms a better access to Mexico over the Japanese. likely hurting themselves, since they buy imports more Economic theory, which often lags behind economic phe- expensively from within the PTA. nomena, has now demonstrated (in the work of Gene Gross- The analytically informed proponents of PTAs have man and Elhanan Helpman and Pravin Krishna) that, under claimed that trade diversion is negligible in practice; plausible assumptions, trade diversion provides the key Lawrence Summers, U.S. Deputy Secretary of the Treasury, motivation for opting for preferential-trade agreements. It is has even remarked with characteristic forthrightness that not remarkable, therefore, that those who indulge their pref- he finds it "surprising that this issue is taken so seriously" erence for preferential-trade arrangements are seen also to and that it is a matter to "laugh off." But increasing num- use the preferences to their advantage! bers of empirical studies are now beginning to show that It is also worth noting that the complacency about trade diversion is not a negligible phenomenon. Thus, PTAs not leading to trade diversion has also been aided by World Bank economist Alexander Yeats's 1996 study of the notion that the external trade barriers are no longer Mercosur (including Uruguay, Argentina, Paraguay and very high and that, therefore, preferences could not lead to Brazil) turned up significant evidence of trade diversion.' significant trade diversion away from the non-members. The economists Jeffrey Frenkel and Sheng-Jin Wei have But this is not true for several reasons. also concluded recently that the EU also has been charac- To begin with, even trade tariffs are still very high, both rerized by more trade diversion than hitherto believed. The in developing and in developed countries. In the latter, the diversion of textile trade to Mexico from the Caribbean Uruguay Round has still left several peak tariffs in specific thanks to NAFTA, which excludes the latter and includes products, whereas the tariffication of agricultural support the former, has also been a source of discord. has created truly substantial tariffs. In the developing Indeed, if one examines the administration-released world, countries in South Asia and in Latin America are DRI study of NAFTAs effects, one immediately finds also not free from high trade barriers, making PTAs par- prima facie evidence of trade diversion plaguing Mexico as ticularly dangerous. well. Thus, according to the DRI study, while the peso cri- Besides, the external trade barriers are today only a part sis generally contracted Mexican domestic demand by 3.3 of the protectionist story. "Administered protection," con- percent over the period, U.S. exports to Mexico increased sisting of instruments such as anti-dumping (AD) actions, by 36.5 percent (or $15.2 billion) from 1993 to 1996. U.S. has become the favored policy of protectionisrs who clev- share in Mexican imports thus went up from 69.3 percent erly use the appealing notion of "fair trade" to unfairly gain to 75.5 percent. Equally, since NAFTA was signed, the protection and advantage against successful foreign rivals. Mexican tariffs on U.S. goods were reduced by an average But then these instruments typically yield protection that of 7.1 percentage points, resulting in "a 10 percentage is elastic and selective. Thus, AD duties, which bear little point average tariff advantage over foreign suppliers." The relationship to "predation" in the economic sense and Mexican government should itself worry about this strong hence have in practice no economic justification, are often evidence of trade diversion toward the United States, based on adjusted prices that are estimated in ways calcu- which is celebrated in the DRI Report as if it were evi- lated to fund dumping or on inherently arbitrary "recon- dence of NAFTA's success! In fact, a quick back-of-the- structed costs"; the AD methodology compares not foreign 14 THE CHALLENGES OF REGIONAL INTEGRATION and domestic prices, but foreign costs and domestic prices. The French economist Patrick Messerlin has also writ- Thus, with.in broad margins, it is arguable that AD calcu- ten of the effect the EU association agreements, leading lations and actions will seek to accommodate the needs of toward eventual entry into the EU, had on the Central the protectionist petitioners in the spirit of the story where European countries' acceptance of obligations at the WTO. the interviewing commissar asks candidates what the sum These nations, coming from the collapse of communism of two and two is, and the job goes to the candidate who and the absence of significant protectionist lobbies in their answers: Whatever you want, sir. In addition to their midst, had embraced a strong policy of MFN-based trade being, therefore, elastic, AD actions are selective in the sense liberalization. At the insistence of the EU, which was that they are mounted against specific countries and even interested in getting preferred access to these markets, the specific firms within those countries. Thus, it is possible to eventual acceptance of WTO obligations by these coun- use them tc, zero in against your most potent foreign rivals. tries was marked by a raising of their MFN tariffs! It follows then that, when internal competition among The "Spaghetti-Bowl" Phenomenon: But if individual PTAs members breaks out, the temptation on the part of PTA are flawed in this way, they have raised added concerns members will be to protect each other with administered because of the "systemic" implications when many PTAs protection at the expense of non-members, unless such pro- have emerged on the scene, as they have since the 1980s. tection is severely regulated. In short, protection against These systemic implications arise because such preferen- non-members then becomes endogenous to the PTA. The con- tial-trade agreements magnify the problems that arise, in sequence is that, as trade creation occurs within a preferen- essence, because we try to restrict or liberalize trade on the tial-trade agreement such as NAFTA, the endogenous rais- basis of which product comes from which country, or what ing of protection converts it into trade diversion instead. For I have called the "who is whose" problem. Thus, for example, as Mexico starts crowding out inefficient U.S. pro- instance, as soon as the United States wishes to liberalize ducers, the United States accommodates imports from Mex- preferentially the imports from Israel, it must decide ico by reducing imports from the most efficient non-mem- whether an import coming from Israel is Israeli-i.e., it ber supplier, Taiwan, using AD actions against Taiwan. must establish a "rule of origin," which usually takes the Such a phenomenon is not an idle theoretical specula- form of some sort of "content" rule, such that a product is tion. Important instances of such endogenous raising of considered to be Israeli only if its Israeli content exceeds an protection against non-members have been observed. An arbitrarily specified share in gross value.2 example is the raising of tariffs on more than 500 non- The arbitrariness of this share specification is further com- NAFTA tariffs by Mexico during the 1995 peso crisis pounded by the arbitrariness inherent in computing such con- while the Mexican tariffs on NAFTA items remained on tent. Thus, consider the import of steel ingots that, in con- the downward path. The former could be raised because junction with homogeneous domestically produced ingots, Mexico had bound tariffs above existing levels at the go into producing scissors and forks. How is one to determine WTO; anc[ it could be argued that the reason for taking which of these two products got what share of the imported such weak bindings was precisely that, with NAFTA, Mex- as against the domestic ingots? Again, if forks need to be ico wanted to hold on to the freedom to raise some tariffs coated with plastics, we know that even if the plastics are if deemed necessary, implying that the potential for trade immediately produced at home, their gross value would gen- diversion was implicitly built into the NAFTA treaty. erally include imported intermediates at several stages of Thus, Treasury Secretary Robert Rubin, in claiming (as manufacture, which are impossible, for the same reasons, to does the DRI report) that, unlike 1982, NAFTA had made identify and quantify meaningfully. Again, even if we were to Mexico abstain from the use of trade restrictions to deal estimate such imported shares meaningfully, the imports are with the 1995 peso crisis, was wrong on two counts. Mex- likely in turn to include, in today's globalized production, ico did raise some tariffs, but on other countries and not on intermediates produced by us and used by the producers us, whereas Mexico has been no different from others in its abroad. The difficulties are myriad, even endless. policy reaction to a currency crisis: Trade restrictions as the All of these problems, which inherently lead to absurd knee-jerk, big-bang response to them are generally not arbitrariness in trying to identify the origin of products, seen as appropriate, whether a country has an FTA or not. are particularly acute with an FTA where the different 15 TRADE. TOWARDS OPEN REGIOZNALISM external tariffs inevitably require that the origin be estab- mon external tariff and internal factor mobility, and possi- lished for virtually all traded products. In FTAs, the fear of bly even a parliament and attempts at a unified foreign pol- non-member goods coming into one's territory at a lower icy, then the member nations are forming something close tariff than one's own, simply by entering through another to the federal United States or India. In this case, the trade lower tariff member country, is palpable. dimension cannot be judged in isolation from a very much In reality, FTAs have created yet further problems by bigger picture. Indeed, the overriding political need to having many different rules of origin, varying by products countervail the nascent Soviet threat with a politically (as in NAFTA, for instance) and by FTAs (when, say, the united Europe and the fact that the treaty would aim at EU has FTAs with different rules with several different deep economic integration in Europe and not just trade non-EU countries). The problems inherently posed by the preferences the way an FTA does, were the reasons for the rules of origin are further compounded since FTAs are on benign and supportive U.S. attitude to the 1957 Treaty of different schedules of tariff-cutting by sector and are not Rome. This is true even though our view of PTAs was synchronized (having been negotiated at different points of quite clearly hostile, as it had been about Britain's Imper- time and with different schedules for reaching zero tariff ial Preference during the GATT negotiations and had outcomes), so that we typically find a large and chaotic set indeed been throughout our two-centuries' history. Any- of applicable tariffs on the same good, depending on the one familiar with George Washington's Farewell Address source to which the good is assigned. or with Woodrow Wilson's Fourteen Points will recall that The result is what I have called the "spaghetti-bowl" this is so.4 phenomenon of numerous and crisscrossing PTAs and Second, there is what international economists now call innumerable applicable tariff rates arbitrarily determined the "dynamic time path" justification for PTAs. This has to and often depending on a multiplicity of sources of origin. do with the possibility that, even though we seek world- In short, the systemic effect is to generate a world of pref- wide multilateral free trade, multilateral trade negotiations erences, with all its well-known consequences, which (MTN) at the GATT, now the WTO, may not be the way increases transaction costs and facilitates protectionism. In to get there. We thus need to distinguish between "process the guise of freeing trade, PTAs have managed to recreate multilateralism" and "outcome multilateralism." the preferences-ridden world of the 1930s as surely as pro- The fact remains that the United States, faced in tectionism did at the time. Irony, indeed! November 1991 with opposition by the EU and many And thus in the debates on GATT, international econo- developing countries to the declaration of another MTN mists who have been either ambiguous or complacent in Round to follow the Tokyo Round, felt that it had no regard to PTAs have acquired a proper appreciation of option except to shift to PTAs, legitimate under GATT's Keynes's famous words in the House of Lords, extolling the Article XXIV, as a way of keeping trade liberalization virtues of multilateral MFN-based trade and rejecting the going. Hence came the initiative under U.S. Trade Repre- preferences he had earlier defended.3 sentative Bill Brock for CUFTA between Canada and the The separate blocs and all the friction and loss of friend- United States, the first break in our policy of undivided ship they must bring with them are expedients to which loyalty to MFN multilateralism. It is important to recall one may be driven in a hostile world where trade has ceased that the motive, contrary to what it became later, was not over wide areas to be cooperative and peaceful and where regionalism in the Americas. The motive, rather, was are forgotten the healthy rules of mutual advantage and purely that of trade liberalization; and the intention was to equal treatment. It is surely crazy to prefer that. expand the FTA in any direction, whether in our region or elsewhere. Bill Brock is regarded as having seriously con- Two Qualifiers sidered adding Egypt and the ASEAN countries but found Crazy to prefer PTAs, yes. But still, there are two circum- no takers. We can be certain that if life had been found on stances where a staunch anti-PTA, multilateral free trader the moon, and a government to negotiate with, Brock would make an exception. would have hawked an FTA to it as well. First, if the PTA takes the form of deep integration in This second reason for FTAs then is simply that, if the form of an EU-core style Common Market, with com- MTN is not working to liberalize trade worldwide, we will 16 THE CHALLENGES OF REGIONAL INTEGRATION take the PTA route to freeing trade worldwide, inefficient preferred access-as in the case of NAFTA and, more as PTAs can be en route. In short, if the turnpike is not important, Central European countries that were already open, we take the dirt road. cited. Again, Sebastian Edwards of UCLA, who was the World Bank's Chief Economist for Latin America, No Excuse Now for PTAs by the United States recently argued in The Wall Street Journal that Chile's There is really no reason now for us to keep going on the progress toward MFN liberalization had slowed down as dirt road. The turnpike is open. The Uruguay Round con- a result of the different regional PTAs it was getting cluded successfully. GATT, pronounced by the summarily involved in. Thus, to those who say cutely that we will do skeptical Lester Thurow to be dead, did die but joined the well by following a "GATT plus" approach, using PTAs blessed twice-born, appearing as the much strengthened alongside multilateralism, I say: It is likely to work as and institutionalized WTO. The unfinished business of the well as the "marriage plus" approach worked with Demi Round, set on the negotiating agenda, has also been pro- Moore in the film, Indecent Proposal. ducing results: the telecommunication pact, the informa- tion technology agreement, and pretty soon, without What about Regionalism for the Americas? doubt, there will be a multilateral agreement on financial So, clearly there is no good case for us today to be pursuing services. These multilateral agreements happen to be trade liberalization through the NAFTA extension and backed by strong export lobbies that will profit from them, through other FTAs. Indeed, FTAA remains at the and they have no countervailing import-competing lobbies moment the only "big-ticket" item on the PTA agenda arrayed against them. So, in fact, the renewal of fast-track to with a possible political future. get them through Congress is not necessary. Indeed, unless Thus, the Asian members of APEC have refused to date one goes for a new MTN Round, which the Clinton admin- to turn it into an Article XXIV-sanctioned FTA despite istration has not committed itself to, there seems to be no our clear desire in that direction. Instead, they have compelling reason to immediately renew fast track legisla- insisted on undertaking trade liberalization under APEC tion as far as multilateral trade treaties are concerned. auspices only on an MFN basis. Partly this is because the There is also little evidence for the argument, often successful Asian exporters have never wanted to go regional advanced, that we should go for both MTNs and PTAs or subregional in trade; the whole world was their market. because the latter will help the former move forward. Fred Partly, the politics of it would be absurd for them since an Bergsten argues that the Seattle summit of APEC moved FTA would then exclude the EU, and hence the previous the EU intc settling the Uruguay Round by threatening colonial powers with whom many Asian countries have a Europe with an American trade alternative of regionalism special relationship. Just imagine President Salinas enter- that would exclude it from the trade game. But this is fan- ing into an FTA that excluded the United States. Then ciful since there are plenty of indications that the EU was again, the Trans-Atlantic Free-Trade Area (TAFTA), about to settle anyway; besides, we made the key trade con- floated by Foreign Minister Klaus Kinkel of Germany and cession-in agriculture-which finally broke the logjam. embraced by the Atlanticist Foreign Minister Michael Equally, APEC is cited as having led to the information- Rifkind in Prime Minister Major's government, has given technology agreement three weeks before the first WTO way now to a non-FTA transatlantic initiative. ministerial meeting in Singapore; but this is nothing but a In fact, a few weeks ago the European Council of Min- "post-hoc-eigo-propter-hoc" fallacy. Our intensive work isters, responding to the growing concerns about the pro- would have produced the agreement at Singapore even if liferation of PTAs and its own role in creating this prob- there had keen no APEC, for sure. Besides, APEC is a lem, resolved that the "current architecture" of its trading "regional" arrangement but is not a PTA. system would be frozen. In practice, it means that no new On the other hand, there are disturbing examples of PTA initiatives will be undertaken without far more acute how PTAs have undermined the multilateral liberaliza- examination and skepticism than was the case to date. The tion process instead, by holding up concessions by coun- WTO also has an active Committee on Regionalism whose tries on an MFN basis because partner countries, espe- task it is to analyze the working of Article XXIV in light cially hegemnonic ones, in the PTA look for and get of the growing tide of skepticism about PTAs. 17 TRADE ITOWARDS OPEN REGIONALISM One would think that the Clinton administration An Alternative for the Americas: Return to would take heed as well and desist from further indul- Camelot gence of FTAs, reverting instead to multilateral trade ini- The policy option for the United States is entirely clear. tiatives. But it has not. Part of the reason, of course, is We should revert to exclusive focus on multilateralism and intellectual laziness. Partly it is also the fact that no inter- MFN-based trade liberalization, asking the South Ameri- national economists of repute have ever had access to the can nations to join in the multilateral opening of markets White House, where the main concern has been the poli- through a variety of initiatives, while pursuing "regional- tics of its trade policies, not their economics. Then again, ism" and its separate objectives through non-PTA means. there is the inertia of past decisions: The FTAA is in effect In short, we should renounce the FTAA gracefully, easing the legacy of the Baker transformation of the open, non- into an Americas Initiative that focuses, like APEC, on regional FTA initiative into a closed, regional initiative issues like security, democracy, human rights, drug traf- for the Americas. ficking, customs procedures, and a whole host of issues of Matching the American folly is much of South Amer- hemispheric interest, while becoming a regional platform ica's own misguided enthusiasm for the FTAA. It is for launching multilateral trade liberalization initiatives. remarkable that this continent, so rich in its literature, has This is, in fact, the policy framework that President been so deficient in its trade policy in the postwar period, Kennedy had when he stayed loyal to non-discrimination having successively succumbed to the destructive strategy in trade, eschewing our use of Article XXIV-sanctioned of import substitution, then fallen for the Generalized Sys- PTAs while launching the Alliance for Progress for South tem of Preferences for developing countries, which many America for several non-trade objectives. Thus, in pursu- economists today believe to have been a mistake, and who ing free trade through non-discriminatory trade liberaliza- are now enamored of the PTAs. It is sad to see the huge tion and regionalism through the Alliance for Progress, he enthusiasm for free trade that has finally broken out there implicitly understood the chief lesson of policymaking- being channeled not into the multilateral freeing of trade, that two objectives are usually achieved best with the use but into the politics and economics of PTAs. of two policy instruments. As our forefathers put it, one The South American enthusiasm is largely responsive to cannot (generally) kill two birds with one stone. It is surely our own, seeing access to our markets as the gain they get not too late to return to Camelot. from our preferred access to theirs. It also reflects the sense of solidarity, as against hostility, that is now more manifest Eliminating Preferences in these countries toward us. "Regionalism" is then seen as On the general question of PTAs, some would even like to a cementing bond, a glue to launch joint efforts for human roll back existing PTAs to rid the world trading system of rights, democracy, et al., in the Americas. It must also be this pox. That is utopian, but a standstill, a freeze, is more recognized that the politics of PTAs in the Americas is also doable. There is, however, a simple solution to ridding the divisive and sidetracks free-trade sentiments into unpro- world of the preferences that existing PTAs burden us ductive politics. This is true of the well-known rivalry with. Since preferences relative to zero are zero, one can between NAFTA and Mercosur as the leading lights of the effectively eliminate them also by reducing the MFN tar- trade-liberalizing movement in the Americas. Brazil, the iffs down to zero. Hence, many economists have allied dominant player in South America, sees NAFTA extension themselves to the goal that Martin Wolf of The Financial into FTAA as a hegemonic extension by the United States, Timies first proposed for the trading nations to announce: and Mercosur is, at least in part, in a response to NAFTA. that, by a date such as 2025, the world be rid of all border The creation of APEC with U.S. initiative was, of course, trade barriers. in part a geopolitical response to the Mahathir Asian bloc So far, the U.S. administration has poured cold water on initiative, which, in turn, was a response to our NAFTA. the idea, arguing that concrete trade pacts are necessary, And so, with PTAs, politics tends to hijack and distort the not a vision. That is like saying that it is enough to go freeing of trade, whereas the multilateral system tends to around filling potholes as one finds them, that a road map narrow it down to its economics-and permits more is unnecessary. But then what can one expect from an focused attention to the free-trade objective. administration where trade policy is run by politicians is THE CHALLENCES OF REGIONAL INTEGRATION watching the polls, and where lawyers, with their adver- 4. See the interesting historical account of our commitment to sarial mentality, implement it? "equality of treatment" in trade and hence to nondiscrimination or MFN in trade in David Palmeter (1996). "Some Inherent Problems with Free Trade Agreements," Law and Policy in International Business, Notes The International Law Journal of the Georgetown University Law 1. Mercosur is a customs union (CU), which is different from an Center, Vol. 27 (4), pp. 991-97. FTA, because it also has a common external tariff. A common mar- Palmeter, a distinguished trade lawyer, also expresses astonish- ket is additionally characterized by free mobility of capital and labor ment at the fact that "there has been virtually no debate-vigorous among the members. All are PTAs. or otherwise-on the current retreat from the United States's his- 2. The U.S.-Israel FTA does not have any value as a precedent and toric commitment to nondiscriminatory multilateralism" (page signifies no change in trade policy since it was dictated purely by the 993). Of course, he means a public debate; within scholarly circles in politics of our 'special relationship." the universities, there has indeed been a vigorous debate where I 3. John Maynard Keynes, Speech to Parliament, in Donald Mog- have witnessed an increasing hostility to PTAs emerge in the last gridge, ed., (1979). The Collected Writings of John Maynard Keynes: few years, when there was really little when I started writing in Activities 1944--1946: The Transition to Peace, Cambridge: Cambridge 1991. University Press, pp. 623-24. 19 The New Face of Regional Integration EN R I Q U E V. I G LES IA S R EGIONAL ECONOMIC INTEGRATION HAS PLAYED AN IMPORTANT ROLE IN LATIN AMERICA and the Caribbean's postwar economic history. The 1960s and 1970s saw a number of ambi- tious initiatives inspired by the optimism surrounding the successful Western European experience. There was the formation of the Central American Common Market (CACM), the Caribbean Free Trade Association and later Caribbean Community (CARICOM), the Latin _.American Free Trade Association and subsequent Latin American Integration Association (ALADI), and the Andean Group. At its peak in the late 1960s and early 1970s, the topic of integration was indeed hard to avoid in the discussion of Latin American and Caribbean development. However, some serious disillusionment had clearly set in by the late 1970s, and by the early 1980s discussion of regional integration was nearly silenced by the overwhelming attention required by the external debt problem, adjustment and national crisis management. I imagine that by the mid-1980s many of the region's We also must not lose sight of the fact that the renewed experts on integration had begun to look for new areas of interest in regionalism in Latin America and the Caribbean employment. The topic of integration suddenly re- has parallel unprecedented structural economic reforms, emerged, however, stimulated in part by developments in including dramatic unilateral trade liberalization and a an unexpected place: North America. In the late 1980s, the process of implementing the most ambitious multilateral United States, formerly skeptical of regional integration, round of trade liberalization in history. Thus, my remarks launched an agreement with Canada, which later incorpo- about regional integration in Latin America and the rated Mexico through NAFTA. Caribbean will be couched in terms of broader trade issues The decade of the 1990s has witnessed a wave of and the overall process of policy change in the region. I also regional integration initiatives in Latin America and the want to discuss the role of my institution, the Inter-Amer- Caribbean (14 major agreements-free-trade areas or cus- ican Development Bank. toms unions-since 1990, with a handful more in varying degrees of negotiation). However, this was not just a phe- The Growth in Trade nomenon of Latin America and the Caribbean-regional- ism has more than ever become a global phenomenon. General Trends Indeed, nearly all World Trade Organization (WTO) mem- The 1990s have witnessed a rebound in the region's trade bers are now signatories of at least one preferential-trade after the crisis of the previous decade. Between 1990 and agreement. 1996, the region's exports expanded by 76 percent; Enrique V. Iglesias is the President of the Inter-American Development Bank. 20 THE CHALLENGES OF REGIONAL INTEGRATION imports grew even faster, at 127 percent. Total trade as a between contiguous and non-contiguous countries. In fact percentage of the region's GDP now equals 36 percent, up the boom in intra-regional trade has largely been among from 24 percent in 1990, a clear indication of increased neighboring countries in the region. trade openness. Moreover, throughout the 1990s growth in Relaxation of the external restriction. The decline of the region's trade has consistently exceeded that recorded world interest rates, debt relief and a return of external at the world level: According to WTO estimates, the value capital flows in the 1990s has dramatically raised import of world exports and imports grew by an average 7 percent capacity in the region. Imports have sharply risen across- a year between 1990 and 1996 (compared with 10 and 16 the-board. Since intra-regional imports equal intra- percent for Latin American exports and imports, respec- regional exports, the import boom has been reflected in the tively). Consequently, the region's share in world trade has marked growth of intra-regional exports. expanded, from 4 percent in 1990 to 5 percent in 1996. A Real exchange-rate appreciation. The region's exter- closer look at the region's trade performance in the 1990s nal trade performance has also been influenced by the reveals the following trends: exchange-rate behavior of Latin American and Caribbean * Intra-regional trade has grown more rapidly than countries. The use of exchange-rate anchors in support of trade with countries outside the region. This trend is stabilization programs, as well as significant inflows of for- particularly pronounced in the case of exports. Since eign capital in the early 1990s, have contributed to real cur- 1990, the value of intra-regional exports has grown by rency appreciations in an important number of countries. 18 percent a year on average, compared with 9 percent This has tended to encourage imports while, at the same growth in extra-regional exports. Intra-regional time, posing problems for some of the region's exports. exports now account for 18 percent of total Latin Economic reforms. The structural reforms undertaken American and Caribbean exports, up from 12 percent in the late 1980s and 1990s have energized private-market in 1990. activity, investment and trade. Unilateral liberalization has * Growtht rates for intra- and extra-regional imports have been a key in exposing natural market opportunities for been inore homogeneous. While intra-regional exports to neighboring countries that heretofore were hid- imports expanded by an average 18 percent a year den behind the wall of protectionist policies. between 1990 and 1996, extra-regional imports also Subregional Trade Agreements. The aforementioned grew very fast, by 14 percent a year, reflecting a gen- explosion of subregional and bilateral trade agreements in eralized import boom in the region. The share of the 1990s has stimulated intra-regional trade in many imports originating from intra-regional sources has ways-for example, by: grown f^rom 15 percent of total imports in 1990 to 18 * signaling to the private sector the permanence and percent in 1996. deepening of public commitments to liberalization. * The marked difference in the growth rates of the In some agreements (such as Mercosur), the additional region's overall exports and imports (76 percent and commitment to trade liberalization is accompanied by 127 percent, respectively) reflects a large imbalance in a broader political message pursued at the highest the growth of trade with extra-regional markets, with official levels to promote peace and non-economic imports from these sources expanding nearly twice as cooperation among member countries. fast as exports to extra-regional destinations. * giving the private sector legally binding market Several factors seem to be contributing to these trends. access, which has reduced the risks of trade and I say "seem" because, as I will discuss a bit later, much investment, in contrast to unilateral opening. more empirical work still needs to be done in evaluating * providing the preferences that are part of the agree- the factors behind growing commercial links in the region. ments. It is important to point out that some of the Geography. Neighboring areas dense in capital and preferences (e.g., ALADI's) in fact predate the new population tend to interact and trade. This tendency can be trade agreements of the 1990s and have been eroded enhanced further when income levels, cultures, tastes and by the unilateral liberalization of trade. languages are similar as they are in Latin America and With the progressive stabilization of domestic price when significant differentials exist in transport costs structures in Latin America, there are now signs that the 21 TRADE TOWARDS OPEN REGIONALISM aforementioned process of currency appreciation may be for the 1990-95 period mainly due to a slowdown in trade reaching a plateau. This, coupled with possible further pro- between Guatemala and El Salvador. While intra-regional ductivity gains resulting from structural reforms and new trade has expanded less rapidly than that of other integra- investment, could favor stronger export growth in the tion groups in Latin America, such trade now accounts for coming years and thus facilitate a more balanced overall roughly 20 percent of total CACM exports, up from 16 trade performance vis-a-vis extra-regional markets. Mean- percent in 1990. while, the one-off effects of initial liberalization efforts in The Andean Group. After several years of rapid recent bilateral and subregional trade accords are increas- growth, intra-Andean trade stagnated in 1996, increasing ingly being realized. Future growth in intra-regional trade, by only 1.2 percent compared with 1995. A markcd slow- therefore, will depend on continued deepening of liberal- down in the growth rate of Colombian exports and a sharp ization commitments among Latin American and decline in Venezuelan exports to the subregion were the Caribbean countries, and more effective exploitation of main factors behind this development. Although intra- natural trading opportunities in the region through mea- Andean trade expanded by an average 29 percent a year sures such as the improvement of the network of regional between 1990 and 1995, such trade still accounts for only infrastructure. 11 percent of the subregion's total exports. However, for some countries (notably Colombia and Bolivia), the Andean Cu rrent Trends at the Regional and Subregional Levels market is of much greater significance. Moreover, if petro- In 1996, the valuie of the region's total exports expanded by leum is excluded from total export figures, intra-Andean 11 percent; imports grew by the same rate. This compares exports as a share of total exports represent 18 percent. with growth rates of 21 percent and 10 percent, respec- Group of Three. Trade among G-3 members declined tively, in 1995. Slower export growth in 1996 relative to by 2 percent in 1996, due to a slump in Venezuelan exports the previous year affected all subregions and was mainly a and a significant slowdown in the growth of Colombian result of stagnant or declining world prices for some basic and Mexican exports within this subregion. Due to the rel- commodities, although oil-exporting countries benefited ative insignificance of Mexican exports to its G-3 partners from a sharp rise in petroleum prices, and some countries compared with its overall exports, intra-group trade were able to counterbalance unfavorable commodity price accounts for less than 3 percent of overall G-3 exports, trends with an increase in export volume. In fact, in 1996, although such exports have tripled in absolute value since the volulme of exports expanded at a rate very similar to that 1990. of the previous year (10 percent, compared with 11 percent Mercosur. Mercosur also recorded a much slower in 1995). growth in trade among its member countries in 1996: 12 As for import growth in 1996, it was due almost percent, compared with over 20 percent in 1995. This is entirely to increased volume. Import growth slowed mainly explained by a sharp slowdown in the growth of markedly in five of the region's major economies (Brazil, Argentine exports to Brazil. As in previous years, however, Chile, Colombia, Peru and Venezuela), while Argentina intra-Mercosur trade again expanded at double the rate of and Mexico saw a rebound in imports after recovery in the subregion's exports to third markets, and now accounts their domestic economies. for 22 percent of total cxports, up from 9 percent in 1990. In 1996, for the first time since 1990, the region's intra- NAFTA. Intra-NAFTA exports expanded more rapidly regional exports grew at a slower rate (7 percent) than than NAFTA's total exports, contrary to what had hap- exports to third markets (12 pcrcent). Three subregions- pened in 1995. One of the most notable developments in the Central American Common Market, the Andean 1996 was the recovery of Mexican imports from the United Group and the Group of Three (G-3)-followed this over- States, following a sharp drop in 1995 as a result of the all trend, while NAFTA and Mercosur did not. In all sub- Mexican peso crisis. In 1996, Mexican imports from the regional markets except NAFTA, growth in intra-regional United States grew by 24 percent, not only recovering but trade slowed markedly in 1996. significantly exceeding their precrisis level of 1994. Since CACM. Intra-CACM exports expanded by just 5 per- 1990, intra-NAFTA trade has expanded by almost 11 per- cent in 1996, a growth rate much lower than the average cent a year (faster than the 7 percent increase in NAFTA's 22 THE CHALLENGES OF REGIONAL INTEGRATION trade with third countries) and now accounts for almost 50 The picture for imports is slightly different: Here, percent of total NAFTA exports. While most of intra- extra-hemispheric markets have increased their share as NAFTA tradle happens bilaterally between the United suppliers to the Latin American market. In 1995, almost States and Canada, and the United States and Mexico, 35 percent of the region's imports were supplied by coun- respectively, Mexican-Canadian trade has increased seven- tries outside the hemisphere (mainly the European Union, fold since 1990, albeit from a very low base. Japan and the newly industrialized Asian countries), com- CARICOMA. In 1990-95, intra-CARICOM exports pared with a 33 percent share in 1990. It should be noted grew much faster (8 percent) than CARICOM exports to that, among the various subregions in Latin America and the rest of the world (5.5 percent). Like other subregions in the Caribbean, Mercosur is the one least dependent on the the hemisphere, CARICOM countries thus seem to be tak- North American market. In 1995, more than 50 percent of ing increased advantage of regional export opportunities. its trade was with extra-hemispheric markets. Nevertheless. intra-CARICOM exchanges still represent Intra- and extra-regional exports from Latin America only around 13 percent of the subregion's exports, and are and the Caribbean display marked differences in terms of heavily dominated by the oil trade of one country, Trinidad their product structure, with manufactures accounting for and Tobago. In that sense, the small island economies have a much larger share of intra-regional commerce. This pat- not yet reacaled an important degree of integration. It tern is evident even if Mexico-whose maquila trade with should be noted, however, that trade with neighboring the United States accounts for a large share of Latin Amer- countries of the Association of Caribbean States (ACS), ica's overall exports-is discounted from the regional aver- while still modest, has increased rapidly in recent years, age. Excluding Mexico, manufactures account for approxi- even faster than intra-CARICOM trade. (According to mately 50 percent of intra-regional exchanges, compared IMF data, CARICOM exports to these countries have with around 23 percent for extra-regional exports. In both grown by almost 16 percent a year since 1991.) ACS mar- cases, the share of manufactures in the total has expanded kets may, therefore, represent a potential area of new slightly since 1990, although for most countries of the opportunities. region, extra-regional exports are still heavily dominated The rapid growth in intra-regional trade among Latin by basic commodities, and hence vulnerable to fluctuations American andl Caribbean countries in recent years, coupled in world commodity prices. with the dramatic increase in Mexican trade with its NAFTA partnoers, has meant that more than 70 percent of Why the IDB Supports Regional Integration Latin American and Caribbean exports are now destined to The Inter-American Development Bank (IDB) was created markets within the Western Hemisphere, compared with in 1959 with the specific mandate to promote regional around 60 percent in 1990. U.S. and Canadian share in the integration and economic and social development in Latin region's overall exports has increased, but this is largely America and the Caribbean. This dimension of regional explained by their trade with Mexico. Throughout most of integration is unique among development banks and has Central and South America, the United States and Canada been one of the defining characteristics of the Bank. When have become relatively less important export destinations shareholders added $40 billion to the Bank's capital in in recent years, although in absolute figures, trade with 1994, the mandate to support "regional integration and these countries has increased significantly. Meanwhile, the modernization" was reaffirmed. share of exports going to extra-hemispheric markets has Why is the IDB especially disposed to support regional declined for all Latin American and Caribbean subregions integration? One fundamental reason is the political values except Central America. From the U.S. viewpoint, the of many of its shareholders: Latin Americans share a com- Latin American and Caribbean region has constituted a mon heritage, language and culture. There is a side of us market of growing export opportunities; this is the case Latin Americans that makes us want to be more together- even when its rapidly expanding trade with Mexico is not sentiments extending back to Bolfvar and independence. considered. UJ.S. exports to non-NAFTA members in the The phenomenon is sometimes hard for non-Latinos to hemisphere now account for 8.4 percent of the total U.S. understand and could seem contradictory given the long exports, up from 6.5 percent in 1990. history of serious political disputes among Latin American 23 TRADE TOWARDS OPEN REGIONALISM nations. But the fact remains that the centrifugal forces of * The first and most important level of liberalization disagreement have coexisted with, and often have been has been through unilateral measures to open up overcome by, the centripetal forces of a common heritage economies. During the late 1980s and early 1990s and culture. These opposite forces, of course, can be espe- practically all countries of the region moved to reduce cially intense in the various geographic subregions of Latin trade barriers. This effort is reflected in the fact that America. the average tariff in Latin America and the Caribbean More to the point, however, regional integration is a has declined from 45 percent in the second half of the subset of, and complement to, a phenomenon of integration 1980s to 13 percent in 1995, accompanied by a sharp of a much broader scope that is inherently part of the devel- reduction of tariff dispersion as well. Furthermore, opment process that the Bank supports.1 In effect, the over the same period the share of the region's imports dynamics of economic development induce the integration subject to non-tariff barriers has declined from 31 of markets, regions and people. The larger markets and percent to 11 percent. Specific tariffs have virtually competition brought on by this process enhance specializa- disappeared even while they are still common in the tion, efficiency and growth with corresponding welfare industrialized economies. I think most will admit gains. that this truly has been an impressive transformation. Even though processes of integration are a natural part * The second level of external opening is multilateral. of growing market economies, this development can be The region was a very active participant in the blocked by political, institutional and economic barriers. Uruguay Round, and assumed the new disciplines The Bank's programs in support of development and eco- that emerged from this exercise. It was the only devel- nomic integration are designed to help countries overcome oping region to bind 100 percent of its tariffs. With these barriers. Panama joining the WTO in 1995, all of the region The task of the Bank has been greatly facilitated by the is now subject to the rights and obligations of this dramatic shift in the 1980s regarding development in world organization. Latin America, away from state-dominated and protected * The third tier of opening is through regional integra- economies, and toward open, private market-driven tion. It is often overlooked that in the new context of regimes. Indeed, nearly all the countries of the region have policy change in Latin America, regional integration undergone a profound process of structural economic is an additional instrument to open economies to reforms. The IDB has been actively supporting this competition and complements levels one and two of process, and, of course, our host, the World Bank, as well the trade-liberalization process. Indeed, it is the as the International Monetary Fund, have been at the fore- broader commitment to general economic liberaliza- front in providing support for reforms not only in Latin tion that is one of the major driving forces behind the America but throughout the developing world. recent spurt of regional integration initiatives in the The reform process in Latin America's economies has 1990s and helps give the processes the character of energized markets and entrepreneurial activity and natu- open regionalism." rally increased the region's interaction with the world This, of course, was not always the case. In the first economy. It also has induced spontaneous regional market decades after World War II, Latin America and the integration, which, as mentioned, explains in part the Caribbean pursued a number of ambitious economic inte- growth of intra-regional trade. gration initiatives. However, the regional integration of The IDB is supporting deliberate regional integration that era inserted itself into the prevailing development initiatives in the broader context of its support for the strategy of import substitution that worked so well in the process of structural economic reform. In the initial stages interwar period and in the 1950s, but that had been of their development, regional integration arrangements exhausting itself in the face of the postwar renewal of pri- link up with the overall reform process, most obviously vate market activity, liberalization and the initiation of the through the trade-liberalization component. In effect, we postwar phase of globalization. Indeed, the integration view regional integration as a third tier of a three-tier lib- schemes of the period were designed in part to rescue the eralization process. import substitution model through a strategic and selec- 24 THE CHALLENGES OF REGIONAL INTEGRATION tive expansion of a highly protected market. While the * New domestic investment and foreign investment, integration initiatives achieved some important results- both of which are channels for the transfer of new for example, the significant liberalization of trade in Cen- technology, enhanced international competitiveness tral America-outcomes fell far short of objectives. On the and global economic convergence, brought about by one hand, the strong national political commitments to the forces of competition and the larger market with domestic protection made opening up even among associ- guaranteed access. ate countries a very laborious negotiating process that * The emergence of new exporters of manufactured rarely achieved more than very partial results. On the other goods because of the security of subregional market hand, the costs of trade diversion were amplified due to the access, preferences and the familiarity of neighbor- general presence of extraordinarily high and sometimes hoods. The learning curve associated with subregional increasing barriers to third parties. export experience-which, as we saw, heavily involves The traditional model of development in the region manufactured goods-can serve as a platform for new changed in the 1980s, and this, in turn, has dramatically international exports. This is an important considera- changed the face of regional integration itself. The regional tion, because history has shown that developing coun- integration of today has inserted itself into the broader tries can achieve new dynamic comparative advantage overall strategy of opening up to the world economy. Not on the road of their long-term convergence with surprisingly, the new subregional integration initiatives industrialized countries. have mirrored the relatively aggressive tone of the overall * Authorities' commitment to liberalizing policies sig- opening-up strategy observed in the region. Countries have naled by a regional arrangement, and locked-in com- entered into multiple arrangements that are eliminating mitments that otherwise are more easily reversible. tariffs among partners in substantially all trade within 10 North-South agreements in particular are often cited to 15 years and often involve other commitments that even for these confidence-building effects. A good example go beyond the WTO's trade-related disciplines. is the incorporation of southern Europe into the Euro- pean Union. NAFTA may similarly serve Mexico, and Some Expected Benefits of Regional Integration we expect that the FTAA will also contribute to lock- Without being exhaustive, here are some of the positive ing in more mature trade and investment policies in (and interrelated) impacts on development that we hope to Latin America. derive from regional integration initiatives: * Increasing confidence from the enhanced international * The introduction of a lowering of the average level of competitiveness brought about by regional integra- protection vis-a-vis the status quo, which raises com- tion, which also should prepare countries for further petition and promotes specialization in the subre- advances in unilateral and multilateral liberalization. gional market. This could be done through further In this latter case, regional integration can also be unilateral opening, too. However, in certain conjunc- viewed as a way to move ahead with liberalization tures authorities can find that a partial opening in a while the region awaits consensus on development of subregional arrangement can meet less political resis- a new round of reciprocal disciplines. tance (and even be relatively popular as in the case of All these aspects of regional integration are, of course, Mercosur). This may be because of a number of asso- potential developments rather than guaranteed outcomes; ciated factors such as public sentiments about "get- what happens in practice depends on the nature of policy ting together" with a well-defined neighbor, reciproc- implementation, a point I will return to. Moreover, the ity with guaranteed market access, and more limited distribution of these impacts will, unless compensation impacts on fiscal income. mechanisms are available, tend to be uneven among * The expansion of market size to facilitate greater spe- partners. cializat:ion and industrialization through economies of scale and possibilities to exploit economies associated Some Expected Costs of Regional Integration with the forces of location and the agglomeration of The potentially positive aspects of regional integration are production activity. accompanied by costs, too. These are frequently cited costs: 25 TRADE TOWARDS OPEN REGIONALISM * Preferences in regional trading arrangements can A case in point is Mercosur. This agreement is barely divert trade away from possibly more efficient firms five years old, and still only in the initial stages of forma- that are located in non-member countries. This has tion. Nevertheless, it has already been under siege as a costs for consumers in the partner economies and for threat to itself and the world for allegedly being a hotbed the non-member countries that lose market share. of trade diversion that is creating a "fortress Mercosur." The trade diversion risks locking the country into The fact of the matter is that the subregion has never in its patterns of inefficient production. recent history been so open to world trade. Due to unilat- * Regional integration agreements can improve the eral liberalization, average external tariffs on manufactures terms of trade of member countries at the expense of have fallen from 25 percent in 1990 to 12 percent in 1995. non-member countries, and give rise to incentives for This, coupled with the program of internal elimination of maintaining or increasing preferences and protection. tariffs, has created more competition in these economies * When there are serious asymmetries in average tariff than they have seen in more than 60 years. The opening has levels among prospective partners of an integration contributed to a surge of imports across the board (a rise of agreement, the loss of tariff revenue in the liberaliza- 22 percent per annum since 1990), new investments and tion process can have serious redistributive effects important gains in productivity. While imports from Mer- among the countries. cosur partners have modestly increased their relative share * Successful integration could, under some assumptions of total imports, shares have also risen for most of Merco- of opportunistic behavior, discourage or delay interest sur's external partners. Moreover, a preliminary examina- in unilateral and multilateral liberalization. tion by the IDB's Integration, Trade and Hemispheric Issues Division of shifts in intra- and extra-Mercosur par- Putting Costs and Benefits into Perspective ticipation in total imports at the three-digit standard There is no doubt that integration has costs. For instance, international trade classification (SITC) level between some trade diversion is inevitable, and the initial redistrib- 1990 and 1995, found significant trade diversion in only utive effects from lost tariff revenue can be important, one sector (petroleum) and about 15 percent of the product especially in North-South agreements like NAFTA and the categories. Indeed, the overall picture is indicative of con- prospective Free-Trade Area of the Americas (FTAA). But siderable trade creation. 2 we all know that any process of economic transformation, Mercosur is young and suffers from many start-up prob- or reform, has costs, and these costs are typically dispro- lems that concern the IDB. But we certainly see the glass portionately concentrated up front. The costs, however, are more as half-full than as half-empty. Indeed, it is our view justified by the greater benefits that are expected, but these that so far Mercosur has been a major catalyst of modern- are typically spread out over a longer period of time. An ization in the subregion that has been characterized by seri- integration scheme is clearly a long-term project in which ous political conflict, closed economies and fears of liberal- countries accept short-term costs in order to capture some ization with each other and the rest of the world. The Bank of the long-term dynamic benefits outlined above. is committed to encouraging Mercosur-and other integra- Many of the recent critiques of integration have especially tion agreements in Latin America and the Caribbean-to focused on the trade diversion issue. It is certainly important evolve in a way that maximizes the benefits for development to be vigilant about the costs involved in regional integra- and minimizes the costs to itself and the rest of the world. tion. But to cast one's vote against a new regional integration initiative due to risks of trade diversion is probably overly Future Challenges and the Role of the IDB rash. Judgment should await the evaluation of the direct The Bank's programs in support of regional integration development impact which is a medium- to long-term gen- work on multiple fronts in an attempt to help its member eral equilibrium phenomenon that is admittedly not easy to countries overcome what it views to be the major chal- untangle. In any event, it is only with the emergence of the lenges in the development of welfare-enhancing arrange- longcr-tcrm dynamic impacts on growth and development ments in the region. Typically, our strategies of assistance that one can judge whether the initial costs were worth it for for integration aim to encourage a policy of open regional- the country and for the world community at large. ism, which minimizes the risks of unacceptable costs and 26 THE CHALLENGES OF REGIONAL INTEGRATION amplifies potential benefits. Some of the major challenges structure. For the new schemes of the 1990s the task that must be confronted are: is just the opposite: to fortify incipient institutional 1. Progressive elimination of imperfections in subre- arrangements so that instruments are compatible gional integration schemes, including: with objectives. The Bank has programs in the * Working to ensure full implementation of agree- region that are working in both of these directions. ments and effective enforcement. * Creating effective mechanisms for the coordination * Developing politically feasible formulas to eliminate of regional infrastructure networks that allow the remaining exceptions to agreed trade liberalization. private sector to fully exploit the natural advantages This is difficult but often worth the effort because of location and mobilize financing, including an the opening up of sensitive sectors is usually very enabling environment to allow for private sector rich in trade-creation effects. participation. * Identifying and eliminating or harmonizing the * Participating in infrastructure projects. This is an numerous non-tariff measures still blocking intra- area where the Bank has been very active. In regional trade. 1990-96 the Bank lent more than $11 billion to the * Liberalizing services-an important new area for public sector for infrastructure projects. The Bank is attention. Fortunately, services are on the agenda of also now supporting private investment. For exam- a nurnber of agreements in the region. The Bank ple, our Private Sector Department has organized alreacy has a program in place to support liberaliza- more than $700 million of Bank loans and guaran- tion of services in Central America. tees for the finance of infrastructure projects worth * Dealing with cumbersome and sectorally restrictive more than $2 billion. Bank financing of infrastruc- rules of origin, which are present in some of the ture in our host area of Mercosur was $4.5 billion region's agreements. Simplification is certainly pos- over the same period, and an additional $4 billion of sible through more exclusive reliance on tariff-shift loans are programmed. criteri.a. Restrictive sectoral criteria should be grad- * Supporting the private sector's participation in trade ually relaxed.3 The IDB is actively helping countries and investment, especially small- and medium-sized to empirically evaluate the impact of different types enterprises, which generate much employment but of rules of origin on their trade, and the Bank hopes are not always fully equipped to take full advantage multilateral talks in the WTO will provide insights of regional integration. The Bank has active pro- on how to improve these regimes. grams in support of these enterprises in integration * Promoting the development of modern trade disci- processes and indeed is now in the process of sup- plines in integration agreements, which are WTO- porting a second phase of a project in Mercosur. consistent or WTO-plus. Of particular concern is * Improving official mechanisms for the interchange dispu-e settlement. When states dominated the of information and analysis on macroeconomic economies of the region, it was perhaps effective to developments in the subregions and monitoring resolve commercial disputes diplomatically (the tra- processes of convergence. ditior. in the region). However, now that private * Developing effective ways to monitor the distribu- markets are the driving force of the economy it is tion of benefits within integration schemes and to necessary to make integration arrangements trans- create collective mechanisms to correct politically parcnc-ly rule-based: Only in this way will the mem- sensitive and potentially destabilizing imbalances. ber countries encourage the investment that is so The Bank is helping subregional agreements to con- important for the efficient specialization that is at front these issues and is also supporting initiatives to the heart of successful integration agreements. help the smaller economies in the FTAA process. * Rationalizing regional institutions. In the case of traditional integration schemes which modeled 2. Consolidation and deepening of structural economic themselves after Europe, the task is to downsize an reforms-including unilateral trade liberalization-which overdi.mensioned and underfinanced institutional have been underlying the success of intra-regional trade 27 TRADE: TOWARDS OPEN REGIONALISM and the progressive improvement of international compet- activities in the region regarding compliance with the itiveness. The Bank is actively supporting developments in WTO disciplines and is exploring setting up a program to this area through national lending programs. allow countries to evaluate their own progress in comply- Continued gradual opening of the economies to the ing with Uruguay Round disciplines. One hopes that the world is central to the process. While there has been much industrialized countries are monitoring their own progress progress in unilateral liberalization, the average tariff level on this front. of 13 percent in the region is still high. External tariffs (and hence preferences in integration schemes) should be 5. Preparing for scheduled multilateral negotiations at gradually reduced for efficiency reasons and for maintain- the end of the century on key areas such as agriculture, ser- ing international legitimacy for open regionalism. Fiscal vices and intellectual property, and participating actively reform is needed to compensate for lost tariff revenue, to in the WTO's built-in trade agenda. While the Uruguay generate public savings for infrastructure development and Round still remains to be digested, the region should also to assist in long-term macroeconomic stability. Enhanced begin strategic thinking about the possibilities of initiat- social services are essential for the human capital required ing new multilateral talks early next century, which could to compete internationally and to ensure that the gains multilateralize many of the deeper commitments emerging from trade are distributed more equally. in regional integration. I believe that most Latin American The region must develop more direct instruments to countries would eventually support such an initiative: deal with fiscal disequilibriums and stabilization problems. Developing countries have the most to gain from multilat- Some recent disruptions in the trade-liberalization process erally agreed rules of behavior and market access in the do not reflect a reversal of commitments to open regional- world trading system. ism but rather "over-committed" policy instruments that have induced countries to confront fiscal and balance-of- 6. Attempting to ensure consistent, fair and transparent payments disequilibriums with temporary increases in tar- multilateral monitoring of integration arrangements under iffs and non-tariff barriers. This has unfortunately distorted Article XXIV and Article V of the GATT. However, the trade regimes that have previously undergone valuable article and its implementation, even with the important reforms. With the development of more direct instruments clarifications of the understanding attached in the Uruguay for domestic policy, trade instruments could be used exclu- Round, still suffers from a degree of imprecision. The sively for trade policy. doubts that sometimes are raised about regional integra- tion could be more constructively dealt with in the context 3. Preparation for the FTAA, which promises to be an of effective Article XXIV reviews with multilaterally overarching hemispheric agreement that should absorb the agreed criteria and strong empirical foundations. We will simple free-trade arrangements and coexist with those follow with interest developments in the WTO's Commit- agreements that have more extensive or deeper commit- tee on Regional Agreements, which is examining the rela- ments. As mentioned, integrating with the large open tionship between regional trading arrangements and the markets to the north should have many tangible long-term multilateral system. benefits for Latin America, but there will be important costs, and these must be managed in a socially efficient way 7. Last but not least, developing better data collection. if commitments are to be sustained and full benefits real- Both the advocates and critics of regional integration are ized. Thc Bank has been providing its member countries prisoners of the serious shortcoming of our theory and with technical support during the FTAA deliberations, models regarding the phenomenon of integration and the both through its working groups and individually, and is poor state of data availability for testing and delimiting the always on call to consider adjustment assistance when assumptions in those models.4 Our models have yet to ade- needed. quarely capture the full dynamics of the integration process which, as I mentioned, involve very complicated medium- 4. Ensuring compliance with Uruguay Round disci- to long-term general equilibrium issues. Data-availability plines. The Bank is supporting training courses and related is troublesome; for example, systematic information on 28 THE CHALLENGES OF REGIONAL INTEGRATION intra-regional investment flows and firm specialization in than alternatives in the pursuit of more liberal and open the region--central to integration-is highly deficient, as trade."5 are data on evolution of preferences, application of rules of Not everyone may yet share this optimism. Indeed, con- origin and non-tariff measures, as well as data on other dis- structive critiques can be a useful filter for processes that ciplines. Several programs in the Bank that target this can potentially go awry. Nevertheless, I would think that problem, including technical support to the FTAA, are oversight of regional integration would benefit if it were beginning to improve data-availability, but the bulk of the done more effectively at the multilateral level where there work remains to be done. Until we are more successful in can be some agreed criteria based on the needs of world overcoming empirical problems for evaluating regional trade and development. Further, constructive evaluation of integration, the waters of debate will remain very muddied regional integration requires improvement of our empiri- indeed. cal observation of the phenomenon and of the techniques for its measurement and interpretation. This World Bank ConclusioIIs Conference is fortunately a step in this direction and I am As long as regional integration in Latin America and the pleased to be part of it. Thank you. Caribbean continues to be an integral part of an overall policy framework of structural economic reform and suc- Notes cessfully confronts the challenges that I have outlined, 1. In L. J. Garay and A. Estevadeordal (1995). "Protection, Pref- there is good reason to believe that processes of integration erential Tariff Elimination and Rules of Origin in the Americas," willb e an-ffective instrument of growth and develop- Inter-American Development Bank, Department of Integration and will. boreover, an nce commitments among neighbors can Regional Programs, Division of Integration, Trade and Hemispheric ment. Moreover, since commitments among neighbors can Affairs, there is an analysis of the systems that are applied in the be uniquely deep and extensive as well as driven by forces hemisphere. that are often more than economic, the multilateral system 2. Inter-American Development Bank (1996). "Trade Deviation will coexist with them. While Latin America and the Compared with Trade Creation: The Case of Mercosur," in Integration Caribbean pursues regional integration, it also has a vested and Trade in the Americas, Department of Integration and Regional Programs, Division of Integration, Trade and Hemispheric Affairs interest in ensuring that regionalism is consistent with a' and the Unit of Statistics and Quantitative Analysis, Washington, healthy ancL progressively more liberalized rules-based D.C., periodical note, Annex 1. world trading system; only with such a system will the 3. See Nancy Birdsall and Robert Devlin (1997). "Regional Inte- region capture the full benefits of economic liberalization. gration and the Inter-American Development Bank," World Markets Fortunately, there is growing consensus among economists (May). and policymakers about the potentially positive contribu- 4. The brilliant physicist Richard Feynman believed that, even if tion of the new open regional integration to the world the scientific imagination is a powerful instrument, what scientists believe should correspond to reality. See James Gleick (1992). trading syst em. The trend is well captured in a recent Genius: The Life and Science of Richard Feynman, New York: Pantheon study by the WTO Secretariat, which states: ". ..to a much Books. greater extent than is often acknowledged, regional and 5. OMC (1995). Regionalism and the Word Trading System, Geneva, multilateral integration initiatives are complements rather p. 56. 29 Missing Lessons of East Asia: Openness, Education and the Environment VI NOD THOMA S YAN WANG HE GROUP OF EAST ASIAN ECONOMIES GREW FASTER THAN OTHER DEVELOPING REGIONS over the past three decades. The region is currently undergoing severe financial difficulties, but the experience of the past has nevertheless put the focus on the ingredients for long-term growth. As in other fast-growing countries, several compelling factors emerge as major con- tributors: macroeconomic discipline, outward orientation and human capital investment. Worldwide, a combination of outward orientation and education has interacted positively, raising the returns to education and to all investments. Openness created the demand for high-skilled labor, improved the quality of education and facilitated learning-by-doing, allowing for technological spillover and industrial upgrading. The importance of this virtuous circle of openness and education-the first link explored in this paper-deserves greater attention among the lessons, and has implications for reform pro- grams everywhere. East Asian economies, and other rapidly growing coun- separate actions on environmental problems for fear of los- tries, have also had periods of poor environmental policies. ing competitiveness, there might be a case for complemen- Openness in trade and investment, while conducive to tary regional efforts in implementing environmental poli- rapid economic growth and poverty reduction, has not cies such as taxes and standards. While it may be automatically led to environmental protection (Table 1). In politically difficult to forge regional partnerships, several fact, the combination of outward orientation and environ- promising bilateral initiatives for environmental protec- mental neglect has had an especially negative impact, tion are already being implemented in both East Asia and increasing the urgency for environmental policies. This Central America. second link assessed in this paper is also an underempha- sized lesson from East Asia. The Extent of East Asia's Openness Unilateral actions for trade liberalization have been the East Asian economies expanded exports at twice the aver- basis for exploiting the positive links between education age rate of other developing countries. These economies and openness in East Asia. In this respect, there would more than quadrupled the share of exports in GDP over the seem to be little reason for East Asia to change this past 25 years, from 7 percent in 1970 to nearly 35 percent approach in favor of regional actions on trade. However, to in 1995 (Figure 1). Their share in foreign direct invest- the extent that countries are likely to take joint rather than ment (FDI) to developing countries rose from 16 percent in Vinod Thomas is Director of the Economic Development Institute of The World Bank. Yan Wang is Economist in the Macroeconomic and Policy Division of the Economic Development Institute of the World Bank. 30 THE CHALLENGES OF REGIONAL INTEGRATION TABLE 1 Trade, Growth, Poverty and the Environment (percentages) TRADE GROWTH POVERTY ENVIRONMENT ANNUAL GROWTH ANNUAL GROWTH % POP LIVING ANNUAL CARBON DIOXIDE OF MERCHANDISE IN GNP ON LESS THAN US$1 DEFORESTATION, EMISSIONS, EXPORT (VOLUME) PER CAPITA A DAY (PPP) a % CHANGE INCREASES REGION/COUNTRY 1980 94 1970-95 VARIOUS YR. 1981 90 1980-92 Selected East Asian economies China 12.2 6.9 29.4 (1993) 0.7 79.2 Hong Kong, China 15.4 5.7 < 1 -0.5 81.3 Indonesia 9.9 4.7 14.5 (1993) 1.1 94.7 Korea, Rep. of 11.9 10.0 < 1 0.1 130.2 Malaysia 13.3 4.0 5.6 (1989) 2.1 150.0 Philippines 5.0 0.6 27.5 (1988) 3.4 35.1 Singapore 13.3 5.7 < 1 2.3 66.7 Thailand 16.4 5.2 < 1 3.5 180.0 Average 12.2 5.4 1.6 85.6 Selected Latin American economies Argentina 1.9 -0.4 ... 0.1 9.3 Bolivia -0.3 -0.7 7.1 (1990) 1.2 40.0 Brazil 6.2 ... 28.7 (1989) 0.6 17.9 Chile 7.3 1.8 15.0 (1992) -0.1 29.6 Mexico 13.0 0.9 14.9 (1992) 1.3 28.1 Peru 2.4 -1.1 49.4 (1994) 0.4 -8.3 Uruguay 0.9 0.2 ... -0.6 -16.7 Venezuela 1.1 -1.1 11.8 (1991) 1.2 28.9 Average 4.0 -0.1 0.5 21.2 Note; ... indicates missing. a Uses an international poverty line in the World Doevlopmt indcators 1997. Individual studies may have different estimares for the same counr-Y. Source: World Deveprment Indicators 1997, World Bank, Washington, D.C. 1970 to 55 percent in 1996 (World Bank 1997b). These more, by the end of the 1980s, the protection rates were trade and investment flows were crucial to the transfer of lowered substantially through reforms. technology and the gains in productivity. These economies Since the early 1990s, new waves of trade reforms have are suffering from setbacks in export declines and financial been strong and have resulted in substantial reduction in crises. This, nevertheless, does not call into question the tariffs and quantitative restrictions (Table 2). China has true nature of the East Asian miracle, which quintupled launched several rounds of tariff cuts and eliminated many living standards during the past generation. quantitative restrictions, reducing average tariffs from 43 The East Asian economies have not had the most liberal percent in 1992 to 36 percent in 1995 and 23 percent in trade regimes in the world. Several economies in East Asia 1996. Korea reduced its average tariff rate from 32 percent had moderate import protection in the 1960s and 1970s. in 1982 to 7.9 percent in 1994. Malaysia's overall tariffs In the 1980s, East Asia implemented substantial invest- average about 10 percent on a trade-weighted basis, and ment reform and moderate trade liberalization. As late as the government has cut or eliminated tariffs on hundreds the mid-1980s, the effective protection rate for manufac- of items each year since 1992. The Philippines reduced its turing was nearly 30 percent in the Republic of Korea, 50 tariff rate from more than 40 percent in 1980 to about 20 percent in Thailand, and 70 percent in Indonesia. But a percent in 1994. Taiwan, China implemented a series of key difference in East Asia was that import protection did tariff cuts, increased the import categories exempt from not produce the anti-export bias that it did elsewhere (Dol- controls from 34 percent to 85 percent, and switched to a lar 1992; Bhalla 1993; Thomas and Wang 1996). Further- negative list of controlled goods (all other goods being 31 TRADE TOWARDS OPEN REGIONALISM FIGURE I Exports as a Share of GNP in East Asia and Latin America 40 35 30 25 20 10 15 10 exempt). Thailand launched reforms aimed at tiberalizing East Asia complemented trade reforms by opening its import regime, reducing the number of tariff rates, and investment. Among the first generation of newly industri- eliminating most tariffs that were higher than 30 percent alized economies (NIEs), Hong Kong and Singapore wel- (see Appendix 1). comed FDI, but until the mid-1980s, Korea and Taiwan TABLE 2 Trade Policies of East Asia and Latin America, 1990-93 (percentages) IMPORTS IMPORTS STANDARD AFFECTED BY STANDARD AFFECTED BY MEAN DEVIATION OF NON-TARIFF MEAN DEVIATION OF NON-TARIFF REGIONICOUNTYRY TARIFF TARIFF RATES BARRIERS REGIONI COUNTRY TARIFF TARIFF RATES BARRIERS East Asi a Latin America China 36.3 28.0 11.3 Argentina 9.9 6.9 0.2 Hong Kong 0.0 0.0 0.5 Bolivia 9.7 1.2 2.0 Indonesia 19.4 16.1 2.7 Brazil 1T.1 6.3 1.5 Korea, Rep. of 9.0 r.6 o r a Chile 11.0 0.7 0.1 Ma(ayspa 14.3 A4.) 2.1 Colombia 13.3 4.9 1.7 Philippines 20.0 11.0 ... Mexico 12.6 5.4 3.9 Taiwan (China) 8.9 10.1 35.9 Peru 17.6 4.4 ... Thailand 23.1 16.9 5.5 Uruguay 9.3 7.1 ... Vietnam 12.0 15.5 ... Venezuela 13.4 4.8 2.4 Average 15.9 13.1 9.7 Average 12.0 4s6 1.7 Sotrce: WorldDaflopment Indicators. 1997, World Bank, Washington, D.C. 32 THE CHALLENGES OF REGIONAL INTEGRATION severely restricted foreign investment to encourage domes- sity, ranging from 30 percent in Korea to more than 80 per- tic entrepreneurs. The second generation of fast-growing cent in Malaysia. In China, foreign-funded firms' share of economies in Southeast Asia, and China, have been more total exports rose from 6 percent in 1989 to 39 percent in receptive to FDJ, giving foreign investors some preferential 1995 to 47 percent in 1996 (SSB of China, 1997). FDI treatment (tax and import tariff exemptions). Since the brings new technology and managerial skills, which facili- mid-1 980s, most East Asian economies have been lowering tate learning-by-doing and human capital accumulation. restrictions on FDI and removing restrictions on capital Trade and investment regimes in East Asia, however, are transfers. Major areas of reforms include regulations on sec- not as liberal as those in many Latin American countries tor allocation, ownership, local employment, local content that recently have carried out major trade liberalizations and export obligation (see Appendix 2, Table 1). Many (Table 2). The tariff level in Latin American countries has East Asian economies have also entered into multilateral been lower than 15 percent, and non-tariff barriers are agreements on investment, such as the 1987 Association of minimal. East Asian countries have more protective mea- Southeast Asian Nations (ASEAN) Agreement on Invest- sures, and some sectors (services, for example) are closed to ment, and the 1994 Asia-Pacific Economic Cooperation foreign entry. Government interventions are substantial in (APEC) Non-Binding Investment Principles. setting trade and investment policies, and regulations are Reform of investment regimes has brought big gains: not fully transparent. Net FDI fiows to East Asia soared from $10.2 billion in Before the recent setback, East Asia had the best perfor- 1990 to $61.1 billion in 1996 (Figure 2). FDI has financed mance of any region in terms of export expansion, eco- between 13 and 16 percent of gross domestic investment in nomic growth and poverty reduction (Table 1). The ques- Cambodia, China, Laos, Malaysia, and Vietnam. FDI has tion is why, and whether trade and investment promoted exports and provided access to international mar- liberalizations alone are sufficient for sustainable growth. kets. Firms with U.S. investment have a high export propen- Previous studies have pointed out the importance of open- FIGURE 2 Foreign Direct Investment in East Asia and Latin America 70,000 - 60,000 - 50,000 - East Asian Countries c 40,000 - E m 30,000 - ;) / ~~~~~~ ~ ~ ~~~~~~~~~~~~~Latin Amria Conre 20,000 - 10,000 _ 0 33 TRADE TOWARDS OPEN REGrONALISM ness, broadly defined to include attitudes toward new ideas higher per capita incomes. Primary and secondary repeat and technologies. And East Asia opened up earlier than rates are much lower in East Asia than in Latin America. Latin America in the 1970s and 1980s (World Bank Barro and Lee (1997) studied the effects of family charac- 1993a, 1993b). Further, recent policy reforms in East Asia teristics and school resources on school outcomes (testing have significantly reduced barriers to trade and investment. scores) and found that a major component of East Asia's More importantly, openness, combined with human academic performance is left unexplained by family and capital accumulation and institutional factors, enables a school inputs. country to take off on a growth path and stay on the com- petitive edge. The synergy between openness and knowl- Distribution of Education Affects Development edge accumulation in East Asia was crucial for its coun- A more equitable distribution of education across popula- tries' success in an increasingly competitive world market. tion groups is also crucial for economic development. As with land and physical capital, the equal distribution of Openness and Education Together Had an human capital is important to growth. Human capital, Especially Positive Impact however, is not transferable between individuals-at least Combining education with openness has been the key to not as easily as is physical capital. Using a model reflecting achieving international competitiveness and poverty the non-transferability of human capital, Lopez, Thomas reduction in fast-growing countries. Some East Asian and Wang (1997) estimated Gini coefficients for educa- economies have invested heavily in formal education, and tional attainment for 11 countries and found that educa- in training and learning-by-doing. Educational attainment tional inequality has a negative and significant impact on and enrollment rates are higher in East Asia than in coun- economic growth. tries with similar income levels, with all their well-known Estimated education Gini coefficients for selected East benefits. But it is the combination of the focus on educa- Asian and Latin American countries show a rapidly improv- tion with an openness to the outside world that produced ing distribution of education in East Asia (Table 4). Korea a series of outstanding results, as elaborated on below. has experienced the fastest improvement in education dis- tribution, with Gini coefficients declining from 0.54 in The Quality of Formal Education Is Key 1 970 to 0.24 in 1990, a 56 percent drop. China, too, has Recent evidence shows that the quality of schooling is high seen a dramatic improvement in education Gini coeffi- in East Asia, whether it is measured by test scores, cients, which dropped from 0.52 in 1975 to 0.44 in 1990, teacher/pupil ratios, spending per pupil, repeat rates or a 16 percent decline. Comparators in Latin America showed dropout rates (Table 3). Government spending on primary initial improvement, but deterioration later. India's educa- and secondary students and teachers' salaries is higher in tion Gini coefficients are among the highest in the world East Asia than in Latin American countries that have (0.74 in 1990), and they have improved only slowly. TABLE 3 Trends in Schooling Quality by Region PRIMARY SECONDARY PRIMARY SECONDARY PRIMARY SPENDING SPENDING REPEAT REPEAT DROPOUT REGION YEAR PER PUPIL PER PUPIL RATE RATE RATE East Asia 1960 107 277 ... ... --- 1970 176 385 0 3 6 1980 352 521 i 7 ) 1990 494 530 4 1 13 Latin America 1960 174 435 ... ... 1970 247 580 15 8 46 1980 308 479 12 8 41 1990 272 458 10 8 36 Note .. -Indicates missing values. So=re: Robert Barro and Jong-Wha Lee (1997), "Determinants of Schooling Quality," working paper, Table 2, Harvard University, Cambridge, Mass. 34 THE CHALLENGES OF REGIONAL INTEGRATION TABLE 4 Inequality of Education: Gini Coefficients for East Asia and Latin America EDUCATION GINI COEFFICIENTS CHANGES IN COEFFICIENTS (PERCENT) COUNTRY 1970 1980 1985 1990 AVERAGE 1970-74 1975-79 1980-84 1985-90 India 77.4 75.9 73.1 72.1 74.8 -2.8 0.8 -3.7 -1.3 East Asia China ... 52.7 51.3 44.4 51.4 -7.5 -2.8 -13.4 Korea, Rep. of 53.9 35.5 30.1 23.5 36.8 -23.9 -13.4 -15.4 -21.9 Malaysia 58.3 49.3 46.5 43.6 50.4 -7.1 -9.0 -5.7 -6.3 Philippines 45.9 35.3 35.0 34.0 37.5 -18.5 -5.5 -1.0 -2.6 Thailand 46.5 46.4 42.6 42.1 45.1 2.4 -2.6 -8.2 -1.3 Latin America Chile 36.2 34.8 34.4 34.3 35.3 1.5 -5.2 -1.1 -0.3 Colombia 55.9 50.5 50.6 51.4 51.5 -11.7 2.4 0.1 1.6 Mexico 55.3 52.0 49.0 49.0 51.8 -3.4 -2.6 -5.8 0.0 Venezuela 65.3 41.9 42.0 45.5 51.0 -7.5 -30.5 0.2 8.2 Source: Lopez, Thomas and Wang, 1997. Authors' calculations, based on Barro and Lee's (1997) data on education attainment and school year data from UNESCO. Why has investment in education been more equitable, agriculture fell from 58 percent to 26 percent. In Chile and and with high quality, in East Asia? According to Barro Mexico, the wave of trade liberalization has coincided with and Lee (1997), it may have reflected an "Asian value" increased wage and income inequality. In Mexico's broadly defined as the cultural and religious beliefs that maquiladora enterprises, the ratio of non-production emphasize education. But why is this effect not found in (white-collar) to production wages rose from 2 to 2.5 South Asian countries? An alternative explanation is that between 1985 and 1988. In Chile between 1980 and 1990 the external environment-openness in trade and invest- the wages of university graduates rose by 56 percent rela- ment-has generated higher returns to education and thus tive to those of high-school graduates (World Bank 1995, has given szudents, employees and employers strong incen- pp. 33 and 56). tives to invest in learning. Openness Facilitates Learning-by-Doing Openness Provides Higher Returns to Education Rapid integration into the global market has allowed the Investing in basic education does not by itself guarantee importation of knowledge-enhancing capital, and facili- growth. Openness in trade and investment can provide tated learning-by-doing, technological catch-up and higher earning opportunities and raise returns to educa- industrial upgrading. Openness to international competi- tion. A World Bank study (1995) compared earnings in tion combined with government intervention has enabled highly prorected economies with those in export-oriented these countries to create comparative advantages and main- economies and found that protected economies have lagged tain competitiveness in the world market. Recent studies in manufacturing earnings and in employment growth. have produced evidence showing the importance of FDI in Real earnings per worker grew in Korea, Malaysia and generating technology spillover through training and Thailand fiom 2 to 8 percent annually from 1970 to 1990, transfer of technology.2 The following cases illustrate the while they declined in closed economies. Export-oriented linkage between openness and learning-by-doing. economies have seen wage employment growth far exceed- Companies in many fast-growing countries have ing the population growth, and faster growth of earnings actively used international technology transfers as an for skilled workers. Thus, returns to education are much investment in learning. They have been large importers of higher in terms of both job opportunities and higher earn- modern technology through FDI, subcontracting and ings. In Malaysia, wage employment tripled between 1957 licensing. The training components of technology-transfer and 1989, while the share of the labor force working in agreements with foreign companies frequently cover not 35 TRADE. TOWARDS OPEN REGIONALISM only the acquisition of the skills needed to operate and spillover across firms, as opposed to an infusion of new cap- maintain new facilities (the know-how), but also the skills ital. Finally, the dazzling growth of the coastal regions can and capacities for designing, engineering and managing be explained entirely by their effective use of foreign capi- projects (the know-why) that can be used to generate inno- tal and their export orientation. Husain and Wang (1996) vation and technological change. In 1993 the share of find that per capita FDI is positively and significantly asso- machinery imports of total imports (a proxy measure for ciated with variations in GDP per capita growth among embodied technology) was 42 percent in China and Chinese provinces. Also, the output share of foreign- Indonesia, 54 percent in Malaysia, 45 percent in Thailand, funded firms is positively related to economic and produc- and 32 to 34 percent in Korea and the Philippines. Simi- tivity growth. larly, this share was even higher in Latin America: 43 per- The Case of Chile and Costa Rica. The exports of these two cent in Chile and 50 percent in Argentina. But this num- countries have displayed high growth rates and diversifica- ber was 14 percent in more closed economies such as India tion both in terms of their products and destination mar- and Bangladesh (World Bank 1997a). kets. Both have combined "implicit" export-promotion The Case of Malaysia's Electronics Industry. Malaysia is the policies with imaginative selective and institutional pro- world's leading exporter of semiconductors and the third motion policies. In Chile, a system of returning a percent- largest producer of semiconductors after Japan and the age of the FOB value of exports is especially beneficial for United States. In 1987, electronics exports were Malaysia's small and medium exporters. In addition to trade liberal- top revenue earner, contributing $6.9 billion to the ization and fiscal incentives, export promotion has national account. FDI and technology and skills transfers included an institution-building component. Public and from multinationals made this achievement possible. The private institutions have played a fundamental role in Investment Incentives Act of 1968, the establishment of export promotion through programs to identify markets free-trade zones in 1972, and specific incentives for the and to provide technical assistance in the production, sales, electronics industry attracted the influx of foreign compa- and marketing of goods (Inter-American Development nies. Foreign investment transformed the industry from a Bank 1996, p. 100). labor-intensive operation in the early 1970s to a capital- These cases show, to some extent, that fast-growing and technology-intensive operation today. countries are able to combine openness and learning, form- Rapid technological change and significant skills trans- ing a virtuous circle: Openness creates demand for learn- fer and capacity building characterize the electronics ing, and learning and institution-building make a coun- industry in Malaysia. Continuing investments by multina- try's export sector more competitive. The cases also show tionals have resulted in the establishment of local research that learning and knowledge accumulation involve a local- and development capability and a gradual expansion of ized process. In other words, one firm or one country can production into areas of higher technology. Many multina- learn to learn, or to become more efficient in managing tionals have set up formal apprenticeship programs, schol- knowledge accumulation in the course of production and arships, and skills development courses. An example of the trade (Stiglitz 1987). successful transfer of skills and technology is the Penang Skills Development Center, a public/private joint venture. The Two-Way Link: Education and Openness The Center offers training courses to the entire manufac- Recent economic literature generally supports the above turing sector and has emerged as one of the country's lead- observations and suggests two-way links between educa- ing training institutes (World Bank 1994). tion and openness. First, knowledge accumulation influ- The Case of China. In China, export expansion and FDI ences a country's trade performance and competitiveness have promoted knowledge spillover and productivity (Grossman and Helpman 1989, 1990; Romer 1990); and growth. Using data from 434 urban areas, Wei (1993, second, trade enhances knowledge-accumulation, espe- 1996) found that export expansion from 1980 to 1990 was cially through imports (Ben-David and Loewy 1995; Coe positively associated with higher growth rates. In the late and Helpman 1993; Keller 1995; and Padoan 1996). Lucas 1980s, however, FDI became the main engine of growth. (1993) pointed out that to sustain any kind of knowledge Its contribution came from technological or managerial accumulation, a country would have to be outward ori- 36 THE CHALLENGES OF REGIONAL INTEGRATION ented and a significant exporter. Young (1991) and Keller and a more open economy have a 3 percentage points (1994) noted that trade per se cannot be the engine of higher rate of return than those that had only one or the growth; rather, it must be through some mechanism, such as other (Figure 3). the formation of human capital, that it affects growth. The The detailed findings of this study suggest that impact of trade openness on long-term growth, therefore, increases in education and openness might have interacted depends on the absorptive capacity of local human capital. to promote the returns to investment. First, we find a pos- Beyond the case studies from East Asia and elsewhere, itive and significant relation between changes in the level there is broader evidence of the education/openness link. of education and lending project performance, whether For example, an earlier study found that in a sample of 60 measured by the probability of receiving a satisfactory rat- developing countries from 1965 to 1987,3 economic ing or by economic rate of return. Second, in the regres- growth rates were especially high when there was a com- sions, the variable for the interaction between education bination of a high level of education and macroeconomic and openness also has a positive and significant association stability and openness (World Bank 1991, p. 5). More nar- with project performance. The findings are consistent with rowly, Thomas and Wang (1997) looked at the possible the observations in the case studies and suggest that open- interaction of openness and education and its impact on ness facilitates the importation of knowledge-enhancing the performance of the World Bank's lending projects.4 capital, promotes learning-by-doing, and provides higher On average, countries with a more educated labor force payoffs to education. FIGURE 3 Education and Openness Interact and Increase Investment Returns c17.0 /F^ ^ 16.0 ' 14.0 Low educatio < w | , 0 < I\i~~~~~~~ore open High education~~l Less open Note: Economic rates of return are from the evaluation database of the Operations Evaluation Deparrment. Education is measured by the average level of schooling of the labor force, and openness is measured by the logarithm of the foreign exchange parallel market premium. SoTrce: Aurhocs' calculations. 37 TRADE TOWARDS OPEN REGIONALISM Openness and Environmental Neglect Had an reduction in poverty. Environmental losses have been seri- Especially Negative Impact ous-pollution, congestion, deforestation and loss of biodi- East Asia's environmental record stands in contrast to its versity. The Central American economies have grown phenomenal economic growth and poverty reduction. In slowly for a variety of economic and sociopolitical reasons. the past quarter century, income grew at an average rate of An exception is Costa Rica, a country with a strong record 5 percent a year. The incidence of poverty fell sharply, of promoting human development. But more generally, the declining by an estimated 50 to 70 percent in Indonesia, economies of Central America have been dominated by tra- Malaysia and Thailand (Johansen 1993). At the same time, ditional exports, which have faced declining terms of trade; environmental losses and degradation in East Asia have by highly unequal income distributions; and by inadequate surpassed those of any other region in the world. educational investment, exacerbated by political instability. Nine of the world's 15 cities with the highest levels of Because growth has been slow, poverty levels have remained air pollution are located in East Asia. About 20 percent of stubbornly high. Environmental quality has deteriorated, all vegetated land suffers soil degradation from waterlog- too, as evidenced by extensive deforestation, soil degrada- ging, erosion and overgrazing. About 50 to 75 percent of tion, overfishing and water pollution in coastal zones. coastlines and protected marine areas are classified as hav- Thus, both slow- and fast-growing economies can suffer ing highly threatened biodiversiry (World Resources Insti- from severe environmental degradation. Within East Asia, tute 1996). The region has some of the highest deforesta- there was not much difference in water and air pollution tion rates in the world. The countries that implemented levels and congestion in Manila, where growth was slow in trade reforms in the 1980s, such as China, Malaysia and the 1980s, and the rapidly growing Bangkok area. Like- Thailand, have seen some measures of pollution double or wise, natural resource indicators, such as freshwater with- triple after reforms. From 1980 to 1992, carbon dioxide drawal per capita, appear to be comparable in East Asia and emissions rose from 1,489 million tons to 2,668 million Latin America (World Bank 1997a). tons in China, from 28 million tons to 70 million tons in Growth per se, therefore, cannot be blamed for environ- Malaysia, and from 40 million tons to 112 million tons in mental degradation. When the sources of environmental Thailand (Table 1). problems-underpriced resources (be it forests, water or air), weak institutions, unclear property rights and the Is Growth Primarily to Blame? neglect of externalities-are not adequately addressed, Openness in trade and investment, though conducive to rapid growth seems to worsen the problem. However, rapid growth and poverty reduction, is not an automatic growth and high incomes can be harnessed to mitigate ally of the environment. Rapid growth has come at the environmental degradation and improve resource use if expense of the environment and has generated attendant accompanied by timely environmental actioins. social costs, such as the negative effects of air and water pol- Most rapidly growing countries have taken the lution on human health.5 Rapid growth does not automat- approach of growing first and cleaning up later. Evidence ically improve the environment-environmental policies shows that this is a costly strategy socially and ecologically, must be put in place simultaneously. Many growth-induc- and could threaten the sustainability of growth itself. The ing policies, such as clarifying property rights, investing in damages are much costlier to address later, if not irre- sanitation, improving education and ensuring sound macro- versible in some instances. Furthermore, new approaches economy, help to improve resource use and contribute to a and technologies are now becoming available that make it better environment. But in crucial areas government worthwhile socially to address growth and environmental actions, such as imposing taxes and standards, investing in sustainability together rather than in sequence. technology and improving production methods, are neces- Many Asian and Latin American economies are begin- sary. Rapidly growing economies are learning this lesson ning to take corrective actions. In East Asia, subsidies on the hard way, and some are now taking corrective actions. gasoline, diesel and kerosene fuel are being removed. Latin It is interesting to compare Southeast Asia (Indonesia, America has begun to use market-based instruments. The Malaysia, Singapore and Thailand) and Central America. Philippines has developed a unique system of national Southeast Asia has grown rapidly and achieved a sharp parks and protected areas. Brazil, Colombia and Venezuela 38 THE CHALLENGES OF REGIONAL INTEGRATION charge a forestry tax for tree-harvesting undertaken with- lateral agreements (for discussions see Appendix 2; Bhag- out equivalent reforestation. Public participation and com- wati and Panagariya 1996; Srinivasan 1996; Winters munity involvement can be effective, especially where 1996). If some regional agreements can successfully institutions are weak and enforcement is expensive. For address environmental concerns, it would be easier to example, in Japan the local governments and resident achieve a global agreement on environmental standards. groups negotiate with firms to arrive at a detailed written East Asia, it would seem, has little reason to pursue agreement on emission levels. Between 1971 and 1991, the regionalism, especially discriminatory regionalism, in the numberofagreements increased from approximately 2,000 areas of trade and investment. East Asia did not pursue to 37,000. Once standards were agreed on, they were effec- regionalism during the first wave in the 1950s. Reasons for tively implemented. such lack of regionalism include: first, historically, the major players in the region used to be political rivals; sec- Are Environmental Standards a Rationale for ond, these countries had very different levels of protection, Regionalism? and large adjustment costs can arise for less developed and Such unilateral actions by individual countries may not be more protective economies if a free-trade area is formed; sufficient, however. No country is isolated in the global third, since the size of most economies was still small, environment. One country's actions in resource use and trade relations were traditionally oriented toward distant pollution can affect the welfare of its neighbors. Thus, a markets, particularly North America, Europe, and Japan first argument for taking regional environmental actions is (see Appendix 2). based on externalities. Because of the negative spillover In a rapidly integrating global economy, there would effects of one country's environmental neglect, regional seem to be a generally declining need for regionalism. East actions might be necessary. Asian economies have experienced rapid integration, with The second argument is incentive-based. If one country their trade accounting for more than one-third of all devel- unilaterally implements environmental standards, it might oping country trade. And these economies remain com- fear losing its competitiveness. A common agreement for mitted to the principle of multilateralism and to keeping all countries in a region to act on the environment might their regional arrangements nondiscriminatory.6 This reduce this concern Thus, there is an incentive for coun- determination was reconfirmed by commitments under- tries to participate in collective action and to implement taken at the APEC meetings in Tokyo and Subic Bay. environmental standards in the context of a regional agree- There might be a rationale for regional efforts to imple- ment. The commitment by all countries in a region would ment environmental protection in view of the above argu- motivate each individual country to take action. ments. To the extent that countries are likely to take joint The degree of optimal action by each country varies, rather than separate actions on environmental problems for since the benefits and costs of environmental controls vary fear of losing competitiveness, there are strong incentives across countries. Regional agreements, therefore, might for regional action. Concerted regional action on environ- ideally be differentiated across countries on the economic ment policies and standards might be encouraged, which grounds of benefits and costs. But on political and admin- may well be the building blocks for multilateral agree- istrative grounds, the standards might have to be uniform ments. The most efficient means to combat environmental across countries. To balance these two opposing factors, problems are direct policies such as environmental stan- regional agreements might be uniform for a threshold dards or taxes. East Asia, in view of its economic success, (minimum) degree of environmental controls, with higher can create a model for achieving environmental protection standards on top of the minimum standards implemented by using instruments such as taxes and standards, while by individual countries in the region. maintaining open trade regimes. Third, global agreements on the environment might be difficult to achieve. They might take years to negotiate, as Bilateral Initiatives for Environmental Protection did the Uruguay Round, which took eight years. Regional Regional and multilateral actions on environmental stan- agreements on tougher environmental standards might dards are promising, but reaching consensus on this scale is expedite this process, forming building blocks for multi- not likely to be easy. Thus, there is a need to develop bilat- 39 TRADE TOWARDS OPEN REGIONALISM eral initiatives in parallel. Both East Asia and Central But openness in trade and investment and rapid eco- America have taken the lead in this regard. An agreement nomic growth did not automatically lead to environmental between NEES (a power company in the United States) and protection. Experience showed the need for complementary the Innoprise Corporation in Sabah, Malaysia, aims to pro- actions to protect the environment. The implication is not mote reduced impact of logging in tropical forests. The to stop liberalizing trade and investment regimes, but incremental costs associated with this environmentally rather to implement strong environmental policies along friendly logging technique are borne largely by NEES. But with liberalization. both partners stand to benefit-Malaysia through the The international experience shows that trade and preservation of its soils, water and biodiversity, and NEES investment liberalization produces especially strong results by accumulating carbon offset credits. when coupled with learning and capacity building in the In Central America, Costa Rica has been most active in course of the reforms. No group action by economies was achieving forest protection by developing a market mech- needed in these respects, however. On the other hand, to anism of payments for environmental externalities. The the extent that countries are likely to take joint rather than pilot phase of Activities Implemented Jointly (AIJ), a separate actions on environmental problems for fear of los- recent agreement with Norway, helps to protect large ing competitiveness, there might be a case for regional tracts of natural forests in Costa Rica. The country benefits action to directly address environmental protection via the conservation of its forests, while Norway (and the through environmental policies. world community at large) gains through carbon seques- tration and mitigation of the threat of global warming (see References Kishor and Constantino 1994). Barro, Robert and Jong-Wha Lee (1997). "Determinants of Schooling These agreements demonstrate the power of interna- Quality," Working Paper. Harvard University, Cambridge, Mass. tional trade in environmental services to benefit both the Ben-David, D. and M. Loewy (1995). "Free Trade and Long-Run developing and developed countries. In addition, bilateral Growth," CEPR Working Paper No. 1183. Cencer for Economic partnerships for environmental protection need not be Policy Research, Stanford, CA. restricted to forestry-they have been successful in the Bhagwati, J. N. and A. Panagariya (1996). "Preferential Trading Areas and Multilateralism: Strangers, Friends or Foes?" in Bhagwati and areas oenrygnrtoancosrainaPanagariya (eds.), Free-Trade Areas or Free Trade? The Economics of ling air and water pollution, as well. However, it must be Preferential Tradinig Agreements. Washington, D.C.: AEI Press. noted that the success of such international partnerships Bhalla, S. S. (1993). "Free Societies, Free Markets, and Social Wel- depends on economic and political stability, and an out- fare." July 15, 1993. Washington D.C. Working Paper. ward-looking development policv in the host countries. Blomstrom, Magus, Robert Lipsey and Mario Zejan (1992). "What Clearly, once these preconditions are satisfied, bilateral Explains Developing Country Growth?" NBER Working Paper No. 4132. National Bureau of Economic Research, Cambridge, schemes hold strong potential for environmental protec- Mass. tion in developing countries. Borenzsrein, E., J. de Gregorio and J. Lee (1995). "How Does Foreign Direct Investment Affect Growth?" NBER Working Paper No. Conclusion 5057. National Bureau of Economic Research, Cambridge, Mass. Investing in education was a key to rapid development in Coe, D. and E. Helpman (1993). International R&D Spillovers," many parts of the world. Openness and competition cre- NBER Working Paper No. 4/144. National Bureau of Economic ated the demand for high-skilled tabor, raising the returns Research, Cambridge. Mass. aredthe emad fo hig-sklledlabr, risin th retrns Dasgupta, Partha and Joseph Stiglitz (1988). "Learning-by-Doing, on education and learning. In addition, effective learning Market Structure, and Industrial and Trade Policies." Oxford Eco- improved the competitiveness of export industries. The nomic Papers, 40 (1988):246-68. link between openness and the accumulation of human Dean, Judith M., Seema Desai and James Riedel (1994). "Trade Pol- capital raised returns on all investment. This was the basis icy Reform in Developing Countries Since 1985: A Review of for the rapid growth and poverty reduction in East Asia. Evidence," World Bank Discussion Papers No. 267. World Bank, The financial crisis in East Asia does not call into question Washington, D.C. DeRosa, D. (1993). "Regional Trading Arrangements among Devel- the true nature of East Asian development, which quintu- oping Countries: The ASEAN Example," mimeo. International pled living standards during the past few decades. Food Policy Research Institute, Washington, D.C. 40 THE CHALLENGES OF REGIONAL INTEGRATION Dollar, David (1992). "Outward-Oriented Developing Economies Stiglitz, Joseph E. (1987). "Learning to Learn: Localized Learning Really Do Grow More Rapidly: Evidence from 95 LDCs, and Technological Progress." In Partha Dasgupta and Paul Stone- 1976-85." Economic Development and Cultural Change, 40 (April) man (eds.), Economic Policy and Technological Development. Cam- 523-44. bridge: Cambridge University Press. Gale, Lewis R. and Jose A. Mendez (1988). "The Empirical Rela- Stiglitz, Joseph E. and Marilou Uy (1996). "Financial Markets, Pub- tionship between Trade, Growth and the Environment," Interna- lic Policy, and the East Asian Miracle," The World Bank Research tional Review of Economics and Finance (U.S.), Vol. 7, No. 1, pp. Observer, 11(2): 249-76. 53-61. Thomas, Vinod and John Nash (1991). Best Practices in Trade Policy Grossman, Gene M. and Elhanan Helpman (1989). "Growth and Reform. A World Bank Publication. New York: Oxford Univer- Welfare in a Small Open Economy," NBER Working Paper No. sity Press. 2970. National Bureau of Economic Research, Cambridge, Mass. Thomas, Vinod and Tamara Belt (1997). "Growth and the Environ- (1990). "Comparative Advantage and Long-Run Growth." * ment: Allies or Foes?" Finance and Development, June. American Economic Review, 80: 796-815. Thomas, Vinod and Yan Wang (1996). "Distortions, Interventions, Grossman, Gene M. and Ann Krueger. (1995). "Economic Growth and Productivity Growth: Is East Asia Different?" Economic Devel- and Environment," Quarterly Journal of Economics, May, pp. opment and Cultural Change, 44(2): 265-88. Husain, Ishras and Yan Wang (1996). "China's FDI and Economic (1997). "When Learning Makes Lending More Productive," Growin, Ishrath." d Revu d'Ecno d1996 D "veCohpmna' FDIandEconomWorking Paper, Economic Development Institute. World Bank, Growth." Reavme d'Economie dze De'eloppment.Wahnt,D.,Ferry Tnter-Americani Development Bank (1996). Economic and social Washington, D.C., February. Ul Haque, Irfanm (ed.) (1995). Trade, Technology, and International Progress inLatin America: 1996 Report. Washington, D.C.: Johns Competitiveness. Economic Development Institute. World Bank, Hopkins University Press. International Monetary Fund (various years). Exchange Arrangements Washington, D.C. I atnd alngestariins (vannua reprs). Washington, DnC. U.S. Department of State (1997). National Trade Data Bank: Country and~ ~ ~, Exhag Retitos(nulrprs. Wahntn .C. Reports on Economic Policy and Trade Practices, Washington, D.C., Johansen, Frida (1993). "Poverty Reduction in East Asia: The Silent Reporti Revolution," World Bank Discussion Papers No. 203. World January. Bank, Washington, D.C. Viner, J. (1950). The Customs Union issue. New York: Carnegie Keller, W. (19)94). "Absorptive Capacity: Understanding the Cre- Endowment for International Peace. ation and Acquisition of Technology in Development," mimeo. Wei, Shang-jin (1993). "Open Door Policy and China's Rapid Yale University, New Haven. Growth: Evidence from City-Level Data." NBER Working Paper (1995). 'International R&D Spillovers and Intersectoral Trade No. 4602, December. National Bureau of Economic Research, Flows: Do They Match?" mimeo. Yale University, New Haven. Cambridge, Mass. Kishor, Nalin and Luis Constantino (1994). "Sustainable Forestry: (1996). "Foreign Direct Investment in China: Sources and Can It Cornpete?" Finance and Development, December. Consequences." In Takatoshi Ito and Anne 0. Kreuger (eds.), Leipziger, Danny and Vinod Thomas (1993). The Lessons of East Asia: Financial Deregulation and Integration in East Asia. Chicago: Uni- An Overvierv of Country Experience. World Bank, Washington, D.C. versity of Chicago Press. Lopez, Ramon E., Vinod Thomas and Yan Wang (1998). Working Winters, L. Alan (1996). Regionalism vs. Multilateralism. Discussion Paper. World Bank, Washington, D.C. papers series, No. 1525. Center for Economic Policy Research, Lucas, Robert E. (1993). "Making a Miracle." Econometrica, 61(2): Stanford, CA. 251-72. _ (1997). "Regionalism and the Rest of the World: Theory and Padoan, Pier Carlo (1996). "Trade and the Accumulation and Diffu- Estimates of the Effects of European Integration." Review of Inter- sion of Knowledge," Policy Research Working Paper No. 1679, nacional Economics, forthcoming. International Economics Department, International Trade Divi- World Bank (1991). World Development Report 1991: The Challenge of sion. World Bank, Washington, D.C. Development. New York: Oxford University Press. Panagariya, A. (1994). "East Asia and the New Regionalism." World _ (1993a). The East Asia Miracle: Economic Growth and Puhlic Pol- Economy, 17(6): 817-39. icy. New York: Oxford University Press. Romer, Paul (1990). "Endogenous Technological Change."Journal of (1993b). The Lessons of East Asia. A series of papers on individ- Political Economy, 98(5): 71-102. ual country experience of East Asia. East Asia and Pacific Srinivasan, T. N. (1996). "Developing Countries and the Multilateral Region.Washington, D.C. Trading System: From GATT (1947) to the Uruguay Round and _ (1994). East Asia's Trade and Investment: Regional and Global the Future Beyond." Paper prepared for the Economic Develop- Gains from Liberalization. East Asia and Pacific Region. Washing- ment Institute, World Bank, Washington, D.C., January. ton, D.C. State Statistical Bureau of China (1997). "Statistical Bulletin on _ (1995). World Development Report: Workers in an Integrating National Economic and Social Development in 1996." People's World. New York: Oxford University Press. Daily, 4/07/1997, p. 2. _ (1996). Managing Capital Flows in East Asia. East Asia and 41 TRADE TOWARDS OPEN REGIONALISM Pacific Region, WashingEon, D.C. 3. Updated to 1994 later. _ (1997a). World Development Indicators, 1997. Washington, D.C. 4. The cross-country, project-level data set includes variables on _ (1997b). Global Development Finance, 1997. Washington, D.C. education, per capita income, openness, government expenditure and World Resources Institute (1996). World Resoerces 1996-97. New project performance. The project data cover 3,590 lending projects York: Oxford University Press. in 109 countries evaluated by the Operations Evaluation Depart- Young, Alwyn (1991). "Learning by Doing and the Dynamic Effects ment (OED) between 1974 and 1994, with the OED rating of over- of International Trade." Quarterly Journal of Economics, 56: all performance (satisfactory/not) and economic rates of return. 369-406. 5. Grossman and Krueger (1994) found that economic growth brings an initial phase of environmental deterioration that is fol- Notes lowed by a subsequent phase of improvement. The turning point 1. Calculation of Gini coefficients for education takes two steps. usually comes when a country reaches a per capita income of $8,000. First, Lorenz curves are drawn, which are similar to that of income, Using a revised model, Gale and Mendez (1998) found that with percentage of population aged 15+ on the horizontal axis, and increased economic activity has a negative effect on the environ- percentage of cumulative education on the vertical axis. Gini for edu- ment, while changes in per capita income have a positive, and lin- cation is then calculated by the ratio of the area between the diago- ear, relationship with the environment. The trade policy measure is nal and the Lorenz curve divided by the total area of the half-square not significant, and its effect is ambiguous a priori. And pollution in which the curve lies. rises with the capital abundance of a country (since this favors cap- 2. Borenzstein, de Gregorio and Lee (1995) find that, beyond a ital-intensive and dirtier industries) and falls with increases in labor minimum threshold, higher FDI inflows ace associated with higher and land abundance. productivity of human capital, indicating that FDI may have positive 6. Using a computable general equilibrium model, it was shown spillover effects through the training of workers. Other studies have earlier that East Asian countries would benefit the most if they take found that FDI in developing countries is associated with the transfer concerted unilateral trade liberalization on a most-favored-nation of technology (see, for example, Blomstrom, Lipsey and Zejan 1992). (MFN) basis (World Bank 1994). 42 THE CHALLENGES OF REGIONAL INTEGRATION Appendix 1 its 8,500 import categories by implementing a negative list. This list increases the percentage of import categories A Review of the Recent Reforms in Trade and exempt from controls from 34 percent to 85 percent. Investment Regimes* Thailand has launched a flurry of legislative and regu- latory reforms aimed at liberalizing the domestic market. Reforms in Trade A wider reform of the import regime reduced the number China has launched bold trade reforms, including several of tariff rates and eliminated most tariffs above 30 percent. rounds of tariff cuts and elimination of quantitative restric- In 1993, Thailand began implementing the AFTA tariff rions. In 1992, the government announced sizable reduc- reductions, which will gradually eliminate the exclusion tions in tariff levels on 225 tariff lines, and the abolition of list of protected items and generally expand AFTA into a import regulatory duty. In 1994, the foreign-exchange real free-trade area. retention system and mandatory import plan were abol- ished, and import licensing requirements and quota con- Reforms in Investment Regimes trols on 320 items were eliminated. In 1995, import China: Special economic zones were established in the restrictions on 367 tariff lines were abolished. The average early 1980s and preferential treatments were given to for- tariffs were reduced from 43 percent in 1992 to 36 percent eign-funded enterprises. In 1988, the government adopted in 1995, ancl again to 23 percent at the beginning of 1996. a new Chinese-foreign cooperative joint ventures law. In In the Republic of Korea, a five-year tariff reduction 1990, an amendment to the 1988 law was adopted which plan ended in 1994, when the average Korean tariff rate simplified the approval procedures for new foreign invest- was reduced from 32 percent in 1982 to 7.9 percent. As ment enterprises. In January 1994, the official exchange part of the Uruguay Round settlement, the government rate and the swap market rate were united. Profit repatria- will continue to reduce tariffs and to eliminate most of its tion is subject to some restrictions, and foreign-exchange import restrictions by 1997. balancing is required. Malaysia's overall tariffs average about 10 percent on a Indonesia: By enacting a new deregulation package in trade-weighted basis, and import licenses are required for June 1994, Indonesia took a big step forward in improving a small range of products, such as tobacco and plastic its investment environment. The package, known as PP20, resins. The government lowered or eliminated tariffs on eliminated initial foreign equity requirements and signifi- over 600 items in the 1993 budget, and on more than 500 cantly lowered divestiture requirements. Both solely for- items in the 1994 budget. In the 1995 budget, import tar- eign-owned projects and joint ventures with a minimum iffs on another 2,600 items were reduced, largely to meet domestic equity of 5 percent are now provided for by its commitments in the Uruguay Round and the ASEAN Indonesian law. Several previously restricted sectors are Free Trade Area (AFTA). opened to foreign investment, including telecommunica- The Philippines reduced its tariff from more than 40 tions, electricity generation, airlines and railways. percent in 1980 to about 20 percent in 1994. The Ramos Korea: Changes in regulations announced in June 1994 Administration has expanded reforms by liberalizing trade, resulted in a streamlining of foreign investment applica- foreign exchange, and investment regimes; reducing entry tion procedures and easing of a number of barriers to direct barriers in vital industries (most recently in banking, foreign investment. At the same time, the opening up of telecommur[ications and insurance); and encouraging pri- several sectors that had previously been closed to foreign vate sector investments in much-needed infrastructure. investors was accelerated. Changes in laws and regulations The country has embarked on a path of sustained strong also make it easier for foreign companies to purchase land growth. for business purposes. Taiwan, China has implemented a series of tariff cuts. Malaysia: Malaysia began deregulating slowly in the In the course of GATT negotiations, Taiwan has commit- late 1960s. The Investment Incentives Act of 1968 and the ted to liberalize its trading regimes in many sectors, such establishment of free-trade zones in 1972 encouraged an as manufactured products and agricultural products and influx of foreign companies. In the 1980s, freer capital services. In July 1994, it simplified import procedures for movements were allowed and large FDI flows led to the 43 TRADE TOWARDS OPEN REGIONALISM cxpansion of a number of export-oriented industries. In munication sector has been steadily liberalized. In 1990, 1992, the guidelines on foreign equity capital ownership the maximum limit on the proportion of foreign share- were liberalized. Foreign investors are permitted to hold holding in the local banks was raised to 40 percent from 20 equity of up to 100 percent if they expDrt 80 percent or percent. more of their production. For projects exporting less than Taiwan: Remaining investment barriers have been 80 percent, the share of foreign equity ownership depends reduced both as part of its WTO accession process and as on such factors as the level of technology, spin-off effects, part of its drive to become a regional operations center. size of the investment, location, value added, and the use of Local content requirements have been phased out for most local raw materials and components. Projects producing manufacturing industries. In addition, Taiwan has high-technology products or priority products for the removed many discriminatory limits on foreign securities domestic market, as determined by the Malaysian govern- firms, insurance companies and banks. ment from time to time, may be allowed foreign equity Thailand: In May 1990, the government announced a ownership of up to 100 percent. series of measures to significantly liberalize the exchange The Philippines: The Foreign Investment Act of 1991 control regime. The central bank also raised limits on cap- allows full foreign ownership of companies engaged in ital transfers abroad and allowed free repatriation of invest- activities not covered by investment incentives. (There was ment funds, dividends, profits and loan repayments. a 40 percent ceiling imposed by previous investment reg- Vietnam: A new law on foreign investment was ulations.) In addition, the "negative list" of sectors, where approved in December 1987, providing for three forms of foreign ownership is either banned or limited, was greatly foreign investment including categories of full foreign reduced. In 1994, the government liberalized the entry and ownership. In 1992, an amendment to the 1987 law was scope of operations of foreign banks in the Philippines. approved. Under the amendment, further liberalization Singapore: To attract foreign investment, many of its measures included the extension of the duration of joint public policy measures are tailored to ensure an environ- ventures to 50 years from 20 years (and, exceptionally, to ment conducive to their efficient business operations and 70 years), and that domestic private firms were allowed to profitability. While seeking to develop more high-tech participate with foreign firms in investment projects. industries, the government does not impose restrictions on production standards, requirements of local purchases, or Sources: U.S. Department of State 1997, and IMF various percentage of output exported. Since 1989, the telecom- years. 44 THE CHALLENGES OF REGIONAL INTEGRATION Appendix 2 trade area may end up being an instrument of protection rather than liberalization; (b) There are incentives for trad- Background on Regionalism vs. Multilateralism ing blocs to raise rather than lower tariffs on outside coun- in East Asia tries; and (c) When the choice of regionalism is offered in the battle against protectionists, the political process may Literature review take the regional route to the exclusion of multilateral free Regionalisrn refers to discriminatory trade liberalization by trade. a group of countries. The definition of multilateralism, Winters (1996) first classifies the theoretical contribu- however, varies. Winters (1996) proposes that a country's tions on the issue into five broad groups. These models multilateralism index is a positive function of (a) the suggest that the relationship between regionalism and absence of discrimination in its trade policy; and (b) the multilateralism can go either way. He then evaluates the closeness of- its trade regime to free trade. The basic static actual experience, which is very limited. The tentative con- welfare of discriminatory trade liberalization can be clusions he draws include that regionalism may: (a) help to explained by the concepts of trade creation and trade diver- liberalize highly restrictive trade regimes; (b) increase the sion, introcluced by Jacob Viner (1950). A free-trade area vulnerability of less restrictive regimes to break down; and (FTA) is likely to be beneficial if, on balance, it gives rise to (c) be more likely to be a danger if governments are subject more trade creation than trade diversion. Ceteris paribus, to sector-specific lobbying forces. Winters (1997) examines the higher the initial tariff, the lower the difference between the effects of regional integration on excluded countries' the prices of the two supplies of imports; and the larger the welfare and surveys various ex ante estimates of the impact economic size of the union, the more likely the free-trade of European integration. These estimates suggest that area will improve efficiency-that is, there will be more neighboring countries linked tightly to the European trade-creation than trade-distortion generated. economy could lose significantly from the latter's integra- Bhagwari and Panagariya (1996) focus on the dynamic tion, but that for other countries the losses are likely to be time-path issue and ask the question whether regional very small. arrangements are "building blocks" or "stumbling blocks" to multilateral free trade. "Regionalism vs. Multilateral- A Brief Review of the History of Regionalism in East ism" switches the focus of research from the immediate Asia consequences of regionalism on the welfare of the partici- pating countries to the question of whether regionalism The Early Years. After the Second World War, East Asia will lead tD worldwide free trade more quickly or more did not court regionalism, while the first wave of regional- slowly. Economists are very divided on this issue. ism, launched with the founding of the European Commu- Frankel and Wei (1995) carefully review the arguments nity (EC) in 1957, swept large parts of Western Europe, about whether regional arrangements will be building Africa and Latin America. There was only one regional blocks or stumbling blocks to multilateral free trade. arrangement founded during this period: the Association Regional FTAs could be building blocks to global free trade of South-East Asia Nations (ASEAN). Even this arrange- by the following arguments: (a) If groups are organized ment did not create significant trade preferences among its into custorns unions, the multilateral negotiations will be member nations. There were six members of ASEAN: made more efficient; (b) It may be easier for the political Brunei, Indonesia, Malaysia, the Philippines, Singapore leadership in a country to get support for liberalization and Thailand. Excluding Singapore, which is a free-trading within a regional context than in a multilateral or unilat- country, intra-ASEAN trade accounted for less than 5 per- eral context; (c) The threat of regional integration may cent of the countries' trade. Of this, less than 5 percent was facilitate the liberalization among groupings; and (d) Pro- subject to any kind of trade preferences. gressive expansion of regional groupings may eventually There are three reasons for such lack of regionalism in lead to multilateral free trade. East Asia. First, historically, the major players in the region Regional agreements could be stumbling blocks to global have been political rivals. Second, these countries have very trade integration by the following arguments: (a) A free- different levels of protection and are at very different stages 45 TRADE TOWARDS OPEN REGIONALISM of development. Hence, large adjustment costs can arise to This development has manifested itself in the recent sign- less developed and more protective economies if a free- ing of the ASEAN Free-Trade Area (AFTA). At the Sep- trade area is formed. Third, since the size of most tember 1994 meeting in Chiang Mai, AFTA members economies is still small, trade relations have traditionally agreed to reduce the 15-year implementation schedule to been oriented toward distant markets, particularly North 10 years, gradually to eliminate the exclusion list of pro- America, Europe and Japan. tected items, and generally to expand AFTA from a tariff- reduction scheme into a real free-trade area. New Wave of Regionalism. The second wave of region- However, regional trading integration among develop- alism started in 1989, three years after the Tokyo Round, ing countries with high external trade barriers has been when the lUnited States was unsuccessful in persuading the tried since the 1950s. It has served primarily to extend the EC and developing countries to undergo another round of national import substitution policies to the regional level, multilateral trade negotiations (MTNs). Believing that an and failed to deliver faster growth. So far, both qualitative ever-expanding set of free-trade areas (FTAs) could achieve and quantitative analyses lead to the conclusion that worldwide free trade, U.S. Ambassador William Brock regional trade arrangements such as AFTA and SAPTA are turned to regionalism as an alternative instrument for sus- not particularly desirable. Gains from such arrangements, taining the movement for free trade. The switch in tactics if there are any, will come from their contribution to the proved a turning point in the history of regionalism and speeding up of nondiscriminatory liberalization in the par- launched what Jagdish Bhagwati called the Second ticipating countries. Regionalism. Pursuit of regionalism in North America and The World Bank (1994) considered several options for Europe has led a dramatic pursuit of regional pacts around further liberalization in East Asia. Unilateral reform and the world. Between 1989 and 1994, there were 33 regional preferential blocs were found to benefit the country or group agreements. With regard to Asia, in 1990 the East Asian of countries engaging in liberalization, but they offered Economic Caucus (EAEC) was formed. In 1992, members small benefits or possible losses to trading partners. Con- of ASEAN signed an agreement to form the ASEAN Free- certed most-favored-nation (MFN) trade liberalization by Trade Area (AFTA) by 2007. In 1995, south Asian coun- East Asian countries was found to offer greater welfare gains tries announced plans to form a South Asian Preferential to the region, and to major trading partners. The report Trading Area (SAPTA). argued that East Asia is well positioned to take a leading role in concerted multilateral, non-preferential trade and invest- The Future of Liberalization in East Asia. Until now, ment liberalization, which is a win-win solution to world the impact of regionalism on East Asia from its own trade tensions. Leaders of the Asia Pacific Economic Cooper- regional arrangements has been minimal. How should Asia ation Forum (APEC) have been moving in this direction-a respond to the growing regionalism around the world? concerted multilateral liberalization. Today, the issue of regionalism has resurfaced in East Asia. 46 THE CHALLENGES OF REGIONAL INTEGRATION APPENDIX 21TABLE I Trade and Irivestment Policies in East Asia 11996) MEMBER OF MULTILATERAL/ EXPORT PROMOTION REGIONAL TRADE EXCHANGE CONTROL FDI POLICIES POLICY ORGANIZATION Cambodia Two exchange rates; Eliminated the foreign exchange restrictions N/A license required for applying to investors in Cambodia; profits repa- engaging in foreign triation is permitted in accordance with relevant exchange transactions laws and regulations China One exchange rate Full ownership allowed under restricted condi- Some incentives and determined in the inter- tions; local content required on a case-by-case special benefits bank market basis; restriction apply to transfer of profits and foreign exchange balancing is required Hong Kong, Fixed exchange rate; no Few restrictions; no restrictions on transfer of No direct government WTO China exchange controls profits support Indonesia Managed float; no for- Full ownership permitted in certain sectors if the Preferential creatment WTO; ASEAN eign exchange control investments meet certain criteria; barriers in ser- on import duties since 1972 vice sectors; divestment is eventually required for foreign ownership; no restrictions on transfer of profits Korea, Rep. of Market average rate sys- Set-up of local partnership is no longer required; Tax privileges have been WTO tem; no direct foreign no restrictions on transfer of profits continuously reduced exchange controls Malaysia Market determined flex- Full ownership allowed if export 80% or more Export allowance and ASEAN; ratified ible rate; some foreign their production; most service sector protected; tax incentives to Uruguay Round Agree- exchange controls no restrictions on transfer of profits exporters ment in 1994 Philippines Marker-determined flex- Full ownership allowed for companies engaged in Various incentives WTO; ASEAN ible rate; most foreign activities not covered by investment incentives, granted to investment exchange restrictions but "negative list" of sectors remains; no restric- in export industries were liberalized since tions of transfer of pmofits 1992 Singapore Managed float; no for- No restrictions on FDI, except some restrictions No exports subsidy, WTO; ASEAN ma1 exchange controls in the service sector; no restrictions on transfer of although the govern- profits ment actively promotes it Taiwan, China Market-determined Some restrictions on FDI; significant incentives Generally no export floating exchange rate; to attract FDI in export-oriented industries; no subsidy or tax incentive some foreign exchange restrictions on transfer of profits to exporters, but excep- controls tiors do exist Thailand Managed float; in 1990, Few restrictions on FDI, certain sectors are Various incentives and WTO; ASEAN a series measures to lib- closed; no restrictions on transfer of profits special benefits eralize the exchange controls Vietnam Exchange transactions No maximum limit on foreign capital share, but N/A ASEAN permitted for six cur- a minimum of 30% share is required. No sectors rencies and trading are closed. Some de facto restriction of profit must take place within repatriation the official range Sources: IMF, Exchange Arrangtents and Exchange Restrictions, 1988-96; and U.S. Depactment of State, National Tirade Data Bank: Counrry Reports on Economic Policy and Trade Practices. 1997. 47 TRADE: TOWARDS OPEN REGIONALISM APPENDIX 2 TABLE 2 Evolution of Foreign Direct Investment and Trade Performance in East Asia and the Pacific vs. Latin America and the Caribbean EAST ASIA & PACIFIC LATIN AMERICA & CARIBBEAN FDI IN EAP EXPORTS AS A FDI IN LAC EXPORTS AS A YEAR (NET IN US$ MILLIONS) PERCENT OF GNP IN EAP (NET IN US$ MILLIONS) PERCENT OF GNP IN LAC 1970 201 7.59 1,091 12.40 1980 1,312 21.19 6,148 17.84 1989 8,330 24.99 8,138 18.28 1990 10,179 27.28 8,121 18.18 1991 12,706 28.66 12,504 17.44 1992 20,923 30.23 12,740 16.83 1993 38,128 31.48 14,066 16.22 1994 44,105 33.00 24,238 16.52 1995 51,776 33.51 22,897 19.32 1996 Preliminary 61,120 31.17 25,925 20.39 Source: World Bank. 1997. Global Development Finance 1997. NWVash-gron, D.C. 48 II. Assessment of R egional Integration Assessing Regional Integration L . AL AN WI NTER S I T IS FLATTERING TO BE ASKED TO PRESENT AN OVERVIEW PAPER AS THE SPEAKER FOR THIS HIGH- quality panel, but it is not easy. The conference organizers have commissioned a series of excellent papers and collected a set of excellent participants to explore the past and future success of regional integration in this region. The last thing we need at this stage is someone skimming rapidly over the whole field purporting to offer definitive answers to the questions we are about to tackle. Thus this paper is not an assessment of regional integration within Latin America and the Caribbean, but rather a discussion of how we might go about making such an assessment. I hope that it will help to define an agenda and some rules of debate for the conference rather than the outcome of the conference. I also, of course, hope that it will help to define the future research agenda for this important, but complex, topic. First, some terminology: Economists and policymakers I should also note that while most RIAs are regional in have increasingly come to recognize the economic benefits the geographic sense, this is not necessary for most analyt- of integration between the economies of different ical discussion. Thus, I shall use the term RIA for any rec- nations-e.g., Sachs and Warner (1995), World Bank iprocally preferential or discriminatory arrangement (1996). Thus the issue is not whether regional integration between two or more countries. per se is desirable, but whether regional integration explic- This paper is restricted to the "direct effects" of RIAs- itly pursued and fostered by discriminatory policy inter- those that follow immediately and causally from the initi- ventions is desirable. For this reason I shall talk of ation of the RIA rather than those that occur indirectly Regional Integration Arrangements (RIAs) and try to do so because the RIA induces other changes in economic policy. in studiedly neutral terms. As Martin Wolf has noted, ter- Thus, I consider so-called static effects, through which minology can affect the way we think about something: RIAs affect trade patterns, the changes in economic welfare Names matter. Who but a staunch protectionist emanating from such effects, and dynamic effects on eco- could have anything against a "free trade agree- nomic growth felt through factors such as investment and ment"? "Preferential trade agreements" sound less the relocation of industry. What I exclude is the fascinat- benign, while "discriminatory trade agreements," yet ing and rapidly growing analysis of the effects of RIAs on another name for the same thing, sound nasty..." other countries' trade policy and on the multilateral trad- -Financial Times, ing system. This limitation is primarily for reasons of space Tuesday, October 29, 1996, p. 14. and because I have analyzed "Regionalism vs. Multilateral- L. Alan Winters is Research Manager in the International Economics Department of the International Trade Division of the World Bank. I am grateful to Robert Devlin, Ricardo Ffrench-Davis, Guillermo Perry and Maurice Schiff for comments, Anju Gupta for research assistance and Audrey Kitson-Walters for logistical support in preparing this paper. It forms part of the output of a program of research on "Regional- ism and Development" coordinated by Maurice Schiff and myself. 51 TRADE: TOWARDS OPEN REGIONALISM ism" elsewhere (Winters 1996a). It also, however, reflects of the bloc can easily become the touchstone of loyalty for the suspicion that for individual governments and policy- members regardless of their view of the substantive issue.3 makers, the direct effects, especially those on their own Second, regional blocs generally have fewer policies and countries, are likely to be the more important: That is, only sources of identity than do individual countries, and thus rarely will systemic considerations reverse decisions made adverse comment on one of them is more threatening to on the basis of direct effects.' the blocs' existence than it is for a country. To attack Swiss It is useful Eo set out the agenda and some debating agricultural policy does not undermine the concept of a rules for at least two reasons. First, discussions of RIAs are Swiss state, but, particularly in the early years, to advocate always complex and frequently diffuse. There need to be at abolition of the CAP was to attack one of the few concrete least three countries, and everything occurs in a world that manifestations of European unity and thus to threaten the is "second-best" in the technical sense that distortions and European dream. In Winters (1996a) I explore more fully policy interventions are present in every outcome we con- the dangers of identifying integration with progress on sider specifically, restrictions on trade with non-member particular policies. countries.2 As a consequence, there are few general and As blocs mature, and as they acquire broader and deeper unambiguous economic results to appeal to: each case is sui sources of identity, it should be possible to discuss individ- generis. Additionally, many of the alleged costs and bene- ual policies more calmly. Analytical discussions should fits of RIAs are in areas that economists understand rela- seek to reach this stage from the outset by being explicit tively poorly even without regional complications-for that it is not the existence of regional arrangement 'x' that example, growth, technology transfer, political economy is under discussion, but the policies it pursues. That dis- and political benefits. Moreover, there are few cases of sig- criminatory trade policies might be harmful does not nificant regional integration on which to base empirical undermine the desirability of cooperating with and inte- studies (Foroutan 1997), which increases the burden borne grating naturally with neighboring countries. by our already inadequate theory. A second reason for suggesting some rules of debate is Standards of Comparison that RIAs seem to arouse more passion than do most other We all know the joke about the economist who, when issues of international economic policy. Thus, for many asked, "How is your wife?" replied, "Relative to what?" He years, to attack the Common Agricultural Policy (CAP) worked on RIAs. Perhaps the most difficult part of assess- was to be held to be attacking the European Community ing the effects of an RIA is to work out what would have itself. Even in the 1990s, academic worries about whether happened in its absence-the anti-monde, or counterfactual. the European Union (EU) had the best of all possible trade In fact, for ex post studies using the residual imputation policies and growth records (Winters 1993) were greeted method, this is the only question. These studies take actual by comments that the analysis showed the state of the data on some phenomenon after an RIA has bccn creatcd author's "mind rather than his heart about EC [European and attribute the difference between them and an anti- Community) integration" (Sapir 1993, p. 230). Passion of monde to the RIA. Anti-mondes vary from the simple (and this kind is not helpful analytically, and I offer my guide- frequently implicit) "no change from pre-RIA values" to lines as a possible means of preventing its onset. complex model-based predictions based on exogenous data. An interesting question is why RIAs stir up such emo- For ex ante studies of the sort that must be used for dis- tion. After all, individual countries' agricultural policies cussions of policy options, the anti-monde has to be con- have been criticized as roundly as has the CAP, and indi- structed, but again "no change" is the most frequent vidual countries' trade policies as roundly as has Merco- choice. Thus, for example, modelers such as Haaland and sur's, without such outbursts. I suspect two basic causes. Norman (1992) or Harrison, Rutherford and Tarr (1997) First, no criticism is more wounding than one that identi- calibrate their models to actual data for a historical (base) fies a failing that one has recognized oneself but decided to year and then shock it with changes to represent the RIA live with; since regional blocs' policies are compromises they are investigating. Thus, their estimates are as if negotiated between partners, one partner is always in this "1992" happened in 1985, and the anti-monde is the base- position. Moreover, defending the compromise in the name year actual values. 52 ASSESSMENT OF RECIONAL INTEGRATION If economies are developing very rapidly, "no change" is partners." If this term has any role in the assessment of not adequate, and it is necessary to be more explicit about RIAs, it requires (a) that we can, as a general proposition, the anti-monde. This inevitably involves judgment. Thus, equate "natural" with "desirable," and (b) that, for the RIA for example, if we are assessing the effects of introducing free in question, we cannot make a direct assessment of its trade within Mercosur over the period 1991-95, what do we desirability, for otherwise we could jump straight to the assume about internal tariffs in the absence of Mercosur? We latter without worrying about "naturalness." As a practical could assume that Mercosur has made no difference to the matter of discourse we also need (c) to be able to define and evolution of cariffs, so that if it had not come about, intra- measure "naturalness." Mercosur trade would have paid the same as extra-Mercosur In fact, neither requirement (a) nor (c) is very secure. trade (anti-monde = 1995 actuals on non-members); alter- Some commentators define "natural" in terms of outcomes natively, we could assume that all tariff changes would have (volumes of trade between potential partners)-Summers stopped in 1991 (anti-monde = 1991 actuals), or that some (1991>-while others focus on transportation (or transac- most-favored-nation (MFN) liberalization would have con- tions) costs-Krugman (1991). In the former case, large tinued (anti-monde based on trends of liberalization over, flows are far from sufficient to render preferential liberal- say, 1988-91). It makes a huge difference. ization beneficial (Panagariya 1997; Schiff 1996), while in Once we move beyond simple tariffs, anti-mondes the latter, relative transportation costs may not matter become even more complex. If we believe that RIAs foster (Amjadi and Winters 1997) or may have perverse effects the credibility of policy reforms, how do we represent cred- (Nitsch 1996a). Overall, therefore, I believe that we would ibility in their absence? be better off not using the term "natural trading partners" My conclusion from this is that we should always be or "natural integration," except in the sense I used in the explicit about the anti-monde and be prepared to justify our Introduction as meaning without any policy inducements. choice (if only within our own heads). The rest of the paper considers some of the arguments A second issue of standards of comparison concerns the that have been advanced for and against creating RIAs. It indicator we examine. An economics profession of the cur- obviously cannot do them justice in a single paper, but I rent size can be justified only in terms of making life bet- hope to give a feeling for the nature of the arguments, ter-that is, in terms of someone's welfare. We might still identify their crirical components-particularly in the debate whose welfare-i.e., what welfare weights to use dimensions just noted-and comment briefly on empirical (including those of future generations)-and whether and or practical aspects of their application. I start with the how to incorporate supposedly non-economic dimensions simple static effects of regional integration which, despite of welfare, such as security, culture, etc. We should also be a huge and venerable literature, still appear to generate clear whether results refer to the welfare (suitably aggre- some confusion. gated) of the citizens of one or all of the member countries in the RIA, those of non-members, or those of the world as Static Analysis a whole. But in the end, I believe that we must come back to welfare. In practice, however, many discussions of RIAs A Priori Analysis are conducted in terms of intermediate variables such as The traditional tools for RIA analysis-Viner's (1950) aggregate output (GNP) or levels of trade. This is legiti- trade creation and trade diversion-are marvelous heuristic mate, for we must simplify to make problems manageable, devices for illustrating the fundamental effects of RIAs. If, but it requires care. The mere observation that interna- as a result of a preferential-tariff reduction, imports from a tional trade has increased-still less that intra-RIA trade partner displace higher-cost local production, real has increased-is not sufficient to declare that an RIA is resources are saved in satisfying local consumption, and the desirable; think of the CAP again. Thus we should always domestic economy benefits accordingly. This is trade cre- bear in mind the links between the variable under examina- ation. If, alternatively, the preferences allow partner sup- tion and welfare. plies to displace those from non-members that would oth- A particular way in which intermediate stages of analy- erwise have been purchased because they were cheaper sis can confuse us is the use of the term "natural trading when both faced equal tariffs, the domestic economy ends 53 TRADE: TOWARDS OPEN REGIONALISM up paying more real resources for imports and loses accord- plier is unique and hence faces a downward-sloping ingly. This is trade diversion, and it amounts to diverting demand curve; thus, all changes in tariffs or other trade part of what would have been tariff revenue on imports barriers are partially passed through to consumers, result- from non-members to producers in the partner country ing in changes in consumer welfare and, through changes who now face no tariffs. in consumption bundles, pressures on other (domestic and Unfortunately, except in the very simple case explored non-preferred) suppliers to adjust their prices and quanti- by Viner, the mapping from trade creation and diversion to ties. This gives much more scope for RIAs to have real wel- economic welfare is treacherous.4 For example, pure trade fare effects than do homogeneous good models in which diversion can be beneficial if its effect of lowering con- world prices, assumed fixed, tend to anchor domestic sumer prices is sufficiently important to welfare. The ben- prices. A problem with many differential goods models, efits of trade creation for a single country can be out- however, is that through their assumptions of symmetry weighed by the losses of tariff revenue on the pre-RIA (all goods compete equally with all others), they ignore volume of imports from the partner country, i.e., by return- comparative advantage and/or "home preference," and ing to partner producers the tariffs on those imports pur- hence overemphasize trade diversion see Srinivasan chased from them even in the absence of preferences (Pana- (1993) and Jones (1993). Symmetry also means in these gariya 1997). Thus, ultimately, simple creation and models that results depend on relative intra- and extra-bloc diversion exercises are indicative rather than definitive. transport costs rather than absolute transport costs. Amjadi Nonetheless, because they are so pervasive and intuitively and Winters (1997) suggest that for the Mercosur coun- appealing, it is worth thinking briefly about their empiri- tries and Chile this relativity is sufficiently small that only cal basis. Before doing so, however, I consider two other minor benefits from RIAs are likely. aspects of static behavior. Once products are differentiated we open the way to One of the early puzzles about why countries adopted consider models of RIAs under imperfect competition. RIAs was that in neoclassical analysis they could always do Starting from Smith and Venables (1988), much literature better by an MFN liberalization. Wonnacott and Wonna- has emerged in which RfAs are generally predicted to have cott (1981, 1992) offered a solution by observing that, if much larger effects than perfectly competitive models sug- trade with the rest of the world was costly because of their gest.6 Market integration changes the nature of competi- tariffs or transactions costs, two partners might be able to tion-almost always reducing market power by enlarging make terms-of-trade gains by removing tariffs on their the scope of each market and removing some domestic mutual trade and trading only with each other. This players' special privileges. This generates additional gains requires that trade with the rest of the world cease; other- to the usual allocative ones. The existence of rents in wise, internal prices would remain anchored to world imperfect competition, which get shifted about by trade prices plus the tariff. Overall, such changes in trade pat- changes, also make RIAs more strongly redistributive, terns do not seem very realistic.5 Amjadi and Winters usually with larger benefits to small previously oligopo- (1997) explore the possibilities of these sorts of gains in lized economies. Several authors in Winters (1992) report Mercosur. Transport costs with the rest of the world are important advances in this genre, and Baldwin and Ven- high enough for there to be substantial savings by switch- ables (1996) give a more recent survey. ing to intra-bloc trade, but the requirement that trade The imperfect competition literature has given RIAs a with the rest of the world completely cease seems unlikely new lease of intellectual life (as it has international trade to be fulfilled in any but a few commodities. theory in general), but it should not be accepted com- Transportation costs also matter in the models based on pletely uncritically. For example, the permanence of oli- differentiated goods, such as Frankel, Stein and Wei (1994) gopolistic positions is not guaranteed even without RIAs, and Nitsch (1996b). As in Wonnacott and Wonnacott, this especially in the face of a determined government. If RIAs is because they rely on corner solutions, in which there is dominate MFN liberalization-as is possible in these mod- no trade with the rest of the world; this means that even els-it is at the expense of the rest of the world. It is not marginal changes in trade result in changes in internal guaranteed that intra-bloc competition really will be prices. In models of this kind, each good from each sup- fiercer-see Haaland and Wooton (1992) in theory and 54 ASSESSMENT OF REGIONAL INTEGRATION Jacquemin and Sapir (1991) in practice. The latter found thus is closer to non-discriminatory trade than to the mem- that French, German, Italian and United Kingdom profit bers' pre-Mercosur situations. margins were significantly squeezed by extra-EC imports A problem with anti-mondes based just on changes in but not by in:ra-EC imports. It is not clear that the EU has import (or export) shares is that they fail to take account of really shifted from being 15 segmented markets to a single changes in the relative sizes of economies as they grow at dif- unified one as is assumed in most analyses, and it is cer- ferent rates. For example, the huge growth of the East Asian tainly not clear that Mercosur or the Andean Pact would. economies in the last three decades would lead us to expect Finally, we have, to date, no empirical tests of these mod- that, ceteris paribus, their shares in partners' trade would els at all. grow strongly. One solution is to look at trade intensity ratios-the share of i in j's imports relative to the share of i Estimating S5tatic Effects Ex Post in the world's imports. The level of this index is obviously The simplest approach to ex post estimates of creation and influenced by factors such as geographical proximity and i's diversion is to look at the evolution of partner and non- and j's comparative advantages as well as by trade policies, partner shares of total imports. The typical anti-monde is but changes in the index will primarily reflect changes in the that, absent the RIA, these shares would not change. But last because the other factors are so static. Anderson and even putting aside the potential weaknesses of the anti- Norheim (1993) show how dramatically these indices monde, this approach cannot separate trade creation from changed during European integration, and Yeats (1997) trade diversion. Writing MN for imports from non-mem- shows dramatic changes for Mercosur countries after 1991. bers and MP for imports from partners, one is examining Changes in trade intensity indices do not identify actual MP1(MP+M?M). Clearly this can increase if MP rises in isola- trade diversion due to Mercosur-the increased intra-bloc tion (creation) or if MP rises as MN falls by the same trade could all be at the expense of domestic sales-but if amount (diversion). Mercosur and non-Mercosur goods do actually compete, and To identify creation and diversion, one clearly requires if shares in world markets reflect general competitiveness, an extra piece of information. One approach, emanating then changes in intensity indices do reflect trade diversion from Truman (1969), is to add sales of domestic output (D) relative to a situation of non-discriminatory liberalization. to the equation and look at shares in apparent consump- The third way to establish an anti-monde for an RIA is tion. Thus if Sp= MP/(MP+MN+D) rises at the expense of by reference to the trade patterns of other countries. The SD D/(MP+:MN+D), we have creation, whereas if it rises locus classicus for this is Aitken's (1973) use of the gravity ar the expense of SN we have diversion. For applications of model on the EEC, but see also Braga, Safadi and Yeats this approach on Europe see, for example, Winters (1993) (1994) on Latin America. In Aitken's model it is again and Sapir (1992). At an aggregate level these shares are impossible to separate trade diversion from trade creation often approximated by corresponding shares in GDP (rely- (that, ceteris paribus, imports from partners exceed those ing on the fac- that MP+MN N X, where X is total exports). from non-partners is consistent with either), but later A novel alternative approach to identifying trade diver- scholars have done so by defining "normal" trade indepen- sion is due to Yeats (1997). He recognizes that intra-bloc dent of the RIA members' behavior and then checking imports, MP, equal intra-bloc exports, Xp, and uses XN to whether MP is significantly higher than the norm (creation) suggest an an-i-monde for XP. Yeats illustrates his method and/or MN significantly lower (diversion). Recently, Bay- on Mercosur, and focuses on the compositions of Xp and oumi and Eichengreen (1996), Frankel and Wei (1996) and XN. He finds that the strongest increases in Xp occur in Sapir (1997) use this sort of approach to suggest that the products for which performance in XN is very weak. Since creation or enlargement of the EC did, in fact, cause mate- members' exports are competing with the same third- rial amounts of trade diversion, a significant change in the country exports in P and in N, Yeats infers that the much conventional wisdom. greater relative success for some of them in P is due to pref- erences. That is, X, (_ MP) displace MN in Mercosur A Diversion on the Rest of the World because of preferences-i.e., there is trade diversion. Note The discussion so far has focused on the effects of RIAs on that the anti-monde here is based on world markets and their members, but equally important is their effects on 55 TRADE. TOWARDS OPEN REGIONALISM non-members. Even the simple comparative static issue of other countries is essentially a comment on the economic how an RIA affects non-member trade and welfare has size of the RIA. If the RIA is small in the technical sense caused a good deal of confusion. The traditional approach of having no perceptible influence on the prices at which has been to consider the evolution of the share of non-mem- other countries trade, non-members should be indifferent bers in each member's to imports-i.e., MN/ (MN +M,)-or to its creation. Although if the RIA cuts its exports to the apparent consumption-i.e., MN/(MN+ MP+D). This is not rest of the world (RoW), the shortfall must fall somewhere an implausible approach to the trade effects. The anti- (not every member of RoW can replace the lost imports), monde is that, ceteris paribus, shares would have been con- the loss is so marginal that it is effectively zero.' Thus, a stant, which is probably reasonably accurate over the short critical question about the effect of an RIA on non-mem- run. As the time period of investigation increases, however, bers must be its effect on prices, or the terms of trade. it becomes less relevant, as factors such as different partners' Again, to my knowledge, no exercises of this sort have been ability to supply imports and the different income elastici- published, although recent work by Gupta and Schiff ties of demand for different imports become more signifi- (1997) and Winters and Chang (1997) is starting to cant. In these circumstances-just as for members' own address this lacuna. The former argue that unusual tastes or trade shares-a more sophisticated view of anti-monde barriers to trade with RoW can make even a small RIA shares is required, such as provided by trade intensity large to certain neighbors, and it illustrates the possibili- indices, trade propensity indices, more complex import ties with beef trade between Argentina, Peru and functions or a gravity model. Venezuela, with the formation of the Andean Pact, and While the shares of different suppliers in members' with white maize trade when the Central American Com- total imports offer some insight into the effect of RIAs on mon Market broke down. The latter authors examine the members' welfare as well as on their trade, they offer vir- relative prices of EC and non-EC exports of machinery to tually none into non-member welfare (Winters 1997a, b). Spain over the Spanish accession in 1986; it finds evidence First, even if non-member exports to the RIA were the cor- consistent with non-EC suppliers receiving lower prices. rect driver of non-member welfare, it is the absolute level, Further work on the price effects of RIAs seems a high pri- not its ratio to total RIA imports, that affects the evolution ority, although as Winters and Chang observe, it is not easy of welfare through time. The ratio is informative only rel- to do.8 ative to the gains that non-members would have made if the same expansion of members' imports, had been made Deep Integration and Security of Access in a non-discriminatrry fashion-that is, not relative to The standard discussion of RIAs proceeds as if tariffs were the initial situation but relative to an MFN liberalization. the only barrier to trade, and as if once these had been set Second, and much more important, the level of their to zero within the bloc, that was the end of the story. Of exports is a poor indicator of non-members' welfare. At the course, this is far too simple. First, there are myriad other extreme, if ceteris are strictly paribus, more exports implies costs to international trade other than tariffs-for example, fewer goods for non-member consumption, i.e., it corre- currency exchange and its associated uncertainty, customs lates with lower welfare! Of course, exports are actually a formalities, industrial standards, and unfamiliar legal means to obtaining imports, which do raise welfare, but structures. Unfortunately, we know very little about these then we should consider the effects of the RIA on non- even for industrial countries, but experience suggests they members' imports and/or their terms of trade. To my knowl- are real enough. Herin (1986) estimates that such costs edge these have never been investigated, although in Win- average 3 percent of the gross value of a trade flow, and ters (1985) I noted that British exports of manufactures to most analyses of the EC's "1992" program assume that they certain non-EEC markets fell below expected levels follow- are 2½/2 percent (see, for example, Smith and Venables 1988 ing British accession to the EEC. But even this is not par- or Winters 1992). McCallum (1995) shows that trade ticularly informative if British exports were replaced by between Canada and the United States, among the most other supplies of equivalent quality and price. integrated of partners, is subject to additional costs equiv- The observation that a non-member's trade with mem- alent to adding an extra 2,000 miles to an equivalent intra- bers of an RIA might easily be substituted by trade with U.S. trade. 56 ASSESSMENT OF REGIONAL INTEGRATION Reducing transaction costs should be an important to be beneficial, it is clearly desirable to answer such ques- objective of policy, because they not only segment markets tions before embarking upon it. and reduce competition, but typically waste real resources One final question warrants some thought: Are trade in doing so. In many cases, however, this seems feasible preferences necessary to the achievement of deep integra- only in a cliscriminatory-i.e., geographically limited- tion? In some cases, clearly not: Countries already share way. As long as different currencies and standards exist some standards (e.g., electrical power, and compatible rail- within the world economy, for example, harmonizing with way gauges and timetables) but nonetheless impose tariffs one partnet automatically precludes harmonization with on each other's goods. In other cases, probably not: Even if another. Thus, this sort of "deep integration" seems ideally countries use the same currency, they could tax mutual suited to lUAs. "Discriminatory" deep integration can trade. In yet other cases, preferences are necessary: It is hard have trade-creation and trade-diversion effects just as can to abolish frontier formalities if you wish to levy taxes on discriminatory tariff reductions, but with one crucial dif- cross-border trade. ference: If trade barriers entail the expenditure of real Deep integration concerns facilitating market access, but resources rather than the creation of rents, reducing them in the real world the security or reliability of access is nearly saves the real resources and is thus, at least in simple mod- as important-for both exporters and importers. RIAs els, welfare-improving even if it diverts trade. It is just that enhance security by making it more difficult to increase tar- you get lower gains if there is diversion than if the reduc- iffs on partner trade. Even a bound tariff can be raised under tion is non-discriminatory. Overall, therefore, discrimina- GATT rules so long as the increase is negotiated and com- tory deep integration seems unlikely to be harmful except pensation is offered, and an RIA may offer more direct routes in the sense of forgoing the greater gains from non-dis- for retaliation than does the GATT. But, in fact, raising criminatory integration.9 bound tariffs is relatively rare under the GATT because gov- If deep integration is so benign, why do we not see more ernments can achieve similar results without renegotiation of it? The reason is that it is formidably difficult and sen- via safeguards actions and without either renegotiation or sitive to achieve. It requires great mutual trust and confi- compensation via anti-dumping actions. How do RIAs dence between partners as well as a perception that their affect these instruments? interests in various issues are compatible. For example, Within the EU (as between U.S. states) allegations of achieving a single currency entails surrendering a funda- dumping would be investigated under the auspices of com- mental element of sovereignty, and forgoing a major tool of petition policy. This imposes more rigorous standards of (short-run) economic management. Harmonizing stan- proof than does anti-dumping law-so rigorous that mis- dards can entail agreeing about the trade-offs between, say, conduct is rarely claimed or identified-and also is admin- costs of production and safety, and always appears to favor istered by a super-national body, eliminating the strong the partner whose standard is adopted by the group. Expe- "them vs. us" rhetoric of dumping. Under these circum- rience in the EU amply illustrates the difficulties and slow stances, RlAs more or less assure market access, although pace of deep integration, although it may be easier to on occasion pragmatic restrictions on access have emerged achieve among developing countries with less long-lived within the EU-e.g., on cars made by Japanese firms in traditions of this sort of economic management and more EU countries. Also, EU members exercise anti-dumping fluid institutions. against third countries collectively through EU organs as is In assessing RIAs ex post deep integration is not con- required by a common trade policy. I believe that Merco- ceptually difficult except when one tries to separate the sur intends to move to the EU situation, but it has not yet integration that would occur "naturally" and via global created the necessary institutional framework. forces (e.g., through network economies or via the World Among less deep RIAs, to my knowledge only three Trade Organization) from that which is induced by the have eschewed anti-dumping: Australia-New Zealand, RIA. In ex-ante assessments, however, one needs to exercise Canada-Chile, and Iceland-EU. In NAFTA and in the EU's the self-discipline of asking exactly how and in what areas agreements with Eastern European and Mediterranean deep integration will be achieved-who will surrender countries, the northern partners have maintained the right what and how. If deep integration is necessary for an RIA to institute anti-dumping actions against their partners. In 57 TRADE TOWARDS OPEN REGIONALISM both cases the institutional arrangements have been Dynamics changed somewhat, but not in ways that seem to have Dynamics play an almost mystical role in many discussions obvious effects on the frequency or rigor of the use of the of economic integration. Having found that the static ben- instrument. In the EU cases, the parties have to refer cases efits are usually rather small or possibly even negative, of alleged dumping to association councils-relatively advocates of regional integration arrangements (RIAs) typ- high-level joint political bodies-to seek a "solution ically appeal to the dynamic benefits. However, what these acceptable to the two parties." This approach arguably constitute and how they come about are frequently rather encourages managed trade outcomes to complaints of vague, and the evidence linking dynamic benefits to par- dumping, but possibly gives the developing country part- ticular instances of integration very difficult to pin down. ner more power in negotiation."0 While the EU's use of This section defines dynamics as anything that affects a anti-dumping against Eastern European countries seems to country's rate of economic growth over the medium term. have declined since the RIAs were signed, it is still the Thus, it includes both permanent increments to the rate of subject of bitter complaint among the latter. growth and temporary but long-lived increases of, say, NAFTA, building on Canada-U.S. Free Trade Agree- more than five years, as countries move from one growth ment (CUSFTA), has established a new review process for path to another. It then briefly introduces some of the argu- anti-dumping actions, based on bilateral panels. However, ments made about dynamics and seeks to isolate their crit- the panels' terms of reference are linked to procedural rather ical components. As with the previous section, the aim is than substantive issues. NAFTA is too recent for its effects to try to put discussion of dynamic costs and benefits on a on anti-dumping to be isolated, but Bond (1997b) has more objective basis-to force us to be explicit about what studied the effect of CUSFTA on U.S. actions against Cana- we expect of RIAs. dian exporters and finds no statistically significant changes. Improvements in growth stem from accumulation Security of market access figured prominently in the either of a factor of production-specifically capital, phys- aims of the developing partners to NAFTA and the Europe ical or human-or of knowledge. Traditional growth the- Agreements; the issue of anti-dumping was negotiated ory suggests that accumulating factors will eventually run vigorously in both cases. In neither case did the developing out of steam, for as one adds ever more units of physical partners gain much on paper, but one might plausibly and human capital to a fixed stock of land and laborers, assume that their treatment has become a little gentler de their rates of return decline to such an extent that further facto-certainly it is not likely to have become worse. On accumulation appears unprofitable. Some recent endoge- the other hand, the one formal test extant has failed to nous growth theorists have sought to avoid these con- identify such benefits. For prospective members of RIAs straints by arguing that production exhibits constant the brutal truth is that the major industrial users of anti- returns to scale in capital and increasing returns to scale dumping are unlikely to give up much of their discretion, overall. On the whole, however, this is difficult to believe, and thus that one should not expect too much in this direc- and recent work by Jones (1995) has offered a refutation in tion even from far-reaching RIAs. For RIAs among devel- the case of the United States. The accumulation of knowl- oping countries, which are newer converts to the delights edge, on the other hand, could lead to permanent increases of anti-dumping, it may be feasible to reach mutual disar- in rates of growth, for as an economy grows, the return to mament agreements, but it is probably necessary that these an increment in knowledge (which is basically a public be formally part of the agreement before much reliance can good) increases, and as knowledge accumulates, the base be placed on them. for further advances expands. Finally, some commentators have hoped that RIAs will The direct evidence that RIAs stimulate growth is actu- also earn their members more sympathetic treatment under ally rather weak. Henrekson, Torstensson and Torstensson sanitary and phyrosanitary-for example, Chilean views on (1997) use a cross-section regression to suggest that Euro- joining NAFTA (see Schiff and Sapelli 1996). Except in so pean integration has enhanced members' growth rates, but far as an RIA opens new routes to disputes-settlement or other commentators-e.g., deMelo, Panagariya and facilitates closer cooperation in standards and testing, the Rodrik (1993) and Vamvakidis (1997)-have failed to do evidence we have to date is not encouraging to this view. so with similar approaches. Part of the difference lies in the 58 ASSESSMENT OF REGIONAL INTEGRATION treatment of openness. Vamvakidis identifies an EU Since this paper is concerned with economic integration, growth effect if he excludes openness from his equation, I shall concentrate on the work of just one economist-Ben- but he finds it insignificant if openness is included. He David (1993, 1994, 1995, and 1996). Ben-David (1993) does find beuseficial effects on a country's growth rate from offers very striking evidence that after they signed RIAs, large open neighbors, but this is quite independent of the EEC, the European Free Trade Association and the RIAs. To date no one has identified any growth effects from United States and Canada displayed a marked increase in non-European RIAs (see the brief discussion in Baldwin trade between members and a dramatic fall in the standard and Venables 1996). deviation of incomes per head across countries. Ben-David Attractive as the sweeping generalizations from cross- (1994) suggests that this convergence is upward-i.e., it section regressions are, they are not wholly persuasive. increases the growth rates of poorer members. They offer no information on the mechanisms through Ben-David (1995) extends the analysis to explore the which growi:h occurs, and may well be subject to simul- role of trade in creating so-called convergence clubs, small taneity problems. Thus the rest of this section looks at groups of countries (not necessarily linked by RIAs) that alternative, less direct arguments and evidence. appear to display convergence. He finds that if we create Whether we are talking about permanent changes in clubs between countries that are each other's major trading the rate of growth or only temporary ones, we must partners, convergence is a very common phenomenon, remember the distinction between output and welfare. whereas if we create similarly sized clubs by drawing coun- Accumulation is not free, so while an RIA might raise the tries randomly, convergence is very rare. This appears to growth rate of output, total economic welfare will not reinforce the view that international trade is the mecha- increase by as much, because the investment in plant, edu- nism through which convergence occurs, although it cation or knowledge has to be paid for in terms of forgone might equally well be other elements of international com- consumption. Not every increase in the rate of growth- merce, such as foreign direct investment (FDI), that are still less in in-vestment-is welfare-improving. strongly correlated with trade. Finally, Ben-David (1996) uses the same set of countries Convergence and Spillovers as the previous article to demonstrate that the convergence Convergence is the phenomenon by which countries' clubs appear to owe more to convergence in rates of total incomes per head converge toward each other over time. factor productivity growth than to convergence in rates of Over the last 10 years a huge amount of effort has been investment. This suggests that convergence arises from devoted to describing and explaining convergence. The contact and spillovers rather than from incentives to accu- classic references include Baumol, Blackman and Wolff mulate physical capital. It might also be due to the stimu- (1989); Lucas (1988); Mankiw, Romer and Weil (1992); lating effect of overseas competition on economic per- and Barro and Sala-i-Marrin (1995). The evidence suggests formance, to the direct consequences of technology rather strongly that the world's economies do not converge improvements through FDI, or even to the mobility of unconditionally, but many economists maintain that highly skilled labor, be it permanently or temporarily. economies do convergence conditionally, such that, after It would be comforting to infer from all this that a allowing for a series of other explanatory variables, the developing country has only to increase its trade with a remaining unexplained component of growth shows poorer richer country in order to experience a rapid catch-up. economies gtowing more rapidly than richer ones, and thus Unfortunately, however, this would be premature. First, converging. The explanatory variables are used either to while convergence appears between OECD countries, it is explain the steady-state levels of income per head toward not generally evident between them and other countries which each cDuntry is converging, or to explain differences (Ben-David 1994). Second, Ben-David's recent work in growth rate for a given level of steady-state income. includes very few developing countries-all middle- Among the variables that are used for these purposes are income countries in Latin America. Thus, we do not know human capital, the rate of investment, political stability how well the experience will translate to developing coun- and the openness of the economy. (For a recent survey see tries in general. Third, the extent to which preferences lie de la Fuente 1995.) behind an RIA's or a convergence club's increase in trade 59 TRADE TOWARDS OPEN REGIONALISM rather than, say, the reduction in real trade costs clearly suppliers than the EU, it could be important. The static matters. Overall, however, Ben-David's work does offer analysis of RIAs has long observed that trade diversion is some encouragement for expecting dynamic benefits, espe- potentially harmful. What I have described here is a cially from North-South RIAs. dynamic version of trade diversion. A separate stream of work that also shows the role of international trade in convergence is Coe, Helpman and Accumulation Hoffmaister (1997), who seek to explain the rate of increase I have argued that growth results from accumulation. in total factor productivity across a wide range of countries. Hence, if RIAs affect the incentives for accumulation they They construct an index of total knowledge capital (mea- will have at least temporary effects on growth. An early sured by accumulated investment in R&D) in each indus- and striking application of this insight is Baldwin (1989, trial country, and then assume that trading partners get 1992), who models the effects of integration on accumula- access to a country's stock of knowledge in proportion to tion to generate a "medium-term growth bonus." An RIA their imports of machinery and transport equipment from makes trade easier and hence tends to raise the returns to that country. Using import-weighted sums of industrial all factors of production. If the cost of capital is unchanged, countries' knowledge stocks to reflect develocountrountries' the response to increasing rates of returns is to invest more access to foreign knowledge, they seek to explain the latter's and thus to increase the capital stock. This leads to a tem- total factor productivity (TFP) growth. They find that porary, but generally long-lived, increase in growth rates as access to foreign knowledge is statistically significant. the accumulation shifts the economy onto a higher trajec- ln the case of developing countries, Coe, Helpmani and tory. Once the new steady-state level of capital stock has Hoffmaister find TFP growth is related to the interaction been achieved, there will be higher levels of output per between the openness of the economy (imports/GDP) and head, but growth will return to its original level. Baldwin its access to foreign knowledge. Thus an economy benefits (1989) suggests that the medium-term bonus could double from foreign knowledge first according to how open it is, or treble the static efficiency effects of an RIA on output. and second according to whether it is open to those coun- If we accept the basic model, the first question to ask is tries that have the largest knowledge stocks. These results whether an RIA will raise or lower a developing country's are intuitively very attractive and suggest, again, that trade rate of return to capital. A simple-minded application of is a major conduit for spillovers between countries. Unfor- the Heckscher-Ohlin model might lead us to expect that tunately, however, the conclusion is not wholly secure the latter would fall in a North-South arrangement. For because Coe, Helpman and Hoffmaister assumed rather example, comparing the EU and, say, Morocco, the EU is than tested that imports from industrial countries provide capital-abundant and Morocco capital-scarce. Since inter- the correct weights with which to combine stocks of national trade tends to reduce the returns to the scarce fac- knowledge in order to reflect importers' access to foreign tor, increased trade with the EU seems likely to reduce the knowledge. Keller (1996) has suggested that, in fact, the returns to capital in Morocco." This is a salutary thought, results are little better than would be obtained from a ran- but maybe the basic Heckscher-Ohlin model is too simple dom weighting of countries' knowledge stocks. for this purpose. A further implication of the Coe, Helpman and Hoff- First, it applies only to a so-called square model with maister model, if we believe it, is that trade policy that equal numbers of factors of production and goods; there is switches a country's imports of machinery and equipment no indication that this is the way the real world is. Second, away from the United States and Japan, which have the the Euro-Med Agreement is not a complete liberalization highest stocks of knowledge, toward other economies, of trade but a partial liberalization, which could have dif- which have lower stocks, may be harmful to the rate of TFP ferent effects. Third, the Heckscher-Ohlin model presumes growth. Winters (1996b) explores this possibility for an homogeneous products, whereas experience suggests that EU-Lebanon RIA and finds, fortunately, that, provided it many markets are better represented by a model of differ- increases general openness a little, its net effect will be pos- entiated products and intra-industry trade. In the latter itive despite the trade diversion. For other RIAs, however, case it becomes very important how substitutable domestic where trade may be diverted to less knowledge-intensive and foreign goods are for each other. Building on these 60 ASSESSMENT OF REGIONAL INTEGRATION complications Baldwin and a number of collaborators have inflows of FDI are to be beneficial, such advantages are suggested a number of reasons why one might expect eco- essential, for FDI is an extremely expensive way of borrow- nomic integration to raise the rates of returns on capital in ing capital if that is all you get. Also, while FDI might both partners regardless of capital abundance. (See Bald- boost GDP, it does less for GNP, because many of the ben- win, Forslid and Haaland 1995, Baldwin and Seghezza efits of FDI will accruc abroad. If the local cconomy is to 1996a, b, c and Baldwin and Forslid 1996.) I will briefly reap a benefit from FDI, we have to identify the route sketch four of them. through which incomes, as well as output, are stimulated. First, an RIA typically reduces the transactions cost on Among the most obvious of such routes are the ability to tradable goocds more than those on non-tradable goods. If, tax foreign companies (although sometimes governments as is commonly believed, non-tradables are labor intensive are prone to forgo this right precisely to attract the com- and tradables capital intensive, trade liberalization will panies); benefits to employment if there are pre-existing increase the (lemand for capital relative to labor, and thus unemployed resources (but then we need to know why increase the rate of return on capital. Relatedly, increased labor that was potentially employable was unemployed in competition in tradable goods sectors may induce the first place); and spillovers, mainly of the sort which we improvements in efficiency and declines in markups in this discussed briefly under the heading of convergence. sector. This will cause increased demand for inputs into the The literature on spillovers from FDI is ambiguous (see tradable sector, and thus reinforce the effect above.12 Blomstrom and Kokko 1997a), but it is fair to say that Second, an RIA may reduce tariffs and trading costs on there is at least some evidence in favor of their existence. imports of capital equipment. This would reduce the prices FDI appears in many cases to stimulate productivity in industry has to pay for imported investment goods, and local firms. In the same industry as the FDI, this may be may stimulate the domestic capital goods industry to due to demonstration effects (local firms see that things can greater efficiency and/or less monopolistic behavior, or be managed better than they currently are) or a competi- both. tion effect (local firms are obliged to adjust or go under). In Third, an RIA that goes beyond tariff reductions could other industries it may reflect forward linkages (other improve efficiency in the financial sector. If this led to forms using better inputs from the multinational com- reductions in lending margins, it would stimulate invest- pany) or backward linkages (the multinational company's ment by reducing the cost of funds. requiring and helping local suppliers to produce better Fourth, an RIA may improve the atmosphere for invest- inputs for all their customers). It may also arise from the ment by inducing greater credibility in the government's mobility of labor as the multinational corporation trains willingness or ability to pursue sound policies. labor and managers who then move on to other parts of the There is at least circumstantial evidence that RIAs can economy. The fact that the multinationals seem to pay effi- generate investment booms, as is evidenced in the invest- ciency wages-that is, wages above the going market ment data for the "Six" after the creation of the EEC, the rate-suggests that they are conscious that such spillovers Iberian enlargement of the EC, EC "1992," NAFTA and are possible and that it is to their advantage to try to hang Mercosur. Thus there is some support for this as a route to on to the skilled workers they have trained. a temporary spurt of growth. However, given the well- The evidence suggests that the principal requirement known possibilities of rising investment in the presence of for attracting FDI is sound policies at home-including other distortions, one needs to be clear that such invest- macroeconomic stability, good labor relations and open ment is economically warranted before declaring such borders. One has only to consider China and Indonesia to RIAs successful. realize that RIAs are not necessary for attracting FDI, and An important element of accumulation in some RIAs is Greece to realize that they are not sufficient. believed to be FDI. Many economists see inflows of FDI A simple RIA may reduce FDI flows between member first as harbingers of confidence in the economy, and second countries because they make trade a more attractive option. as the route through which an economy can modernize On the other hand, FDI from outside the bloc may increase for example, through access to modern technology, modern as foreigners seek to exploit new investment opportunities management, marketing networks and sources of inputs. If and to use one member as a platform for serving the whole 61 TRADE TOWARDS OPEN REGIONALISM bloc. Blomstrom and Kokko (1997b) suggest that while wanese (China) firms than German ones, but the ROOs in the U.S.-Canada FTA had little investment effect, Merco- the Euro-Med Agreements will favor the latter. sur and NAFTA (Mexico) both stimulated FDI inflows. It is easy to argue that RIAs will generate new investment Intra-RIA flows may be stimulated-as, for example, with and, in particular, increased inflows of FDI, but it is difficult the EC's Iberian expansion-by flows in non-traded sectors to do so with great confidence in either the prediction or its (e.g., several services) and, if they are made, by provisions desirability. RIAs have clearly been associated with such for capital inflows, the repatriation of profits, and enhanced changes but to establish the advantages of the RIAs on these dispute settlement. grounds one needs to show: that the RIA and not, say, MFN The changing patterns of FDI under an RIA sometimes liberalization generated the incentives, that the investment is attract the sobriquets "investment creation" and "diver- not immiserising, that spillovers occur, and that "better" sion." As mere descriptors of rising and falling shares these investment was not displaced by "worse" investment. terms are unexceptional, but then we hardly need new vocabulary for the phenomena. However, as parallels to Credibility'3 trade creation and trade diversion, and especially their wel- A common claim for RIAs is that they enhance policy cred- fare implications, these terms are dangerous. Advocates of ibility (Whalley 1996, for example). On the narrow issue the view that RIAs increase FDI rarely argue that it dis- of the instruments negotiated under the RIA-e.g., tariffs places domestic investment (the parallel with creation) or on intra-bloc trade-this seems reasonable, although one that member flows displace non-member flows (diversion). must still consider the incentives and ability for one part- Still less, too, do they argue that member flows are less ner to discipline another. An RIA probably focuses the desirable than non-member flows. Moreover, new FDI incentives better than the GATT. Under the GATT there from any source could go into the production of goods for is a large "public goods" element to retaliation-if A retal- trade diversion and thus worsen the RIA's welfare overall. iates against B for increasing a tariff, (a) other suppliers Thus, my advice is to drop the terms "investment creation" benefit directly from the lower tariff, and (b) if, as a result, and "diversion" from our vocabulary and spell out directly B becomes less prone to such anti-social behavior in future, the phenomenon we are interested in. the likely major beneficiaries are C, D.... Within an RIA In immediate violation of my own advice, let me iden- the spillover is smaller, being restricted to other member tify one very close investment parallel to trade diversion. countries. On punishment, if an RIA delivers benefits Rules of origin (ROOs) within FTAs require substantial beyond the WTO in the form of, say, investment, punish- shares of the value of a good to originate in member coun- ment for defection can be correspondingly heavier. tries before the good qualifies for the RIA's trade prefer- Whether RIAs discipline trade policy toward third par- ences. This severely handicaps potential investors from ties is moot. Clearly if trade policy is common to the RIA, outside the bloc who would, at least initially, probably individual countries have less ability to impose their own wish to source considerable shares of their inputs from barriers, but, on the other hand, any bad policy that does their traditional (non-bloc) suppliers. That is, ROOs give emerge is spread more widely-e.g., Brazil's tariffs on cap- member investors an artificial advantage. If members are ital goods (see Winters 1994). Where members set their less suitable sources of FDI than non-members, we have own policies on third-country trade, there is no formal dis- essentially investment diversion. This might be the case if cipline (or spillover), and two offsetting forces are at work. members are less efficient producers than non-members. If tariffs on partner imports are fixed at zero, those on other Consider, for example, the advantages that U.S. and EU suppliers will be lowered by the incentive to minimize dis- auto producers receive in Mexico and Eastern Europe rela- tortions between different sources of imports and raised by tive to Japanese firms. It could aLso occur if member firms the desire to protect domestic output more vigorously offer fewer opportunities for spillovers than non-member against supplies from third parties if one cannot protect it firms. The ability of local firms to assimilate information against those from partners. The latter effect depends on may decrease as the gap in knowledge and experience goods being differentiated by supplier, but, given this, between partners widens. Thus, for example, Mediter- which effect dominates depends on the degrees of substi- ranean producers may learn more from Korean or Tai- tutability between different goods. 62 ASSESSMENT Of REGIONAL INTEGRATION It is equally difficult to resolve the matter empirically, domestic rhetoric, and constitutional limitations. These do because the anti-monde for policy is so difficult to define. not have the trade-diverting risks of RIAs. For example, Mexico responded to the peso crisis of 1994/95 To summarize, locking-in and credibility of policy by raising tariffs on 500 items against non-NAFTA suppli- reform do seem to have been closely linked with RIAs over ers. Bhagwati and Panagariya (1996) see this as diverting the last few years. And it is true that many countries-not protectionist pressure onto third parties. Others argue that least in Latin America-have liberalized both globally and previous crises have witnessed far worse protectionism and regionally. However, I doubt that this confers on RIAs an thus that NAFTA has induced restraint. Bhagwati and independent ability to grant credibility. Greece has been a Panagariya respond that the intellectual atmosphere is far member of the EU since 1981 and yet subject to serious more liberal now than previously-consider, for example, macroeconomic shortcomings, showing that RIAs are not India-and, less convincingly, that vigorous protectionism sufficient for sound policy and hence presumably not par- such as seen in the 1980s sows the seeds of its own demise, ticularly good sources of credibility in themselves.'4 so that by permitting a partial response now, NAFTA could Hence, overall, if we are to believe that a prospective RIA slow down ultimate progress toward multilateral liberalism. enhances a government's credibility, I believe we should Turning to policies that are not part of an RIA, it is spell out clearly why and how. more difficult to see where credibility comes from. Fernan- dez (1997) identifies two possibilities. First, an RIA may Location raise the cost: of macroeconomic laxity because it typically A major issue in RIAs among developing countries has increases marginal leakages to imports, although, as Fer- been the location of industry, often loosely equated with nandez notes, the RIA also increases the (temporary) the location of the benefits of integration. Indeed, in previ- returns to competitive devaluation which pushes the oppo- ous decades, when governments felt they could and should site way. Second, if entering an RIA entails (political) sunk manage location, this was sufficiently contentious to costs, and if it requires liberal or sound policies to make destroy some RIAs. Now governments profess greater faith sense, entry provides the government with a signaling in the market to distribute industry efficiently, but they device, for only a government with liberal intentions still worry about outcomes and wonder whether they can would sign. Thus, in the presence of asymmetric informa- nudge the process their way. tion about the government's type, an RIA could improve Recent analysis has suggested that there may be some credibility. truth to worries that RIAs and the deepening of integra- I find the last argument a persuasive explanation for the tion within them could lead to some regions losing, or at recent interest by governments of developing countries in least failing to gain. These analyses depend on the interac- RIAs. There has clearly been a shift in perceptions of the tions between economies of scale and transactions costs. policy requirements for economic growth, and after several Krugman and Venables (1990) provided the first, very ele- decades of pursuing different policies, governments clearly gant, example, which showed how industry's preferred require means of signaling genuine changes in attitudes. location between a high-cost large-market "center" and a Whether RIAs are the best means of such signaling, how- low-cost small-market "periphery" would vary with trans- ever, and whether signaling is sufficient to justify RIAs, is actions costs between them. Briefly, at very high (prohibi- less clear. First, there is an element of circularity in the tive) transactions costs, industry locates in both places just argument. If an RIA is genuinely liberal-i.e., if it increases in order to serve both. At intermediate costs, it is feasible domestic comnpetition and creates trade-it is beneficial to trade, and if scale effects are strong enough, it pays to (relative to the status quo) in those terms alone. Thus, cred- concentrate production in one place and export to the ibility is just icing on the cake. If, however, the RIA is not other; obviously if it is not too much more expensive in liberal, it will presumably not induce much credibility. production terms, firms choose the large-market center Thus, credibility is, in some sense, not an independent char- because there they avoid transactions costs on the larger acteristic of IRIAs. Second, there are other means of signal- part of their sales. At zero transactions costs, location is ing and winning credibility-for example, binding trade determined solely by production costs, so now firms prefer policy under the WTO, accepting Article VIII of the IMF, the low-cost periphery for all their production. 63 TRADE: TOWARDS OPEN REGIONALISM This model had a salutary effect in simply disproving These models are stimulating and challenging, for they the notion that low-cost peripheries would automatically capture better than most the common perception that attract industrial activity. However, it offered little practi- countries compete and that development can be uneven. cal guidance for predicting the effects of RIAs. First, it was However, they do not yet constitute practical tools for pre- partial equilibrium in nature-costs of production were dicting the effects of RIAs, for they lie so far from measur- independent of the volume of industry. Second, the exercise able reality. Thus while it is now clear that an RIA cozld is comparatively static; it does not actually refer to firms have locational effects we are not able to predict these ex relocating. Third, we have no idea whether the irreducible ante nor, with confidence, identify them ex post. minimum transactions cost between two countries-per- haps best thought of as that between two states in the Politics United States-is high, low or intermediate. Brulhart and Regional Integration Agreements are frequently argued to Torstensson (1996) have found suggestive evidence for the be an important element of diplomacy. The classic case is EU that industry has concentrated and along the lines pre- the precursor to the European Economic Community, the dicted by Krugman and Venables, but, on their own European Coal and Steel Community, 1951, which was admission, it is not strong. explicitly seen as a way of reducing Franco-German ten- Economic geography has advanced a lot since Krugman sions-making war not only unthinkable but materially and Venables, but it is still a long way from producing impossible. Similar objectives are said to be present, if not testable or quantitative predictions because the models it so centrally, in Mercosur and ASEAN. These cases raise at uses are so far removed from reality. Among the most rele- least two sets of questions: Why use trade as the diplomatic vant advances are by Puga and Venables (1997a, b) who tool? And what implications does this have for the RIA have a model in which industry is both attracted to large itself? markets to avoid sales costs and repelled from them by On the latter, Schiff and Winters (1997) have recently their higher production costs. They use this in the latter offered a theoretical analysis. Taking as given the premise paper to explore how industrialization in two identical that trade increases understanding and harmony between developing southern countries might be affected by various partners, they show that an FTA could be an optimal pol- forms of integration arrangements. Among their conclu- icy if actors value these things. They also show that as the sions are that even if the two countries liberalize their trade FTA grows, and if it pursues deep integration, the optimal at the same rate, they will take off at different times, degree of preference for intra-bloc trade decreases. This because, at low levels of southern industrial development, translates into declining external tariffs-exactly the pat- agglomeration economies make it inefficient to split indus- tern observed in the EU. They also examine the optimal try between two sites. Although eventually both develop- degree of preference as a bloc widens its membership. ing countries fare the same, there are clear benefits to being There is a weak presumption that this too could lead to the first to industrialize in discounted terms. Moreover, the declining tariffs, but this is easily overturned by changes in lagging developing country can actually lose before it particular circumstances. industrializes itself. Thus developing countries are at least On the question of why use trade diplomacy-i.e., partly competing against each other. whether trade is the appropriate or optimal tool-little Puga and Venables also suggest that while unilateral lib- analysis exists. Mansfield (1993) relates politics and RIAs, eralization by a developing country stimulates its industri- but with the former "causing" the latter and a requirement alization, signing an RIA does better (for the member (for his argument to make sense) that the RIA be welfare developing country). While precise results depend on para- enhancing. Fawcett and Hurrell (1995) also see causation meter values and the level of remaining trade costs, about running in this direction, with RIAs as part of the response both of which we have almost no information, their results of nation states to the erosion of their power by globalism. suggest that North-South RIAs dominate South-South That is, states pool sovereignty in an RIA in order to RIAs. Multilateral liberalization dominates all options for a achieve sufficient leverage to manage global pressures. developing country except for a North-South RIA outside From this perspective it is not clear whether RIAs are which the excluded developing country refuses to liberalize. desirable for developing countries, for it is not clear that 64 ASSESSMENT OF REGIONAL INTEGRATION managed solutions are better than just accepting the role of References a small operl economy. Aitken, N. D. (1973). "The Effect of the EEC and EFTA on Euro- In the case of European integration Milward, Brennan pean Trade: A Temporal Cross-Section Analysis," American Eco- naomic Review, Vol. 63(5), pp. 881-92. and Frederico (1992) have advanced similar arguments. noiReewVl.65)p.8192 Alogoskoufis, G. (1995). "The Two Faces of Janus: Institutions, Pol- These clearly have some validity, for Europe has generated icy Regimes and Macroeconomic Performance in Greece," Eco- material economic power in those areas in which it can act nomic Policy, Vol. 20, pp. 147-92. in a concerted fashion. One should also observe, however, Amjadi, A. and L. A. Winters (1997). "Transport Costs and 'Natural' that in the European case several other forms of diplo- Integration in Mercosur," Policy Research Working Paper Series, . . ~No. 1742, World Bank, Washington D.C. macy/integration were tried and failed-e.g., the Political No. A., Ld Bank, angts D.C. and Defense Communities in the 1950s. Thus, in a sense Amjadi, A., L. A. Winters and A. Yeats (1995). "Transportation and DfsCmuteite 0Th,iCosts and Economic Integration in the Americas," Swiss Journal trade diplornacy was all that was left. of Economics and Statistics, Vol. 13(3), pp. 465-88. The political dimension of RIAs does seem an impor- Anderson, K. and H. Norheim (1993). "From Imperial to Regional tant one, but one over which the analyst needs to exercise Trade Preferences: Its Effect on Europe's Intra- and Extra-regional caution. Cooperation without tariff preferences seems per- Trade," Weltwirtschaftliches-Archiv; Vol. 129(1), pp. 78-102 fectldfree trade does nor guar- Baldwin, R. (1989). "The Growth Effects of 1992," Economic Policy, fecely possible if one wishes It, and free War, does was Vol. 9, pp. 247-81. antee peace. Witness the American Civil War, which was ____ (1992). "Measurable Dynamic Gains from Trade," Journal of partially caused by disagreements over trade policy. Thus, Political Economy, Vol. 100, pp. 162-74. to justify an RIA on political grounds one needs to show Baldwin, R. and R. Forslid (1996). "Trade Liberalization and how preferences contribute to the political rapprochement Endogenous Growth: A q-Theory Approach," National Bureau of and to show that such a rapprochement is valuable. The Economic Research, Working Paper Series, No. 5549. latter requir-es (a) that rapprochement is real and (b) that it Baldwin, R. and E. Seghezza (1996a). "Testing for Trade-Induced Investment-Led Growth," Centre for Economic Policy Research, would not have happened anyway. Discussion Paper Series, No. 1331. (1996b). "Trade-Induced Investment Led Growth," mimeo, Conclusion Graduate Institute of International Studies, Geneva. The body of' this paper has identified several potential ben- - (1996c). "Growth and European Integration: Towards an efirs from RIAs and a few possible drawbacks. The chal- Empirical Assessment," Centre for Economic Policy Research, lenge is not in thinking up how RIAs could he beneficial, Discussion Paper Series, No. 1393. Baldwin, R. and A. Venables (1996). "Regional Economic Integra- but in offeri.ng convincing reasons why they will be or have tion," in Grossman, G. and K. Rogoff (eds.), Handbook of Interna- been. I have argued elsewhere (Winters 1996a) that, as well tional Economics, Vol. III, Amsterdam: North-Holland. as generating trade diversion, regionalism can be harmful Baldwin, R., R. Forslid and J. Haaland (1995). "Investment Creation in terms of arresting progress toward liberal multilateral and Investment Diversion: A Simulation Study of the EU's Sin- outcomes. Hience, while they may well provide benefits tel- gle Market Programme Centre for Economic Policy Research," ative to the status quo, RIAs also entail apotentialdown Discussion Paper Series, No. 1308. ative to the status quo, RIAs also entail a potential down- Barro, R. J. and X. Sala-i-Martin (1995). Economic Growth, New side risk. York: McGraw Hill. In discussing RIAs I believe that there is no alternative Baumol, W. J., S. A. B. Blackman and E. N. Wolff (1989). Produc- to the careful exploration of the various sources of costs and tivity and American Leadership: The Long View, Cambridge, Mass.: benefits. By this I am most definitely not saying that we MIT Press. should condut carful rade reatin an divesion xer- Bayoumi, T. and B. Eichengreen (1996). "Is Regionalism Simply a Diversion? Evidence from the Evolution of the EC and EFTA," cises and ignore everything else on the grounds that it can- Centre for Economic Policy Research, Discussion Paper Series, not be measured. Rather, we owe it to each other to spell No. 1294. out the benefits that we have in mind and their channels of Ben-David, D. (1993). "Equalizing Exchange: Trade Liberalization causation, so that their bases can be discussed and exam- and Income Convergence," Quarterly Journal of Economics, Vol. ined explicitly. Even if we cannot quantify and add up the 108, pp. 653-79. (1994). "Income Disparity Among Countries and the Effects of various effects, we can at least be clear and identifyareasof Freer Trade," in Pasinetti, L. L. and R. M. Solow (eds.), Economic agreement and disagreement. I hope this paper has helped Growth and the Structure of Long Run Development, London: Macmil- us to do this during the rest of this conference. lan, pp. 45-64. 65 TRADE TOWARDS OPEN REGIONALISM _ (1995). "Convergence Clubs and Diverging Economies," Frankel,J. A. and S-J. Wei (1996). "Regionalization of World Trade Foerder Institute Working Paper, pp. 43-95. and Currencies: Economics and Politics," Chapter 7 of Frankel _ (1996). "Trade and Convergence Among Countries," forth- (ed.), The Regionalism of the WorldEconom', Chicago: Chicago Uni- coming: Jotsrnal of International Economiecs. versity Press. Bhagwati, J. N. and A. Panagariya (eds.) (1996). The Economics of Frankel, J. A., E. Stein and S-J. Wei (1994). "Trading Blocs: The Preferential Trade Agreements, Washington D.C.: The AEI Press, Natural, the Unnatural, and the Super-Natural," University of and College Park, Maryland: Center for International Economics, California at Berkeley Center for International and Development University of Maryland. Economics Research Working Paper: C94-034. Blomstrom, M. and A. Kokko (1996). "Multinational Corporations Gupta, A. and M. Schiff (1997). "Outsiders and RTAs among Small and Spillovers," Centre for Economic Policy Research, Discussion Countries: The Case of Regional Markets," (mimeo). Interna- Paper Series, No. 1365. tional Trade Division, World Bank, Washington D.C. _ (1997a). "How Foreign Direct Investment Affects Host Coun- Haaland, J. I. and V. Norman (1992). "Global Production Effects of tries," Policy Research Working Paper Series, No. 1745. World European Integration," in Winters, L. A. (ed.), Trade Flo's and Bank, Washington, D.C. Trade Policy After 1992, CEPR and Cambridge University Press, _ (1997b). "Regional Integration and Foreign Direct Invest- pp. 67-88. ment: A Conceptual Framework and Three Cases," Policy Haaland, J. I. and I. Wooton (1992). "Market Integration, Competi- Research Working Paper Series, No. 1750. World Bank, Wash- tion and Welfare," in Winters, L. (ed), Trade Flo's and Trade Pol- ington, D.C. icy After '1992,' CEPR and Cambridge University Press, pp. Bond, E. W. (1997a). "Using Tariff Indices to Evaluate Preferential 125-146. Trading Arrangements: An Application to Chile," Policy Harris, R., and D. Cox (1996). "Qualitative Assessment of the Eco- Research Working Paper Series, No. 1751. World Bank, Wash- nomic Impact on Canada of Sectoral Free Trade w:th the United ington, D.C. States," Canadian Joirnal of Elconomics, Vol. 19, pp. 377-94. _ (1997b). "The Impact of Canada-U.S. Free Trade Agreement Harrison, A., T. F Rutherford and D. Tarr (1997). "Trade Policy on Antidumping Filings and Decisions," (forthcoming) Policy Options for Chile," (forthcoming) Policy Research Working Research Working Paper Series. World Bank, Washington, D.C. Paper Series. World Bank, Washington, D.C. Braga, C. A., R. Safadi and A. Years, (1994). "Regional Integration Henrekson, M., J. Torstensson and R. Torstensson (1997). "Growth in the Americas: Deja Vu All Over Again?" World Economy, Vol. Effects of European Integration," European Economic Review, Vol. 17(4), pp. 577-601. 1134. Brulhart, M. and J. Torstensson (1996). "Regional Integration, Scale Herin, J. (1986). "Rules of Origin and Differences between Tariff Economies and Industry Location in the European Union," Cen- Levels in EFTA and the EC," EFTA Occasional Paper, No. 13, tre for Economic Policy Research, Working Paper Series, No. Geneva: EFTA. 1435. Jacquemin, A. and A. Sapir (1991). "Competition and Imports in the Coe, D. T., E. Helpman and A. W. Hoffmaister (1997). "North- European Market," Winters, L. A. and A. J. Venables (eds.), Earo- South R&D Spillovers," Centre for Economic Policy Research, pean lntegration: Trade and Indaistry. Cambridge, New York and Discussion Paper Series, No. 1133. Melbourne: Cambridge University Press, pp. 82-91 de la Fuente, A. (1995). "The Empirics of Growth and Convergence: Jones, C. (1995). "Time Series Tests of Endogenous Growth Models," A Selective Review," Centre for Economic Policy Review, Dis- Qnarterly3Jsornal of Econorsics, Vol. 110(2), pp. 495-525. cussion Paper Series, No. 1275. Jones, R. (1993). "Regionalism vs. Multilareralism: Analytical de Melo,J., A. Panagariya and D. Rodrik (1993). "The New Region- Notes' Comment," in de Melo, J. and A. Panagariya (eds.), New alism: A Country Perspective," in de Melo, J. and A. Panagariya Dimensions of the Regional lntegration, New York: Cambridge Uni- (eds.), Neu Dimensions in Regional Integration, New York: Cam- versity Press, pp. 79-83. bridge University Press, pp. 159-92. Keller, W. (1996). "Are International R&D Spillovers Trade- European Commission (1996). "Economic Evaluation of the Internal Related? Analyzing Spillovers among Randomly Matched Trade Market," Ezeropean Economy, Reports and Studies, No. 4. Partners," mimeo, Department of Economics, University of Wis- Fawcett, L. and A. Hurrell (1995). Regionalism in World Politics: consin-Madison. Regional Organization and International Order, New York: Oxford Krugman, P. (1991). "The Move Towards Free Trade Zones" in Pol- University Press. icy Implications of Trade and Currency Zones, a symposium sponsored Fernandez, R. (1997). "Returns to Regionalism and Evaluation of by the Federal Reserve Bank of Kansas City. Jackson Hole, Non-Traditional Gains from RTA," (forthcoming) Policy Wyoming, pp. 7-42. Research Working Paper. World Bank, Washington, D.C. Krugman P. and A. J. Venables (1990). "Integration and the Com- Forouran, Faezeh (1997). "Does Membership in an FTA Make a petitiveness of Peripheral Industry," in Bliss, C. andJ. de-Macedo Country More or Less Protectionist?" mimeo, International Trade (eds.), Unity with Diversity in the European Commanity: The Comma- Division, World Bank, Washington, D.C. nity's Sozsthern Frontier, Cambridge: Cambridge University Press. 66 ASSESSMENT OF REGIONAL INTEGRATION Lucas, Robert E. (1988). "On the Mechanics of Economic Develop- macy," Policy Research Working Paper Series. World Bank, ment," Journal of Monetary Economics, Vol. 22, pp. 3-42. Washington, D.C. Mankiw, G., 1). Romer and D. Weil (1992). "A Contribituion to the Smith, A. and A. J. Venables (1988). "Completing the Internal Mar- Empirics of Economic Growth," Quarterly Journal of Economics, ket in European Community: Some Industry Simulations," Euro- Vol. 107, pp. 407-37. pean Economic Review, Vol. 32, pp. 1501-25. Mansfield, E. J1993). "Effects of International Politics on Regional- Srinivasan, T. N. (1993). "Regionalism vs. Multilateralism: Analyt- ism in International Trade," in Anderson, K. and R. Blackhurst ical Notes' Comment," in de Melo, J. and A. Panagariya (eds.), (eds.), Regional Integration and the Global Trading System, New Neuw Dimensions of the Regional Integration, New York: Cambridge York: St. Martin's Press, pp. xxiv, 502, p. 1919. University Press, pp. 84-9. Mazumdar, J. (1996). "Do Static Gains from Trade Lead to Medium- Summers, L. (1991). "Regionalism and the World Trading System," Run Growth?" Journal of Political Economy, Vol. 104 (6), pp. in Policy Implications of Trade and Currency Zones, a symposium 1328-37. sponsored by the Federal Reserve Bank of Kansas City. Jackson McCallum, J. (1995). "National Borders Matter: Canada-U.S. Hole, Wyoming: 295-302. Regional Trade Patterns," American Economic Review, Vol. 85 (3), Truman, E. M. (1969). "The European Community: Trade Creation pp. 615-23. and Trade Diversion," Yale Economic Essays, No. 9, pp. 201-57. Michaely, M. (1976). "The Assumptions of Jacob Viner's Theory of Vamvakidis, Athanasios (1997). "Regional Integration and Eco- Customs UJnions," Journal of International Economics, Vol. 6, pp. nomic Growth," mimeo, International Trade Division, World 75-93. Bank, Washington, D.C. Milward, A. S., G. Brennan and R. Frederico (1992). The European Viner, J. (1950). The Customs Unions Issue. Carnegie Endowment for Rescue of the Nation-State, Berkeley: University of California Press. International Peace, New York. Nitsch, V. (1996a). "Natural Trading Blocs: A Closer Look," pre- Whalley, J. (1996). "Why do Countries Seek Regional Trade Agree- pared for European Economic Association Meeting, Istanbul, ments?" National Bureau of Economic Research, Working Paper August. Series, No. 5552. (1996b). "Do Three Trade Blocs Minimize World Welfare?" Winters, L. A. (1985). "Separability and the Modelling of Interna- Review of International Economics, Vol. 4, pp. 35 5-63. tional Economic Integration: UK Exports to Five Countries," Panagariya, A. (1997). "Preferential Trading and the Myth of Natural European Economic Review, Vol. 27, pp. 335-53. Trading Pa.rtners," Japan and the World Economny. Vol. 9, pp. 471-89. _ (1992). Trade Flows and Trade Policies after '1992,' Cambridge Puga, D. and A. J. Venables (1997a). "Preferential Trading Arrange- University Press. ments and Industrial Location," (forthcoming)Journal of Interna- _ (1993). "The European Community: A Case of Successful Inte- tional Economics. gration," in de Melo, J., and A. Panagariya (eds.), New Dimensions (1997b'. "Trading Arrangements and Industrial Development" in Regional Integration, New York: Cambridge University Press, (forthcoming), Policy Research Working Paper Series. World pp. 202-28. Bank, Washington, D.C. (1994). "The EC and Protection: The Political Economy," Sachs, J. D. and A. Warner (1995). "Economic Reform and the European Economic Review, Vol. 38, pp. 596-603. Process of Global Integration," Brookings Papers on Economic Activ- _ (1996a). "Regionalism versus Multilateralism," World Bank ities, No. 1. pp. 1-118. Policy Research Working Paper Series, No. 1687. Sapir, A. (1992). "Regional Integration in Europe," EconomicJournal; _ (1996b). "Lebanon's Euro-Mediterranean Agreement: Possible Vol. 102 (415), pp. 1491-506. Dynamic Effects," mimeo, The Lebanese Centre for Policy Studies. (1993). "The European Community: A Case of Successful Inte- (1997). "What Can European Experience Teach Latin America gration? Discussion," in de Melo, J. and A. Panagariya (eds.), about Integration?" Working Paper Series 215, Washington, New Dimensions in Regional Integration, Cambridge University D.C.: Integration and Regional Programs Department, Inter- Press, pp. 230-33. American Development Bank. (1997). "Domino Effects in West European Trade: 1960-92," _ (1997a). "Regional and the Rest of the World: The Irrelevance CEPR Discussion Paper Series, No. 1576. of the Kemp-Wan Theorem", (forthcoming) Oxford Economic Papers. Schiff, M. (1996). "Small is Beautiful: Preferential Trade Agreements _ (1997b). "Regional and the Rest of the World: Theory and and the ]:mpact of Market Size, Market Share, Efficiency, and Estimates of the Effects of European Integration," (forthcoming) Trade Policy," Policy Research Working Paper Series, No. 1667. Review of International Economics. World Bank, Washington, D.C. Winters, L. A. and W. Chang (1997). "Regional Integration and the Schiff, M. and C. Sapelli (eds.) (1996). Chile en el NAFTA. Centro Prices of Imports: An Empirical Investigation," (forthcoming) Internacional para el Desarrollo Econ6mico. Policy Research Working Paper Series. World Bank, Washing- Schiff, M. and L. A. Winters (1997). "Regional Integration as Diplo- ton, D.C. 67 TRADE: TOWARDS OPEN REGIONALISM Winters, L. A. and Z. K. Wang (1994). Eastern Esrope's International 4. There is some dispute about what Viner actually said-see Trade, Manchester University Press, distributed by St. Martin Michaely and Viner and the references therein in theJournal of Inter- Press in U.S. and Canada, New York, p. xvi, 189. national Economics, February 1976. Wonnacort, P. and R. Wonnacott (1981). "Is Unilateral Tariff 5. Also note that Wonnacott and Wonnacott's gains depends only Reduction Preferable to a Customs Union? The Curious Case of on there being transport costs on the rest of the world, not, as they the Missing Foreign Tariffs," Amierican Economic Review, Vol. 71 claim, on the relative size of these and intra-bloc transport costs (4), pp. 704-14. (Amjadi, Winters and Yeats 1995). Wonnacott, P. and R. Wonnacott (1992). "The Customs Union Issue 6. 1 discount an early effort by Harris (1986) because it faiLed to Reopened," Manchester School of EconoZmics and Social Studies, Vol. provide a coherent view of imperfectly competitive pricing. See 60(2), pp. 119-35. Winters (1992). World Bank (1996). Global Economic Prospects, World Bank, Wash- 7. The loss of imports is matched by a loss of exports, so the wel- ington, D.C. fare loss arises from the (small) loss in the efficiency of transforming Yeats, A. (1997). "Does Mercosur's Trade Performance Raise Con- one into the other, not from the loss of the whole bundle of imports. cerns about the Effects of Regional Trade Arrangements?" Policy 8. Models of differentiated products all predict price effects Research Working Paper Series, No. 1742. World Bank, Wash- (because all demand curves are downward sloping), but this has not ington, D.C. yet been exploited empirically. 9. This statement is true of simple cases with constant returns to Notes scale and no other distortions. Once these are introduced, deep RIAs 1. This is not saying that countries ignore the effects of their poli- can be harmful-see, for example, Haaland and Wooton (1992) and cies on other countries, only that the feedbacks via systemic effects the section on dynamics (subsection on location) where costs and are so small and uncertain that few small- or medium-sized countries economies of scale interact. It is also true that where deep integration will internalize them in policy decisions. involves harmonizing standards or policies, they need to be harmo- 2. It is true that all policy advice occurs in a second-best context, nized at "sensible" levels if doing so is to be beneficial. but in other areas the connection, or degree of interaction, between 10. Also, signing a Europe Agreement has been the signal for the the remaining (immovable) distortions and those whose removal is EU's treating an eastern partner as a market, rather than a non-mar- being contemplated is rarely as close as that between, say, tariffs on ket, economy. These issues are not formally linked, however, and are wheat imported from Argentina and wheat imported from Canada. also conceptually separate. Thus, generally speaking, second-best is a bigger problem in discus- 11. Mazumdar (1996) shows a similar problem if liberalization sions of RIAs than elsewhere. favors the labor-abundant commodity. 3. The force of this argument was brought home to the author 12. In the case of efficiency increases, this effect depends on the when trying to explain the limitations of the first Europe Agree- elasticity of demand being high enough to increase demand by more ments to a meeting of senior officials and negotiators from the EU than efficiency improvements reduce the inputs required per unit of and the Visegrad countries just after negotiations had finished. output. The reception was universally and vigorously hostile, despite the 13. This section draws freely on Fernandez (1997). fact that, subsequently, many of the problems forecast arose and 14. In fact, Alogoskoufis (1995) suggests that EU membership the Agreements were modified. See Winters and Wang (1994, worsened Greece's macroeconomic policy by subsidizing its profli- Chapter 3). gacy-fortunately not a likely outcome in most other developing country RIAs. 68 Comment ROBERT D E V LI N _I - HIS IS AN EXCELLENT PAPER, WHICH CRITICALLY REVIEWS WHAT HAS BEEN DONE IN THE area of integration and how it has been done. In the introduction of his paper Alan Winters states that his objective is not so much to make an assessment of regional integration as such, but rather to focus on "how" one might go about making such an assessment and also to pro- vide directions for future research. I would like to concentrate my very limited time on this central issue and make some practical suggestions. Evaluating regional integration processes and their costs Third, as Alan points out, regional integration is very and benefits is no easy task. Part of the problem is the much a second-best world where generic prescriptions can nature of the subject matter. be especially dangerous. First, regional integration is a complex general-equi- These characteristics place great burdens on analysts. librium phenomenon with dynamic processes, making it Regional integration is often evaluated in light of what difficult to dissect causal explanation. As Alan points out, would have happened in its absence. Moreover, economists the process involves issues that link growth to technol- are interested in measuring changes in welfare (i.e., being ogy, learning, externalities, political economy and poli- better off); given the complications of defining this for a tics, all of which economists have trouble grappling with particular subregion, they often use a proxy expressed in a at a national level-not to mention among several coun- summary statistic reflecting growth or trade. tries simultaneously. A further serious complication one However, it is well known that counterfactual analysis faces finds in Lacin America is that the integration processes a daunting epistemological problem: Contrary to fact condi- are an integral part of profound structural reforms that tionals can never be verified by realizing their antecedent (the have touched virtually all levels of the economy and cre- "if " clause); thus, the resulting explanation is never correct or ate big changes. Moreover, initial conditions, and the incorrect, but rather only persuasive or not persuasive. phases and sequencing of these reforms, are usually quite We also know that counterfactuals are more likely to be different arnong the partner countries of an integration persuasive: (a) the more simple the causal process studied; agreement. (b) the shorter the time period in question; (c) the smaller Second, regional integration is a medium- to long-term the changes considered; and (d) the less analysis turns on process. When successful, one expects to see initial costs exact magnitudes. Reflecting back on the characteristics of compensated by benefits that play out over the medium our subject matter, one sees that counterfactual analysis of and long term. integration is challenged on all these counts. Robert Devliri is the Chief of the Division of Integration, Trade and Hemispheric Affairs of the Inter-American Development Bank. 69 TRADE TOWARDS OPEN REG[ONALISM Another problem is that conclusions about integration our present state of ignorance of Latin American integra- rarely are based on the entire story. Much of the debate cen- tion, instead of examining what would have happened in ters on the static trade-creation and trade-diversion effects the absence of integration, we might want to spend more first pointed out by Viner. This is partly because many time discovering what is actually happening and how it is economists consider these effects to be the fundamental happening. In effect, one would examine the different dimension for evaluating regional integration. One prob- objectives of a specific integration arrangement, assess lem, however, is that the static analysis frequently uses a whether these different objectives are being realized, and partial-equilibrium framework to jump to general conclu- begin to catalogue the causal factors contributing to devel- sions about a process that is a general-equilibrium phe- opments, without necessarily being overly concerned about nomenon. Worse, the mere existence of trade diversion precise weights. For example, one frequently stated objec- alone in the new integration agreements (never mind the tive of regional integration is to enhance competition; net effects with trade creation) has sometimes been the hence, we can examine how sectoral markets are changing basis for categorically negative evaluations of a regional their competitive structure and the forces behind that. Is agreement, and regionalism more generally. intra-industry specialization increasing in the subregional More importantly, trade creation and diversion is clearly market? Are the different parameters of the integration only part of the story, and many other economists would agreement stimulating firms to invest? Are firms' technol- argue that it is not the major part. This is because the net ogy and cost structures improving in the direction of benefits of the dynamics of integration are suspected to be greater international competitiveness, and is there room to several times larger than their static reallocation effects. reduce preferences? Are new international exports and Problems exist here, too, because our models of dynamics comparative advantages emerging out of experiences in the and empirical foundations for testing them are very defi- subregional market? cient, so much so that Alan has characterized analysis in this This type of research is at "ground zero" and examines area as "mystical." It is true that the empirical foundation the integration agreement from the bottom up. Field of dynamic analysis is still weak. Nevertheless, the models research does not generate elegant and timeless analytical of dynamics are sufficiently specified to suggest that the structures. It is time-consuming and expensive, often benefits behind the dynamics of integration are potentially requiring the building of primary data bases. It also will large. Therefore, it is worth the effort to go beyond static not generate summary statistics of welfare or permit cate- trade creation-diversion analysis (which has its ambiguous gorical evaluations of integration processes. Nevertheless, dimensions as well) to begin to better understand, even if it has four potential benefits: only very imperfectly, the longer-term dynamics. 1. It will allow for better observation of what is actually The empirical bottlenecks to understanding Latin happening in the different dimensions of integration. American integration should not be underestimated. Even The analyst gets "inside" the process where the action basic data such as the evolution of preferences, rules of ori- is and examines the dynamics of sectoral markets and gin, non-tariff measures, intra-regional investment flows firms that actually move the process forward. and firms' cost structures, etc., are unavailable or incom- 2. By working at relatively low levels of aggregation plete. The many gaps sometimes induce questionable ad one might be able to identify causal factors that are hoc compromises in our analytical techniques or cause us to not easily captured in existing theory or more aggre- ignore phenomena altogether through the convenient use gated analysis. of ceteris paribus. Better data development and more field 3. While such analysis will not permit the adding up of research will not eliminate the debate over regionalism, effects into a summary statistic of welfare, the exam- but it would certainly help to ground the debate more in ination of multiple disaggregated dimensions of an reality, and it probably would help narrow our differences. integration process will permit a series of analytical The starting point is to better complement our powers vignettes that, taken together, can build a tentative of scientific deduction with much more empirical field picture of whether the integration process at least is work and case studies of the disaggregated dimensions of moving in the direction of achieving its goals in the dynamics of regional integration. In other words, in strategic areas. 70 ASSESSMENT OF REGIONAL INTEGRATION 4. The empirical work will feed our economic modeling structive analysis that sheds light on them also should be of integration with better informed assumptions and welcome. Moreover, vigilance about costs is extremely better data for testing, important. On the one hand, while fashionable, not all In sum, this approach does not suggest that we abandon integration arrangements make economic sense. On the modeling ancd counterfactual analysis, which are key tools other hand, even those that do can potentially go awry. of our profession. Rather, it suggests that more intensive Finally, a word about semantics. As Alan points our, interaction between deductive and inductive methods will words often carry labels; for example, "discrimination" enhance our powers of discovery and evaluation of a process (bad); "neutral" (good) and "natural" (good). However, in a that is ever mlore present in the world economy. second-best world, one should be careful about placing a Until we have a better ability to measure and evaluate priori normative labels on our terminology. Discriminatory the full story of integration, we should be careful about our policy can be good or bad depending on the circumstances conclusions. Any major transformation has costs, usually and its effects. For instance, areas dense in capital and pop- concentrated up front. Therefore, it is no surprise that inte- ulation tend to invest and specialize relatively intensively. gration has costs. For instance, regional integration is a If such areas have activity truncated by borders and other strategic compromise among different countries in differ- transactions costs, it may be useful to use discriminatory ent economic conditions; for this reason trade diversion is preferences to accelerate commercial interaction and cap- hard to avoid, and the initial redistributive effects from ture the externalities of location and agglomeration. It is lost tariff revenue can be important, especially in North- true that this might occur anyway, but it may bc at a slow South agreements such as NAFTA. However, countries jus- and uncertain pace with costs for delays. Neutrality sounds tify these costs by the greater benefits that are expected, good, but since initial conditions among countries may which are spread out over a longer period of time. When differ, some discriminatory policy may be justified to avoid examining these up-front costs, analysts should be careful a potentially destabilizing or inequitable distribution of to interpret them as only a piece of a story that plays out gains. Natural market forces sounds good, but we know over a longer term, and should refrain from categorical that these can lead to bad things such as monopoly, which overall assessments, except in the most extreme cases. can cause a society to discriminate against successful firms. Meanwhile, since regional integration is a strategic deci- In sum, Alan's warning about the power of language is sion, participants should have their objectives clearly artic- certainly another useful suggestion for assessing regional ulated. One objective is to minimize costs; thus, any con- integration. 71 Comment R I C AR D O FF RE N C H - D AV I S IT T IS A PLEASURE TO BE HERE AND TO HAVE THE OPPORTUNITY TO COMMENT ON THE PAPER BY Alan Winters. It is an excellent paper, well balanced, that surveys many issues related to regional integration. He emphasizes the need of making explicit the benefits and costs of regional inte- gration; actually, anybody who does economic policy should try to make some sort of measuring of the benefits and costs of any economic policy. It is difficult, but one should try to think about what one is building with economic policies, and that, of course, includes regional integration. We should try to be the most rational possible, given the limitation of information and the complex issues we are deal- ing with. I want to make three points of strong agreement with percent (quantum) per year on average. That is 1.5 times Alan Winters. Then I will present five points that I think faster than world trade in that period. However, GDP bring forth a balance in favor of regional integration, but growth was only 1.6 percent. It matters how exports were with some qualifications, on the road to a global economy. fostered and what sort of exports were growing-not only The three points of agreement: The first is related to the quantity, but the quality of exports. trade creation and trade diversion. I agree that they are Let us try to move toward the effects on economic wel- merely indicative, rather than definitive concepts. As he fare and not take a given reform at face value, whether it is says, it is risky to jump to economic welfare conclusions. PTAs, unilateral import liberalization or other reforms. They are useful indicators, but we have to go beyond that. The relevant question is whether they represent inputs for Creation is not completely positive, and diversion is not economic welfare growth. In our area of concern today, how completely negative (with a little bit of positive effect in do unilateral trade liberalization and PTAs affect the the consumer level). There is much more beyond them. I sources of growth? would like to add that we know that trade creation may Third, growth depends on accumulation of human and have negative effects in the transition, if it is abrupt; and physical capital-combined with more efficiency in a diversion may end to be positive under cost-reducing framework of more complete markets. One of the problems economies of scale and learning-by-doing. of developing economies is that they have several missing The second point is that we must go beyond trade itself. markets. Markets of knowledge, labor training or long- We must not measure the quality of trade policy simply by term capital are incomplete or are missing in many what happens directly on trade. We should look more economies. So we have to build those markets, or simulate broadly-to economic welfare. For instance, it is a fact that or complement them. We will argue below that some sorts exports grew vigorously in Latin America in the 1980s, 7 of regional integration can help to build some of these mar- Ricardo Ffrench-Davis is a Regional Adviser in the Office of the Secretariat of the Economic Commission for Latin America. 72 ASSESSMENT OF REGIONAL INTEGRATION kets, PTAs being in some relevant cases superior to unilat- obvious conclusion that overall unilateral liberalization is eral liberalization, the optimal national policy and better than any PTA, par- The five points that I will mention, with analytical and ticularly among developing countries. policy implications for PTAs, are associated with incom- Why, then, do so many nations want to be involved in plete markets, cost of information, a role for generating integration processes, even in these times of fashionable systemic comperitiveness and within that for the acquisi- free trade? Regional integration builds on strategic consid- tion of comparative advantage. So let us see these points erations arising from imperfect and incomplete markets at and the framework within which we place them. home and abroad, which handicap the spread of efficiency gains in certain sectors and the development of new pro- Analytical lBasis for PTAs ductive patterns with progressively higher degrees of value Deep trade reforms have been undertaken in the Latin added. The five issues that follow, related to trade in goods America and Caribbean Region (LAC) as part of a broad- and services, provide analytical bases to support regional ranging process of change, in which international compet- integration arrangements with preferential import itiveness and exports play a leading role. Most countries are regimes. One crucial assumption we adopt is that regional seeking an export-led development. Nonetheless, in con- integration takes place in a framework of open regionalism, trast with the experience of East Asian nations, the main with "moderate" external tariffs. instrument of trade reform has been a rather indiscriminate First, world markets are not widely open and stable. and rapid liberalization of imports (see Agosin and Nonetheless, they are broad, growing 50 percent faster than Ffrench-Dav+s 1995; ECLAC 1998, Chapter V). Most GDP in the last half century, and reaching one-fifth of countries in the region introduced reforms that could be world GDP. They are dynamic, but not completely open described as drastic and sudden. Generally speaking, the and stable. On the other hand, LAC's exports are concen- tariff protection provided at present differs considerably trated in natural resource-based primary and semi-manufac- from its pre-reform levels, and the spread of rates of effec- tured commodities. Thus, with or without participation in tive protection have diminished substantially. For instance, PTAs, world markets have been and will continue to be cru- the simple average external tariff was reduced from 45 per- cial for LAC's traditional exports; instability actually pre- cent by the rnid-1980s to 13 percent today. No country has vails in those markets, but it refers more to prices than to yet adopted a zero tariff rate. These regional trends in trade access (or volume). However, for many nontraditional prod- policy have been complemented by a drive toward imple- ucts (including nontraditional natural resources), access to menting bilateral or multilateral free-trade agreements, markets is more limited and unstable. These other products covering a wide spectrum of items. The fact that tariffs are tend to be a crucial part of the growth process, and, actu- different from zero but with moderate levels leaves space ally, they were very significant in the development drive of for reciprocal tariff preferences but with more limited trade East Asian nations. It is for these type of products that PTAs diversion than in earlier integration programs. become relevant to foster a diversifying growth of exports. The conventional literature on the benefits and costs of Regional integration can be important for these nontra- economic integration focuses on tariff preferences in a ditional products, locking in better access to markets. It framework of optimal competitive equilibrium. This equi- can be a device to foster a diversification of exports, toward librium is assumed to be disturbed only by the existence of output more connected to overall competitiveness of our import restrictions. In this framework, integration is ben- economies. eficial only if it implies a move toward free trade-that is, Second, given those distortions or restricted access in if the effects of trade creation (shift toward cheaper sources world markets, economies of scale and specialization are of supply) are larger than those of trade diversion (shift more difficult to secure for a country or producer in the toward more costly sources of supply). The crucial issue, process of learning to export. To lock in improved access to however, is how costs are measured; in the standard regional foreign markets helps to make use of those approach it is at actual market prices net of tariffs, assum- economies, and in fact this achievement has been a leading ing away transitional costs and incomplete markets, as well target of policymakers and a force encouraging regional as acquirable competitiveness. The assumptions lead to the integration. As a consequence, in the face of economies of 73 TRADE TOWARDS OPEN REGIONALISM scale, what otherwise would be a costly trade diversion can tion; thus, in the transition period, they provided a net become a cost-reducing and welfare-enhancing trade diver- positive balance of pulls for the domestic output of trad- sion (Corden 1972; Ffrench-Davis 1980). ables (encouraging use of capacity, and of investment to Third, domestic factors markets are incomplete or dis- increase that capacity). Given LAC's option for an import- torted. Labor training, technology and long-term capital led reform, a parallel process of regional PTAs becomes are scarce, with nonexistent or infant markets in LAC. more attractive, in order to increase the efficiency of the These market failures are more significant for nontradi- productive transformation (ECLAC 1995, Chapter V). In tional exports of differentiated products, whether of nat- fact, PTAs add a compensatory ingredient to unilateral ural resources, manufactures or exportable services. These import liberalization (and more so if the exchange-rate had markets have been improving, but they are still quite far appreciated in the process), fostering reciprocal exports in from maturity. If access to external markets is improved for tandem with reciprocal imports. Hence, the doses of posi- the exportables intensive in those ingredients of develop- tive and negative pulls (impulses) to economic activity and ment, it can strengthen the effectiveness of efforts to com- investment are more balanced with PTAs than is the case plete markets and dilute segmentation. in pure unilateral import liberalization. Nontraditional exports, differentiated products and All the restrictions are severe obstacles to the expansion products more value-added and intensive in knowledge are of production, and trade in goods and services relatively more typical of the export basket to regional integration more intensive in knowledge content and longer learning than the export basket to the rest of the world (see ECLAC curves, elements which are now recognized as key compo- 1998, Chapter III). nents of the growth process. Regional integration can be a FoUrth, infrastructure, rrade financing and knowledge of strategic tool to partially overcome these obstacles (see markets (marketing channels, organized transportation, Devlin and Ffrench-Davis 1998) by: standards, etc.) are often biased against intra-regional trade v Expanding market size to facilitate greater specializa- in LAC. All these special "factors" of trade traditionally tion and industrialization through economies of scale have been more developed for deals with the "center" while and possibilities to exploit economies associated with they are nonexistent or more rudimentary for trade among the agglomeration of productive activity; LAC's neighbors. This is a significant variable explaining A Enhancing the forces of competition, enlarging a mar- why intra-regional trade has been lower in LAC than what ket with guaranteed reciprocal access and intensifying the gravity of geography suggests. Thus, PTAs can help to the specificity of information flows, all of which in put things in a more competitive stage by removing these turn should induce new domestic investment and per- unnatural, artificial obstacles to reciprocal trade, being a mit better conditions to attract efficient foreign direct second-best, efficient move toward the potential frontier. investment (FDI); So, preferential treatment in regional integration might * Creating the security of subregional market access and help to leave the gravity of economy to operate more naturally. exploiting the familiarity of neighborhoods, which Fifth, in economies like those in LAC which have been combine to accelerate the emergence of new producers reforming trade policies, sliding away from excessive and and traders of nontraditional exports. In fact, the arbitrary protection for import substitutes and inputs of learning curve associated with experience in subre- exportables, there tends to emerge significant transitional gional markets can serve as a platform for new inter- costs. These are enhanced if the exchange rate happens to national exports. The expected enhanced international appreciate, as it has been the case in most LAC countries in competitiveness, brought about by regional integra- the 1990s. tion, should build confidence and prepare countries East Asian nations minimized transitional costs in the for globalization and further advances in multilateral 1960s and 1970s with an export-led strategy for opening liberalization. to the world economy (see Agosin and Ffrench-Davis 1995). That is, in their opening processes, nations like Intra-Regional Trade Japan, Republic of Korea and Taiwan, China put a stronger Total intra-regional exports almost tripled between 1990 emphasis on export promotion than on import liberaliza- and 1996. Initially it was principally a recovery from the 74 ASSESSMENT OF REGIONAL INTEGRATION sharp drop of the 1980s. However, given a notably rapid To appreciate the strategic dimension of integration, it growth, shorcly the prior peaks were reached. A record was is necessary to examine the profile of intra-regional exports. achieved in 1992, with an additional jump in 1993-94. Intra- and extra-regional exports from Latin America dis- Subsequently, the Tequila crisis transitorily reduced the play marked differences in their product structure and share of intra-regional exports, particularly those to Argen- technological content, with manufactures accounting for a tine and Mexican markets. Nonetheless, Mercosur shows a much larger share of intra-regional commerce, as shown in persistently rising share of reciprocal trade among partner the table below. countries: it jumps from 9 percent in 1990 to 21 percent The profile of intra-regional trade contributes to a drastic in 1996. change in the composition of LAC's exports: The predomi- It is interesting to compare GDP growth, total exports nance of primary exports was partially replaced by manufac- and intra-regional exports (all in real terms). GDP of Latin tures, which now account for one-half of intra-regional trade. America grew 20 percent between 1990 and 1996, while This notable increase in manufactured exports corresponds total exports rose 59 percent. Within these, intra-regional especially to new industries, including both labor-intensive exports expanded 160 percent, while to extra-regional and capital-intensive activities. It is highly relevant to notice markets they rose 44 percent. That is faster than world that border trade (with neighboring countries) represents trade, with a gross GDP elasticity of 2.2. These data sup- the bulk of the intra-regional trade, a new proof that geog- port the hypothesis of open regionalism, with trade grow- raphy matters (Devlin and Ffrench-Davis 1998). ing fast with all markets, but with a rising share to the To sum up, intra-regional trade, because of its charac- partner's destinations. teristics, complements LAC's linkages with the global Latin Americai 114 Countriesa: Composition of Exports by Destination, 1970-74 and 1995 (In percentages) INTRA LATIN AMERICA WORLD 1970-74 1995 1970-74 1995 A. Primary commodities 51.0 19.7 53.6 32.2 1. Agricultural products 11.7 10.3 29.9 16.0 2. Mining pcoducts 1.0 2.4 6.2 4.0 3. Energy products 38.3 7.0 17.6 12.2 B. Industrialiized products 48.8 79.7 46.0 65.6 R. Semi-marnufactures 23.3 29.9 33.6 30.5 1.1 Based orl agriculture and labor-intensive 7.5 10.3 9.5 11.9 1.2 Based ott agriculture and capital-intensive 3.1 4.8 6.0 5.2 1.3 Based orl minerals 6.4 8.4 9.2 9.1 1.4 Based oni energy 6.2 6.5 8.9 4.4 2. Manufactured goods 25.5 49.8 12.4 35.1 2.1 Traditional industries 7.2 12.4 5.2 8.4 2.2 Basic-input industries 4.8 10.2 1.9 7.2 2.3 New labor-intensive 7.0 10.7 3.0 8.3 a) Medium technological content 4.1 6.3 1.6 4.5 b) High technological content 3.0 4.4 1.4 3.8 2.4 New capital-intensive 6.5 16.6 2.3 11.2 a) Medium technological content 5.0 14.7 1.7 10.1 b) High technological content 1.5 2.0 0.6 1.2 C. Other 0.2 0.5 0.4 2.2 Total 100 100 100 100 So,rce: ECLAC (-998), Table 111.9, on the basis of offici.l data. aArgentina, Bolvia, Brazil, Chile, Colombia, Costa Rica, Ecuador, Guatemala, Honduras, Mexico (excluding maquila), Paraguay, Peru, Uruguay and Venezuela. Exports with low technological coatent are summed up with traditional industries. 75 TRADE TOWARDS OPEN REGIONALISM economy and provides a dynamic context of technological Devlin, R. and R. Ffrench-Davis (1998). "Towards an Evaluation of apprenticeship, leading to greater international competi- Regional Integration in Latin America in the 1990s," Occasional tiveness and a more diversified, balanced pattern of spe- Paper No. 1, Integration, Trade and Hemispheric Issues Division, Integration and Regional Programs Department, Inter-American cialization. We are in the middle of straightforward open Development Bank. The Netherlands: FONAD (forthcoming). regionalism (ECLAC 1994). ECLAC ( 994). Open Regionalism in Latin America and the Caribbean, United Nations publication, Santiago. References _ (1998). Policies To Improve Linkages with the Global Economy, San- Agosin, M. and R. Ffrench-Davis (1995). "Trade Liberalization and tiago, 1995, and updated edition in Spanish with Fondo de Ca/l- Growth: Recent Experiences in Latin America," Journal of Inter- tara Economica, Santiago. American Stadies and World Affairs, Vol. 37, No. 3, Fall. Ffrench-Davis, R. (1980). "Distorsiones del mercado y teorfa de las Corden, M. (1972). "Economies of Scale and Customs Union The- uniones aduaneras," Integraci6n Latinoamericana, No. 44, Buenos ory," Jo.rnal of Political Economy, Vol. 80, No. 3, Chicago, Aires, marzo. May-June. 76 III. Andean Pact Trade Flows in the Andean Countries: Unilateral Liberalization or Regional Preferences? J UA N J O S E EC H AVARRi A L l ATIN AMERICA DID IN 10 YEARS "THE EQUIVALENT OF ABOUT FIVE GATT ROUNDS, SIX Gramm-Rudmans, and more deregulation than had been accomplished by the Carter and Rea- gan administrations together" (Williamson 1990, p. 356). Tariffs were slashed, quotas elimi- nated, labor and financial markets deregulated and government-owned business privatized. Important complementary reforms were adopted in such areas as infrastructure and foreign -di investment. The drive toward "open regionalism" was always seen as another central element of this broad strategy, and it adolpted two related paths. On the one hand, great efforts were made to reintroduce trade and com- petition as the key variables behind healthy integration (de-emphasizing industrial planning and sectoral development). On the other hand, new agreements were signed with other countries and regions (i.e., with Mexico, Chile and CARICOM), promoting additional trade and reducing trade diversion in the region. The common unilateral domestic reforms emphasizing steps toward more challenging future hemispheric negoti- trade and open markets were the main factors behind the ations. Exclusion from large blocks could be very costly, impetus reached by the integration schemes, with "lock- and trade wars and domestic pressures could lead to larger in" effects sometimes sought (i.e., the signing of the protectionism in the United States. The small Andean Common External Tariff, or CET). However, the pattern, countries were rightly concerned with the quality and reli- speed and characteristics of Andean integration were also ability of future access to the larger hemispheric markets shaped by 'larger hemispheric developments. In particu- (Whalley 1996). lar, the signing of Mercosur and NAFTA in 1991 and This paper tries to assess the impact of the unilateral 1992, the IMiami Summit of 1994 and the Trade Minis- and multilateral integration reforms on the evolution of terials held in Denver, Cartagena, Bello Horizonte and regional trade flows. Did reforms matter? If so, was the Costa Rica in 1995-98 stimulated the deepening of large expansion of exports toward partner countries largely Andean integration. determined by lower tariffs and apertura (opening the mar- The agreements with Mexico, Chile and CARICOM ket) along a "natural" pattern, or by regional preferences had their own merits, but were also seen as pretiminary and artificially created trade deviation? Juan Jose Echavarrfa is a consultant in the International Trade Unit of the Organization of American States (OAS). The author wishes to thank Larisa Caicedo for his able research assistance, the advice given by the members of the Trade Unit at the OAS and the financial sup- port of the Andean Finance Corporation. The collaboration of the Cartagena Agreement Board, the Latin American Integration Association and the Inter-American Development Bank were crucial in the completion of this paper. 79 TRADE: TOWARDS OPEN REGIONALISM The next section briefly describes the relative (that is, Trade flows were very dynamic in the period 1986-96, compared with other Latin American countries) depth and especially trade in manufactures, with a large geographical characteristics of the unilateral reforms adopted in Bolivia, reorientation toward the Andean markets in 1990-95. The Colombia, Ecuador, Peru and Venezuela in the financial fourth section offers a brief presentation on the main charac- and labor markets, and in the areas of taxes and privatiza- teristics of trade flows in the region; it shows that the effect tion. It reviews the depth and speed of trade reforms. of unilateral liberalization with regional preferences was The third section presents a brief history of the main important in 1993-95, increasing Andean trade by a maxi- institutional developments in the Andean Group, the char- mum of 60 to 80 percent. Interestingly, the exercise based on acteristics of the Common External Tariff (CET) signed in the so-called "gravity" equation, does not support Yeats's con- November 1994, and the process of regional liberalization tention that trade expansion in Mercosur was artificially cre- that took place between 1990 and 1992. The Andean coun- ated by regional preferences. The section also shows that tries created "the most sophisticated integration scheme to trade creation significantly dominated trade diversion in foster import substitution at the regional level (and with the 1986-95 for all sectors in Colombia and for most of them in most complex set of regulations to implement the agree- Venezuela. It specifies import functions and suggests that the ment)" (Nogues and Quintanilla 1992, p. 285) which, of rise in regional trade flows was determined by unilateral lib- course, never worked. Among the reasons for this failure are eralization rather than by regional preferences. that the goals were extremely ambitious, the countries were The final chapter of this paper further characterizes poor, and intra-regional trade represented less than 2 percent Andean trade flows in terms of the level of intra-industry of total trade at the time. Industrialization became the trade (IIT), revealed comparative advantage (RCA) and main-really the only-goal of integration, and industrial labor skills. The participation of intra-industry trade planning was considered to be the ideal mechanism to com- increased to levels close to 40 percent in regional trade, pensate the small countries for their lack of new industry. which suggests that the adjustment to large changes in Chile withdrew in 1976, and liberalization agreements trade flows took place inside the firms and did not imply were systematically violated by the remaining five partners large realignments in domestic production. We also show during the 1980s. It was initially thought that more import that the most dynamic exports to the region took place substitution and protectionism was needed (Nogues and partially because of geographical reorientation, but that Quintanilla 1992, p. 285), but it was soon realized that they were not especially far from the "average" in terms of those "minor" reforms were worthless; a whole new strategy capital intensity or revealed comparative advantage. was required. "Open regionalism" was implemented after the end of the 1980s. Trade and competition were seen as Unilateral Liberalization central channels to increase welfare, especially for the small countries;' industrialization was the result and not the main Broad Reforms factor behind economic growth. Sebastian Edwards (1995b) considers that the fundamental It was easier to agree on common objectives in this new forces behind the economic policy shift were mainly scenario. It is no coincidence that countries signed the CET "local," rising from "the soul-searching that began in Latin and dismantled all regional exceptions during the 1990s. America in the early 1980s." This, in turn, was partially The goals of the Cartagena Agreement had been finally caused by the failure of the heterodox programs imple- achieved after a delay of 12 to 15 years. mented in the mid-1980s, the example of the East Asian The second part of Section III considers the character- countries with outward-oriented policies and the reinter- istics of the CET and of those exceptions to regional lib- pretation of the Chilean experience. We should not forget eralization in place before 1990. Tariff preferences cer- that Latin America's participation in world trade had been tainly decreased during the period 1986-95 (and decreasing decade after decade, from levels falling from 1990-95) with a huge reduction in tariff (and para-tariff) close to 11 percent in 1948 to less than 3 percent in 1990 levels. The exceptions dismantled were also important, (Norheim, Finger and Anderson 1993).2 but did not change the whole trend produced by unilat- Stephan Haggard (1995) gives a larger role to "external" eral liberalization. forces like the international debt crisis of 1982 and the 8() ANDEAN PACT policies of the international financial institutions.3 All of 1985 or 1995, and the index for slow reformers was below these forces played a role, but it is likely that "local" forces the mean both in 1985 and in 1995. The authors show a played a larger role in the Andean countries in their condi- wide diversity of experiences in the Andean region. Bolivia tion as "latecomers." The average debt-to-GDP ratio was was one of the four early reformers of the region (with 0.9 when the early and second-wave reformers (Edwards Argentina, Chile and Jamaica); Peru is a recent reformer; 1995a, Table 1.1, pp. 2-3) undertook their reforms, and and Colombia a gradual reformer. Finally, Ecuador and 0.54 when the third-wave reformers adopted theirs. Obvi- Venezuela are considered slow reformers. ously, lower debt ratios reduced the leverage of the inter- This classification could be useful when discussing the national institutions. Debt and external forces were not the future of "open regionalism" in the Andean countries since, main factors behind the new policies adopted in Venezuela, as we argue in this paper, "open regionalism" is the out- where reforrns started in 1989, Colombia (1990-91), come of unilateral liberalization. It should concern us that Ecuador (1990)4 or Peru (1991). Instead, innovative ideas the two largest countries of the region are in the group of were forcefully pushed by newly elected presidents and by gradual (Colombia) or slow reformers (Venezuela). technocrats. Looking at the different areas of reform, we see that only Lora and Barrera (1997) argue that the structural Bolivia undertook "very deep" reforms (ranking lower than reforms of the last decade increased Latin America's per 5) mainly in the areas of exchange-rate differentials (rank- capita income 12 percent, and that the average growth ing of 2), taxes (3) and privatization (4). Only in a few cases potential of GDP in the region would have been only 1.9 do the Andean countries qualify as strong reformers (i.e., percent without the reforms. It is also likely that with the rankings between 1 and 10): Colombia and Venezuela in reforms trade will regain some of the importance and the area of exchange controls, Ecuador in taxes, and Peru impact it once had (see above). and Venezuela in the area of privatizarions. The deepest reforms occurred in the realm of trade and Hufbauer and Schott (1994, Chapter 5, especially Table finance. Multiple exchange rates were dismantled, and 6.1, p. 102), built "readiness indicators" for integration most restrictions on capital flows and requirements for the with each country's neighbors for 23 countries of the region expatriation of export revenues were eliminated. The aver- in 1990-92. They considered seven variables: price stabil- age differential between the market price of foreign ity, budget discipline, external debt, currency stability, exchange and the official rate was 2 percent in 1995, as market-oriented policies, reliance on trade taxes and func- compared with 72 percent in 1989. Venezuela was the only tioning democracy. Policies were graded from 1 to 5, with Latin American country that reversed the exchange-rate higher grades for more "neutral" or "orthodox" policies. liberalization process, doing so temporarily in 1994. The Andean countries again get a combined grade of Other reforms were less ambitious: Value-added taxes 3.4, better than the grade for the Central American Com- replaced more distortionary taxes, and marginal income mon Market (CACM, 2.7) and Mercosur (3.1), but worse taxes were reduced, but the effectiveness of the tax systems than for NAFTA (4.4), Chile (4.4) and CARICOM (3.7). remained low due to problems of evasion and lack of good Venezuela (3.9) does a little bit better, and Peru much administration. Privatization was highly concentrated in worse (2.1) than the other three countries (3.7 for Colom- few countries, and only five of them (out of 26 considered) bia and Bolivia; 3.4 for Ecuador). Schott and Kotschwar showed privatizations averaging more than 1 percent of (1996) repeated the exercise for more recent years with a GDP a year during the 1990s. Finally, reforms in the labor large improvement of the grades given to Peru after its market were limited, with only five countries carrying out reforms in 1992. Rajapatirana (1998) finds some negative significant reforms. aspects of economic policy in Colombia in 1994-97, Based on relative rankings for the initial (1985) and mainly in the area of trade, with increased protection for final (1995) years, the authors (Lora and Barrera) divide agriculture and more use of anti-dumping policies. their sample into "early," "recent," "gradual" and "slow reformers." The first group had a high level of "neutrality" Tariffs (high index'l in economic policy both in 1985 and 1995, Figure 1 summarizes the evolution of trade-weighted bor- the recent and gradual reformers had a low index either in der taxes (tariffs and para-tariffs) in the Andean countries 81 TRADE TOWARDS OPEN REG[ONALISM FIGURE 1 Ad-Valorem Tariffs in the Andean Countries, 1986-95 (weighted averages) 70 - 70 - ~~~~ ~~ ~~~~~~~~~~~~~~~~64 64 60 - 1986 1992 1988 1994 -ll 50 - Ej1990 F~ 1995 40- 37 ;~~~~~~~~~~~~~~2 29i, D 30 -~2 10 Bolivia Colombia Ecuador Venezuela Peru Andean Group So-rs: Table A. 1, LAIA and Data Inral Aletohdiogy Para-rarifs included, imports of 1993 used as weights between 1986 and 1995, the solid line representing the compared with the three or four levels for the CET. The level of the CET. Information is also provided on tariff lev- standard deviation decreased year after year, from levels els in 1995 for Bolivia and Peru, the two countries that did that were particularly high in Ecuador and Venezuela. The not sign the CET (see below). Table A-I (in the Appendix) maximum level reached by total border taxes in 1986 was gives additional information on the number of tariff levels 169 percent in Ecuador, 138 percent in Peru, 135 percent and the standard deviation of border taxes and tariffs. in Colombia, 113 percent in Venezuela and 38 percent in Total border taxes decreased in the Andean Group from Bolivia. The maximum tariff was reduced to 35 percent 36 to 37 percent in 1986-88 to less than 12.8 percent in with the adoption of the CET. 1995 (because of the CET's Decision 3705), the initial lev- All countries, except Venezuela, had important para- els being much higher in Peru (64 percent) and Colombia tariff, "disguised" border taxes before 1992, representing (45 percent) than in Ecuador (32 percent), Venezuela (26 84 percent of the official tariff in the year in which they percent) or Bolivia (21 percent).6 This is consistent with were definitely removed in Bolivia (1986), 63 percent in the view that, with the exception of Bolivia, the Andean Colombia (1988), 56 percent in Peru (1990) and 24 per- Group countries were "late-comers" in shifting economic cent in Ecuador (1992). policies. On the other hand, trade reforms were completed Did tariffs increase as a result of the adoption of the in less than three years in most cases, which means that the CET? We will explore this issue further in the section speed of trade reforms was faster than in many other Latin below, but the global averages contained in Figure 1 sug- American countries.7 gest that changes were rather small. Tariffs did not increase Dispersion was also reduced drastically. In 1986 there in Venezuela, and increased only marginally in Colombia were 73 border-tax8 levels (46 tariff levels) in Peru, 61 (36) (from 11.0 percent to 12.8 percent) and Ecuador (from in Ecuador, 54 (45) in Bolivia, and 41 (41) in Venezuela, 10.1 percent). Bolivia did not adopt the CET, and Peru 82 ANDEAN PACT will do it on:y in the year 2006. Again, we must empha- among the countries, nor did it assure-by means of size that the average tariff adopted after the signature of investment-planning on a regional scale-the equitable Decision 370 could be lower since, according to the agree- distribution of the benefits of integration" (Ffrench-Davis ment, countries could adopt lower tariffs for goods nor pro- 1977, p. 138). Integration was the panacea, and it was duced in the region. thought that its benefits should go to domestic producers Information on quotas and non-tariff barriers is always rather than to foreign enterprises. In such a dream it did difficult to gather and evaluate, although it is clear that not matter much that countries were poor, entrepreneur- they were pervasive during the 1980s in most Latin Amer- ship lacking and economic interaction practically nonexis- ican countries. Thus Edwards (1995a, Table 5.2, p. 126) tent; intra-regional trade represented less than 2 percent of shows that in 1985-87 non-tariff barriers covered 73 per- total trade at the time.10 cent of imports in Colombia and were also widespread in "Export pessimism," the United Nations Economic Com- Ecuador (59 percent), Peru (53 percent), Venezuela (44 mission for Latin America and the Caribbean (ECLA), and percent) and Bolivia (25 percent). the so called "dependency school"-which in its more We do not know much about the global or sectoral extreme forms argued that trade caused underdevelopment effect of import restrictions on protection, but the scarce and poverty-did not create a healthy environment to pro- evidence seerns to suggest that sectors highly protected by mote trade and commercial integration. Socialist President quantitative restrictions also had the highest tariffs. Thus, Salvador Allende in Chile and the left-leaning military gov- raw materials and capital goods were much easier to import ernment in Peru were not defenders of trade and free markets. than finished goods, food and agricultural products. The Bogota Declaration in August 1966 and the Decla- Most non-tariff barriers were eliminated in the region ration of the American Presidents in April 1967 were the and, as shownl by Edwards (1995a, p. 126), the coverage of two central preliminary steps toward the creation of the non-tariff barriers was 0 percent in Bolivia, Peru and Andean Group. The Bogota Declaration considered the Ecuador in 1991-92, 1 percent in Colombia and 5 percent creation of subgroups with similar levels of development in in Venezuela. No (or very few) non-tariff barriers have been LAFTA; and the Declaration of the American Presidents adopted by the Andean countries after 1992 according to contemplated the possibility of speeding up subregional the yearly evaluation undertaken by LAIA.9 integration processes and of going deeper than LAFTA (Lozano and Zuluaga 1998). The issue was officially settled The Andean Community when LAFTA accepted Venezuela's accession to the Andean Pact in 1973 (Guerrero 1979, p. 218). It became irrelevant Early Histoiy in 1980 when LAFTA evolved into LAIA (Latin American The movement toward regional integration began in Latin Integration Association), which allowed countries to sign America during the 1950s and took concrete form with the subregional agreements without extending concessions to signature by 11 countries of the Latin American Free Trade other partners (Decision 44). Association (LAFTA) in 1960. Integration arrangements The Andean Group was created by the Cartagena Agree- were supposed to overcome the limits imposed by small ment, signed by Bolivia, Chile, Ecuador, Colombia and Peru domestic markets on an import-substitution development on May 26, 1969. Venezuela joined in 1973," and Chile strategy, allowing economies of scale and providing com- withdrew in 1976. The two principal organisms responsible petitiveness and a training ground for local firms (De la for designing and implementing policies were the commis- Torre and Kelly 1992). Important results were obtained in sion (executive body) and the junta (technical body)."2 Deci- 1960-66 but decreased markedly after 1967 (Aitken and sions taken by the commission were adopted automatically Lowry 1973: Wonnacott and Lutz 1989, Table 2.2, p. 76). in the countries, not requiring congressional ratification or Many thought at the time that the lion's share of the approval. The Andean Court of Justice was created in 1979, benefits generated by LAFTA would go to countries al- inspired by the European institutions, though countries have ready relative well-off-i.e., Mexico, Brazil and Argentina- been reluctant to settle disputes formally through this and that LAFTA was too "trade oriented": "It did not mechanism. It seems that only recently has it been instru- include measures to guarantee balanced development mental in the area of intellectual property rights. 83 TRADE. TOWARDS OPEN REGIONALISM The policy instruments of the Andean Group were: Besides, all products being negotiated were placed in * a liberalization program among the member countries "limbo" and could not be liberalized while negotiations * the CET continued. * the sectoral development programs During the first 10 years the technical body of the * a common policy toward direct foreign investment. agreement presented proposals for eight industrial pro- Both the CET and regional liberalization were supposed grams, but countries only approved those for petrochemi- to be in place in 1980, 10 years after the agreement was cals and machine tools (the other programs were automo- signed (see below), but there were multiple delays, and the biles, fertilizers, steel, electronics, chemicals and initial agreements were re-written in 1976 and 1978 (Lima pharmaceuticals). and Arequipa Protocols), and again in 1987 (Quito's Pro- A "typical" program specified the list of products (tariff tocol).'3 Only in the 1990s were the two goals achieved. items) included in the program, a very high CET,'4 the Regional obstacles were finally removed between 1990 and path of regional liberalization and the list of products 1992, and the CET was finally signed in 1994. We explore assigned to each country. Production of these goods rarely these two issues below. materialized. It was too difficult to produce some of the goods in countries like Bolivia, and the larger countries Sectoral Development Programs. Looking at the never fulfilled their commitments not to allow production Andean experience, Hirschman argued that every time a of the "new" sectors in their domestic markets (Urrutia new development model failed in Latin America, a more 1981, pp. 76-7). ambitious model was tried (quoted by Urrutia 1981, p. 74). Trade and integration were the central elements in the orig- Foreign Investment. The initial idea was to have a com- inal conception pushed by Colombia and Chile, but things mon policy in the area, but the final result was a policy changed later on. Sectoral development programs, industri- designed to exclude multinationals from the benefits of the alization and industrial planning received much more industrial programs. Decision 24 in December 1970 pro- attention during the first two decades of the agreement. hibited investment in some areas (namely, mining, public The emphasis on industrial planning and industrializa- utilities and the financial sector), created strict controls on tion was the key to most of the problems suffered by the technological transfer and patents, considered limits on Andean integration scheme during almost three decades. foreign ownership,'5 and restricted profit remittances to 14 Guerrero (1979, p. 359), for example, argues that the percent of total investment. It was adopted against the will design and implementation of the industrial programs of the private sector in all five countries (Guerrero 1979, p. received almost all the attention from the organisms cre- 239). ated in Lima, with very meager results. No country wanted For Carlos Diaz Alejandro (1976) Latin America would to be left out of any industrial program, and the technical have done much better during the 1930s had integration body of the Cartagena Agreement ended up distributing schemes been in place at the time, and he argued that inte- little bits of each program to all five countries (Guerrero gration schemes were really designed for that purpose. 1979, p. 223). Unfortunately, history did not pass this test, and regional The two small countries, Ecuador and Bolivia, pushed agreements did even worse in periods of international cri- this approach with the support of Venezuela and Peru. sis, when countries drastically cut all imports from their Ecuador and Bolivia considered industrial planning the neighbors of "non-essentials." Regional trade flows were ideal tool to counterbalance their lack of new industry and greatly reduced during the 1980s, and the Andean Group to tame the undesirable effects of market forces. In the virtually collapsed (Ocampo and Esguerra 1995, p. 124). same direction, Venezuela's private sector argued that they could not compete in a situation of an "overvalued" Open Regionalism exchange rate because of large oil exports (see Urrutia The new and crucial developments of the Andean Group 1981; Guerrero 1979). Even worse, these countries condi- after 1989 were part of the same new global strategy, tioned any agreement on liberalization and trade to the emphasizing free markets and trade. The crucial turning industrial programs (Guerrero 1979, pp. 234; 261-63). points were the presidential summits in Galapagos in 84 ANDEAN PACT 1989, in La ?az in 1990 and in Caracas and Cartagena in * Goods included in the "reserved list" (nmina de 1991, with important advances in the areas of trade liber- reserva); alization and foreign investment, intellectual property * Products totally liberalized after February 28, 1971, rights and transportation. not produced in the region; At the Galapagos Summit the presidents agreed to * Products liberalized after April, 1970, included in the eliminate adrninistered trade and to create a free-trade zone Common List in LAFTA; by December 1993 (1995 for Bolivia and Ecuador) and a * Products whose tariffs were linearly (and automati- customs union by December 1997 (1999 for those two cally) reduced after December 31, 1970; small countries). Those dates were accelerated in La Paz to * The Steel Program: (a) products not liberalized 1991 and 1995, and again in Cartagena to 1991. Bolivia because they have not been assigned to any country; was allowed to keep its flat tariff. (b) products totally liberalized after September 1972, After four years of intense and complex negotiations, a whose production was assigned to a particular coun- final decision was finally adopted in 1994 (Decision 370) try; (c) products for which liberalization began after and a CET was in place by February 1995. "Geographic the secretariat verified that national production isolation" andl its small size were the arguments used to started in a particular country; and allow Bolivia to keep the flat rate adopted in 1985/86, but * The Petrochemical Program: (a) products included Peru was nor a small country with special geographical under Agreement No. 6 in LAFTA, totally liberalized conditions. After years of negotiation and "transition peri- by Colombia; (b) products not liberalized to a partic- ods," Peru decided to keep its flat rate and officially with- ular country because the product was not assigned to drew from the Andean Group after the presidential sum- that country; (c) products totally liberalized after Sep- mit in Sucre in April 1997. Recently, there was an tember 1975; (c) products liberalized linearly and agreement that will allow Peru to keep its "special status" automatically. in the Andean Group until its full adoption of the Andean The truth is that only some few "enlightened" individ- CET in the year 2006. uals understood what was going on-mainly bureaucrats The treatment of foreign investment was drastically who themselves negotiated the agreements, and rarely peo- modified in IMarch 1991 (Decision 291): Limits to profit ple from the private sector. This made it difficult to do remittances were eliminated, as were excluded sectors. business in the region. Major changes were also undertaken in the area of intellec- Here is our best effort to describe the liberalization tual property (Decisions 313 and 344), where national and process adopted after 1990.17 We must basically consider most-favored--nation treatment were given, patents were products in four categories: (a) goods included in liberal- protected for 20 years and "parallel imports" allowed. Most ization agreements negotiated under LAFTA; (b) goods not relevant, goods could be imported from the partner coun- produced in the region-and not included in the industrial try when the product was registered but not used. New programs; (c) goods produced under industrial programs; negotiations are currently underway in the areas of trade in and (d) other goods. services and government procurement. LAFTA defined a list of products (the "Common List") to be liberalized in different subperiods and using different Regional Liberalization modalities. Products covered in the First Tranche of the Common List were liberalized in 1969 (when the agree- Dismantling Regional 'Exceptions' after 1990. It is not ment was signed). The rest of the Common List and those easy to present a clear picture of the liberalization process goods not produced in the region (category b) were slated undertaken in the Andean Group after 1990 because of the to be liberalized before December 1971. This did not hap- many exceptions and special programs designed by differ- pen, and their status remained in limbo."8 Goods included ent governments in the past.16 Just to have a rough idea of in category d should be fully liberalized by 1980. The orig- what is involved, let us consider the different categories inal commitments were never implemented, and periods utilized by Garay (1981, p. 108) in his analysis of regional were extended in Lima (1976), Arequipa (1978) and Quito liberalization in Colombia during the 1960s: (1987). 85 TRADE: TOWARDS OPEN REGIONALISM As for the industrial programs (category c), we turn first tions for Colombia, Ecuador and Venezuela, and for the to the Petrochemical Program, which was started in 1970 three countries as a group. and was partially negotiated before the creation of the Close to 25 percent of imports from the region in 1992 Andean Group under the so-called Complementarity Pro- were exempted from regional liberalization, exceptions gram in LAFTA (No. 6, which had Peru, Bolivia, Chile and being higher in industry than in agriculture or mining, Colombia as subscribers). Production rights were assigned and higher in "transport equipment" (83.7 percent of to different countries, and the commitment was made not imports exempted), "textiles, footwear and clothing" (60.6 to promote investment in the other countries. The pro- percent), "consumer goods in general" (31.4 percent), and gram was adopted in 1975, and full regional liberalization "capital goods" (53.3 percent). of these products was reached in May 1991 by Bolivia, Exceptions represented close to 15 percent in Colombia, Colombia, Peru and Venezuela."9 and 25 percent in Ecuador and Venezuela, but the struc- The Machine Tools Program started in August 1972, ture was relatively similar for the three countries. The but the withdrawal of Chile and the admission of Spearman rank correlation for the first classification of the Venezuela forced new negotiations, which were finally con- table was 0.96 for Colombia-Ecuador, 0.87 for Colombia- cluded in 1979. A new agreement was reached in May Venezuela and 0.96 for Ecuador-Venezuela. 1991 to fully liberalize regional trade by the end of that year. The Common External Tariff There was agreement on the list of products considered The Cartagena Agreement of 1969 stated that the CET in the Steel and the Automobile Programs, but the pattern should be adopted before December 1980, with an addi- of regional liberalization was only agreed upon at the end tional period of five years for Ecuador and Bolivia. A min- of the 1980s; full liberalization took place after 1992. imum Common External Tariff was designed as a interme- Colombia, Ecuador and Venezuela signed on the Automo- diate step, where tariff levels varied with such criteria as tive Agreement in 1993, fixing not only the CET and the labor-intensiveness, technological sophistication, the uses pattern of regional liberalization, but also considering the of the product and the existence of domestic production. minimum value-added content that was needed to benefit The CET was finally signed in November 1994 (Decision from low tariffs on inputs and parts (CKD-completely 370), and started being applied in February 1995. The knocked down). adoption of the CET took 15 years more than initially In summary, few of the initial agreements were imple- agreed, and it is only partial, because Bolivia (permanently) mented. Full regional liberalization was supposed to hap- and Peru (until 2006) can keep their flat tariffs.21 pen by 1980, but it only occurred in 1991. The agreements The Andean presidents defined the main characteristics only included Venezuela and Colombia since Ecuador of the CET in Galapagos in 1991: It would have four lev- withdrew from the negotiations at the end of the '80s and els-5 percent, 10 percent, 15 percent and 20 percent- only participated in the Automotive Program. However, and the 20 percent level would disappear in one year. In Ecuador rejoined the process later on and fully liberalized practice, capital goods and raw materials not produced in trade with its partners in January 1993. the region got the lowest tariffs (5 percent), semi-elabo- rated products got tariffs of 10 to 15 percent and final con- The Weight of the Exceptions to Regional Liberaliza- sumer goods got 20 percent. Each country can adopt a low tion. We base our calculations on the list of items included 0 to 5 percent tariff if the goods are not produced in the in the Automotive, the Machine Tools (Decision 300), and region, even if the tariff contained in Decision 370 is the Steel Programs (Decision 299); and of those included higher. Finally, Decision 370 does not include the Auto- in the "Reserved List" and the "Exceptions List" (Decision mobile Pact, which has a maximum tariff of 35 percent 263). Together, they account for all the exceptions at the (maximum tariffs in Decision 370 are 20 percent).21 time (for a description of each list see Guerrero 1979). Annex 2 of Decision 370 allows Ecuador to have lower Table 1 presents the relative importance of those goods tariffs (5 points lower) for 990 items (harmonized eight excepted from the liberalization compromises by the end of digits), and Annex 3 allows zero tariffs for 32 items (goods the 1980s. We show results for BEC and ISIC classifica- related to health, education and massive communications). 86 ANDEAN PACT TABLE I Exceptions to Regional Liberalization, Prior to 1990 (percentages) IMPORTS EXCEPTED/TOTAL IMPORTS COLOMBIA, ECUADOR AND VENEZUELA COLOMBIA ECUADOR VENEZUELA I. Broad Economic Categories I Food and Beverages 3.2 4.7 10.7 0.9 2 Producers' Materials 22.6 12.8 24.6 20.8 3 Oil and Lubricants 4 Capital Goods (exc.Transport Equipment) 53.3 6.7 57.0 46.9 5 Transporc Equipment 83.7 95.9 63.4 64.8 6 Consumer Goods 31.4 5.6 13.1 35.9 II. ISIC rev. 2 11 Agriculturil & Livestock Production 0.2 0.2 0.0 0.2 12 Forestry 0.0 0.0 0.0 0.0 13 Fishing 0.0 0.0 0.0 21 Coal Mining 0.0 0.0 0.0 22 Crude Petroleum & Natural Gas Producrion (gas only) 0.0 0.0 0.0 23 Iron Ore Mlining 0.0 0.0 0.0 29 Other Minerals 0.2 0.0 2.8 0.0 31 Food, Beverages and Tobacco 3.4 4.0 10.7 1.3 32 Textiles., Footwear and Clothing 60.6 43.3 101.3 62.6 33 Wood Products 22.1 0.0 53.6 23.2 34 Paper and Printing Products 41.6 17.0 97.4 36.6 35 Chemicals 7.5 1.6 29.9 3.9 36 Non Metallic Minerals 23.0 0.9 57.8 17.0 37 Iron and Steel 26.7 39.1 11.5 3.4 38 Metal Procucts and Machinery 46.2 79.4 29.3 16.4 39 Other Manufacturing Industries 3.3 0.3 25.9 0.6 Source: Data Intal and Cartagena's Agreement Board Methodology: Impor:s from the Andean Region, 1992. Exceptions Included: Metal-Mechanical Program, Steel Program, Automobile Program, "List of Exceptions" and 'Reserved List" Finally, Annex 4 permits transitory exceptions for Colom- Venezuela protects "food and beverages" relatively more bia and Verezuela (230 items each) and Ecuador (400 than Colombia and Ecuador, and "transport equipment" rel- items). All those exceptions should disappear in 1999. atively less. Also, and for obvious reasons, Venezuela tends to Table 2 compares the CET with the national tariffs of protect "oil and lubricants" less than the other two countries. 1992, when most reforms had already taken place. The Ecuador gives less protection to raw materials in general. results show that both the CET and the tariff schedules of Did tariffs increase as a result of the adoption of the Colombia, Venezuela and Ecuador protect more "transport CET? We suggested before that this was the case, but equipment," "food and beverages" and "other consumer changes were relatively mild for the aggregate. The detailed goods," and protect less "oil and lubricants," "other raw figures of the table show that the adoption of the CET materials" arid "other capital goods." Tariff structures did increased protection in all subsectors; even in "transport differ somewhat among countries, however. The Spearman equipment," if we consider the correct figures (see above). rank correlation between the tariff schedule of Colombia Was it convenient to adopt a CET in the Andean and the CE1' is 0.9 and the same figure is obtained for Group? We think so. To start with, even if tariffs increased Ecuador. But the rank correlation for Venezuela-CET is over the levels of 1992, they were much lower (and homo- 0.7, which suggests that Venezuela had to introduce more geneous) than in any other decade of the past. Thus, the changes than the other two countries in order to adopt the adoption of the CET gives the correct long-run signals to CET. the private sector. 87 TRADE TOW'ARDS OPEN REGIONALISM TABLE 2 The CET and National Tariffs (percentages) CET 1992 U.N. BROAD ECONOMIC CATEGORIES BOLIVIA COLOMBIA ECUADOR PERU VENEZUELA I Foods and Beverages 15.8 10.0 14.8 13.5 17.6 16.1 6 Other Consumer Goods 17.6 10.0 15.0 14.5 20.7 14.4 3 Oil and Lubricants 9.7 10.0 9.7 9.2 15.0 2.2 2 Other Raw Materials 11.9 10.0 10.2 8.7 16.1 10.1 5 Transporr Equipment 16.4 9.7 20.6 150 15.2 15.4 4 Other Capital Goods 11.9 8.8 9.2 72 15.2 8.4 7 Others 18.3 10.0 14.0 17.2 15.5 9.2 TOTAL 12.8 9.4 11.1 10 2 16.2 11.8 Source; LAIA, Carragena Agreement Methodology: Werighted averages, total imports (1 countries, 1992) as weights; para-tariffs included. The figures of this Table are based on Decision 370 of 1994, wh,ch does not include the Automotive Pact The average unweighted tariff for rhe 22 tariff items negotiated under tha- agreement was 26.8 percent, while the CET (Dec,sion 370) was 18.2 percent for those same Items. On the other hand, averages for the CET could be lower since countries can apply lower tariffs for goods nor produced in the reg.os. Second, the discussion on the merits and costs of free- 1986 and 1990, but decreased in 1990-92 and remained trade areas and custom unions seems to be shifting in favor stagnant in the whole period 1990-95. of the latter. Thus, Wonoacott and Lutz (1989, p. 67) What is remarkable is the expansion of mnanufactured argued initially that a free-trade area generally dominates a exports and the stability of manufactured imports. They customs union," but Wonnacott (1996, p. 92) has given a also fell during the first years of the 1980s, but their recov- more balanced view recently. He mentions two main ery started in 1982-84, much faster than for all the other advantages of the CET: The hub and spoke problem does variables considered in the figure. not arise; and rules of origin are not required. Similar con- Figure 3 shows that an increasing proportion of those clusions are presented by Krueger (1993 and 1995). manufactured exports went to the Andean markets after Finally, the argument that a free-trade area would imply 1990. The figure also shows that the relative weight of intra- lower tariffs does not seem to apply to Latin America. In regional trade in total trade has increased rapidly since 1990 fact, the opposite seems to have happened, and the CET has to levels similar to those in Mercosur, but it is still much been useful to check those pressures to increase the tariff to lower than in NAFTA (30 percent) or in Europe (higher those levels prior to reforms. It could be true, however, that than 70 percent). Again, what is really remarkable is the the adoption of the CET reduced the probability of the importance of the Andean countries as markets for their own adoption of a flat tariff in some of the countries. manufactured exports, with a participation higher than 30 percent in 1995. This level is much higher than in Merco- Reforms and Trade Flows sur, and similar to the one reached in NAFTA. The role played by the Andean countries as markets for The Broad Picture their manufactured exports was especially marked during Figures 2 and 3 present a broad picture of the evolution of 1990-95 when exports to the region grew five times faster trade flows in the Andean region between 1970 and 1995. than manufactured exports to the world. The boom of It is clear, to start with, that total imports and exports were 1986-90 spread across all markets, but the boom of not very dynamic. After a vigorous expansion in 1970-80, 1990-95 was concentrated in sales to the Andean markets.23 total exports and imports deteriorated abruptly during the Colombia and Venezuela, the largest two economies in first part of the 1980s, with sales to the Andean countries the group, accounted for an increasing proportion of being hit especially hard. Total exports and imports in exports to the Andean region (68.1 percent in 1986; 82 1986 represented less than half of those reached in 1980, percent in 1995) and for a decreasing proportion in Andean and they were lower in 1995 than in the previous peak of exports to the world (79 percent in 1986 and 71.4 percent 1980. Total real exports increased very quickly between in 1995). 88 ANDEAN PACT FIGURE 2 Exports and Imports, 5 Andean Countries (constant US$ of 1990, billions) 50 9 Total Exports Manufactured Exports (right side scale) 4 5 .. . . . . . ............ 40 7 35~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2 5 Manufatred Imports 20 ............... AmTotal Im ot 15 .. .... 10.... ..2 5 1. 0 IO 0 O1 ON O N O N O 0 N O N N O N N X m t o co tE X r % % tt > a 3 a a I I GE Gr GE G01 C1 GN C, G O\ GE N G", CN C\ O\ Ge , G\ Sources: Data Intal: Ocarnpo and Esguerra (1995); IMF Financial Statistics; Manufactures: SITC 58; all figures in current US$, deflated with U.S. GDP The fall in manufactured exports during 1980-86 was Krugman (1998, p. 167) gave new theoretical justification common to four of the countries (except Venezuela), and to the model, arguing that the validity of the model is the global expansion of 1986-90 was especially marked in closely tied to the presence of increasing returns. Venezuela, Peru and Ecuador. In fact, it is not clear why Equation (1) presents the gravity equation following Colombian exports to the region stagnated in a period in Frankel's formulation (Frankel 1994, pp. 11-21).24 Exports which its total manufactured exports were growing rapidly. from country i to country j will be positively related to GNP (Y) and to per capita income (Y/N, N for population) Did Trade Reforms Matter? Gravity Models in each country, and negatively related to distance. The How much did trade flows increase as a result of unilateral product specification implies that trade will be larger when liberalization cum regional integration? To answer the GNPs are similar, a feature of the new trade models under question we use the so-called gravity model, which says imperfect competition. It is also expected that countries that trade between two countries is proportionate to the trade more when Y/N is large, since they tend to specialize product of tlheir GNPs and inversely related to the distance more and trade more as they develop. We include dummy between them, by analogy to the gravitational attraction variables for 'adjacency," and for three common markets, between two masses in physics. Unexplained trade is the Andean region, Mercosur and Central America. attributed to factors not included in the model such as rel- ative prices andeconomic policy.(1) Log (X 1)= a + ft Log (Y .X'1) +I,2 Log (Y INj. Yj/N1) ative prices and economic policy.'' Despite some recent criticisms (Bhagwati and Pana- -3 Log (Distance) + + 71.Dadjacent ± 72 Dandean ± Y3 gariya 1996, p. 34) and its "dubious" theoretical heritage Dace--ca + 74 Dmercosur (Deardorff 1984), the main virtue of the model is its capac- The 21 countries included in the sample are the United ity to fit the data remarkably well. Besides, Helpman and States and those considered by Aitken and Lowry (1973) for 89 TRADE TOWARDS OPEN REGIONALISM FIGURE 3 Exports to Partners/Exports to World (percentages) I. Total Exports 35 - 30- NAFTA 25- 20- 105 10 C 8 ~~~~~~~Mercosur / 5 - c (N N< RN kG ( GN c a o (N N> \ t R cc 3 03 0G 0 0 0 cc 03 CO GC G C C> C> C.> C>5 C> C> C> C> C> C.> C> C> C> C> C> C> C> C> II. Manufactures 3 5 30 25 NAFTA 20 - 15 10 5_ I I I IID I I, 00 0C 03 CC 03 X3 03 0c 00 tl C> Cc c> C> \> C> C> C> C> C> C> C> C> C> C> C> C> C> C> > G\ C> Sources: Data Intal and authors calculations. Methodology: Manufactures SITC 5-8 90 ANDEAN PACT their analysis of the impact of LAFTA and the Central Amer- manufactured exports, which represent less than 65 per- ican Common Market during 1959-67.25 The period of cent of Mercosur's exports. analysis is 1986-95, and the number of observations varies The results of Table 3 allow us to explore further the between 180 in 1987 and 316 in 1995. Results are presented impact of economic policy on trade flows. It basically in Table 3. The global fit of the regressions seems to be satis- quantifies the additional trade flows corresponding to the factory, with coefficients (R2) typically superior to 0.7. dummy variables Dandean (72) and Dc.america (73) for the period As expect:ed, the coefficients of Y..Y. and YI/N.. Y./N. 1993-95, the only years for which 72 results are significant. are positive, and the coefficient of distance negative, all of We report the corrected coefficients in all cases. Mercosur them significant at the 99 percent level; the dummy for is not included since 74 is not significant. "adjacency" .is also highly significant (95 to 99 percent) and The first three columns of Table 4 present the confi- positive. The coefficient of Y. .Yj (I1) oscillates between dence interval for 72 andy 3 in equation (1); the mean value 0.56 and 0.68, a result consistent with the range 0.53 to coincides with the regression coefficient in Table 3. Col- 0.75 obtained by Frankel for the hemisphere. This means umn (4) shows the average trade flow in each region for the that trade increases with size, but less than proportionally, respective year, and Columns (5) to (8) the gross trade cre- and that the small economies are generally more open to ation derived from the different values of y. Thus, 72 is 0.78 trade. B2 fluctuates between 0.34 and 0.51 (for those coef- in 1995, with a confidence interval of 0 to 1.58, which ficients significant at the 99 percent level), a range consis- means additional trade flows of US$73 million (79.3 per- tent with Frankel's results for the period 1965-80 (though cent of the average trade flow for that year). his coefficients become extremely low, 0.06 to 0.09 for In summary, our figures indicate that unilateral liberal- 1985 and 19)90). Finally, P3 (distance) fluctuates between ization cum regional preferences could have increased -0.58 and -1.06, and for Frankel between -0.35 and -0.68. trade a maximum of 60 to 79 percent for the case of the The regression results seem to suggest that apertura cum Andean Group, and 94 to 95 percent for Central America. regional integration were central factors behind the expan- No significant impact was found for Mercosur's regional sion of trade flows in Central America in every year after agreement. 1988, and also in the Andean Group after 1993. However, we do not firsd significant coefficients for the case of Mer- Trade Creation vs. Trade Diversion cosur (Dme,rt,s,r). In a sense, this result contradicts Yeats Trade creation occurs when new imports from the partner (1997), maybe because the author centers his analysis in country substitute domestic production, and trade diver- TABLE 3 Gravity Equation Andean Countries, Central America and Mercosur Dependent Veiriable: Imports (current US$) 86 87 88 89 90 91 92 93 94 95 C -8.91' -11.1' -11.4' -10.81' -9.29' -9.08' -11.4' -11.9' -12.3' -11.9' Distance -0. 58' -0.66a -0.61' -0.76' -0.85' -0.92' -0.94a -0.91' -0.95' -1.06' yyj 0.61a 0.65' 0.61a 0.56a 0.58' 0.61a 0.59' 0.56' 0.63' 0.68' (YjINj).(YJ/N) 0. 6 0.25b 0-34 0.48a 0 37' 0.35' 0. 5' 0.59a 0.46' 0.41' D andean -0.36 -0.24 -0.01 -0.05 -0.11 0.23 0.16 0.47' 0.67' 0.78' D central america 0.41 0.64 0.85b 1.09' 1.03' 0.95' 1.33' 1.44a 1.62' 1.54' D mercosur 0.39 0.08 0.17 0.11 -0.12 -0.18 -0.27 -0.24 -0.07 0.03 D adjacent 0.58' 0.67' 0.76' 0.68' 0.54' 0.51' 0.62' 0.75' 0.57' 0.38 •2 0.62 0.67 0.68 0.73 0.70 0.66 0.73 0.73 0.73 0.73 N.Obs 185 180 210 247 247 250 282 287 308 316 Yii: GDP countries i and j , crrent US $; N,,: Population; D: dummy variables All variables (except the dummies) in logs; a and b represent coefficients which are significant at the I and 5 percent significance level, respectively; White's heteroscedasticity cor- rected standard errDrs in all cases Countries included: the same countries considered in Aitken and Lowry (1973) plus the United States. These are five countries in the Andean Group; five countries in central Amer- ica; four countries in Mercosur; plus Chile, Mexico, Dominican Republic, Jamaica, Panama, and Trinidad-Tobago. We excluded records with export levels lower than US$863,000, the value of the first quartile for exports in the sample. 91 TRADE TOWARDS OPEN REGIONALISM TABLE 4 Gravity Equation. Derived Trade Flows AVERAGE IMPORTS CONFIDENCE INTERVAL FOR y INTRA-PARTNERS GROSS TRADE CREATION USS THOUSANDS US$ THOUSANDS (%) MIN MEAN MAX MEAN MAX MEAN MAX (6)/(4) (7),(4) (1) (2) (3) (4) (5) (6) ,7) (8) Dummy Andean Group 93 (0.00) 0.47 0.94 46,630 17,492 28,422 37.51 60.95 94 0.00 0.67 1.36 63,577 31,346 47,237 49.31 74.30 95 (0.00) 0.78 1.58 92,967 50,682 73,734 54.52 79.31 Dummy Central America 93 (0.00) 1.44 2.89 37,798 28,867 35,687 76.37 94.42 94 (0.00) 1.62 3.25 43,261 34,752 41,587 80.33 96.13 95 0.00 1.54 3.09 53,316 41,929 50,884 78.64 95.44 Ssarce, oad Methodology: see Table 3 Confidence Interval: 5 t ,,.S(y), w-tls: t student coefficienr; S(Y): standard error of y sion when they substitute previous-and cheaper- only have information on manufactures, the sector where imports from "abroad." In the restricted scenario specified most trade diversion could have taken place (see Figure 3 by Viner (1950) a customs union is welfare-increasing above). when trade creation dominates trade diversion, but this is Our results suggest that trade creation was much more not necessarily true under different assumptions. Pure important than trade diversion, with an increasing ratio in trade diversion can be beneficial, thanks to the effects of Colombia from 18.4 percent in 1986 to 19.3 percent in lower prices on consumers; trade creation can be detrimen- 1990 and to 25.5 percent in 1995; and also in Venezuela tal to welfare once we consider its effects on tariff revenues. (23.2 percent, 25.6 percent and 29.4 percent, respectively). That is why Winters (1997, p. 9) considers that the mea- Trade creation seems to dominate trade diversion in most surement of these effects should be indicative rather than cases. The relation between imports and apparent con- definitive. sumption increased in Colombia in all sectors, both in Of course, the complete evaluation of the impact of 1986-95 and in 1990-95. Trade creation also dominates trade flows on welfare should also include output and scale in Venezuela, though important amounts of trade diversion effects, important when there are increasing returns to took place in "Metal Products and Machinery" both in scale and imperfect competition. The output effect arises 1986-95 and, even more, in 1990-95; in "Non-Metallic when there is a change of output in industries where price Minerals" in 1990-95; and in "Food, Beverages and differs from average cost, and the scale effect measures the Tobacco" in 1990-95. If we look at the percent change in reduction of average costs when output expands. We the ratio, we conclude that trade creation was especially should also include variety and accumulation effects (Bald- marked in "Textiles, Footwear and Clothing" and in win and Venables 1995, pp. 1600-02). Terms of trade "Wood Products" in Colombia and Venezuela; and in "Iron effects should also be included, but they do not seem to be and Steel" in Venezuela. The first two sectors were very relevant in our case. closed in both countries before liberalization, and they Table 5 quantifies trade creation and diversion in remained relatively closed after liberalization. Colombia and Venezuela (which represent close to 70 per- cent of the Andean trade flows) for 1986, 1990 and 1995, Liberalization or Regional Preferences? measuring the relation between imports from "abroad" Both unilateral liberalization and regional preferences (non-Andean countries) and apparent consumption 26 We could have been responsible for rapid export growth to the 92 ANDEAN PACT TABLE 5 Trade Creation-Trade Diversion in the Andean Region (imports from "abroad" (non-Andean countries)/apparent consumption) COLOMBIA VENEZUELA ISIC REV2 1986 1990 1995 1986 1990 1995 31 Food, Beverages and Tobacco 1.6 1.5 2.7 5.0 9.2 8.5 32 Textiles, "sootwear and Clothing 2.9 3,6 13.1 8.7 15.9 18.3 33 Wood Products 2.0 1.8 9.4 1.8 2.5 5.5 34 Paper ancl Printing Products 12.1 12.3 15.7 15.4 19.0 25.2 35 Chemicals 23.2 24.4 27.5 20.9 14.5 37.6 36 Non-Metallic Minerals 3.8 3.8 6.7 5.6 8.5 7.0 37 Iron and Steel 31.4 35.4 36.5 12.2 24.1 36.5 38 Metal Products and Machinery 43.6 40.6 48.1 48.9 58.7 41.8 39 Other Manufacturing Industries 14.4 30.3 32.3 24.0 24.9 46.8 TOTAL 18.4 19.3 25.5 23.2 25.6 29.4 Sosrces: UNIDO, Yearbook of Industrial Statistics; Data Intal and author's calculations region, but the implications of each alternative differ radi- The number of observations is 48, and the adjusted cally. The first one implies a healthy expansion of exports, coefficient of determination (R2 adj) is relatively satisfac- caused by liberalization along a "natural" pattern, while tory for a cross section-time series pool: higher than 0.6 for the second could be due to artificially created trade diver- absolute imports, and close to 0.3 for shares. Both appar- sion, although not necessarily. ent consumption and the real exchange rate appear with It is clear that regional preferences decreased in the correct sign and are significant at the 99 percent level Venezuela during the period, from levels of 14.2 percent in all cases. Import elasticities relative to apparent con- for agriculture in 1988 to 12.1 percent in 1994; from 12.4 sumption are close to 0.6, and the elasticity imports-real percent for mining in 1988 to 3.7 percent in 1994; and exchange rate between 4.81 and 5.52. from 21.3 percent to 13 percent in the same period. Pref- The sign of (1 ±ta)/(1l+t,0 .) results negative in all cases, erences fell in most ISIC two-digit sectors, except in "Fish- both for nominal imports and for shares, which suggests ing" (ISIC 13, which increased from 7.5 percent to 15 per- cent) and in "Other Manufactured Industries" (ISIC 39, TABLE 6 from 12.7 percent to 18 percent). Regression Results: Venezuela's Imports from Colombia Table 6 specifies import functions for manufactures in IMPORTS (M) SHARES: M Venezuela using a pool of cross-section (ISIC three digits) FROM FROM and time series (1990 and 1995) data. Those were the COLOMBIA COLOMBIA/M TOTAL group of products and the period for which regional reori- DEPENDENT VENEZUELA VENEZUELA entation was largest (see Figure 3). As dependent variables DEPENDENT VARIABLE (1) (2) (3) (4) we use imports (in current U.S. dollars) from the partner Constant 25.67' 25.12' 19.63b 18.37b country (coLumns (1) to (2)) and the share of the partner's ( +ta)61 +tir_a -3.16 _ -1.64 -2.30 (1 + t') -.9 23 imports on total imports (columns (3) to (4)). As indepen- (1+t;i) 2.36 0.01 dent variables we consider the real exchange rate (real Real Exchange Rare -5.52' _537b -5.13' -4.81' Apparent Consumption 0.58b 0.57 b national ctarrency per real U.S. dollar, negative sign AdjR 2 0.62 0.62 0.33 0.33 expected), apparent consumption (positive sign expected), N.Obs 48 48 48 48 and the effect of tariffs on relative prices, (1 +ta)/(l +t int ); ta r': external tariff; t.,,,: tariffs with the partner country; Real Exchange Rate: Real corresponds to the import weighted external tariff and National $lReal US$; Apparent Consumption: Production + Imports - Exports All variables in logs; a and b represent coefficients which are significant at the 10 and tirca to the 'regional tariff. 27 The sign of (1 + ta)/(1 + t it.) is 1 percent significance level, respectively; White's heteroscedasricity corrected standard not clear a priori, and depends on how important was uni- Soerres i Daa Intas; UNIDO, Statistica Yearbook; IMF Financiai Statistics; World lateral liberalization (-) or regional preferences (+) in the Bank; Cenrral Bank of Venezuela Mlethodology: A cross section-time series pool was built with observarions for each ISIC explanatiort of import growth. (rev 2), 3 digits Sectors, for 1990 and 1995. 93 TRADE TOWARDS OPEN REGIONALISM that tariff reductions, and not preferences, were the central Intra-industry trade is also intense when barriers to factor behind the large dynamism showed by trade in trade are low. Thus, it is directly correlated with distance 1990-95. and transport cost, and it is higher when countries form Equations (2), (4), (6) and (8) consider separately (1 +t') customs unions. (Loertscher and Wolter 1980). Intra- and (1 +rt..). The first variable appears with the correct industry trade in Argentina, Brazil and Mexico is higher negative sign in all cases, though it only appears significant than in other NICs, something that some authors attribute for the absolute levels (both in Colombia and in to the existence of integration schemes in the region Venezuela); (I+tintra) appears with the wrong sign, and it is (Havrylyshyn and Civan 1985, p. 262). significant for Colombia. The relative weight of intra-industry trade increased in the world during the 1960s and 1970s, but it remained Characteristics of Andean Trade: stagnant during the 1980s (Globerman 1990). In Latin Intra-Industry Trade, Factor Intensity and America it has increased decade after decade, even during Comparative Advantage the 1980s (Baumann 1992). Table 7 presents the Grubel-Lloyd weighted index for Intra-industry Trade the United States and for the Andean region (Peru Most world trade is intra-industry today, and takes place excluded) between 1990 and 1995, both for total exports between rich countries with similar levels of development (without mining) and for manufactures.29 It shows that (Helpman 1987). Thus, Havrylyshyn and Civan (1985, p. intra-industry trade continued rising during the 1990s: 259) found that 58.9 percent of total trade among indus- The index increased from 0.61 to 0.65 for trade between trialized countries in 1978 was intra-industry (Grubel- the United States and Europe, and from 0.69 to 0.73 for Lloyd index of 0.589; the index takes the value of I when the United States and NAFTA. Figures are very similar for all trade is intra-industry, and 0 when all trade is inter- manufactures and for total exports excluding mining. industry)28 with a much lower figure for trade among the Intra-industry trade also rose in the Andean Group, NlCs (42 percent), and even lower among the developing from levels close to 0.35 to 0.41 (though it fell somewhat countries (22.6 percent). between 1992 and 1994) and it is higher for intra-Andean TABLE 7 Intra-Industry Trade (IIT) in the Region and in the United States, 1990-95 ANDEAN COUNTRIES EXPORTS AND IMPORTS TO/FROM U.S. EXPORTS AND IMPORTS TO/FROM ANDEAN ANDEAN COUNTRIES MERCOSUR NAFTA EUROPE ASIA TOTAL COUNTRIES EUROPE NAFTA MERCOSUR I. TOTAL EXPORTS (MINING EXCLUDED) 1990 0.35 0.21 0.19 0.15 0.14 0.26 0.14 0.61 0.69 0.32 1991 0.32 0.21 0.18 0.12 0.11 0.25 0.15 0.63 0.70 0.35 1992 0.39 0.22 0.18 0.11 0.08 0.25 0.15 0.64 0.70 0.37 1993 0.36 0.16 0.16 0.10 0.07 0.24 0.16 0.63 0.70 0.36 1994 0.35 0.17 0.18 0.11 0.08 0.26 0.18 0.63 0.72 0.35 1995 0.41 0.19 0.19 0.13 0.08 0.29 0.19 0.65 0.73 0.35 II. MANUFACTURES 1990 0.32 0.19 0.13 0.15 0.09 0.24 0.16 0.63 0.73 0.32 1991 0.31 0.18 0.15 0.13 0.08 0.26 0.17 0.65 0.73 0.34 1992 0.37 0.20 0.14 0.10 0.05 0.25 0.16 0.67 0.73 0.38 1993 0.34 0.16 0.13 0.09 0.05 0.23 0.17 0.66 0.73 0.36 1994 0.34 0.16 0.16 0.10 0.06 0.26 0.21 0.66 0.75 0.34 1995 0.39 0.20 0.15 0.13 0.05 0.30 0.20 0.68 0.70 0.35 S,oarc, Data-lItal and authors calcularions MAlth,doIzKl: lIT Andean Group: The Grubel and Lloyd weighted Index was calculated for Bolivma, Colombia, Ecuador and Venezuela; we then weighted the different countrres using exports to the Andean Group as weights. Mining: Sector ISIC 2 (res 2); Manufacturing: ISIC 3 (rev 2) 94 ANDEAN PACT trade than for trade with Mercosur and NAFTA, and even The definition of the index comes in the notes to Table more than for trade with Europe and Asia.30 8, and here we just mention that the ROR gets high val- It occurs in sectors where scale economies and product ues when products sold to the region are not sold to other differentiation are important. Investment goods have a markets. On the other hand, the RCA takes high values for much higher intra-industry trade index (39.8 percent) than those products exported by the Andean countries and not intermediate goods (31 percent) or consumption goods, by the world. though the highest IIT coefficients occur in chemicals, Among the SITC (three digits) sectors with the most machinery and transport equipment (Havrylyshyn and dynamic exports to the region (category 1, "very high" in Civan 1985, pp. 264-65; see also Ethier 1982 and Krug- the table) were some mineral products (coal mining and man 1994, p. 231). petroleum), but also sawmills, metal industries and This was also our case, the highest level of IIT being in machinery (electric and non-electric). Among the group industrial raw materials both for trade with the subregion with "high" export growth were agricultural products and (IIT index of 0.79) and with the world (0.66). But other food, some chemicals and petroleum refineries, iron and sectors ("Foodls and Beverages") also show high levels of IIT steel industries, and metallic products (see Table A-2 in the in the Andean region. Finally, IIT is high for "Other Cap- appendix). ital Goods" in the Andean trade among partners, but not The results in column (1) indicate that the regional in trade with the world. market played a crucial role in the expansion of exports, since there was a close link between export dynamism and Reorientation, Intra-Industry Trade, Factor Intensity geographical reorientation. Those with the highest export and Comparative Advantage growth have a ROR index of 9.11, and those with the low- In this part we closely follow Yeats's analysis for Mercosur est growth have a ROR index of 4.37. and provide further insights into the characteristics of the There is a positive relation between intra-industry trade trade flows in the region. As in the case with the author, we and export growth to the region, though intra-industry are interested in insights "concerning the extent to which trade is also very important in group 3 ("low," 0.55). Inter- a regional arrangement diverts trade from patterns estingly, there seems to be a negative relation between both expected on the basis of efficiency conditions and compar- variables when exports go to "other" markets; the index is ative advantage" (Yeats 1997, p. 1). 0.30 for group 1 ("very fast") and 0.4 or higher for groups Yeats comnbines the Regional Orientation (ROR) Index 2 to 4. with indexes of Revealed Comparative Advantage (RCA) The dynamic sectors were not in the top list of revealed and factor intensity to characterize those sectors where intra- comparative advantage, something that should not worry regional trade was most dynamic. He concludes that there is us too much, because they are in second place, with an no evidence that Mercosur's intra-regional trade is evolving average index much higher than for categories "high along lines consistent with efficiency conditions. Rather, growth" and "very low growth." This pattern is consistent "the products recording the largest shift toward the region for the years 1990 and 1995. The most dynamic sectors are are those for which Mercosur has not demonstrated an abil- in an intermediate position considering capital intensity, ity to export competitively elsewhere" (Yeats 1997, p. 17). though the picture is not neat: They are in a low ranking Table 8 divides the sample of sectors (ISIC, 3 digits) (third place in the four groups) for value added per worker, into four groups, according to export growth to the sub- in second place for the other two variables. region (Table A-2 in the appendix presents the sectors In summary, it is clear that the dynamic sectors (high included in .ach group).31 The variables considered are the export growth to the Andean region) were dynamic partly Regional Orientation Index, Grubel's Intra-Industry Trade because they reoriented their exports to the Andean group Index (both for exports to the Andes, and for exports to the and, as expected, intra-industry trade was generally higher world), and a Revealed Comparative Advantage Index. It in the dynamic sectors. However, the products exported to also presents some standard proxies for capital intensity: the region are not very capital-intensive and show some value added per worker, capital stock per worker, and the degree of comparative advantage. In some sense, then, the participation of wages in value added. findings of this section confirm the central conclusion of 95 TRADE: TOWARDS OPEN REGIONALISM TABLE S Main Characteristics of the Dynamic Sectors (arithmetic averages) EXPORT GROWTH TO REGIONAL THE REGION. ORIENTATION IIT, X AND M 1990-95 INDEX (ROR) TO THE REGION IIT, TOTAL X AND M (1) (2) (3) 1.VeryHigh 9.11 0.53 0.30 2. High 5.05 0.39 0.40 3. Low 4.10( 0.55 0.43 4. Very Low 4.37 0.42 0.42 REVEALED COMPARATIVE ADVANTAGE (RCA) CAPITAL INTENSITY (4) (5) (6) (7) (8) 91() 95 VALUE ADDED/L K/L WAGESNVA (AVG=100) (AVG=100) 1. Very High 8.61 4.20 75.6 113.8 35.3 2. High 2.65 1.76 195.4 125.9 28.1 3. Low 14.47 40.13 46.3 68.4 42.7 4. Very Low 1.07 1.96 82.7 91.8 29.9 IIT: Intra-Industry Trade Index; VA: value added; K: capital stock; L: employment. Sources: Dara Intal and aurhors calculations; Banco Central de Venezuela and Staristical Office, Colombia; U.N. Yearbook of Industrial Staristics; GTAP Database. alethsodology: ISIC (3 digit) sectors wirh "low" values (less than rhe first quartile) of exports to the Andean markets excluded; "very high," "high," etc. correspond to quartile values for export growth to the region in 1990-95. RORi= (S, anIiS,h,e m,ke)' the share of good i in exports to the Andean markets/share of good i In exporrs to other markets. IIT: Intra-Industry Trade, Grubel and Lloyd weighted index. IIT. =1- x E - , I Yx(xA + M,) RCAi= S J/S world (share of good i (ISIC, 3 digits) in total exports made by the Andean countries/share of good i in total exports made by the world). USA's exports used as proxy for world's exports. Capital Intensity: average for the four groups = 100; L was calculated as the sum of operatives and employees, adjusting for differences in average wages; K: capital srocks. The information for most secrors was obtained from the U.N. Yearbook of Industrial Statistics, but we also used the inputioutput matrix of Venezuela (1986) and Colombia (1992) 1 the paper: Trade flowed to the Andean region mainly due Conclusions to the influence of "natural" forces and market opening. A very dynamic and healthy picture emerges from the These results are entirely consistent with the domestic experience of "open regionalism" in the Andean countries firms' learning how to export in the regional markets, and in the period 1986-95, and it is paradoxical that some then competing later on in the more challenging interna- political obstacles still remain when the agreement is tional scenario. How could this happen if production for finally working. the regional markets were extremely inefficient? In his The Common External Tariff was adopted in 1995, not study on this crucial aspect Buitelaar (1993) found that 64 as bad a decision as some seem to suggest. The tariff struc- out of the 106 firms included in his sample jumped from ture was not modified substantially. and the absolute level exports to the regional markets in 1979 to exports to out- is one of the lowest reached in the region ever. Those who of-region markets. This pattern was even more marked for attack the CET because they think that more liberalization firms with a high relation between exports and domestic is in order do not know about the political economy of pro- sales (see Echavarria 1996). tection in Latin America. In fact, there is no doubt that the 96 ANDEAN PACT adoption of the CET has already influenced the govern- reason alone plainly justifies the recent Chilean debate on ments of Colombia and Venezuela in the right direction. the merits of integration-industrialization in a very open The adoption of the CET provides long-term signals to economy. Indeed, trade liberalization may prove to be the a private sector looking for stability, and the alternative- successful path to the long-standing goal of competitive namely, the rules of origin signed in NAFTA or negotiated industrialization in Latin America. by Mexico with Costa Rica, Colombia and Venezuela-are perceived as being more difficult to implement than a CET. References The new literature comparing costs and benefits for free- Aitken, N. D., and W. R. Lowry (1973). "Cross-Sectional Study of trade areas and customs unions is not so clearly in favor of the Effects of the LAFTA and the CACM on Latin American free-trade areas. One factor in this analysis is that tariffs Trade," Journal of Commzon Market Studies, Vol. 12, June, pp. have dropped quickly in most Latin American countries 326-336. Balassa, B. (1973). "Tariffs and Trade Policy in the Andean Common and are unlilcely to be further lowered in the medium term. Market,"Journal of Common Market Studies, Vol. 12, pp. 173-95. The CET levels in the Andean Group and in Mercosur Baldwin, R. E. and A. J. Venables (1995). "Regional Economic Inte- most likely ireflect the tariff levels that will remain in place gration," in G. M. Grossman and K. Rogoff (eds.), Handbook of over the medium term. International Economics, Vol. III, pp. 1597-644. Amsterdam: We show in the paper that reforms had an important North Holland. influence on trade flows, and that unilateral liberalization Baumann, R. (1992). "An Appraisal of Recent Intra-Industry Trade for Latin America," Cepal Review, Vol. 48, pp. 83-94. (and not regional preferences) was the central force behind fo Lai Amrc,'ea eiw o.4,p.8-4 Bhagwati, J. and A. Panagariya (1996). "Preferential Trading Areas the very high dynamism shown by total and manufactured and Multilateralism-Strangers, Friends or Foes?" in J. Bhagwati exports to tlhe region. The paper also shows that trade cre- and A. Panagariya (eds.), The Economics of Preferential Trade Agree- ation amply dominated trade diversion, though further ments, pp. 1-78. Washington, D.C.: American Enterprise Institute. research is obviously needed to determine the relation Buitelaar, R. (1993). "Dynamic Gains from Intra-Regional Trade in between the new trade flows and economic welfare. Latin America," in R. Belous and J. Lemco (eds.), NAFTA as a Model of Development: The Benefits and Costs of Merging High and Low Intra-indiastry trade between the Andean countries rose to 40 percent in 1995, a figure much higher than for De Wage Areas. Washington, D.C.: National Planning Association. De la Torre, A. and M. R. Kelly (1992), "Regional Trade Arrange- exports to other regions and countries. Finally, we show that ments," International Monetary Fund Occasional Papers, Vol. 93. geographicat. reorientation was an important factor behind Deardorff, A. (1984). "Testing Trade Theories and Predicting Trade export growth to the region, and that the most dynamic Flows," in R. W. Jones & P. B. Kenen (eds.), Handbook of lnterna- sectors are not very far from the average in terms of capital tional Economics, op. cit., pp. 467-517. intensity or revealed comparative advantage. These findings OiDfaz Alejandro, C. (1976). Foreign Trade Regimes and Economic Devel- opment: Colombia. New York: Narional Bureau of Economic are entirely consistent with the fact that many firms learn to Research. export in the regional markets and compete later in the Echavarrfa,J.J. (1996), "El G3. En el Camino de la Integracihn Con- more challenging international markets. tinental," in R. Lipsey and P Meller (eds.), NAFTA y Mercosur: We know that scale economies and market segmentation Un Dialogo Canadiense-Latinoamericano, pp. 129-63. Santiago: are crucial factors behind intra-industry trade, and we also CIEPLAN-Dolmen. know that its distributional impact is much lower than for (1997). "Producers, Unilateral Liberalization and Regional traditional trade. That explains why producers are so satis- Integration in the Andean Countries," mimeo, Trade Unit, Orga- traditional t:rade. That explains why producers are so satss- nzto fAeia tts nization of American States. fled with the Andean Group (Echavarrfa 1997). But is not Edwards, S. (1995a). Crisis and Reform in Latin America: From Despair a protected inarket any more, and it is no coincidence that to Hope. Washington, D.C.: World Bank. the averages obtained for the CET coincide with the levels _ (1995b). "Comments," in S. Haggard (ed.), Developing Nations recommended by Balassa (1973) more than 20 years ago. and the Politics of Global Integration, pp. 123-43. Washington, A final cDmment could nicely close our discussion on D.C.: The Brookings Institution. Ethier, W (1982). "Dumping," Journal of Political Economy, Vol. 90, integration in the region. Ir IS paradoxical that govern- ment areachivingindutriaizaton troug intgraton, pp. 487-506. menlts are achieving industrialization through integration, Pfrench-Davis, R. (1977). "The Andean Pact: A Model of Economic the main objective sought in the old schemes-but Integration for Developing Countries," World Development, Vol. 5, through much more dynamic and competitive paths. This pp. 137-53. 97 TRADE: TOWARDS OPEN REGIONALISM Frankel, J. A. (1994). "Trading Blocs: The Natural, the Unnatural, tion and International Trade, pp. 55-75. Readings in Economics. and the Super-Natural," Center for International and Development Cambridge, Mass.: MIT Press. Economnics Research C94-034. Nogues, J. and R. Quintanilla (1992), "Latin American Integration Garay, L. J. (1981). El Pacto Andino: Creacidn de mn Mercado para and the Multilateral Trading System," in New Dimensions in Colombia? Bogota: Fedesarrollo. Regional Integration, pp. 278-318. Cambridge: Cambridge Uni- Globerman, S. (1990). "Recent Trends in Intra-Industry Trade and versity Press. Their Implications for Future Trade Liberalization," Norheim, H., K.-M. Finger and K. Anderson (1993). "Trends in the Weltwirtschaftliches Archiv, Vol. 124, pp. 25-49. Regionalization of World Trade, 1928-1990," in K. Anderson Grubel, H. G. and P. J. Lloyd (1975). Intra-industrj Trade: The The- and R. Blackhurst (eds.), Regional Integration and the Global Trad- ory and Meeasrement of International Trade in Differentiated Produects. ing System, pp. 436-86. New York: St. Martin's Press. New York: Wiley. Ocampo, J. A. and P. Esguerra (1995). "The Andean Pact and Latin Guerrero, C. (1979). 10 Anos del Grnpo Andino: Memorias doe n Pro- American Integration," in R. Bousass and J. Ross (eds.), Economic tagonista. Bogota: Centro Internacional de Investigaciones para el Integration in the Western Hemisphere, pp. 122-45. Notre Dame: Desarrollo, CIID. University of Notre Dame Press. Haggard, S. (1995). Developing Nations and the Politics of Glohal Inte- Rajapatirana, S. (1998). Colombian Trade Policies and the 1996 WTO gration. Washington, D.C.: The Brookings Institution. Trade Policy Review, mimeo. Havrylyshyn, 0. and E. Civan (1985). "Intra-Industry Trade among Resnik, S. A. and E. M. Truman (1973). "An Empirical Examination Developing Countries," Journal of Development Economics, Vol. 18, of Bilateral Trade in Western Europe."Journal of International Eco- pp. 253-71. nomics, Vol. 3, pp. 305-35. Helpman, E. (1987). 'Imperfect Competition and International Schott, J.J. and B. Kotschwar (1996). "Western Hemisphere Readi- Trade: Evidence from Fourteen Industrial Countries," Journal of ness Indicators: An Interim Report," mimeo. the Japanese and Internrational Economies, pp. 62-81. Urrutia, M. (1981). "Viabilidad del Modelo Andino de Integracion," Helpman, E. and P. R. Krugman (1998). Market Structutre and Foreign Coyunucra Economrica, Vol. 11, pp. 69-86. Trade. Cambridge, Mass.: MIT Press. Viner, J. (1950). The C7estoms Union Issue. New York: Carnegie Hirschman, A. (1978). The Turn to Authoritarianism in Latin America Endowment for International Peace. and the Search for its Economic Determinants. Paper prepared for the Whalley, J. (1996). "Why Do Countries Seek Regional Trade Agree- Working Group on the State and Public Policy of the SSRC, ments?" National Bureau of Economic Research Working Paper 5552. mimeo. Williamson, J. (1990), Latin American Adjustment: How, Much Has Hap- Hufbauer, G. C. and J. J. Schott (1 994). Western Hemisphere Economic Inte- pened? Washington, D.C.: Institute for International Economics. gration. Washington D.C.: Institute for International Economics. Winters, L. A. (1997). Assessing Regional Integration Arrangemnents, Krueger, A. 0. (1993). "Free Trade Agreements as Protectionist Paper Presented at the third World Bank Annual Conference on Devices: Rules of Origin," Cambridge, Mass.: National Buerean of Development in LAC, mimeo. Economic Research Working Paper 4352. Wonnacott, R (1996). "Beyond NAFTA: The Design of a Free-Trade __ (1995). "Free Trade Agreements vs. Customs Unions," NBER Agreement of the Americas," in J. Bhagwati and A. Panagariya Working Paper 5084. (eds.), The Econom7ics of Preferential Trade Agreements, pp. 79-107, Krugman, P. R. (1994). Rethinking International Trade, Cambridge, op. cit. Mass.: MIT Press. Wonnacott, P. and M. Lutz (1989). "Is There a Case for Free-Trade Laban, R. and P. Meller (1996). "Estrategias Alternativas de Comer- Areas?," in J. J. Schort (ed.), Free-Trade Areas and U.S. Trade Pol- cio para un Pais Pequefno: El Caso Chileno," in R. Lipsey and P. cy, pp. 59-84. Washington, D.C.: Institute for International Meller (eds.), NAFTA y Mercosur: Un Dialogo Canadiense-Lati- Economics. noamericano op. cit., pp. 163-94. Yeats, A. J. (1997). "Does Mercosur's Trade Performance Justify Loertscher, R. and F Wolter (1980). "Determinants of Intra-Indus- Concerns about rhe Global Welfare Reducing Effects of Regional try Trade: Among Countries and Across Countries," Trade Agreements? Yes," mimeo. Weltwirtschaftliches Archiv, Vol. 116, pp. 280-293. Lora, E. and F. Barrera (1997). "A Decade of Structural Reform in Latin America: Measurement and Growth Effects," Washington, Notes D.C.: Policy Discussion Paper, Inter-American Development 1. This is a standard proposition in the Hecksher-Ohlin theory of Bank, mimeo. trade, but is also valid for models with imperfect competition. In this Lozano, 0. L. and S. Zuluaga (1998). Estruictura, Estrategia e Impacts case "world real income will definitely increase with trade, as will the de las Negociaciones de la Comunidad Andina de Naciones. Washing- real income of the smaller country. The monopolist in the large coun- ton, D.C.: Inter-American Development Bank. try may however reduce production relative to autarky, which could Markusen, J. R. (1993). "Trade and the Gains from Trade with possible result in netative gains from trade for the large country" Imperfect Competition," in G. Grossman (ed.), Imperfect Competi- (Markusen 1993, p. 70). 98 ANDEAN PACT 2. The relation between total trade and GDP fell from 31 percent produced in the other two countries. But the program failed. Bolivia to 25 percent in the same period. produced two out of the 10 "products" assigned and failed in both 3. Haggard (1995) also mentions the pressures by the industrial cases when Venezuela violated the original agreement and started countries through retaliatory commercial policies; and the changes in producing them. See Urrutia (1981), p. 76. the regime governing international trade. 20. Bolivia's tariff had two levels of 10 percent and 19 percent in 4. We assign rhat year to Ecuador (based on our observation of 1988, and 5 percent and 10 percent in 1994. The Peruvian tariff had tariff reforms in the section on unilateral liberalization), though two levels: 15 percent and 25 percent in 1994. Edwards classifies Ecuador as a non-reformer. 21. The average unweighted tariff for the 22 tariff items negoti- 5. Decisioti 370 allows countries to apply lower tariffs when the ated under the "Automotive Agreement" was 26.8 percent, while the goods are not produced in the region. CET's (Decision 370) was 18.2 percent. 6. These levels were already low compared with previous decades. 22. With trade diversion occuring, and Partner A buying from Thus, Balassa (1973, p. 177) reports the following unweighted tariff Partner B a product previously bought "abroad," Partner A can levels for 1973: Bolivia (54 percent), Chile (172 percent), Colombia decrease the costs in an FTA with lower tariffs. This "unilateral deci- (70 percent), Ecuador (106 percent) and Peru (90 percent). sion" is not possible in a customs union. 7. Tariffs were reduced in 10 years in Brazil, Paraguay and 23. NAFTA lost participation in our exports of manufactures but Uruguay. Argentina's tariffs were reduced drastically between 1986 gained participation in exports of non-manufactured goods. The par- and 1988, but increased somewhat in 1994. ticipation of the EEC and Asia decreased for almost every type of 8. Tariffs and para-tariffs. product. 9. Rajapatirana (1998) argues that new quotas ("absorption con- 24. We also tried, unsuccesfully, to include relative prices and tracts") have been introduced recently in Colombia for agricultural exchange rates. It would be interesting to see whether our results imports- wouild change if we excluded oil imports from the calculations. 10. To be sure, these goals were shared by other integration Frankel (1994, pp. 20-1) tried some additional variables but the agreements such as the Central American Common Market. How- results were not affected. He found only a "bit" of support for ever, while regional trade represented 10 percent of GDP in Central Heckscher-Ohlin type variables: differences in capital/labor ratios, America, it only represented 1 to 2 percent of GDP in the Andes and educational attainment levels and land/labor ratios. in the LAFTA countries (Wonnacott and Lutz 1989, p. 76). 25. Five countries in the Andean Region, four countries in Mer- l 1. Venezuela participated in the negotiations but did not sign cosur, five countries in the Central American Common Market, two the agreement in 1969, mainly because the private sector was NAFTA countries (Mexico and the United States), plus Chile, opposed to it. Jamaica, Panama, the Dominican Republic and Trinidad-Tobago. 12. There is also the Consultive Committee, created to promote 26. This is the metholodgy followed by Resnik and Truman communication between the countries and the Junta; and the Advi- (1973) and suggested by Winters (1997, p. 12). The ratio is a proxy sory Committee on Social Issues (Comite Asesor Econ6mico Social) for the relation between imports and GNP, assuming that X=MP+M, composed of workers and the private sector. where X corresponds to exports, and M and M. to imports from part- 13. The Lima Protocol asks for a process of regional liberalization ners and from non-members. See Winters 1997, p. 12. for 1979-84; and the Arequipa Protocol changed the period to 27. T.,,, is either 0 (fully liberalized) or t* (excepted) for each 1980-84. The Quito Protocol demanded that regional liberalization harmonized item, eight digits. However, it can take any value once should be reached by 1997. we get weighted averages. 14. Urrutia (1981), p. 78, for example, mentions that the CET 28. See Grubel and Lloyd (1975). The index is built as negotiated for the metal-mechanical program was 51 percent, com- 5;l X- M. where i is normally the 3-digit industrial classifi- pared with a Colombian tariff of 29 percent for the same products. (Xi+ M,) cation (ISIC). It is expected that i is a homoge- The average external tariff agreed for automototive parts was 49 per- neous sector, with similar capital-labor ratios for cent, compared with 19 percent for the Colombian tariff. the different subsectors. 15. Local participation had to be higher than 50 percent for new 29. Our original data is at the level ISIC 4 digits. We then build foreign investment (except if the firm exported more than 80 percent the Grubel and Lloyd index. of total prodcution) and for firms exporting to the Andean region 30. There are some common features, but also important differ- with low tariffs. ences among the five countries: 1IT is higher for trade with the 16. The first part of this section is based on Urrutia (1981) and Andean partners in Colombia and Venezuela and today also in Lozano and Zuluaga (1998). Ecuador (not in 1990 or 1991, when Ecuador showed the fastest 17. Based on Lozano and Zuluaga (1998). growth in IIT). On the other hand, Peru is the only country for 18. The status of the Common List remained unclear when which IIT with the Andean region has decreased through time. LAFTA evolved into LAIA in 1980. Finally, IIT levels are very low for Bolivia, even lower than for 19. The fPetrochemical Program assigned 55 products to one Bolivia's trade with Mercosur. This suggests, again, that distance is country and 15 to two countries; two additional products could be a key variable in the explanation of trade flows in the region. 99 TRADE TOWARDS OPEN REGIONALISM 31. Following Yeats, we excluded from the analysis those sectors marginal products. This reduced our sample from 41 lSIC three- with very low exports to the Andean markets in 1990 (i.e., we digit sectors to 31 sectors. We divide the sample into four groups exclude the lowest quartile) to prevent the results being biased by according to quartiles for export growth. 100 Appendix I'ABLE A-I Tariiffs and Para-Tariffs in the Andean Countries, i9896-9 WEIGHTED (IMPORTS) AVERAGES LEVELS STANDARD DEVIATION COLOM- ECUA- VENE- ANDEAN COLOM- ECUA- VENE- ANDEAN COLOM- ECUA- VENE- ANDEAN BOLIVIA BIA DOR ZUELA PERU GROUP BOLIVIA BIA DOR ZUELA PERU GROUP BOLIVIA BIA DOR ZUELA PERU GROUP A. Total Border Taxes (Tariffs and Para-Tiariffs) 86 21.2 44.0 32.5 24.2 58.2 36.0 54 29 61 4L 73 52 4.6 16.9 36.4 28.A 24.7 22.2 88 17.5 45.3 32.0 25.8 63.5 36.8 3 35 64 42 76 44 1.9 17.6 35.9 29.7 25.1 22.1 90 14.3 21.2 29.1 18.9 63.9 29.5 3 15 30 11 61 24 2.7 14.1 20.5 17.1 21.2 15.1 92 9.4 11.1 10.2 11.8 16.2 11.8 3 9 13 7 3 7 0.9 6.0 5.9 6.1 4.5 4.7 94 9.3 11.0 10.1 13.2 15.3 11.8 /I 8 10 9 3 7 1.1 6.1 6.3 6.2 3.6 4.7 95 9.3 12.8 15.3 12.8 4 1 4 3 4 1.1 4.8 3.6 3.8 B. Tariffs 86 11.5 29.3 22.4 24.2 41.3 25.8 45 26 36 41 46 39 1.8 16.6 30.8 28.4 23.9 20.3 88 17.5 27.8 21.7 25.8 40.4 26.6 3 25 34 42 48 30 1.9 17.3 30.1 29.7 24.0 20.6 90 14.3 21.2 21.4 18.9 41.2 23.4 3 15 15 11 41 17 2.7 14.1 19.6 17.1 23.4 15.4 92 9.4 11.1 8.2 11.8 16.2 11.3 3 9 10 6 3 6 0.9 6.0 5.9 6.1 4.5 4.7 94 9.3 11.0 10.1 13.1 15.3 11.8 4 8 10 8 3 7 1.1 6.1 6.3 6.2 3.6 4.7 95 9.3 12.8 15.3 12.6 4 4 3 4 1.1 4.8 3.6 3.8 Para-Tariffs: Bolivia: Tasas por Servicias Prestados Colombia: Impuesto Adicional (Art. 9 Ley 50/84) mainly Impuesto Promocic6n de Importaciones para el fomenro de Exportaciones Impuesro pama IFI, Caja Agraria y *lesoro Nacional Ecuador: Recargo Arancelario Recargo de Estabilizaci-o Moneraria Petu: Sobretasa Soarce: LAIA and author's calculation Methodology: Imports of 1993 (in each country) used as weiglhts; fig.res for Colombia, Ecuador and Venezuela correspond to the Common External Tariff (Decision 370). z >- Z > TRADE: TOWARDS OPEN REGIONALISM TABLE A-2 Dynamic and Not-So-Dynamic Sectors (exports to the subregion) GROUP 1. VERY HIGH EXPORT GROWTH GROUP 3. LOW EXPORT GROWTH 210 Coal mining 321 Textiles 220 Crude petroleum & natural gas production (gas only) 322 Manufacture of wearing apparel, except footwear 290 Other minerals 352 Other chemical products 331 Sawmills, planing & other wood mills 356 Plastic products 354 Manuf of miscellaneous products of petroleum and coal 361 Manufacture of pottery, china and earthenware 372 Non-ferro-s metal basic industries 390 Other manufacturing industries 382 Non-electric machinery 383 Electric machinery GROUP 4. VERY LOW EXPORT GROWTH 312 Food products exc. beverages GROUP 2. HIGH EXPORT GROWTH 313 Beverages 111 Agricultural & livestock production 323 Tanneries & leather fiorshing 311 Food products exc. beverages 341 Manufacture of pulp, paper & paperboarcl 342 Printing, publishing & allied industries 355 Rubber products 351 Manufacture of basic industrial chemicals cxcept fertilizers 362 Manufacture of glass and glass products 353 Petroleum refineries 369 Non-metalic minerals 371 Iron and steel basic industries 384 Transport equipment 381 Metalic products exc. machinery 385 Manufacture of professional and scientific, and measuring and controlling equipment, n.e.c. 102 Comment S A R ATH R AJ A PAT I R A N A UAN JOSE HAS WRITTEN AN INTERESTING AND INFORMATIVE PAPER. WHAT IS NEW IN THE paper is the focus on trade flows using a variety of measures to find out what has happened to them in the last decade. His principal finding is that the manufactures trade among the Andean Group countries has increased rapidly compared with the trade with the non-member countries. In addi- tion, he finds that the most rapid growth was in intra-industry trade. What is, however, most inter- esting is that he finds the reason for the increase in manufactures trade was unilateral trade liber- alization among the members of the group rather than the exchange of preferences. His main evidence is the reduction in tariff preferences at the time the rapid growth in trade was taking place in response to the liberalization of the trade regimes, particularly in Bolivia, Colombia and Peru and, to some lesser extent, Venezuela-and, to an even lesser extent, Ecuador. In the paper Juan Jose draws a sharp contrast between described in Alan Winters's survey for this conference that the Andeart Group of the 1960s and the one that was the purpose of thc conference is to consider ways of think- reborn in the late 1980s. As the Vice Minister for Trade at ing about regional integration agreements rather than to the time when Colombia liberalized its trade regime give definite answers.1 It is in this spirit that I raise these rapidly, Juan Jose is eminently placed to speak to the issue issues. of the contrast in approach and intent for the new Andean Group. There are interesting observations in the paper Trade Flows and Welfare about the negotiations related to the adoption of a Com- Measuring trade flows does not answer the welfare ques- mon External Tariff (CET) that reveal the different inter- tion-whether the Andean Group members are better off ests of the member countries. It is also interesting to note with the Andean Group or with unilateral liberalization, that the CET that emerged showed the closest affinity to either as a group or in terms of individual countries. While the Colombian and Ecuadorian tariff structures than to the paper does not claim that it is attempting to answer those of Bolivia and Peru. The latter two countries have what it calls this "classic question," it is important to bear lower and rmiore uniform tariffs. Peru, as we know, opted to in mind that trade flows in general and export growth in leave the Andean Group and, more recently, to rejoin the particular cannot answer questions of welfare. The paper Group. Bolivia was permitted to keep to its two 5 percent does not cast the inquiry in terms of the Vinerian concepts and 10 percent tariffs, which conform to the lower two of trade creation and trade diversion. In fact, part of the rates of the CET. increase in trade can well be trade diversion rather than I raise a number of issues for discussion on the paper. I trade creation. Even so, it is important to note that the evi- note that Juan Jose's paper attempts to meet the ideal dence shown in the paper supports the view that unilateral Sarath Rajapatirana is an Adviser to the Operational Services Team of the World Bank. 103 TRADE: TOWARDS OPEN REGIONALISM trade liberalization is the main factor in the rapid growth they have to be taken into account. It is easy to see that the of manufactures exports. This reduces the possibility of export and import demand functions have many right- trade diversion. That unilateral trade liberalization leads to hand-side variables that go beyond trade. So one can imag- more intra-regional trade has been observed in the case of ine the difficulties in predicting trade flows from past the ASEAN countries, which had exchanged weak prefer- relationships. cnces with one another. Remaining Agenda for Trade Reform in the Measures of Trade Flows Andean Group My second comment relates to the different measures used Fourth, despite the strong trade reforms undertaken in the in the paper such as the Intra-industry Trade Index (the last decade and the reduction of trade barriers within the Grubel weighted index), the Trade Orientation Index region, there is an important unfinished agenda for reform (used by Yeats), and the Revealed Comparative Advantage in the Andean Group. Achieving this agenda will enable Index of Bela Balassa. All these measures are static mea- the realization of more gains. I highlight five elements of sures in the sense they take the past pattern of trade as that agenda: given. While they are convenient starting points to pose (a) Agriculture protection has continued in the form of questions regarding the effects of different trading the agriculture price band system. This provides pro- arrangements on trade flows, they are not measures on tection to agriculture, although the system has been which policy issues can be analyzed. The main problem justified on the basis of stabilizing agricultural with them is that they provide little guidance to the prices. The long five-year memory of the prices car- proper counterfactual for analysis. For example, if a coun- ried in the system makes price bands unresponsive to try has liberalized its trade, we cannot expect trade flows secular price declines. to be predicted by these indexes with a reasonable degree (b) The automobile sector continues to remain pro- of accuracy. We would be forcing a counterfactual of a tected above the CET rate. There is little justifica- given history that, in most cases, may not be the appro- tion to continue with it. Also, overall protection lev- priate counterfactual. This is because the indexes by and els can be reduced further: The four rates can be large reflect past trade policies and all the other factors reduced to two rates as in Bolivia, or a uniform tar- that impinge on trade flows of the past. Particularly at a iff like the CET can be adopted, which Peru has been time when patterns of demand and consumption, access to contemplating. inputs and access to new markets are changing rapidly, the (c) Although the Andean Group is a customs union indexes may well be understating the potential for the with a CET, there are rules of origin that act as pro- expansion of trade. New export avenues and opportunities tective device. Until all the exemptions from the arise in unpredictable ways when countries open up. That CET are rescinded, this type of trade impediment has been the common experience. They are hardly ever will continue. Given that Andean Group members captured by indexes of this type. At best, these indexes have individually signed different bilateral agree- indicate potential; but that potential could be larger than ments, they have different rules of origin tied to dif- what the indexes indicate. ferent timetables for achieving a full customs union. For example, Bolivia has signed 26 agreements in Trade Policies vs. Other Determinants of Trade all, each with different tariff rates, different excep- Flows tions and different rules of origin tied to different Third, it is important to note that trade flows are affected timetables with 6,000 tariff lines. This creates by many factors-namely, exchange rates, global economic opportunities for customs officials to engage in rent- conditions, terms of trade, foreign investment, and other seeking. But these rules of origin may be less severe variables such as transport costs. Trade flows are themselves than in NAFTA. In fact, in the absence of these determined by factors that are outside of trade. To the rules, these countries could have experienced faster extent these factors are present at the time of unilateral growth in exports and imports and increased the trade liberalizations and in the exchange of preferences, tradability of these economies. 104 ANDEAN PACT (d) A near free-trade situation exists among some of the sons for the latter are not immediately known, and it is one Andean Group countries with common borders. For that future research should attempt to understand better. example, between Colombia and Venezuela, this Despite the increase in trade flows we are not in a position trade .has grown rapidly-perhaps the most rapid for to make a welfare assessment of these developments. An any two counties in South America. This creates important policy agenda remains to be addressed in the challenges for macroeconomic coordination. Earlier, Andean Group, to secure greater gains from trade liberal- when the trade links were not so strong, these two izations. These include lowering the CET, reducing agri- countries could well ignore their relative macroeco- culture protection, eliminating the special automobile nomic stances; but no longer. regime, recognizing the macroeconomic challenges that (e) Finally, anti-dumping has increased rapidly in the come along with greater integration, addressing remaining past three years among regional groups.2 This is to barriers that arise from rules of origin, and bringing anti- be expected as other barriers are removed. dumping under control. To summarize, Juan Jose has provided us an interesting Notes paper. Its main finding that manufactures exports grew 1. 1. Alan Winters. "Assessing Regional Integration," page 49. most rapidly within the group compared with the trade in 2. Guasch, Luis and Sarath Rajapatirana (1997). "Anti-Dumping these goods with other countries and groupings is an and Competition Policies in Latin America and the Caribbean." important finading. The other important finding is that the Institute of the Americas, University of California, San Diego, April. growth was even stronger in intra-industry trade. The rea- 105 IV. NAFTA NAFTA: An Interim Report J E F F REY J. SC HO T T J NNJUNE 1991, THE UNITED STATES, CANADA AND MEXICO LAUNCHED NEGOTIATIONS ON A NORTH American Free-Trade Agreement (NAFTA). The core agreement was signed by the three heads of state in Washington on December 17, 1992; the pact was then supplemented by three "side agree- ments," covering environment, labor issues and import surges, which were concluded in August 1993. The agreements entered into force on January 1, 1994. In 1996, the NAFTA partners had a combined GDP of about $8.5 trillion and a population of about 380 million. The U.S. economy dominates the NAFTA region: The United States accounted for 89 percent of the NAFTA GDP and about 70 percent of the population. NAFTA represents the culmination of decades of These arrangements were developed to support the "silent" integration of the three economies of North growing regional network of trade and investment and to America, reinforced by the negotiation of numerous manage the bilateral disputes that inevitably arise as the bilateral trade agreements. By the time Mexico an- volume of trade expands. Interestingly, U.S. bilateral ini- nounced its; interest in free-trade talks with the United tiatives with both its neighbors began with sectoral issues States in 1990, the pathway to a prospective NAFTA had (autos, energy subsidies) but evolved soon after into already been substantially cleared by prior trade accords, broader trade and investment initiatives, reflecting the including: depth of common trade and investment interests among * the U.S.-Canada Auto Pact of 1965; the North American countries. * the U.S.-Mexico Understanding on Subsidies and By 1990, U.S. trade with Canada (merchandise and Countervailing Duties of 1985; commercial services) surpassed $200 billion, U.S.-Mexico * the U..S.-Mexico Framework of Principles and Proce- bilateral trade approached $75 billion, and Canada-Mexico dures For Consultation regarding Trade and Invest- trade ran at about $2.5 billion. U.S. and Canadian compa- ment Relations of 1987; nies invested heavily (combined about $95 billion) in each * the Canada-U.S. Free-trade Agreement (CUSFTA) of other's economy,' and U.S. investors accounted for about 1989; $10 billion in foreign direct investment (FDI) in Mexico. * the U.S.-Mexico Understanding Regarding Trade and The United States already accounted for the predominant Investment Facilitation Talks of 1989; and share of the trade of both Canada and Mexico, and of FDI * ten Canada-Mexico bilateral accords, including a in those countries. In short, North American economic Framework for Trade and Investment Consultations, integration was proceeding long before NAFTA negotia- concluded in March 1990. tions were even broached, building on the new opportuni- Jeffrey J. Schott is a Senior Fellow at the International Institute for Economics. 109 TRADE TOWARDS OPFN REGIONALISM ties emerging from the domestic economic reforms in Mex- January 1, 1994. Throughout the first year of NAFTA, the ico, introduced by the de la Madrid and Salinas govern- Mexican government was buffeted by a wave of political ments, and the inflation- and budget-cutting initiatives in shocks that further complicated policymaking in that the United States and Canada. important election year. The story has been well told else- NAFTA today is a work in progress. Many of its provi- where (see DeLong et al. 1996; Lustig 1996). Suffice it to sions have only recently been put into effect, and many of say that a combination of loose fiscal policy (statistically the important trade reforms have yet to be implemented. To concealed in the public accounts), low private savings and date, NAFTA has required few changes in U.S. trade prac- an overvalued currency culminated in the peso crisis of tices and regulations, although some obligations have been December 1994, and in the subsequent sharp recession of deferred (e.g., trucking regulations). Mexico has faced the 1995 with GDP falling by 6.2 percent. more challenging task of introducing trade liberalization In most respects, the Mexican crisis was homemade. and regulatory reforms right from the start, and has perse- NAFTA reforms bore scant responsibility for the develop- vered despite the peso crisis and severe recession that ments that led to the liquidity crisis, except to the extent erupted less than a year after NAFTA went into effect. that expectations of continued capital inflows prompted by the "promise" of NAFTA-induced growth led Mexican NAFTA Objectives officials into being unduly complacent about the overval- The preamble and Article 102 of NAFTA spell out the ued peso. However, NAFTA was critical to the effective concrete objectives that the three countries sought to response to the crisis, in two important respects. achieve through the trade pact. In brief, the three countries First, NAFTA obligations led President Zedillo to fol- sought to: low a highly orthodox recovery program: fiscal discipline, * promote increased regional trade and investment; ultra-tight money and sharp exchange-rate devaluation. * "create new employment opportunities and improve This program differed markedly from the extensive trade working conditions and living standards in their and capital controls usually deployed by developing coun- respective territories"; tries in response to balance-of-payments problems. The • provide a framework for the conduct of trilateral trade pain in Mexico was exceedingly sharp, but the recovery to relations and for the management of disputes; date has been noteworthy, especially compared with the * strengthen and enforce environmental laws and basic prolonged crisis of 1982. In 1996, the Mexican economy workers' rights; and grew by 5.1 percent in real terms and seems likely to • work together to promote "further trilateral, regional exceed 7 percent growth in 1997. and multilateral cooperation to expand and enhance The open access to the U.S. market, reinforced by the benefits of this Agreement." NAF'TA obligations, helped prevent an even more drastic This paper examines the progress to date in achieving recession in 1995: Net exports grew by more than 3 per- these goals. Because it is difficult to parse out the direct cent in 1995 and partially offset the 9.25 percent fall in effects of NAFTA provisions from the "normal" course of domestic demand. The peso depreciation helped fuel the commerce between the three countries, I will generally surge in net exports, but inflation quickly eroded much of report what has happened since NAFTA entered into force the gain to Mexican exporters. Since the first half of 1996, rather than what has happened because of the trade pact. The however, Mexican growth has been led by a revival of implementation of NAFTA provisions is only a small part domestic demand, primarily in the construction sector, of the story; in most respects, macroeconomic policy drives whiich had experienced a 23.5 percent drop in output in integration efforts in the NAFTA region. To put the Mexi- 1995. Net exports increased by 1.1 percent while domestic can experience under NAFTA into perspective, however, I demand grew by 6.2 percent in the second half of 1996. first start with a few brief comments on the peso crisis. Second, NAFTA partnership provided additional impe- tus for the Clinton administration to craft together a finan- The Peso Crisis and Its Aftermath cial rescue package that helped Mexico restructure its Mexico's experience under NAFTA got off to an inauspi- short-term dollar-denominated debt and ease its liquidity cious start with the outbreak of the Chiapas uprising on crisis. The U.S. Treasury loans, totaling about $13.5 bil- 110 NAFTA lion, were aLl repaid with interest ahead of schedule in Jan- ican tariffs on U.S. goods have been reduced from about uary 1997. 10 percent to 3 percent on a trade-weighted average. These reforms have helped raise the U.S. import share Trade and Investment from 69 percent in 1993 to 76 percent in 1996, and at By expandi:ig trade and investment linkages between the least partially insulated U.S. firms from the contractionary industries of the three countries, the NAFTA partners impact of the 1995 Mexican recession (U.S. sales in Mex- hoped to create substantial synergies among the three ico dropped substantially less than those of non-NAFTA economies that could generate important income and partners).3 employment gains. The NAFTA reforms aimed to increase Increased U.S. imports from Mexico resulted less from efficiency by encouraging each country to export those U.S. trade liberalization than from steady growth in the goods and services in which it had a comparative advan- U.S. economy and heightened competition in the Mexican tage, and to promote growth by generating more intense market due to a variety of economic reforms, including competition in each sector of the economy and allowing those resulting from NAFTA, which has forced Mexican firms to take advantage of economies of scale of production firms to shape up and search out new trading opportunities in an expanded North American market. at home and abroad. NAFTA has removed a few barriers to the U.S. market, notably in the textile and apparel sector, Trade but average U.S. tariffs have been cut by less than 2 per- NAFTA has contributed to a sharp expansion of regional centage points from already low levels, and tariffs on the trade, despite a temporary setback in 1995 caused by the most import-sensitive U.S. products (mostly agricultural) Mexican recession. Over the period 1993 to 1996, U.S. will not be removed for another 7 to 12 years. During exports to Mexico increased by 37 percent, and imports NAFTA's first three years, about one-quarter of the U.S. from Mexico by 83 percent. Except for the crisis year of import growth from Mexico has been in motor vehicles 1995, U.S. merchandise trade with Mexico has grown sub- (from $6 to $14 billion), textiles and apparel, and petro- stantially faster than trade with non-NAFTA countries. leum (reflecting higher prices, not NAFTA preferences). Total bilateral merchandise trade reached about $130 bil- Maquiladoras accounted for about 30 percent of the lion in 199'6, more than double the amount transacted growth in total Mexican exports since 1994; maquila when NAFTA negotiations began in 1991 (Table 1).2 exports increased by an annual average of 18.5 percent U.S. exporters have benefited from an acceleration of from 1994 to 1996. Net maquila exports (i.e., less maquila Mexican trade liberalization under NAFTA; average Mex- imports) averaged $5.7 billion over this period. Interest- TABLE I U.S.-Mexico IMlerchandise Trade, 1990-96 (in billions ol0 U.S. dollars) 1990 1991 1992 1993 1994 1995 1996 U.S. Exports to Mexico 28.3 33.3 40.6 41.6 50.8 46.3 56.8 % change from previous year 13.2 17.7 21.9 2.5 22.1 -8.9 22.7 U.S. Exports to the Rest of the World 281.6 303.3 317.0 323.1 347.4 411.2 434.3 % change from previoxs year 8.3 7.7 4.5 1.9 7.5 18.4 5.6 U.S. Imports irom Mexico 30.2 31.1 35.2 39.9 49.5 61.7 73.0 % change from previous year 11.0 3.0 13.2 13.4 24.1 24.6 18.3 U.S. Imports ilrom the Rest of the World 373.7 364.9 398.9 429.6 485.4 536.4 561.9 %changefrom previous year 4.4 -2.4 9.3 7.7 13.0 10.5 4.8 U.S. Balance of Merchandise Trade with Mexico -1.9 2.2 5.4 1.7 1.3 -15.4 -16.2 Note: Canadian trade is excluded from the Rest of World Statistics. Source: U.S. Depa-rmenr of Commerce, Bureau of Erosomic Analysis (Census Basis). 111 TRADE: TOWARDS OPEN REGIONALISM ingly, manufactured exports of non-maquiladoras rose by U.S.-Canada merchandise trade has continued the an annual average of 35 percent, and generated 55 percent robust expansion promoted by the Canada-U.S. Free-trade of the increase in total Mexican exports. Mexico was highly Agreement in 1989. Since NAFTA took effect, U.S.-Cana- dependent on the U.S. market before NAFTA, and it has dian trade has increased by 73 percent and totaled almost become even more reliant on U.S. sales; in 1996, trade $300 billion in 1996. Here again, U.S. trade with Canada with the United States accounted for 80 percent of total has grown faster than trade with non-NAFTA partners Mexican exports and imports (including maquilas), up except in 1995 (Table 2). from 75 percent in 1993. Over the period 1993-96, total CUSFTA and NAFTA reforms clearly contributed to Mexican trade with its NAFTA partners grew two-and-a- the growth in bilateral trade. Industry-specific analysis by half times faster than trade with the rest of the world. Daniel Schwanen (1997) reports that Canadian exports to In contrast, and somewhat surprisingly, U.S.-Mexico the United States grew substantially faster in sectors that services trade (including investment income) has not were liberalized under the trade pacts than sectors that grown significantly, and the growth has been slower than were not affected by the agreements. More important, but trade with the rest of the world. The published data seem harder to quantify, is the impact of tighter Canadian fiscal to indicate that NAFTA reforms in key sectors such as and monetary policy, and the more competitive Canadian insurance and telecommunications have not yet generated dollar, which pressured Canadian firms to look south for new trade flows, although such data are often challenged market expansion and assisted them in doing so. for undercounting trade in business services. However, the sharp rise in regional FDI in these sectors (see below) may Foreign Direct Investment presage substantial trade growth in the coming years. One of the major Mexican objectives for negotiating Canada-Mexico merchandise trade has also expanded NAFTA was to lock in, reinforce and augment the eco- sharply, albeit from a much lower base, and faster than nomic reforms implemented during the de la Madrid and Canadian trade with the rest of the non-NAFTA world. In Salinas administrations, in order to continue to encourage 1996, bilateral trade exceeded $7 billion. During the foreign investment in the Mexican economy. Mexico period 1993 to 1996, Canadian exports grew by 47 per- already had good access to the U.S. market, but the added cent-although growth has been virtually flat over the past "insurance" of inegotiated trade obligations would provide two years (in contrast to strong growth in U.S. sales in additional incentives to domestic and foreign investors 1996). Imports from Mexico have increased by 61 percent alike to expand their presence in the Mexican market as over this period. part of a broader strategy to rationalize production on a TABLE 2 U.S.-Canada Merchandise Trade, 1990-96 (in billions of U.S. dollars) 1990 1991 1992 1993 1994 1995 1996 U.S. Exports to Canada 83.7 85.1 90.6 100.4 114.4 127.2 133.7 to change from previous year 6.2 1.7 6.5 10.8 13.9 11.2 5.1 U.S. Exports to the Rest of the World 281.6 303.3 317.0 323.1 347.4 411.2 434.3 % change from previotes year 8.3 7.7 4.5 1.9 7.5 18.4 5.6 U.S. Imports from Canada 91.4 91.1 98.6 111.2 128.4 145.3 156.5 X change from previois year 3.9 -0.3 8.2 12.8 15.5 13.2 7.7 U.S. Imports from the Rest of the World 373.7 364.9 398.9 429.6 485.4 536.4 561.9 °1 change from previous year 4.9 -2.4 9.3 7.7 13.0 10.5 4.8 U.S. Balance of Merchandise Trade with Canada -7.7 -6.0 -8.0 -10.8 -14.0 -18.1 -22.8 Note: Mexican trade is excluded from the Rest of World Statistics. So-a: 1T.S. DOeparrment of Commerce, Bureau of Economic Analysis (Census Basis). 112 NAFTA regional scale. Mexico agreed to comprehensive investment easier to invest in Mexico by cutting red tape and remov- rules and obligations that compare favorably to the "high ing key ownership restrictions, particularly in the financial standards" that OECD countries hope to establish in their services industries. Nonetheless, companies won't invest ongoing negotiation on a Multilateral Agreement on unless they believe the climate for economic growth (and Investment (see Graham 1996 for analysis of the invest- for making profits) is favorable over a reasonable period of ment provisions of each pact). time. In contrast with the good economics of the Mexican Have companies altered their investment strategies in position, the U.S. debate on NAFTA has been driven by light of NAFTA? A survey of more than 400 firms in Mex- xenophobic concerns resulting from bad economics. U.S. ico (about half of which were U.S.-owned) conducted by opponents of NAFTA raised the specter of a tidal wave of the American Chamber of Commerce of Mexico in Novem- new U.S. investment in Mexico by footloose firms seeking ber 1996 reported that 56 percent of respondents replied to exploit cheap Mexican labor. Ross Perot heard a "giant "yes" to that general question, including 70 percent of the sucking sound" of U.S. jobs being shifted south of the bor- larger firms and two-thirds of the non-U.S. foreign compa- der. These concerns resonated well in the public debate due nies. In contrast, only 46 percent of small companies and to the insecurity of U.S. workers about the massive restruc- 52 percent of Mexican companies indicated that NAFTA turing of major U.S. companies over the past decade and the affected their strategies. Interestingly, two-thirds of the growing adjustment pressures generated by globalization. respondents said NAFTA had not led them to shift produc- In fact, the giant sucking sound has yet to be heard. tion from other foreign countries to the NAFTA region.' NAFTA has had little impact on investment in the United Based on balance-of-payments data reported by the States and reinforced only marginally already small but Bank of Mexico, one can see a strong trend toward significant U.S. investment in Mexico. Data for actual dis- increased FDI in Mexico since the late 1980s. Annual FDI bursements of U.S. funds in Mexico from January 1994 flows averaged $2.9 billion annually over the period through June 1996 indicate that U.S. FDI averaged about 1988-90, $4.5 billion annually from 1991-93, and $9.4 $4.4 billion annually in 1994 and 1995, and was running billion annually under NAFTA from 1994-96. Histori- at about half that rate in the first half of 1996.4 Half of cally, the United States has accounted for more than 60 those invest.ments were in the manufacturing sector; 17 percent of FDI in Mexico, followed by Germany and the percent were in financial services, reflecting perhaps the United Kingdom, with about 6 percent each (Hufbauer accelerated c,pening of that sector instituted as part of the and Schott 1992, p. 74). Since NAFTA entered into force, response to the peso crisis and ensuing recession. Contrary however, the U.S. share in new FDI in Mexico has fallen to to the claims of NAFTA critics, U.S. FDI in Mexico repre- 53 percent. Several factors may be at play. sents a small fraction of U.S. gross investment in plants First, NAFTA clearly has had a positive influence on and equipment (which amounted to about $785 billion in North American investment in Mexico; the NAFTA share 1996). of total FDI invested in Mexico from 1994 through June Similarly, the combined effect of CUSFTA and NAFTA 1996 was almost 60 percent (Riner and Sweeney 1997, on FDI in Canada has mattered less than the progress made Table 6). U.S. and Canadian firms accounted for more than in dampening inflation and reining in the budget deficit two-thirds of the foreign companies operating in Mexico as since the early 1990s. Over the period 1990-95, total FDI of June 1996 (about 80 percent of these firms were major- in Canada rose 28 percent to C$168 billion on a historical ity-owned). Canada has substantially increased its invest- cost basis; U.S. FDI accounted for two-thirds of that total ments in Mexico in both financial services and manufac- in 1995. The stock of U.S. FDI in Canada grew twice as turing, and accounted for 6.3 percent of new FDI in fast as FDI from the rest of the world combined.5 Mexico since the start of 1994 (compared with about 1 per- The Mexican story is more complex and harder to deci- cent pre-NAFTA). pher because of both the peso crisis and 1995 recession, Second, some previously small investors (e.g., India and and because of changes in SECOFI's reporting practices, the Netherlands) made substantial investments in 1995 which make it difficult to compare NAFTA results with that gave them a share of new FDI since NAFTA entered FDI inflows before 1994.6 To be sure, NAFTA has made it into force of 8.3 percent and 6.5 percent, respectively. It is 113 TRADE: TOWARDS OPEN REGIONALISM unclear whether these are one-off projects or whethcr they is further clouded by the peso crisis and severe Mexican portend heightened interest in the Mexican market from a recession of 1995, which resulted in a decline in total Mex- broader array of foreign investors. The peso crisis may have ican imports that year of about 9 percent (although spurred a fire-sale of some Mexican assets in 1995: More imports from its NAFTA partners fell only 2 percent).8 than $10 billion in new investments were approved in that How much of this difference is due to NAFTA preferences record year (Table 3). In that regard, the insurance policy versus other factors requires an industry-specific analysis value of NAFTA probably contributed to the decision to that is beyond the scope of this short paper. invest, despite the economic problems at that time, by Nonetheless, the available trade data allow a rough reinforcing expectations of a durable recovery over the analysis of the impact of NAFTA preferences in one impor- medium term. tant sector. When NAFTA was first negotiated, concerns The sectoral distribution of new FDI in Mexico since about trade diversion focused primarily on the potential NAFTA entered into force seems to have refocused on adverse effect of restrictive rules of origin on third-country manufacturing industries (especially machinery and equip- imports of textiles and apparel sector into the U.S. market. ment, processed farm products and metals) and on financial The U.S. market remains protected by quotas (which will services (which accounts for almost half of all FDI in the not be fully phased out until 2005 under the Uruguay services sector). FDI in manufacturing industries repre- Round reform of the Multi-Fiber Arrangement) and high sented 57 percent of new FDI since 1994, while 21 percent tariffs. Mexico and countries participating in the of new FDI went into services, compared with shares of 52 Caribbean Basin Initiative (CBI) benefited from significant percent and 37 percent respectively for FDI stocks in these trade preferences prior to NAFTA, so the NAFTA prefer- sectors prior to NAFTA in 1993 on a historical cost basis. ences have had only a small additional impact on foreign suppliers already facing discrimination in the U.S. market. Impact on Third Countries However, as a result of NAFTA, Mexico now receives bet- NAFTA was designed to adversely affect the trade of third- ter special treatment in the U.S. market than the CBI country suppliers by strengthening the ability of domestic countries. A quick look at the trade numbers tells the industries to compete in both regional and global markets. story. That said, NAFTA represents only one part, and a small Table 4 reports the growth in the value of U.S. imports one at that, of the package of economic initiatives that each of textiles and apparel from 1993 to 1996. Total U.S. member country has undertaken to improve productivity imports during this period increased by almost $10 bil- and promote higher living standards for its people. Deter- lion; the NAFTA region accounted for 39 percent of this mining the impact of NAFTA preferences on foreign trade growth and the CBI region for 21 percent. Imports from Mexico more than tripled; imports from Canada increased TABLE 3 by a more "modest" rate of 96 percent; and total imports Foreign Direct Investment Inflows into Mexico from the CBI region were up by 50 percent.9 In contrast, (in billions of U.S. dollars) countries with which the United States concluded highly BALANCE OF restrictive bilateral quota agreements-China, Hong PAYMENTS AUTHORIZED Kong, Taiwan and South Korea-experienced flat to BASIS FDI sharply lower sales in the U.S. market. As a result, the 1988 2.9 3.2 NAFTA and CBI regions supplied in 1996 a greater share 1989 3)2 3.6 1990 2.6 3.7 of U.S. imports (27 percent) than China, Hong Kong, and 1991 448 3 6 Taiwan combined (25 percent). 199 44 As shown in Table 5, both Mexico and the CBI region 1994 11.0 8.4 benefited significantly from U.S. trade preferences under 1995 9.5 10.4 the in-bond duty programs (HTS 9802). Mexican ship- 1996 7.6 7.9' ments under this program increased by $2 billion (or 180 ' Based on Januay- June d.ra percent) and accounted for 70 percent of the growth in the Soureo Bank of M-xico (1996,1997) for BoP data. Riner and Sweeney (1997) for data on FDI author,zed by SECOFI. value of U.S. imports of Mexican textiles and apparel over 11 4 NAPTA TABLE 4 data is not crystal clear. More than half of the growth in U.S. Imports of Textiles and Apparel Mexican shipments to the United States over the period (in millions of U.S. dollars) 1993-96 came in three product sectors: men's and women's 1993 1996 CHANGE ) cotton trousers, cotton knit shirts and knit shirts from man- World 36,079 45,933 27.3 made fibers. Mexico accounted for more than 60 percent of the growth in total U.S. imports of cotton trousers (MFA NAFTA 2,392 6,227 160.3 categories 347/348) as sales tripled to $1.3 billion in 1996; Mexico 1,372 4,232 208.5 ctgre Canada 1,020 1,995 95.6 Mexico's import share consequently rose from 12 percent in CBI 4,064 6,107 50.3 1993 to 24 percent in 1996. U.S. imports of these products Dominican Republic 1,458 1,802 23.6 from the three main CBI suppliers (Dominican Republic, Honduras 508 1,223 140.7 Honduras and Guatemala) also increased but at a slower China 4,767 4,892 2.6 pace (14.5 percent); their import market share also fell from 17.5 percent to 16.9 percent. Hong Kong, China 3,957 4,031 1.9 Taiwan, China 2,861 2,733 -4.5 Effects on Jobs and Wages President Clinton sold NAFTA as an engine of job cre- Rep. of Korea 2,477 2,049 -17.3 ation. His expansive claims counterbalanced the provoca- India 1,286 1,737 35.1 tive forecasts by NAFTA critics that the trade pact would result in massive job losses in the U.S. economy. In the Italy 1,048 1,703 62.5 event, neither side was right, nor could they be. Trade Souirce: U.S. International Trade Commission, Annual Report on U.S. Imports of Textiles agreements can affect the composition of employment, and Apparel 1996, April 1997. causing some industries to expand and others to contract, the period 1993-96. Similarly, CBI imports also increased but they will not cause the overall level of employment to by almost $2 billion (or 60 percent), accounting for 92 per- be appreciably higher or lower than the central bank deems cent of the growth in shipments from that region. If these appropriate. trends continue, Mexico could soon replace the CBI region as the major supplier of the U.S. market-provoking CBI TABLE 6 concern that NAFTA preferences, unless offset by "NAFTA Sources of Growth of U.S. Imports from Mexico, 1993-96 parity" legislation for the CBI region, will result in the (in millions of U.S. dollars) diversion of investment in this sector to Mexico. MtFA CATEGORY 1993 1996 INCREASE ($) If trade preferences are the cause of this growth, one 347/348: Men's & Women's would expect that other countries would not sharply Cotton Trousers 459 1,259 800 increase sales to the U.S. market in categories in which 338/339: Men's & Women's Mexican shipments have recorded the strongest growth Cotton Knit Shirts 81 456 375 (Table 6). However, the evidence from these disaggregated 638/639: Men's & Women's TABLE 5 Knit Shirts (man-made fibers) 65 384 319 U.S. Imports of Textiles and Apparel Subject to HTS 9802 647/648: Men's & Women's Preferences Trousers (man-made fibers) 64 257 193 (in millions of U.S. dollars) 649: Brassieres (man-made fibers) 97 167 70 SOURCE 1993 1996 CHANGE (t) 352: Cotton Underwear 38 120 82 World 4,813 8,887 84.6 Subtotal 804 2,643 1,839 Mexico 1,104 3,100 180.8 Other 568 1,589 1,021 Total 1,372 4,232 2,860 CBI Region 3,134 5,009 59.8 Source U.S. International Trade Commission, Annual Report on U.S. Imports of Texti/es Source. U.S. International Trade Commission, Annual Report on U.S. and Appaer': 1996, April 1997. 115 TRADE. TOWARDS OPEN REGIONALISM Since NAFTA entered into force in January 1994, total U.S. critics of NAFTA also argue that the pact will sup- U.S. employment increased by 6.7 million workers press increases in real wages (in both countries) and will (almost all full-time positions) or by about 2.2 million per increase the gap between skilled and unskilled workers. year; employment in the U.S. manufacturing industries Cline (1997, p. 257), however, notes that freer trade (and remained about the same; and the unemployment rate fell immigration) at best may be responsible for 20 to 25 per- below 5 percent (Table 7). Even the auto sector (SIC 371), cent of the increase in the differential between these groups which ran a large trade deficit with Mexico and Canada, in the U.S. economy over the past two decades. To be sure, experienced significant employment gains over this average hourly earnings in U.S. manufacturing has period. At the same time, about 1.5 million workers remained flat since January 1994 (Table 9). However, to annually lost their jobs because of technological change, the extent that NAFTA increased overall U.S. exports, it slack demand, import competition etc. Thus, gross U.S. will have had a small positive effect on U.S. wages, because job creation was 3.7 million per year (3.7 less 1.5 = 2.2). workers in the U.S. export sector earn about 10 to 15 per- What impact has NAFTA had on these developments? cent more than those in non-exporting firms (Richardson Given the size of the U.S. civilian labor force (about 130 and Rindal 1996); this annual export wage premium million), the answer must be "not much." One cannot dif- amounts to more than $4,000 per worker in 1997. ferentiate jobs dedicated to exports from those producing The Mexican experience, in contrast, is dominated by for the domestic market, but the overall impact on U.S. the combined shocks of the peso crisis and subsequent employment of increased NAFTA trade must be minimal recession. Average real wages in the Mexican manufactur- given the magnitude of trade compared with the domestic ing sector fell by about 19 perceint durinig the period economy. The only hard data available are the number of 1993-96, despite impressive 7 percent annual gains in workers certified for NAFTA Transitional Adjustment worker productivity. At their peak in 1994, real wages Assistance because their jobs were abolished by firms that were about 40 percent higher than 1987, but in 1997 were moved to Mexico or Canada, or by firms that lost sales to less than 10 percent higher than the level of 1987. Mexican or Canadian imports. As of March 1997, a total of 117,000 workers were certified under the NAFTA pro- Dispute Settlement gram, or about 39,000 per year-a very small share of the Close friends and trading partners inevitably have differ- annual total of 1.5 million job displacements in the U.S. ences, and the number of trade disputes invariably rises as economy (Table 8). trade volumes expand. This maxim has proved true in the TABLE 7 U.S. Employment, 1994 IQ to 1997 IQ (in millions) TOTAL FULL- PART MANUFACTURING TOTAL RATE EMPLOYED TIME TIME (ANNUAL AVG.) UNEMPLOYED (%U) 1994 (1) 122.0 98.4 23.6 18.32 8.6 6.6 (II) 122.6 99.4 23.1 8.0 6.2 (III) 123.2 100.3 23.0 7.9 6.0 (IV) 124.4 101.0 23.5 7.4 5.6 1995 (1) 124.9 101.4 23.5 18.47 7.3 5.5 (II) 124.7 101.5 23.2 7.5 5.6 (III) 124.8 101.7 23.2 7.5 5.7 (IV) 125.1 102.1 23.1 7.4 5.6 1996 (1) 125.7 102.5 23.2 18.28 7.5 5.6 (fI) 126.4 103.2 23.1 7.3 5.4 (III) 127.0 103.9 23.2 7.1 5.3 (IV) 127.7 104.6 23.1 7.1 5.3 1997 (I) 128.7 105.3 23.5 18.30 7.2 5.3 Sauoer: U.S. Department of Labor, BLS, Enploymert and Earnings (April 1997). 116 NAFTA TABLE 8 that reason, NAFTA incorporated dispute-settlement pro- Workers Certified under the NAFTA Transitional Adjustment visions that help to assuage concerns in those countries Assistance Program about continued good access to the dominant market in the CANADA MEXICO UNIDENTIFIED TOTAL region, the United States. 1994 (1) 543 980 0 1,523 Given the relative openness of the U.S. market, the (II) 1,144 2,599 0 3,743 major source of concern by Canada and Mexico with regard (111) 890 4,530 408 5,828 (IV) 4,162 4,232 1,257 9,651 to market access has been the introduction of new protec- 1995 (1) 1,324 3,124 1,471 5,919 tion through the administration of U.S. unfair trade (11) 3,488 4,228 834 8,550 statutes (especially via anti-dumping and countervailing (III) 2,473 2,475 2,458 7,406 (IV) 3,497 6,564 1,557 11,618 duties). Historically, both countries have been the target of 1996 (1) 1,692 7,418 2,435 11,545 numerous U.S. cases and consequently placed high priority (II) 1,672 9,550 2,903 14,125 (III) 747 6,062 2,645 9,454 on icorporating new disciplies on these practices in the (IV) 941 8,407 757 10,105 NAFTA. For example, Canada faced 70 U.S. anti-dumping 1997 (1) 1,169 12,575 3,305 17,049 or countervailing duty cases over the past three decades, Total 23,742 72,744 20,030 116,516 although their frequency has declined sharply since a rash Note: Assistance provided when jobs lost either because of a shift in production to of steel cases in the early 1990s. Canada/Mexico or because of increased imports from Canada/Mexico (whether or not the increase was at -ributable to cariff cuts under NAFTA). The Canada-U.S. Free-Trade Agreement sought to limit Source: U.S. Department of Labor, Office of Trade Adjustment Assistance database. trade frictions in this area by creating a new dispute mech- anism (in CUSFTA Chapter 19) to review the final deter- NAFTA region, even though the number of disputes and minations of national trade authorities that allegedly did the volume 9f trade affected is relatively small compared not follow the rules of national legislation. This mecha- with overall bilateral trade flows. Yet for Canada and Mex- nism was then incorporated in NAFTA (also Chapter 19) ico, some cases can have a high political profile even if they after Mexico agreed to overhaul its laws and regulations to represent a small share of bilateral trade (e.g., the softwood conform more closely to U.S. and Canadian norms. Note lumber case between the United States and Canada). For that this dispute mechanism does not exempt Canada and Mexico from U.S. anti-dumping cases, but does provide a TABLE 9 more expeditious procedure than lengthy court challenges Average Hourly Earnings: Manufacturing for resolving disputes regarding alleged problems in the administration of those laws in each country. CONSTANT $ Tables 10 and 11 document the cases that have been CURRENT $ (1982) 1994 January 11.96 8.07 brought before dispute panels under the provisions of February 12.00 8.08 NAFTA Chapter 19. From January 1994 through April March 11.99 8.05 1997, NAFTA panels have been convened in 26 cases (of Annufal Avg. 12.06 8.02 which decisions are pending in four cases). The United 1995 January 12.24 8.03 States has been the most active complainant (13 cases) and February 12.24 8.00 the country most subject to complaint (10 cases). Canada March 12.25 7.98 Annual Avg. 12.35 7.99 and Mexico have also been frequently on both sides of the docket. Complaints against Mexico generally cite problems 1996 January 12.64 8.07 February 12.57 8.00 with the administration of the new anti-dumping rules by March 12.54 7.95 the SECOFI bureaucrats, while U.S. and Canadian cases Annual Avg. 12.78 8.04 often involve more arcane problems related to the interpre- 1997 Januaty 13.04 8.08 tation of the dense complex of national anti-dumping and February 13.03 8.05 countervailing duty laws and regulations. Several general March 13.09 n.a. observations on this process bear mention. Source: U.S. Department of Labor, BLS, EmploymentanddEanirgs (April 1997); Table B- First, the cases have been handled expeditiously (com- 17, Average Hourly & Weekly Earnings of Production or Nonsupervisory Worrketm on Private Nonfarm lpayrolls. pared with judicial review of final determinations by 117 TRADE: TOWARDS OPEN REGIONALISM TABLE 10 NAFTA Chapter 19 Panel Reviews, January 1994 to April 1997 CDA: Panel reviews reviewing Canadian agencies' determinations MEX:Panel reviews reviewing Mexican agencies' determinations USA: Panel reviews reviewing U.S. agencies' dererminations DISPUTE ACTION COMPLAINANT OUTCOME CDA-94-1904-01 Cerrain Fresh, Whole, Delicious, Red Delicious Canadian producers Review terminated by joint consent and Golden Delicious Apples, Originaring in or of participants Exported from the United Stares, Excluding Delicious, Red Delicious and Golden Delicious Apples Imported in Non-Standard Containers for Processing (injury determination) CDA-94-1904-02 Synthetic Baler Twine with a Knot Strength of U.S. producer/exporter Unanimously affirmed in part and 200 lbs. or less, from the United States and Canadian importer remanded in part the agency's (injury determination) determination; Determination on remand affirmed CDA-94-1904-03 Certain Corrosion-Resistant Steel Sheet Products U.S. producers Unanimously affirmed in part and from the United States remanded in part the agency's (dumping determination) determination; Determination on remand affirmed CDA-94-1909-04 Certain Corrosion-Resistant Steel Sheet Products, U.S. producers Unanimously affirmed the agency's originating in or exported from the United States determination (injury determination) MEX-94-1904-01 Flat-Coated Steel Products from the United States U.S. producers Unanimously affirmed in part and (dumping determination) remanded in part the agency's determination MEX-94-1904-02 Imports of Cut-length Plate Producers from the U.S. producers Panel majority, with two dissenting United States opinions, remanded the final (dumping determination) determination to the investigating authority, which then issued a new determination that terminated the proceeding. MEX-94-1904-03 Crystal and Solid Polystyrene from the United States U.S. and Mexican prodacers Panel majority affirmed the agency's (dumping determination) final determination USA-94-1904-01 Live Swine from Canada, Canadian producers Unanimously affirmed in part and jCVD Administrative Review) remanded in part the agency's determination; Determination on remand unanimously affirmed USA-94-1904-02 Leather Wearing Appeal from Mexico Mexican producers Unanimously remanded at request (CVD Administrative Review) of agency CDA-95-1904-01 Certain Malt Beverages from the United States U.S. producers Unanimously affirmed the agency's (injury determination) determination CDA-95-1904-02 Fresh, Whole, Delicious, Red Delicious and Golden U.S. exporters Review automatically terminated by Delicious Apples, originating in or exported from the complainant Utited Stares (dumping determination) CDA-95-1904-03 Machine Tufted Carpeting originating in or exported U.S. producers Review automatically terminated by from the United States complainant (dumping determination) I 1i8 NAFTA TABLE 10 NAFTA Chapter 19 Panel Reviews, January 1994 to April 1997, continued DISPUTE ACTION COMPLAINANT OUTCOME CDA-95-1904-04 Refined sugar originating in or exported from the U.S. exporter Unanimously affirmed in part and United States remanded in part the agency's (dumping detetmination) determination MEX-95-1904-01 Seamless Steel Tube originiating in the United States U.S. producers Review automatically terminiated by (dumping determination) complainant USA-95-1904-01 Porcelain-on-Steel Cookware from Mexico Mexican producers Unanimously affirmed in part and (AD administrative review) remanded in part agency's determination; determination on remand unanimously affirmed USA-95-1904-02 Gray Portland Cement and Cement Clinker from Mexican producers Unanimously affirmed agency's final Mexico determination (AD administrative review) USA-95-1904-0' Color Picture Tubes from Canada Canadian exporters Unanimously affirmed agency's (dumping determination) final determination USA-95-1904-04 Oil Country Tubular Goods from Mexico Mexican and U.S. producers Determination affirmed in part and (dumping determination) remanded in part agency's determination; Determination on remand affirmed USA-95-1904-05 Fresh Cut Flowers from Mexico Mexican producers Unanimously affirmed in part and (AD administrative review) remanded in part the agency's determination CDA-96-1904-01 Bacteriological Culture Media from Becton Dickinson U.S. producers Review automatically terminated and Company and Difco Laboratories of the United by joint consent of participants States and from Unipath Limited of the United Kingdom (dumping determination) MEX-96-1904-Cl Cold Rolled Sheet from Canada Canadian producers Review automatically terminated by complainant MEX-96-1904-C2 Rolled Steel Plate from Canada Canadian producers Decision due October 14, 1997 MEX-96-1904-C3 Hot Rolled Steel Sheet from Canada Canadian producers Decision due June 5, 1997 USA-96-1904-0 L Porcelain-on-Steel Cooking Ware from Mexico U.S. producers Review automatically terminated by complainant USA-97-1904-01 Gray Portland Cement and Clinker from Mexico Mexican producers Decision due March 17, 1998 5th AD Administrative Review USA-97-1904-0.2 Gray Portland Ccmcnt and Clinkcr from Mcxico Mexican producers Decision due March 19, 1998 4th Administrative Review Source: NAFTA Secretariat, U.S. Section, Staistical S-mmary of Dispste Settlmenrt Panels anae, the North Amerias FPee-Trade Agrerenent, May 12, 1997 119 TRADE: TOWARDS OPEN REGIONALISM TABLE 1 I NAFrA Disputes under Chapter 19: January 1994 to May 1997 COMPLAINTS BY 1994 1995 1996 1997 TOTAL United States 6 5 2 0 13 Canada 2 1 3 0 6 Mexico 1 4 0 2 7 COUNTRY WHOSE AD/CVD ACTION SUBJECT OF COMPLAINT 1994 1995 1996 1997 TOTAL United States 2 5 1 2 10 Canada 4 4 1 0 9 Mexico 3 1 3 0 7 OUTCOME OF U.S. COMPLAINTS 1994 1995 1996 1997 TOTAL Review Terminated 0 3 2 0 5 Decision Remanded in Whole or Part 3 1 0 1 5 Action Reaffirmed 2 1 0 0 3 OUTCOME OF CANADIAN COMPLAINTS 1994 1995 1996 TOTAL' Review Terminated 1 0 1 2 Decision Remanded in Whole or Part 1 0 0 1 Action Reaffirmed 0 1 0 1 OUTCOME OF MEXICAN COMPLAINTS 1994 1995 1996 TOTALI Review Terminated 1 0 0 1 Decision Remanded in Whole or Parr 0 3 0 3 Action Reaffirmed 0 1 0 1 a. MEX-96-1904-02 decision due; b. USA-97-1904-01 decision due. national agencies) and objectively. Most panels report breaking procedures for the resolution of investment dis- within or close to the requisite time periods, and panel vot- putes, including new rights for private parties to obtain ing has in most instances not fallen along nationality lines, relief directly against governments for NAFTA violations. Second, most of the disputes have been resolved on a "technical" track, without resort to political negotiations NAFTA Agreements on the Environment and (the U.S.-Canada softwood lumber case is a notable excep- Labor tion). The panels do not "rubber stamp" the decisions of To assuage domestic concerns about the impact of NAFTA national agencies; indeed, in about 40 percent of their deci- on labor and the environment, and to bolster support in his sions (nine of 22) the panel has remanded the case in whole own party for the pact, President Clinton insisted that the or part to the national agency. NAFTA partners supplement their free-trade agreement NAFTA contains other dispute-resolution mechanisms with additional obligations relating to labor and the envi- and also allows member countries to choose between ronment."0 The North American Agreement on Environ- regional and multilateral forums. However, to date, only mental Cooperation (NAAEC) and the North American four disputes between NAFTA partners have been brought Agreement on Labor Cooperation (NAALC) were concluded to the WTO (Table 12). Eight cases have been brought in August 1993 and entered into force as side agreements to under the general GATT-like procedures of NAFTA Chap- NAFTA. In addition, the UJnited States and Mexico con- ter 20, involving politically sensitive issues such as truck- cluded a Border Environmental Cooperation Agreement ing, the Helms-Burton sanctions against Cuba and Mexi- that seeks to promote new investment in environmental can tomatoes, which have provoked consultations but few infrastructure (primarily in the border region), and estab- agreements among the partner countries (Table 13). Inter- lishes a North American Development Bank (NADBank) to estingly, the NAFTA investment chapter contains path- channel additional public resources for those purposes. 120 NAFTA TABLE 12 Trade Disputes Brought by the NAFTA Countries in the WT:O CONSULTATION DISPUTE ACTION COMPLAINANT REQUESTED OUTCOME WT/DS 31 Canadian Tax treatment of "split-run" periodicals and United States 11-Mar-96 Panel established 19 June application of favorable postage rates to certain 1996. In March 1997 Canadian periodicals the WTO panel found measures applied by Canada violation of GATT rules WT/DS 49 U.S. anti-dumping investigation on fresh and chilled Mexico 7-Jul-96 In October 1996, the U.S. tomaroes Department of Commerce and Mexican exporters signed an agreement suspending the anti-dumping investigation WT/DS 101 Mexican anti-dumping case on U.S. high fructose corn syrup United States 4-Sep-97 Consultations pending WT/DS 103 Canadian export subsidies on dairy products and Canada's United States 8-Oct-97 Consultations pending TRQ on milk In essence, the NAAEC and the NAALC have two Second, the pacts establish a forum for consultations and broad-ranging objectives. First, the pacts monitor imple- dispute resolution in cases where domestic enforcement is mentation of national laws and regulations in each country inadequate. pertaining to labor and the environment. In effect, the To date, the side agreements do not have a stellar track NAAEC and the NAALC serve a watchdog role, alerting record. Setting up the administrative machinery to imple- countries to simmering abuses of labor and environmental ment the pacts has been a slow and cumbersome process. practices within each country so that the problems can be The dispute-resolution procedures have begun to be used, resolved before they grow into trans-boundary disputes. with some positive results, but the pacts have been less suc- TABLE 13 NAFTA Chapteir 20 Panel Reviews, January 1994 to March 1997 DISPUTE ACTION COMPLAINANT OUTCOME CDA-95-2008-01 Tariffs applied by Canada to certain U.S. government Settled Dec 1996 U.S.-origin agricultural products USA-97-2008-0- Broom Corn Brooms Mexican government Pending MEX Discrimination against U.S. trucking firms U.S. Consultations at NAFTA Commission meeting in 1995; discussions continue between governments USA U.S. Sugar Containing Products Re-Export Program Canada Consultations requested on October 23, 1996; USA U.S. Sugar Containing Products Re-Export Program Canada Consultations requested on October 23, 1996; consultations were held on November 20, 1996 USA Cuban Liberty and Democratic Solidarity Act of 1996 Canada and Mexico Consultations were held on April 20, 1996 and May 28, 1996; on June 28, NAFTA Commission convened USA Provisions on trucking Mexico On January 19, 1996 consultarions were held USA Tariff-rate quota on tomatoes Mexico On January 18, 1996 consultations were held Source NAPTA S-cserar, U.S. Section, Stat/isRl S--rmery of Dispte Settlement Panels se/-er the North AmerUicn Free-Trade Agree-ent, March 1997 121 TRADE TOWARDS OPEN REGIONALISM cessful in channeling new resources into environmental 1993, Table Al, p. 167). Through mid-1997, the projects. The following subsections report notable devel- NADBank has made barely a dent in that demand. The opments in each area. NADBank has reviewed 12 financing proposals that have been certified by the Border Environmental Cooperation Environment Commission as eligible for NADBank loans. Four projects The Council on Environmental Cooperation (CEC) estab- have been approved, involving about $4 million in NAD- lished under the NAAEC has devoted almost half of its Bank financing, but funds have been disbursed for only one $10 million annual budget to joint programs to promote, of them (a water-treatment plant in Brawley, California). inter alia, sound management of chemicals, cooperation Six other projects are in the pipeline, one found alternative on environmental enforcement and the development of a financing, and one project, which sought a grant rather North American pollutant-release inventory. These pro- than a loan, was rejected. grams operate in relative obscurity, and have been over- shadowed by public interest in the dispute procedures. Labor Through April 1997, the CEC received nine submis- In contrast to the NAAEC that has a mandate to monitor sions detailing complaints about the enforcement of envi- national policies, mitigate disputes and develop initiatives ronmental laws and seeking redress under the NAAEC's to upgrade environmental infrastructure, the NAALC has environmental enforcement mechanism. Four of these cases focused primarily on oversight and enforcement of national were terminated because of lack of evidence or because the labor laws and practices. Each country has opened a National issue was already being adjudicated in Canadian courts; Administrative Office (NAO), which serves, inter alia, as a four cases are currently being reviewed by the secretariat to clearinghouse for private sector notifications of abusive labor determine whether to present the case to the CEC council practices in the region. Such petitions can trigger govern- of ministers for action; and one case has advanced to the ment-to-government consultations on the matter by the stage of compiling a factual record of the dispute (which respective NAOs, and subsequently can lead to ministerial presages the convening of dispute-resolution procedures). reviews and the formation of dispute-resolution panels. It If this process seems clumsy and unwieldy, it is-and by has also sponsored four comparative studies of labor market design. Mexico and Canada staunchly resisted the incorpo- conditions in the NAFTA region. ration of new dispute procedures under the side agreements Through mid-1997, the record of the dispute-settle- and insisted that the process be long on consultation and ment process of the NAALC has been mixed. The U.S. short on adjudication. Indeed, the NAAEC dispute-settle- NAO has received six submissions regarding alleged denial ment process can drag out for almost two years, compared of freedom of association and other unfair labor practices by with the 240 days under which disputes are normally set- specified employers; the Mexican NAO has received one rled under NAFTA Chapter 19. submission. The outcome of these cases bear mention. In The NADBank was designed to channel new financial two cases, the petitions exposed problems that were recti- resources to address the burgeoning capital requirements of fied by the national labor authorities: In one instance, the environmental projects in Mexico and the United States. Mexican Labor Board recognized the union subject to dis- The NADBank has been capitalized at $3 billion, with pute; in the other, the U.S. Labor Board ordered a company $450 million paid in by each country over four years and to reinstate with back pay workers who were laid off when the rest callable. At least initially, 90 percent of its funds a plant was illegally closed. Two U.S. cases were termi- have been earmarked for wastewater, water treatment and nated because of insufficient evidence, one case was with- solid-waste projects in the U.S.-Mexico border region; the drawn before the NAO review was completed and the two other 10 percent are targeted for community adjustment other U.S. cases led to ministerial consultations. and investment projects not limited to the border region. In 1993, a variety of government, business and environ- Building on NAFTA mental groups estimated that the cost of needed waste- NAFTA was designed not only to expand trade among water treatment facilities alone in the border region could neighbors in North America, but also to strengthen the total $3 to $6 billion over 10 years (Hufbauer and Schott ability of North American firms to compete more effec- 122 NAFTA tively in both regional and global markets. This is impor- Lessons from the NAFTA Experience tant because, for the foreseeable future, all three countries NAFTA has been in force for only a few short years and will run sizable current account deficits (Mexico's deficit many of its key provisions have not been fully imple- should continue to rise in coming years as its economy mented. Its effectiveness has been impeded by the severe recovers and attracts more capital imports). Obviously, macroeconomic shock that beset Mexico in late 1994. It each country cannot solve its problems primarily by has served as a lightning rod for U.S. concerns about labor exporting more to its regional partners; rather, each needs adjustment and wage suppression, and in so doing has both to export and import more to promote competition in revived latent protectionist pressures in both major politi- its markets and greater efficiency and productivity of its cal parties (and enshrined protectionism as a central tenet industries and workers. of the new reform party). Yet despite the harried experience In that regard, NAFTA is part of the global trade strat- of the NAFTA partners during this period, the trade pact egy of each country. Only two weeks after the announce- has produced tangible results, as documented in this paper. ment of U.S.-Mexico free-trade talks in June 1990, Presi- As the first comprehensive and reciprocal free-trade dent Bush iaunched the Enterprise for the Americas agreement between developed and developing countries, Initiative (EAI). He envisioned the U.S.-Mexico pact as a NAFTA has also illustrated several important aspects of stepping-stone to a broader hemispheric arrangement that free-trade pacts that should help inform other developing would help promote the type of economic and political countries in the Western Hemisphere (or the Asia-Pacific reform that was evolving in Mexico. NAFTA expansion region) that have committed to undertake similar arrange- was seen as one possible model for such a pact, and an ments. I conclude with five critical trade-policy lessons accession clause was included in NAFTA vaguely similar in derived from the NAFTA experience: form to the GATT provisions. 1. Macro matters most. Trade agreements create oppor- At the Summit of the Americas in Miami in December tunities; they do not guarantee sales. To promote sus- 1994, the three NAFTA members committed to negotiate tained growth and take full advantage of those oppor- a Free-Trade Area of the Americas (FTAA) with the other tunities, macroeconomic policy must be prudent-at democratic nations of the Western Hemisphere by the year home and in the partner countries. 2005, and also agreed to open talks with Chile on NAFTA 2. Trade pacts provide an insurance policy against accession. The Chilean negotiations foundered due to the new protectionism at home and abroad. They deter lack of fast-track negotiating authority in the United abrupt policy reversals and help governments with- States, but Canada and Chile signed a bilateral free-trade stand the protectionist demands of their domestic pact at the end of 1996 (Chile and Mexico concluded a lobbies. Mexico's response to the peso crisis is evi- more limited FTA in 1991). dence of this salutary effect. These regional initiatives progressed initially when the 3. Free-trade pacts involve asymmetric obligations Uruguay Rotnd of multilateral trade negotiations was at that fall more heavily on developing than developed an impasse due to U.S.-European differences over agricul- country partners. The benefit for developing coun- tural reforms NAFTA provided new ideas and models for tries is that the pact locks in the domestic reforms several critical GATT accords under negotiation in Geneva needed to reinforce growth, and represents a "Good and served as an important building block for those multi- Housekeeping Seal of Approval" for those policies- lateral deals. Even the negative aspects of NAFTA, such as thus making them more attractive to foreign the restrictive rules of origin for textiles and apparel, con- investors and promoting the transfer of technology tributed to t:he GATT deal by lessening U.S. industry and management skills. opposition to the phaseout of quotas under the Multi-Fiber 4. Trade pacts are not engines ofjob creation, but they Arrangement. On balance, NAFTA has complemented and do support jobs that provide a substantial wage reinforced multilateral efforts to liberalize trade in the premium over earnings in the non-exporting sector. GATT and \WTO; in most respects, it has set high stan- 5. Integration is an iterative process. Not all issues of dards for both trade liberalization and rule-making that importance in bilateral or regional relations are cov- should spur complementary action on a global scale. ered ab initio in trade pacts; but as countries become 123 TRADE: TOWARDS OPEN REGIONALISM more integrated, new issues that span domestic and Schwanen, Daniel (1997). "Trading Up: The Impact of Continental international concerns often are added to the com- Integration on Trade, Investment, and Jobs in Canada," Commen- mon agenda. Indeed, as the Summit of the Americas tar) #89. Toronto: C. D. Howe Institute, March. process has demonstrated, trade talks can serve as a magnet for attracting support on a wide array of ini- Notes tiatives, including strengthening democracy, com- 1. On a historical cost basis, U.S. FDI totaled $65.5 billion in bating drug trade and promoting better environmen- 1990 while Canadian FDI in the United States stood at $27.7 billion. tal conditions and labor rights. 2. These data come from the U.S. Department of Commerce and are not strictly comparable with those published by Mexican sources because of different treatment of maquiladora trade. References 3. These data reflect FOB/FOB trade as reported by the Banco de Cline, William R. (1997). Trade and Income Distribution. Washington: Mexico, and include imports by maquiladoras. Institute for International Economics. 4. These figures were compiled by the U.S. Embassy in Mexico; DeLong, Bradford, Christopher DeLong and Sherman Robinson they are lower than the data reported by SECOFI on registered U.S. (1996). "The Case for Mexico's Rescue," Foreign Affairs, Vol. 75, investments during this period. No. 3 (May/June). 5. Data are from Statistics Canada, Canada's International Invest- Graham, Edward M. (1996). Global Corporations and National Govern- ment Position 1995. ments. Washington, D.C.: Institute for International Economics. 6. Starting in 1994, SECOFI changed from reporting approved Hufbauer, Gary Clyde and Jeffrey J. Schott (1992). Nortk American FDI (which may or may not be undertaken) to registered FDI. Free Trade: Issues and Recommendations. Washington, D.C.: Insti- According to Riner and Sweeney (1997), investments are often reg- tute for International Economics. istered well after they have been made. (1993). NAFTA: Ani Assessment. Washington, D.C.: Institute 7. A summary of the complete results of the AmCham survey is for International Economics, (rev. ed). October. reported in Riner and Sweeney (1997). Lustig, Nora (1996). The Mexican Peso Crisis, The Rescale Package and 8. These figures include trade of in-bond industries, as published Beyond. Washington, D.C.: Inter-American Dialogue, October. by Banco de Mexico (1996, 1997). Richardson, J. David and Karin Rindal (1996). Why Exports Matter: 9. While Mexico has been the fastest growing supplier of these More! Washington, D.C.: Institute for International Economics products to the U.S. market since 1993, the trade of some CBI coun- and The Manufacturing Institute. tries also grew sharply: El Salvador (up 179 percent), Honduras (up Riner, Deborah and John V. Sweeney (1997). The Effects of NAFTA 140 percent). on Mexico. Mexico City: American Chamber of Commerce of Mex- 10. The three countries also concluded an agreement on import ico, processed. surges, which to date has received almost no attention. 124 Comment N ORA LU ST I G I T IS A PLEASURE TO COMMENT ON JEFF SCHOTT'S PAPER, "NAFTA: AN INTERIM REPORT." THE PAPER presents a thorough and clear overview of what has happened to trade and investment flows, employ- ment in the United States and trade disputes since NAFTA's implementation. I fully agree with the author that it is premature-and particularly difficult given the Mexican crisis-to make an assess- ment of the agreement and its impact scarcely three years after it came into effect. Having said that, are there indications that NAFTA is beginning to deliver on its promises of higher trade and invest- ment flows, gains in efficiency and improved living standards (particularly in Mexico)? Have the costs in terms of ernployment losses matched the fears of the opponents of NAFTA in the United States? Has NAFTA become the anticipated stepping-stone for global freer trade and regional economic integration? Strictly speaking, in order to answer some of these ques- tory countries has been expanding to record levels. For rions we wouLd need either an econometric or a computable example, during 1994 (NAFTA's first year), trade between general equilibrium model to be able to isolate the effects the NAFTA countries grew 17 percent, reaching a record of simultaneous events. After all, the implementation of of US$350 billion, of which about US$100 billion repre- NAFTA practically coincided with a 45 percent real deval- sents U.S.-Mexico trade. Taking a longer perspective, all of uation of the peso, a 7 percent drop in Mexican output, and U.S.-Mexico trade surged from US$30 billion in 1986 (a a 22 percent Fall in Mexican real wages during 1995. There year after Mexico began its trade liberalization and when it have also been large variations in the value of the U.S. dol- joined GATT) to close to US$150 billion in 1996.' lar vis-a-vis other major currencies. It is surprising that In 1995 NAFTA appears to have contributed to the despite the large number of models that proliferated before preservation and promotion of trade, despite Mexico's cri- NAFTA was approved, there are no readily available esti- sis. As Schott notes, in contrast with the crisis in 1982, mates on the impact of the various effects so far (at least I when Mexico reintroduced import permits for most prod- have not run across any). In the absence of quantitative ucts, this time it kept its open-economy policies largely in studies, we must circumscribe our analysis to, as Schott place. This might have to do more with Mexico's unilat- puts it, "wha: has happened since NAFTA entered into force eral trade liberalization and GATT-membership than with rather than what has happened because of the trade pact." NAFTA. However, given the large weight of imports from the United States in Mexico's total imports, NAFTA Trade Flows essentially has severely limited the use of protectionist As discussed by the author, there can hardly be any doubt measures to reduce the trade gap. This might partially that since NAFTA's passage, trade among the three signa- explain why imports fell by 12 percent during the 1995 Nora Lustig was a Senior Fellow at the Brookings Institution. She now is a Senior Adviser and Chief of the Poverty and Inequality Advisory Unit at the Inter-American Development Bank. 125 TRADE TOWARDS OPEN REGIONALISM recession, while they contracted by almost 35 percent dur- Mexico's Economic Performance ing the previous recession in 1983. As Jeff Schott points out, NAFTA was not a cause of the Notably, during 1995, U.S. (and to a lesser extent Mexican peso crisis. (I find some discrepancies in the Canadian) exports to Mexico were much less affected than author's analysis of what caused it, but this may not be the those from other areas. While exports from the United place to discuss them.) On the other hand, NAFTA appears States fell by less than 2 percent (and from Canada by 15 to be contributing to Mexico's recovery. The Mexican percent), exports from Central America declined by 44 recovery got under way in the third quarter of 1995, and percent, from South America by 32 percent, from the the 7 percent decline in GDP that year was followed by a European Union by about 26 percent and from Asia by 23 5.1 percent growth in output during 1996 (and a similar percent.2 How much of this is related to NAFTA? Part of growth rate is expected in 1997). As the author noted, the difference in export performance might be attributed exports have been the engine of Mexico's recovery. But here to the depreciation of the dollar vis-a-vis other major cur- again, it is hard to disentanglc NAFTA's contribution from rencies in 1995. However, part of the difference must also Mexico's economic reforms, the devaluation of the peso, be ascribed to the expanded market access for U.S. and the 1995 recession. exporters and the importance of intra-industry bilateral NAFTA's contribution may be more patent in the per- trade between Mexico and the United States, both formance of foreign direct investment. On average, strengthened by NAFTA. between 1994 and 1996 foreign direct investment (FDI) There is, nonetheless, one worrisome aspect. In 1995, was almost double (close to US$8 billion) what it was in Mexican tariffs were increased to their GATT-bound lev- the three years before NAFTA's implementation. Moreover, els in a handful of economic sectors. Because of NAFTA, in 1995, when portfolio flows became highly negative, FDI the United States-as well as other countries that had was 75 percent above the level in the three years before signed free trade accords with Mexico-was spared this NAFTA. A comparison with the previous crisis may be tariff increase. The question remains to what extent these telling. During 1983, FDI fell to one-fifth of what it had changes in trade protection are also responsible for the been in the previous years. Clearly, the business opportuni- striking differences observed in imports into Mexico. ties brought by NAFTA have had a positive impact on FDI The fact that NAFTA appears to have shielded U.S., decisions, the peso crisis notwithstanding. and to a lesser extent Canadian, exports from the brunt of the Mexican crisis and the increase in some tariff lines that Employment and Wages followed is good news for Mexico's NAFTA partners and As Schott points out, the gross disemployment effects due provides pro-NAFTA officials and lawmakers with ammu- to NAFTA are probably negligible. Furthermore, an accu- nition to defend it. However, it does raise some concerns rate assessment of the net employment effect of NAFTA about NAFTA's impact on trade diversion. Perhaps under would need to look at job creation as well. A recent study the prevailing rules at the time of its implementation that attempts to do this analytically found that the overall NAFTA could be shown to create more trade than it impact of NAFTA tariff liberalization on U.S. employment diverted. However, this might not be the case if one fac- has been slightly positive.3 tors in the fact that trade barriers against non-members A more relevant measure of NAFTA-related hardship may be raised while NAFTA rules remain intact when cir- for American workers may be its effect on wages, particu- cumstances change-as happened, for example, during larly for less-skilled workers. Protectionism in the United the Mexican peso crisis in 1995. Worse still, to what States is fueled by the widely held perception that free extent is trade protection against non-member countries trade hurts U.S. wages, particularly unskilled wages. There higher precisely because NAFTA restricts the application is no doubt that real wages for men in the bottom 20 per- of higher barriers toward the NAFTA partners? In order to cent of the wage distribution fell and wage inequality rose, gauge whether NAFTA is really consistent with freer particularly during the 1980s. The question is how much trade on a global scale, it would be useful to monitor and of this trend is attributable to freer trade with developing estimate the impact of changes in trade barriers in mem- countries, and with Mexico in particular. Though ber countries vis-a-vis non-members. NAFTA's actual effect on wages has not been estimated, 126 NAFTA the impact of trade on U.S. wages has been studied exten- repeated concern expressed by U.S. labor leaders is that, sively. The majority of these studies conclude that interna- given the characteristics of Mexico's political system and tional trade explains a relatively small share-about 10 the corporatist role played by the official unions, Mexican percent-of the observed rising wage inequality and workers still find it hard to form and belong to indepen- downward trend in real wages of less-skilled U.S. workers.4 dent unions. This appears to have been the case in several Other studies find the impact to be higher.5 In general, of the complaints brought before the U.S. National though, most studies suggest that technological change- Administrative Office (NAO). But the recommendations with its bias in favor of high-skilled labor-is the main of NAO have been overcautious with this kind of labor- factor behindl the observed trends. rights violations.6 Even if NAO's formal recommendations Even if trade explains a small fraction of the rise in wage on these matters are not the equivalent of rulings and do inequality and deterioration of less-skilled workers' wages, not lead to sanctions under the current terms of the side it is not a negligible one. Protectionism thus becomes an agreement (such as fines imposed on the country where the instrument that can produce tangible benefits, even if it is violations occurred), greater exposure of such violations at the expense of higher future growth, and savvy politi- could eventually lead to a reduction in them. cians know that well. For those who do not share in the ben- On the environmental front, the North American Devel- efits of higher growth, forgoing higher incomes today opment Bank (NADBank) could become an effective tool to makes no sense. That is why identifying ways to address the foster environmental protection and cleanup, especially rising economnic polarization and declining living standards along the U.S.-Mexico border. The bank's capital will soon for the less-skilled deserves greater attention on the part of reach US$2.25 billion. However, as Jeff Schott notes, since policymakers. Otherwise, many lawmakers and politicians its establishment the NADBank has approved only a hand- are likely to continue focusing on the wrong instrument- ful of NADBank loans that together total less than US$5 that is, trade protection-to combat that issue. Likewise, million. One factor behind this low level of activity is that the insistence to include labor standards as part of trade the established procedure, in which projects are certified by agreements will stick. As is true elsewhere in the region, a commission (the Border Environment Cooperation Com- the United States will have to make a major effort to mission) before they are analyzed for their financial viability upgrade the education and skills of the less privileged sec- by NADBank, may be flawed. Switching the order in which tors of the working population. This might be a solution, projects are analyzed and approved may expedite loans. but not in the short-run. In the short-run other mechanisms Another fundamental reason for this low level of lend- to compensate the losers must be implemented. ing is that the NADBank's charter constrains it to operate on strictly commercial terms. Many of the potential bor- Labor and Environment rowers, particularly in Mexico, lack the administrative The NAFTA side agreements on labor and environment experience and the ability to borrow funds or are simply were introduced to gain votes in the U.S. Congress at the ineligible. Projects with high social payoffs, thus, will need time of NAFTA's approval. They imply obligations for sig- some form of subsidization, such as governmental nonre- natory countries to enforce their own laws. However, orga- fundable grants or private donations. Furthermore, NAD- nized labor, many Democrats in Congress and some envi- Bank may eventually need a window of credit on conces- ronmentalists were not satisfied because, in their words, the sional terms for the poorest and most vulnerable agreements lack teeth-that is, there were no real provisions communities. Its portfolio of eligible projects should be to use trade sanctions against violators of the side accords. expanded eventually to include housing, education and The lack of provisions to use trade sanctions within the poverty reduction projects more generally and find ways to agreements is a blessing. Otherwise, environmental and cofinance such projects with the other multilateral devel- labor issues could have been used frequently as excuses to opment banks. introduce protectionist barriers and thereby obstruct NAFTA's irmplementation. Still, more could and should be NAFTA and Open Regionalism done, on both the labor and environmental fronts, by the Essentially, for an agreement to be consistent with the con- governments and NAFTA-created institutions. One cept of "open regionalism," at least two conditions have to 127 TRADE. TOWARDS OPEN REGIONALISM be fulfilled. First, trade barriers against non-member coun- tinue reducing their tariff structure for most-favored- tries should not be higher that those existing before the nation status, should not raise trade barriers against non- agreement. Second, non-member countries that are willing members and should move toward a common external tar- to abide by the agreement's rules should be accepted. How iff to eliminate the conundrum and potential hidden is NAFTA performing on those two counts? protection found in the rules of origin. They should also In a previous study, Jeff Schott, in conjunction with accept the incorporation of willing new members such as Gary Hufbauer, argued that NAFTA rules of origin Chile into NAFTA. implied a higher level of protection with autos and with The political resistance in the United States to NAFTA textiles and apparel. It is important to assess the impact of expansion or NAFTA-like agreements reveals some weak these higher restrictions on non-member countries. In the links in the economic integration agenda. Three main case of textiles and apparel, according to Jeff Schott, the issues require more attention. First, there is the issue of the losers seem to be in Asia. However, the impact on the social costs caused by greater integration and the need to Caribbean Basin Initiative countries appears to be less find adequate mechanisms to compensate the losers and cLear. But this is not what we hear from the governments make the process a socially inclusive one. Second, countries in Central America and the Caribbean, and a more com- should explore how much macroeconomic consultation and plete quantification of the impact of NAFTA provisions on coordination is feasible and desirable so they can avoid this region-including its impact on FDI-is in order. wide exchange-rate movements and sharp crises. Finally, Likewise, the already mentioned changes in trade policy we are in need of recommendations of how, if at all, labor introduced after NAFTA came into effect (such as the and environmental issues should be incorporated into the increase in tariffs observed in Mexico) should be monitored discussion and implementation of trading agreements. and their impact on non-member countries carefully assessed. Notes In terms of allowing new countries to become members, 1. These are Mexican figures and are different from those pre- NAFTAs track-record so far is-to put it mildly-rather sented by Schott. U.S. and Mexican figures differ primarily because of the treatment of maquiladoras. poor. The most striking example is the non-incorporation ofteramntfmquldas 2. Banco de Mexico (1996), The Mexican Economy. 1996: Economic of Chile, a country that has repeatedly manifested its readi- and Financial Developments in 1995, Policies for 1996 (Mexico City), ness to join NAFTA.7 The stumbling block is the United p. 94. States, because the administration has shied away from pre- 3. Hinojosa Ojeda, Raul and others (1996), North American Inte- senting the request for fast-track authority to Congress. gration Three Years After NAFTA: A Framework for Tracking, Modeling Hence, the possibility of serious negotiations has been and Internet Accessing the National and Regional Lahor Market Inhpacts (University of California, Los Angeles, School of Public Policy and postponed at best. Social Research, December). A combination of protectionist forces in both parties, 4. See, for example, Gary Burtless (1995). "International Trade together with the administration's fear of alienating the and the Rise in Earnings Inequality," Journal of Economic Literature Democrats who are pushing for the inclusion of environ- Uune), vol. 33, pp. 800-16, and Robert Z. Lawrence (1996), Single mental and labor standards in the expansion of NAFTA World, Divided Nations? (Brookings Institution Press and OECD (or, for that matter, in any subsequent free-trade agree- Development Centre). 5. Edward Learner (1993), "Wage Effects of a U.S.-Mexican Free ments), has impeded moving forward. The Mexican ceo- Trade Agreement," in Peter Garber, ed., The Mexican-U.S. Free Trade sis, with its consequential fall in U.S. exports to Mexico Agreement (Cambridge: MIT Press); Adrian Wood (1994), North- and the need for a U.S.-led financial rescue package, Souzth Trade, Employment and Inequality: Changing Fortunes in a Skill- undoubtedly made selling NAFTA expansion a politi- Driven World(Oxford: Clarendon Press). cally difficult task. Whatever the reasons, the fact of the 6. Lawrence, op. cit. matter is that NAFTA has not been an "open" agreement 7. More recently, Central American and Caribbean countries changed their stance from requesting NAFTA-parity to requesting that they become NAFTA members (or, alternatively, that they sign To turn NAFTA into a genuine building block for a NAFTA-type agreement with the United States), but the United global free trade, the three member countries should con- States has not agreed to move forward so far. 128 V. Mercosur Mercosur: Objectives and Achievements S AM LA I RD _T - HE SOUTHERN COMMON MARKET (MERCADO COMAUN DEL SUR, OR MERCOSUR) IS A NOTABLE example of renewed worldwide interest in regional trade agreements, although these agree- ments have provoked concerns about the possibility of welfare-reducing trade diversion as well as their systemic implications for world trade. Concern about the possible trade-diverting effects of Mercosur (and other Latin American regional integration agreements) stems mainly from the experience of earlier attempts at regional integration. In the 1960s the regional agreements were an extension of the import-substitution industrialization policies being applied in the individual countries under the influence of Raul Prebisch and United Nations Economic Commission for Latin America and the Caribbean (ECLAC). In reviewing the efforts in tlhe 1960s to establish the Latin American Free Trade Association (LAFTA), the Andean Pact and Central Arnerican Common Market (CACM), de Melo and Dhar (1992) argue that there were several rea- sons for failure: First, reductions were not across the board but on a product-by-product basis that resulted in many exceptions; second, high rates of protection were maintained against third countries; and third, there was little scope for efficiency gains through the exploitation of economies of scale. Langhammer and Hiemenz (1991) reached a similar conclusion. Today, the context is very different. In conjunction These rates were reduced in unilateral reforms up to 1991 with macroeconomic (and political) reforms pursued in when the Treaty of Asunci6n was ratified, and then con- the aftermath of the debt crisis of the early 1980s, there tinued to be reduced as part of the process of convergence has been serious import liberalization in most Latin toward Mercosur. By the time of implementation of the American countries. Edwards (1994), Laird (1995) and Common External Tariff (CET) in 1995, external tariffs Rajapatirana (1994) show that throughout Latin America among members averaged some 10.7 percent. Thus, the tariffs have been substantially reduced and rationalized new interest in regional agreements takes place against a and there has been a substantial reduction in the use of backdrop of increased outward orientation, exhibited in non-tariff nmeasures. For example, prior to these reforms unilateral reforms as well as increased membership and Argentina had a tariff average of some 30 percent in 1989, substantive participation in the work of the General Brazil's rate was 51 percent in 1988 and Uruguay's was Agreemenr on Tariffs and Trade (GATT) and now the more than [00 percent in 1978 (GATT 1992a, b and c). World Trade Organization (WTO). Sam Laird is at the World Trade Organization in Geneva. Helpful comments were received from isidoro Hodara, Homi Kharas, Petros Mavroidis, Dermot McAleese, Paul Meo, Manuel Olarreaga, Marcelo Olarreaga, Sheila Page, Dimitris Papageorgiou, Sarath Rajapatirana, Esteban Ropclo, Malcolm Rowat, Raymundo Vald6s, Jorge Vigano and Alan Winters. Thanks are due to Gerard Durand for assistance with the tariff data. 131 TRADE: TOWARDS OPEN REGIONALISM There is little question that intra-Mercosur trade has November 29, 1991. The Treaty of Asunci6n has been for- grown rapidly during the period from the signing of the mally amended once, in the Additional Prorocol of the Treaty of Asunci6n up to the establishment of the common Treaty of Asunci6n, known as the "Protocol of Ouro external tariff at the beginning of 1995. Table 1 shows that Preto," signed on December 17, 1994. This protocol, exports within Mercosur have more than doubled as a share which concerns mainly institutional issues as well as dis- of total exports since 1990, while imports from within the pute-settlement, also confers on Mercosur a distinct inter- group have also expanded sharply as a share of the total. national legal personality. However, the growth of intra-Mercosur trade cannot be The Treaty of Asunci6n foreshadowed the establishment attributed uniquely to trade diversion resulting from the of a common market among the four countries with free creation of Mercosur because, apart from certain years and circulation of goods, services, capital and workers starting partners, trade also has been growing strongly with other on January 1, 1995 (but not all of this ambitious program countries in the same period.' Moreover, one of the more was achieved, as discussed later). The treaty has 25 articles important and fastest growing areas of trade in Mercosur is in six main chapters covering the purposes, principles and the automotive trade between Argentina and Brazil under instruments of Mercosur, the organizational structure, the arrangements that are not yet part of Mercosur per se, but period of application, accession, denunciation (withdrawal) are among the exceptions (more later). In addition, in 1995 and general provisions. In addition, there are annexes cov- imports from outside the region grew faster than intra- ering the trade liberalization program, rules of origin, dis- regional trade. pute-settlement, safeguards (including safeguards against other members of Mercosur), and the establishment of The Objectives and Achievements of Mercosur technical and policy working groups. Safeguards follow the guidelines of Article XIX of the GATT, but these have not Establishment of Mercosur been allowed on intra-regional trade since the beginning of Mercosur was established under the Treaty of Asunci6n, 1995. The treaty set out the broad principles for dispute- signed on March 26, 1991, by the presidents of Argentina, settlement, and various stages and procedures were elabo- Brazil, Paraguay and Uruguay.2 The treaty was subse- rated in the Brasilia Protocol for the Settlement of Dis- quently ratified by all members and entered into force on putes, signed on December 17, 1991; this is maintained in TABLE I External Trade of Mercosur 1985-94 (U.S. millions of dollars and percent) EXPORTS 1986 1987 1988 1989 1991j 1991 1992 1993 199-4 1995 World (US$ m.) 30,549 34,133 44.875 46,550 46,403 45,896 50,467 54,122 61,893 70,029 Intra-Mercosur 8.6 7.4 6.6 8.2 8.9 11.1 14.3 18.5 19.5 20.5 Total extra-Mercosur 91.4 92.6 93.4 91.8 91.1 88.9 85.7 81.5 80.5 79.5 OtherLAIA 7.8 6.8 7.0 7.2 7.1 9.1 10.5 10.3 9.9 9.8 USA 21.7 24.3 21.8 19.7 20.8 17.1 17.1 17.6 17.5 15.2 EEC 15 27.4 27.6 29.9 29.5 31.7 32.2 30.3 26.7 27.0 25.5 Japan 6.2 5.4 5.9 5.8 6.0 6.6 5.4 5.2 4.9 5.1 Rest of World 28.3 2S.4 28.8 29.6 25.6 23.9 22.4 21.8 21.2 24.0 IMPORTS World (US$ m.) 21,726 24,133 23,126 26,056 29,298 34,264 40,632 48,082 62,218 79,859 Intra-Mercosur 12.3 l0.8 13.2 15.1 14.5 15.3 18.4 19.6 19.9 18.1 Total extra-Merrosur 87.7 89.2 86.8 84.9 85.5 84.7 81.6 80.4 80.1 81.9 Other LAIA 7.2 7.6 6.4 7.1 6.9 7.2 7.1 5.9 5 7 6.1 USA 20.3 18.4 18.5 19.9 19.2 21.2 22.1 22.0 22.0 22.1 EEC 15 25.0 25.8 24.7 22.5 22.2 23.1 23.6 23.9 27.4 27.5 Japan 6.7 6.3 6.9 7.0 7.0 6.4 5.5 5.5 4.6 4.7 Rest of World 28.5 31.1 30.3 28.4 30.2 26.8 23.2 23.1 20.4 21.5 Sosra: UNSTAT Comtrade Database. 132 MERCOSUR accordance with Article 43 of the Protocol of Ouro Preto. (iii) the Mercosur Trade Commission; Brazil notified the GATT of the Treaty of Asunci6n on (iv) the Joint Parliamentary Commission; February 18, 1992, on behalf of the Secretariat of the Latin (v) the Economic and Social Consultative Forum; and American Integration Association (LAIA).3 The notifica- (vi) the Mercosur Administrative Secretariat. tion was macde under the provisions of the Enabling Clause, The role and composition of these bodies, of which the rather than Article XXIV, which contains the main provi- first three (the decisionmaking bodies) are given in hierar- sions concerning territorial application, frontier traffic, chical order, are laid out in the Protocol. The Ministerial- customs unions and free-trade areas. The Enabling Clause, level Council embodies the legal entity of Mercosur and is formally known as the Tokyo Round Decision on "differ- empowered to negotiate and sign agreements on behalf of ential and more favorable treatment, reciprocity and fuller Mercosur with third countries and international organiza- participation. of developing countries," includes a legal tions. Oversight of the management of Mercosur is assured cover for preferential-trade agreements between develop- by meetings of the Common Market Group, the main ing countries, subject to certain conditions, including executive body, every three months. It comprises senior transparency. The essential provisions are that any such officials (four representatives and four alternates from each arrangement should not raise barriers or create undue dif- country) who must include representatives of the Min- ficulties for other contracting parties (i.e., countries or istries of Foreign Affairs, the Ministries of Economy (or trading entities that are signatories to the GATT); that no equivalent) and the Central Banks. Meetings of Ministers criteria are specified for judging the acceptable degree of of Economy and Central Bank Governors constitute the mutual rcduction or elimination of tariffs;4 and that the institutional framework for the exchange and analysis of reduction of non-tariff measures (NTMs) are to be gov- macroeconomic policies. erned by "criteria which may be prescribed by the Con- The Trade Commission has a number of technical work- tracting Parties" (Paragraph 2(c)). On the other hand, ing committees and is responsible for coordinating com- unlike Article XXIV, there is no specification that sub- mon trade policy and implementing the common external stantially all trade is to be covered, but already some 95 tariff. The Joint Parliamentary Commission and the Eco- percent of in,tra-regional trade was duty-free by the end of nomic and Social Consultative Forum are both consultative 1994 (WTC) 1996b, p. 25). bodies. The Administrative Secretariat, which is quite The only significant change in procedures for examining small and based in Montevideo, is to provide operational a regional trade agreement resulting from the creation of support for Mercosur, but it is not intended to be a strong the WTO stems from the requirement of the Agreement policy- or rule-making supranational authority along the on Rules of Origin that a detailed work program for the lines of the European Commission.6 future be established. However, within the WTO, the rel- Essentially, Mercosur is an international treaty subscribed evant working party of the Committee on Trade and by the member states, and the bodies established under the Development, which was established to examine the treaty are inter-governmental, rather than supranational. GATT/WTO consistency of the treaty, did not complete Below the level of the treaty, implementing decisions, reso- its work, and the agreement is now being examined by the lutions and directives (in order of importance) are deter- Committee on Regional Trade Agreements, established in mined, respectively, at the level of the council, the Common February 1996, in the light of the provisions of GATT Market Group and the Trade Commission, but have no force 1994, includling Article XXIV.5 by themselves and need to be implemented by correspond- ing national measures.7 Unlike the Council of Ministers or Institutioncal Structure the Commission of the European Communities, the Merco- Under Article 1 of the Protocol of Ouro Preto, the institu- sur common bodies do not have powers to oblige a member tional structure of Mercosur was established as follows: state to comply with common market rules. (i) the Council of the Common Market (consisting of There is no supranational court through which either a Ministers of Foreign Affairs and Ministers of member state or the Secretariat can enforce treaty obliga- Economy); tions on another member or a private party. There is provi- (ii) the Common Market Group; sion for arbitration under the Brasilia Protocol (1991), but 133 TRADE TOWARDS OPEN REGIONALISM trade disputes are typically resolved by negotiation. By the Overall, the various commissions and committees have beginning of 1997 only one dispute had been sent to a spe- carried out an impressive amount of work, but the weak- cially constituted tribunal or expert panel under the ness in the central institutional structure appears to consti- Brasilia Protocol, and this was settled bilaterally (i.e., "out tute a particular disadvantage for the smaller members, of court") in early April 1997.8 It is clear that the authori- leaving them vulnerable to political pressures. Moreover, ties are still moving forward cautiously in this area. Bouzas (1996) has observed that, while skepticism toward The Common Market Group has two committees (on supranational agencies is comprehensible in the light of rules-norrnalizacidn-and sanitary and phytosanitary mea- earlier experience in Latin America, several issues, such as sures). It is also advised by a number of working groups, or investment and services (discussed further below), need to " subgroups," which examine certain issues relating to the be handled at a supranational level if the union is to integration process. Up to 1995, sub-groups worked under advance and be deepened. the following broad headings: trade issues; customs issues; technical standards; fiscal and monetary matters related to Border Measures and Procedures trade, inland transport; maritime transport; industrial and Article 5 of the treaty sets out the agreed liberalization technological policy; agricultural policy; energy; coordina- program, which was to consist of "progressive, linear and tion of macroeconomic policies; and labor policy. These automatic tariff reductions accompanied across the board sub-groups carry out programmed examination of issues on by the elimination of non-tariff restrictions or equivalent which they make recommendations to the Common Mar- measures ... with a view to arriving at a zero tariff and no ket Group for consideration and implementation (see non-tariff restrictions for the entire tariff area by 31 GATT 1994). In 1995, under Resolution No. 20/95, a new December 1994." After an initial tariff reduction of 47 structure was established, consisting of 10 subgroups: percent in applied rates that followed the ratification of the Subgroup 1: Communications; Treaty of Asunci6n in 1991, individual states generally Subgroup 2: Mining; realigned tariffs on schedule every six months in equal Subgroup 3: Technical Rules; stages up to the implementation of the Common External Subgroup 4: Financial Matters; Tariff (CET) at the beginning of 1995 (except where Subgroup 5: Transport and Infrastructure; increases were implicd). Subgroup 6: Environment; The CET is applied to imports from partners subject to Subgroup 7: Industry; most-favored-nation (MFN) rates. Whereas only the broad- Subgroup 8: Agriculture; est structure was set in 1991 (in order to determine the Subgroup 9: Energy; and extent of the scheduled cuts), finalization was the subject of Subgroup 10: Labor and Social Security. intense internal negotiations that were concluded in late A new Subgroup 11 (Health) was created by the Com- 1994. It was then implemented, with exception lists for mon Market Council at a meeting in Fortaleza in Decem- each Mercosur member, on January 1, 1995. The CET, ber 1966. There are also specialized meetings on science based on the Harmonized Commodity Classification and and technology and tourism, as well as ad hoc groups on Coding System (HS), consists entirely of ad valorem rates, services, institutional matters, Mercosur-LAIA, Mercosur- charged on the CIF (cost, insurance and freight) value of WTO and sugar. the imports. Table 2 gives an overview of the structure in The Mercosur Trade Commission is advised by a Com- 1995, by main sections of the International Standard mittee on the Defense of Competition as well as 10 Tech- Industrial Classification (ISIC), together with the rates to nical Committees: (i) Tariffs, Nomenclature and the Clas- be applied when the exceptions are eliminated by the end sification of Goods; (ii) Customs Matters; (iii) Rules and of the transition periods (up to 2001 for Argentina and Trade Disciplines; (iv) Public Policies that Distort Com- Brazil and up to 2006 for Paraguay and Uruguay). petitiveness; (v) Defense of Competition; (vi) Unfair Prac- The scheduled MFN rates, which individual members tices and Safeguards; (vii) Consumer Protection; (viii) apply as exceptions to the CET in the transition period, Non-Tariff Restrictions and Measures; (ix) the Automotive include sensitive items, and are intended to facilitate struc- Sector; and (x) the Textile Sector, tural adjustment, helping to place the sectors involved in a 134 MERCOSUR competitive position within the region at the end of the Interpretation of Article XXIV of the GATT 1994 states period in question; negotiations are also being conducted that "the general incidence of the duties and other regula- 'with a view to harmonizing public policy in various areas tions of commerce applicable before and after the forma- and establishing a trade regime that will ensure fair com- tion of a customs union shall in respect of duties and petition" (WTO 1995a). Thus, while the CET for capital charges be based upon an overall assessment of weighted goods has been set at 14 percent, Argentine and Brazilian average tariff rates and of customs duties collected."1" tariffs for these products are allowed to converge to that Prima facie, the CET would seem to meet the test of the rate in a linear and automatic manner over the period to GATT: Thus, although Paraguay's average tariff was lower January 2001, while Uruguay and Paraguay have until than under Mercosur, its weight in trade was quite small, January 2006 to achieve convergence. For telecommunica- while the large members previously had higher average tions and inlormation technology equipment the CET is rates. Thus, when the Treaty of Asunci6n was signed at the fixed at 16 percent with convergence by all members by end of 1991, Argentina's average rate was 12.2 percent and 2006. Some 1140 eight-digit tariff items are covered by Uruguay had a 21.5 percent global rate, while Brazil's rate the capital goods exemptions and some 435 in telecommu- was 21.2 percent inJanuary 1992 (GATT 1992a, b and c). nications ancl informatics. However, these are simple average MFN rates, and they do Initially, it was agreed that Argentina, Brazil and not take account of concessional entry.12 Uruguay woald also have the right to have 300 national It is important to note that the applied tariff levels, as exceptions to the CET, while Paraguay would have the shown in Table 2, are considerably lower than the levels right to 39c (WTO 1995a). Items could be eliminated that are legally bound in the WTO. Thus the tariff bind- from these lists, but no new items could placed on the lists, ings, made by the individual Mercosur countries (not the creating a ratcheting down effect. These exceptions could CET on behalf of the customs union, unlike the EU's CET), be higher or lower than the CET, so that for convergence to are mostly at a ceiling level of 35 percent, with the main the CET, upward and downward modifications to national exceptions being tariffed agricultural import barriers. tariffs of the four members are required. For example, in These are allowed to be implemented over 10 years, 1995 the Mercosur countries reported to the WTO that although Argentina implemented these rates fully in 1995. Argentina, Brazil, Paraguay and Uruguay would increase The gap between the applied and bound rates provides a the rates on 84, 123, 214 and 212 items in the Mercosur wide margin for tariff increases without having to renego- common tariff schedule, respectively, while decreasing the tiate concessions under Article XXVIII of the GATT, and rates on 147, 52, 0 and 6, respectively (WTO 1995a).9 must inevitably lead to some uncertainty in the tariff Apart from these items where the national scheduled regime. The fact that Brazil, in June 1995, increased tariffs rates diverge from the CET, each member has its own list to 70 percent for a number of products (in addition to of special concessionary regimes where rates may be introducing quotas on automotive imports) confirms this reduced below scheduled rares-e.g., on investment irems, sense of uncerrainty, albeit within the binding commit- to assure basic supplies of primary products and inputs for ment for 1995 (to be reduced progressively over 10 years to other industries, temporary admission of goods to be re- 35 percent). However, security of access to Mercosur was exported, to allow duty-free entry of goods from the free improved by the substantial extension of the coverage of zones of Marnaus and Tierra del Fuego,"' and so on. These the tariff bindings to all trade, compared with a situation regimes are to be consolidated into a common set of Mer- prior to the round where binding coverage was only 4 per- cosur regimes to replace the national concessions, but a cent of items in the case of Argentina and Uruguay and 6 timetable has yet to be determined. percent for Brazil.'3 Mercosur countries maintain that the CET has been Mercosur's CET is characterized by tariff escalation: Tar- fixed in conformity with Paragraph 5 of Article XXIV of iff protection for raw materials (first stage of processing) was the GATT, which requires that "the duties ... in respect of on average 6.3 percent in 1995, 9.1 percent for semi-man- trade with [non members}... shall not on the whole be ufactures and goods used as inputs for other production higher than.. .the duties... prior to the formation of [the] chains, and 12.5 percent for fully processed goods (see Table union." Elaborating on this, the Understanding on the 3). Escalation was most marked in Brazil and least pro- 135 TRADE TOWARDS OPEN REGIONALISM TABLE 2 Mercosur Tariff Structure, 1995 and Final CEr (2001 or 2006) (percent) ISIC PRODUCT AND CODE PROCESSING DESCRIPTION ARGENTINA BRAZIL PARAGUAY URUGUAY AVERAGE FINAL Total 10.5 11.9 9.4 10.8 10.7 11.2 I Agriculture, hunting, forestry & fishing 7.0 7.0 6.9 6.9 7.0 7.0 111 Agricultural and livestock production 7.0 6.9 6.8 6.8 6.9 7.0 12 Forestry and logging 4.6 4.7 4.4 4.6 4.6 4.6 121 Forestry 5.8 6.0 5.5 5.8 5.8 5.8 l22 Logging 2.0 2.0 2.0 2.0 2.0 2.0 130 Fishing 8.7 8.7 8.7 8.7 8.7 8.7 2 Mining & quarrying 3.4 3.6 3.4 3.4 3.5 3.4 210 Coal mining 0.0 0.0 0.0 0.0 0.0 0.0 220 Crude petroleum and natural gas 0.0 6.8 0.0 0.0 1.7 0.0 230 Metal ore mining 2.2 2.2 2.2 2.2 2.2 2.2 290 Other mining 4.1 4.1 4.1 4.1 4.1 4.1 3 Manufacturing 10.8 12.3 9.6 11.1 11.0 11.5 31 Food,beveragesandrobacco 11.6 11.7 11.5 11.7 11.6 11.6 311 Food products 11.0 11.2 10.9 11.2 11.1 11.0 312 Other food products and animal feeds 11.8 11.8 11.3 11.8 11.7 11.8 313 Beverages L8.1 17.3 18.6 17.6 17.9 18.6 314 Tobacco manufacturing 18.6 18.6 18.6 18.6 18.6 18.6 32 Textile, wearing apparel and learher 17.2 16.9 16.9 16.9 17.0 17.1 321 Textiles 16.8 16.7 16.6 16.7 16.7 16.9 322 Manufacture of wearing apparel 19.9 19.9 19.9 19.9 19.9 19.9 323 Learher products 13.2 12.4 13.2 12.8 12.9 13.2 324 Manufacture of footwear 24.6 19.4 18.8 19.4 20.6 19.4 33 Wood and wood products, inc. furniture 10.8 10.2 10.5 10.5 10.5 10.5 331 Wood and wood producrs, exc. furniture 8.1 7.7 8.1 8.1 8.0 8.1 332 Manuf. of furniture & fixtures exc. metal 18.9 18.0 18.0 18.0 18.2 18.0 34 Paper, paper prods, printing & publishing 11.7 10.7 10.7 10.4 10.9 10.9 341 Paper products 12.1 11.0 11.1 10.7 11.2 11.3 342 Printing, publishing & allied industries 10.1 9.4 9.4 9.0 9.5 9.4 35 Chemicals, perrol, coal, rubber, plastics 7.9 8.2 7.7 7.2 7.8 8.1 351 Industrial chemicals 7.2 7.5 7.1 6.4 7.1 7.5 352 Other chemicals, incl. pharm. 8.6 8.7 8.1 8.3 8.4 8.8 353 Petroleum refineries 0.9 12.7 0.9 0.9 3.9 0.9 354 Manuf. of misc. petrol. & coal prods. 2.0 3.2 2.0 2.0 2.3 2.0 355 Rubber products 16.1 15.4 15.4 15.4 15.6 15.4 356 Manufacture ofplasric products n.e.s. 17.4 17.4 17.4 16.8 17.3 17.4 36 Non-metal mineral prods exc. petrol & coal 10.9 10.5 10.6 10.9 10.7 10.9 361 Pottery and china 15.4 14.8 14.3 15.4 15.0 15.4 362 Manufacture of glass and glass produces 12.6 12.6 12.4 12.6 12.6 12.6 369 Other non-metallic mineral products 9.2 8.6 9.0 9.2 9.0 9.2 37 Basic metal industries 10.9 9.9 9.6 9.3 9.9 9.9 371 Iron and steel basic industries 12.6 11.4 11.0 10.5 11.4 11.4 372 Non-ferrous metal basic industries 8.7 7.9 7.8 7.7 8.0 8.0 38 Fabricated metal prods, mach. & equip. 10.9 15.8 8.0 13.2 12.0 13.3 381 Fabricated metal products 16.0 16.4 14.7 16.0 15.8 16.1 382 Non-electrical machinery incl. computers 7.7 16.8 4.5 12.4 10.4 12.5 383 Electrical machinery 12.1 14.8 9.2 12.6 12.2 12.7 384 Transport equipment 13.5 15.2 11.5 14.6 13.7 14.6 385 Professional and scientific equipment 12.2 14.8 8.8 13.3 12.3 13.4 390 Other manufacturing industries 16.8 16.6 15.8 16.6 16.5 16.6 Notec CET to be completed by Argenrtina and Brazil by 2001 and by Paraguay and Uruguay by 2006. Source: WTO Secretariat calculaticns, based en data supplied by Mercosur 136 MERCOSUR TABLE 3 Tariff Escalation and Tariff Ranges, 1995 and Final CET (2001 or 2006) (percent) ISIC PRODUCT AND CODE PROCESSING DESCRIPTION ARGENTINA BRAZIL PARAGUAY URUGUAY AVERAGE FINAL Total -1st stage of processing 6.3 6.4 6.2 6.1 6.3 6.3 -semi-processed 9.3 9.3 9.0 8.6 9.1 9.4 -fully processed 12.0 14.7 10.2 13.1 12.5 13.3 1 Agriculture -raw materials 7.0 7.0 6.9 6.9 7.0 7.0 2 Mining and quarrying -raw materials 3.4 3.6 3.4 3.4 3.5 3.4 311 Food products -1st stage of processing 7.0 7.0 7.0 7.0 7.0 7.0 -semi-processed 10.3 10.3 10.1 10.3 10.3 10.3 fuoly processed 12.4 12.7 12.3 12.7 12.5 12.3 312 Food manufacturing -Ist stage ofprocessing 7.0 7.0 7.0 8.2 7.3 7.0 semi-processed 16.0 16.0 14.4 16.0 15.6 16.0 -fully processed 13.6 13.6 13.1 13.0 13.3 13.6 313 Beverages -fully processed 18.1 17.3 18.6 17.6 17.9 18.6 314 Tobacco manufactures -fully processed 18.6 18.6 18.6 18.6 18.6 18.6 321 Textiles -1st stage of processing 8.3 6.7 8.3 8.3 7.9 8.3 -semi-processed 16.5 16.5 16.2 16.3 16.4 16.6 -fully processed 18.7 18.7 18.6 18.6 18.7 18.7 322 Clothing -fully processed 19.9 19.9 19.9 19.9 19.9 19.9 323 Leather products -1st stage of processing 10.0 10.0 10.0 10.0 10.0 10.0 -semi-processed 9.7 8.6 9.7 9.2 9.3 9.7 -fully processed 20.0 20.0 20.0 20.0 20.0 20.0 324 Footwear -fully processed 24.6 19.4 18.8 19.4 20.6 19.4 331 Wood products -1st sta,,e of processing 2.0 2.0 2.0 2.0 2.0 2.0 -semi-processed 6.6 5.8 6.6 6.6 6.4 6.6 -fully processed 11.7 11.7 11.7 11.7 11.7 11.7 332 Furniture except metal -fully processed 18.9 18.0 18.0 18.0 18.2 18.0 341 Paper pr-oducts -lst stage ofprocessing 3.6 3.6 3.6 3.6 3.6 3.6 -semi-processed 12.6 11.6 11.4 10.8 11.6 11.6 fully processed 15.5 13.5 14.2 14.2 14.4 14.5 342 Printing -fully processed 10.1 9.4 9.4 9.0 9.5 9.4 351 Industrial chemicals -iscstage of processing 8.6 9.2 8.6 6.0 8.1 8.6 -semi-processed 7.1 7.3 7.0 6.4 7.0 7.4 -fully processed 11.3 11.1 9.3 7.5 9.8 11.3 352 Other chemicals 0.0 -1st stage of processing 11.0 8.7 8.7 8.7 9.3 8.7 -semi-processed 7.8 8.2 7.6 7.4 7.8 8.2 -fully processed 8.9 8.9 8.3 8.6 8.7 9.1 353 Petroleum refineries -1st stage ofprocessing 0.0 12.1 0.0 0.0 3.0 0.0 -semi-processed 3.2 13.4 3.2 3.2 5.8 3.2 -fully processed 0.6 12.7 0.6 0.6 3.6 0.6 137 TRADE: TOWARDS OPEN REGIONALISM TABLE 3 (CONTINUED) Tariff Escalation and Tariff Ranges, 1995 and Final CET (2001 or 20061 (percent) ISIC PRODUCT AND CODE PROCESSING DESCRIPTION ARGENTINA BRAZIL PARAGUAY URUGUAY AVERAGE FINAL 354 Petroleum and coal products -1st stage of processing 0.0 0.0 0.0 0.0 0.0 0.0 -semi-processed 2.0 4.8 2.0 2.0 2.7 2.0 -fully processed 8.0 8.0 8.0 8.0 8.() 8.0 355 Rubber products -1st stage of processing 12.0 12.0 12.0 12.0 12.0 12.0 -semi-processed 14.0 14.0 14.0 14.0 14.0 14.0 -fully processed 17.1 16.1 16.1 16.1 16.4 16.1 356 Plastic products -fully processed 17.4 17.4 17.4 16.8 17.3 17.4 361 Pottery and china -fully processed 15.4 14.8 14.3 15.4 15.0 15.4 362 Glass and products -semi-processed 8.1 7.9 8.1 8.1 8.1 8.1 -fully processed 14.0 14.0 13.8 14.0 14.0 14.0 369 Non-metallic mineral products 1st stage of processing 6.0 6.0 6.0 6.0 6.0 6.0 -semi-processed 6.9 6.6 6.9 6.9 6.8 6.9 -fully processed 9.8 9.1 9.5 9.8 9.6 9.8 371 Iron and steel products -1st stage of processing 0.9 0.9 0.9 0.9 0.9 0.9 -semi-processed 13.0 11.8 11.4 10.9 11.8 11.8 372 Non-ferrous metal -1st stage of processing 3.8 3.8 3.8 3.8 3.8 3.8 -semi-processed 8.9 8.1 7.9 7.8 8.2 8.2 -fully processed 16.0 16.0 16.0 16.0 16.0 16.0 381 Metal products semi-processed 14.0 14.0 14.0 14.0 11.0 1/4.0 -fully processed 16.0 16.4 14.7 16.1 15.8 16.1 382 Non-electrical machinery -semi-processed 12.0 12.0 12.0 12.0 12.0 12.0 -fully processed 7.7 16.8 4.5 12.4 10.4 12.5 383 Electrical machinery -fully processed 12.1 14.8 9.2 12.6 12.2 12.7 384 Transport equipment -fully processed 13.5 15.2 11.5 14.6 13.7 14.6 385 Professional and scientific equipment -fully processed 12.2 14.8 8.8 13.3 12.3 13.4 390 Other manufactured products Ist stage of processing 8.6 8.6 8.0 8.6 8.5 8.6 semi-processed 13.5 13.5 10.0 13.5 12.6 13.5 -fully processed 17.5 17.3 165 173 17.2 17.3 Note: CET io be completed by Argentina and Brazil by 2001 and by Paraguay and Uruguay by 2006. Sosuer: WTO Secretariat calculations based on dara suppled bv Mercosur. nounced in Paraguay in 1995 (the first three lines of Table be obtained from a special study carried out in Brazil 3). For the CET as it will be applied after full implementa- (Kume 1996), showing the current level of tariff protection, tion, escalation will be on average slightly more pronounced as well as the level that is expected to prevail following full than in 1995. Tariffescalation exists in practically all sectors, implementation of the CET in 2001 or 2006 (see Table 4). with petroleum refineries being the sole exception. This latter rate is therefore representative of the effective The presence of tariff escalation means that processing protection that will be available throughout Mercosur. industries benefit from higher levels of protection on their The highest effective protection is accorded to passenger value-added than is evident from the nominal tariffs alone. cars, trucks and buses, reflecting higher nominal tariffs on An indication of the levels of effective tariff protection may motor vehicles and the relatively low value added in the 138 MERCOSUR TABLE 4 with the CET than scheduled. However, in the transition Effective Rates of Protection by Activity in Brazil, 1993.2006 period modifications are being allowed. Thus, in 1995 (percent) Brazil was permitted to exempt from the CET a further list JULY DEC. DEC. CET of up to 150 additional items as part of its stabilization 1993 1994 1995 2006 program for price-control and domestic-supply considera- Agriculture 3.9 3.8 4.4 4.4 Non-fuel mining 0.7 0.9 2.5 2.7 tions. The products maily covered food items such as Petroleum & coal mining -2.0 -1.9 -1.9 -1.7 meat, cheese, rice and barley, but also covered steel prod- NMetallic mineraL products 13.0 11.3 14.3 14.5 ucts and consumer durables, including cars, on which rates Iron and steel 10.7 10.2 12.9 13.8 Non-ferrous metallurgy 6.2 7.6 10.4 10.4 were increased to 70 percent. In early 1996, the list was Other metal products 22.7 19.4 21.3 21.2 reduced and amended, with butter, oil and wine being Electricalnmaterial 2261 26.23 231038 204 added to preclude "undesirable" imports (WTO 1996b). Electronic equipment 23.1 21.3 24.9 13.0 Tariffs on textiles were retained at 70 percent and increased Cars, trucks, buses 129.8 44.6 270.9 53.1 on wine from 20 to 35 percent. When the additional list Other vehicles and parcs 21.3 21.6 21.0 14.4 Wood and furniture 9.7 9.4 12.3 12.4 expired in 1996 safeguard measures were introduced on Cellulose, paper, etc. 8.5 8.0 10.5 12.6 textiles (under the provisions of the WTO Agreement on Rubber products 17.3 15.3 14.6 14.7 Textiles and Clothing), and tariffs were increased from 20 Chemicals 12.8 9.0 5.2 16.1 Petroleum refining 10.5 5.2 8.0 8.8 percent to 70 percent on toys (under the main WTO safe- Various chemical prods. 9.0 5.0 5.8 6.0 guards provisions). The same procedure for modifications Pharmaceutical; & perf. 13.9 2.3 9.8 9.9 Plastic product; 21.5 24.4 23.3 22.3 of the national rate would have provided legal cover for the Textiles 20.6 20.2 23.6 21.5 periodic reintroduction and modification of the Argentine Clothing 23.9 24.5 21.0 22.6 statistical tax (now bound at 3 percent under the WTO in Foorwearand leather 15.1 15.6 21.3 15.8 Coffee industry 12.7 10.1 11.8 11.8 respect of "other duties and charges"); this was used in the Vegetable produ-cts 18.0 17.1 23.2 20.7 past like an import surcharge for fiscal reasons, while pre- Animal slaughter 10.0 7.1 9.7 9.8 Dairy industry 21.6 25.7 24.2 16.5 serving the psychologically important nominal anchor of Sugar 21.2 9.5 16.5 16.8 the national currency to the U.S. dollar.'5 The increases in Vegetable oils 7.6 8.0 8.8 9.3 tariffs and charges by Brazil and Argentina were within the Beverages & food prods 29.7 22.2 23.9 25.1 Other products 23.6 21.0 20.6 19.9 overall binding commitments at the time. It is important to note that there remain a number of Prod, weighted average 14.5 12.3 12.9 15.4 Median 15.1 11.3 14.6 14.4 exceptions to duty-free treatment on trade within Merco- Minimum -2.0 -1.9 -1.9 -1.7 sur. Thus, when the CET was introduced in January 1995, Maximum 129.8 44.6 270.9 53.1 Standard deviation 21.7 9.7 45.9 9.2 the members also agreed on a Regime of Final Adjustment (RWgimen de adecua.ci6n) to the Customs Union, which allows Source: Kume (1996). exceptions to internal duty-free trade. For items covered by industry.'4 Other activities benefiting from higher than this regime, there are to be tariff reductions on intra-group average levels of effective protection include electrical mare- trade in a linear and automatic manner until a zero rate is rials, electronic equipment, the dairy industry, beverages reached. For Argentina and Brazil, the reductions began on and food products, textiles and plastic products. In contrast, January 1, 1995, and are to be completed by December 31, activities wvith effective rates considerably below the aver- 1998; for Paraguay and Uruguay, the reductions began on age, notably non-fuel mining, agriculture and chemicals, January 1, 1996, and are to be completed by December 31, tend to be relatively disadvantaged by the tariff structure; 1999. The main list of goods covered is not extensive, and the negative effective protection estimated for petroleum includes sensitive agricultural and industrial products; it is and coal mining indicates that these activities are even more estimated by Mercosur countries as covering some 5 per- disadvantaged by the existing tariff structure. cent of intra-Mercosur trade. However, this does not take The CET rates may not be modified without the consent account of the exceptions made for sugar or automobiles; of all Mercosur participants, other than to bring rates on the latter is of particular importance in trade between items in the exceptions lists more rapidly into alignment Argentina and Brazil. 139 TRADE: TOWARDS OPEN REGIONALISM Turning to non-tariff restrictions on external trade, a Decision No. 18/96 to continue this situation until distinction is made between those that have to be elimi- December 31, 2000. nated and those that have to be harmonized, such as mea- Concerning procedural customs issues, the meeting of sures relating to plant and animal health, technical stan- the Council and the Common Market Group at Ouro Preto dards, environmental protection and safety This will also on December 16-17, 1994, adopted a series of decisions allow for a common Mercosur policy consistent with WTO and resolutions that, as well as bringing the common provisions; Annex II of the Treaty of Asuncion reproduces external tariff (with the noted exceptions) into effect, the Resolution of the Common Market Group establishing included a number of common trade policy measures nec- the regulatory framework for the elimiination of non-tariff essary for its implementation-for example, a commnoln restrictions and harmonization of measures of a non-tariff system of rules of origin, regulations against unfair trade nature. Under the trade liberalization program of the practices by third countries, a Mercosur customs code and Treaty of Asunci6n (Annex I), non-tariff measures were to a series of harmonized customs regulations. However, the be eliminated on internal trade under the LAIA framework customs code was only ratified by Paraguay and is currently (Economic Complementarity Agreement No. 18). How- being renegotiated. The possibility of integrated customs ever, while a number of non-tariff measures have been controls at the internal frontiers has been discussed, but eliminated, the elimination of others still requires legisla- there is no question of their elimination, even after duties tive approval at the national level. (See also discussion of and other restrictions on internal trade are eliminated. This sectoral policies below). is because duties and domestic indirect taxes will be col- Annex IV of the Treaty for the application of safeguards lected by the member state which is the final destination of follows the guidelines laid down in Article XIX of the imports from third countries, while indirect taxes will also GATT (Emergency Action on Imports of Particular Prod- be collected in imports from other Mercosur countries. ucts), adapting them to the institutional framework of Administrative difficulties have so far limited the Mercosur and the "need to protect the situation of certain implementation of free circulation of goods. Free circula- sectors of domestic industry in some of the States Parties tion would mean that, once imports from third countries to Mercosur" (GATT 1995a). Common Rules on Unfair enter one Mercosur country, they should be able to cross Trade Practices by third countries and Common Rules and internal frontiers into other Mercosur countries simply by Safeguard Measures against third countries were agreed showing that the duty under the CET has been paid at the under Decision No. 17(96 of the Common Market Coun- first port of entry.17 Duties collected in the first country cil in December 1996. Under the decision, a Trade would then be transferred to the authorities in the country Defense and Safeguards Committee was also created, tak- of final destination on the basis of the paperwork provided ing over the relevant functions of the Trade Committee. It by the importer at the port of entry. At present, while free is responsible for safeguards investigations. A safeguard circulation is intended, in practice most goods from third measure may be taken by Mercosur as a group or on behalf countries for transit through one Mercosur country to of a simple member. National action is no longer envis- another travel under international transit arrangements aged in terms of GATT Article XIX, but this may still be (i.e., under customs seal) and pay duty only in the final possible under the special safeguards provisions of the country of destination. (This mainly concerns Paraguay, WTO Agreement on Textiles and Clothing as well as the which is landlocked.) Arrangements for the transfer of Agreement on Agriculture. receipts of customs duties to the final country of destina- Under Article 1 of Annex IV, safeguard actions by one tion still need to be worked out. Thus, shipments that do member state against another have not been allowed since not enter Mercosur under international transit provisions the beginning of January 1995. However, anti-dumping and that pay the duty at the first port of entry could well actions are still allowed under Resolution No. 129/94 of encounter problems at the internal frontier if the shipper the Common Market Group until such time as there is expects not to have to pay duty at that point. agreement on a set of common rules on the defense of inter- Apart from the tax and duty arrangements, and despite nal competition."6 The Council of the Common Market in ongoing efforts to simplify the transit of goods and per- Fortaleza, meeting December 16-17, 1996, decided under sons, there are periodic reports of delays at the frontiers. 140 MERCOSUR The work on rules of origin is intended to lead to har- Committee on Regional Trade Agreements, where Merco- monized rules, since there is no intention of abolishing sur is represented by the presidency pro tempore (which internal frontiers along the lines of the EU's single market. changes every six months). As noted earlier, in the Customs will determine the origin of goods and, hence, the Uruguay Round tariff bindings were made individually, appropriate -level duty to be paid to the fiscal authorities in not as a Mercosur commitment. To the extent that nation- the country of final destination. Duties on goods from ally bound tariffs are currently lower than the CET, mem- third countries that are processed in one Mercosur country bers are to engage in a process of renegotiation of tariff before being re-exported to another will essentially be bindings to align these rates with the CET. Similarly the assessed on the basis of general rules of change of tariff clas- processes of eliminating internal exceptions to duty-free sification and!or 60 percent local content, as well as special trade and convergence to the applied CET will be impor- rules for the capital goods, chemical, telecommunications tant steps toward establishing a common external trade and information-technology sectors. Thus, to a large policy. degree the problems that Krueger (1995) identifies in Currently, common trade rules on anti-dumping have respect of FTAs will not be avoided in Mercosur. been drafted and are being revised to bring them into alignment with WTO rules. In the meantime, members Common Trade Policy are applying their domestic legislation, keeping the Mer- The adoption of a common trade policy toward third coun- cosur Trade Commission informed of any actions. tries is considered by the Mercosur countries to be "an Although no decision has been taken to this effect, when inseparable complement to the implementation of a com- harmonized rules are agreed, it would seem logical that mon external tariff' (WTO 1995a), and Mercosur has these be implemented through the Trade Defense and Safe- defined the main elements of such a policy. While the Ouro guards Committee, as has been agreed for safeguards. Fol- Preto meeting took a number of decision on trade policy lowing that practice, measures could then be imposed in these relate mainly to customs matters, as discussed above, respect of any single member or Mercosur as a whole. and some sectoral questions, discussed later. They do not Article 50 of the Treaty of Montevideo, in accordance refer specifically to trade policy to third countries, with Article XX (General Exceptions) and Article XXI although this is now becoming much clearer. (Security Exceptions) of the GATT, permits LAIA mem- In terms of relations with third countries, Mercosur is bers, including those within the framework of Mercosur, to also intended to be an open regional agreement. Paraguay, apply trade measures for protection of public morality; on behalf of the Mercosur countries, has stated that it is "a human, animal and plant life and health; and national flexible and open process, the opposite of the idea of a security. This also covers regulation of imports and exports 'fortress' reformulating, at the quadripartite level, old iso- of arms, munitions, etc.; imports and exports of gold and lationist concepts" (WTO 1995a). This can mean several silver in bullion form; protection of national treasures of things. On the one hand, Mercosur may take the form of artistic, historical or archaeological value; and the exporta- open regionalism in which liberalization is also applied to tion, use and consumption of nuclear materials, radioactive third countries. On the other hand, the agreement may also products or any other material used for the development be open to membership or to the exchange of concessions and exploitation of nuclear energy. Details are provided in with other countries or regional blocks. To evaluate these Annex I to GATT (1994). possibilities, we look at common external trade policy first In many areas of trade policy, convergence to WTO cri- in the mult ilateral context of the WTO, second in the con- teria, while leaving scope for national variations, will facil- text of the ]LAIA, and third in respect of bilateral relations itate the establishment of a common policy, as well as with other countries and groups of countries. improving integration within Mercosur. The wide sweep of First, in the WTO, despite the goal of establishing a WTO commitments, compared to the GATT, results from common trade policy, the member states of Mercosur con- the obligation under the single undertaking of the round tinue to be represented individually, compared with the of all members of the World Trade Organization to imple- European Union, where the European Commission speaks ment all but the four multilateral accords (the Agreement on behalf olf its members. The exception to this is the new on Trade in Civil Aircraft, the Agreement on Government 141 TRADE: TOWARDS OPEN REGIONALISM Procurement, the International Dairy Agreement and the their Denver meeting in June 1995 and at Cartagena in International Bovine Meat Agreement). Thus, unlike the March 1996, although inevitably the interest of the U.S. Tokyo Round codes in which adherence was voluntary, administration was diverted in 1996, an election year. The members are bound by and cannot opt out of agreements FTAA would extend and consolidate recent autonomous in the areas of anti-dumping, subsidies and countervailing trade reforms in the region, while the inclusion of the duties, customs valuation, etc. Under these agreements United States would also lend considerable credibility to Mercosur countries have, in effect, made commitments to the accord, providing an important guarantee of the out- bring their individual practices into line with the new ward orientation of Mercosur. WTO rules and hence to accept a degree of harmonization Among Mercosur countries, enthusiasm for the FTAA is not apparently yet covered by decisions under the Treaty of not uniform. For example, in a similar context Brazil has Asunci6n. However, the WTO rules on sanitary and phy- argued that joining NAFTA would mean that Latin Amer- tosanitary measures were adopted as Mercosur rules at the ican countries 'would lose their capacity for adopting Fortaleza meeting of the Council of the Common Market. autonomous policies in sensitive areas such as investment, Second, the Latin American Integration Association services and intellectual property."20 Certainly, such an (LAIA) provides the umbrella framework for Mercosur, FTA would also extend liberalization of external barriers even within the WTO, as noted earlier. Indeed, the first beyond the levels currently agreed for Mercosur. In the area priority for Mercosur in relations with third countries is of trade, this includes striving for comprehensive agree- within the LAIA, where other members have also been ments across the hemisphere on tariff and non-tariff barri- undertaking unilateral reforms. In this context, Mercosur ers to trade, agriculture, subsidies, investment, intellectual countries are renegotiating the Economic Complementar- property, rules of origin, anti-dumping duties, sanitary ity Agreements, which individual members have with standards, dispute settlement and competition policy. It other LATA members, to create a single Mercosur agree- also seems likely that the United States would request ment with each partner. In this regard, negotiations are improved investment opportunities in those countries underway to establish an FTA with Mexico (as an LAIA where investment is subject to restrictions. Similarly, if partial-scope agreement) and Panama. NAFTA is seen as a model, there would be a need to apply After five years of operation-i.e., in principle from the principles of MFN and national treatment commit- November 29, 1996-the possibility of admission to Mer- ments across all sectors (with limited exceptions or reser- cosur is open to other members of the LAIA.'5 Bolivia has vations), unlike the more lax GATT commitments made indicated that it would like to become a member and by Mercosur countries. Again, even if discriminatory inter- already participates in some technical groups; in December nal taxes have not been raised as an issue in GATT, it is 1995 it signed an agreement with Mercosur intended to unlikely that they would pass unnoticed in the FTAA lead to a free-trade agreement. In June 1996 Chile also negotiation.2' Overall, import-competing sectors in Mer- signed an agreement, which envisages the comprehensive cosur countries (capital goods, automobiles and even some removal of duties on Mercosur-Chilean trade over a 1 0-year agricultural goods) would come under much greater period, except for a small group of agricultural products on adjustment pressures in such an FTAA. which tariffs are to be phased out over 15 to 18 years.19 Prior to the discussion on the FTAA, Mercosur signed a This agreement also covers rules of origin (many of which framework treaty with the United States in June 1991 to are product-specific), transport investment, services, intel- encourage trade and investment in the region. This treaty lectual property and investment. There is also an interest established a consultative council on trade and investment. in developing a link between the Andean Community and It set out an initial action program that encompassed Mercosur, fostered primarily by Brazil and Venezuela. Uruguay Round cooperation, means to reduce trade and Third, bilateral relations with non-LAIA countries are investment barriers in the Americas, access to technology, now dominated by work on the proposed Free-Trade Area intellectual property rights, access to markets for goods of the Americas (FTAA), announced at the Summit of the and services, sanitary and phytosanitary regulations in Americas held in Miami December 9-11, 1994. Work on agriculture and the need for a transparent safeguard regime the FTAA was consolidated by the work of ministers at in conformity with the GATT. 142 MERCOSUR Under an inter-regional cooperation agreement signed harmonization, which would see the elimination of Brazil- in December 1995 with the European Union, a joint sub- ian farm support policies (Lopes 1996). committee on trade began work in 1996. In the longer Details of the Argentina-Brazil arrangements in the term, the agreement envisages the opening of negotia- automotive sector are shown in the box below. Effectively, tions on tariffs and technical standards in the year 2002 this is a managed trade arrangement, in which the sector with the possibility of establishing free trade between the benefits from local content plans, allowing concessional two blocks by the year 2010. To some extent this agree- entry on vehicles and parts, as well as export-balancing ment is seen as counter-balancing the influence of the requirements. The arrangements, which have been partic- United States in the FTAA. There have also been indica- ularly advantageous for the parts industry, are to remain in tions of interest in establishing closer trade relations with force until a common sectoral policy is established in 1999. APEC. The present arrangements are covered by a waiver from the provisions of Mercosur under Council Decision No. 29/94. Sectoral Policies To be compatible with the WTO Agreement on Trade The coordination of sectoral policies, which is a feature of Related Investment Measures (TRIMs), it will be necessary a deeper level of integration than the WTO requirements to eliminate the local content and export-balancing for a customs union, has been the subject of discussions at requirements by the year 2000. This results from the pro- the technical level in the areas of agriculture, industry, vision of the TRIMs Agreement that stipulates that all energy, transport and labor. The results of these activities such measures were to be eliminated immediately in 1995 are being incorporated in decisions or resolutions of the unless notified to the WTO, while those measures that higher-level bodies of Mercosur, with the intention that were notified must be eliminated within five years (i.e., by these will lead to greater harmonization of domestic sec- the year 2000) by developing countries. Argentina made toral policies and will further the integration process. such a notification for its automotive arrangements. The Negotiations are taking place in Mercosur's technical status of the Brazilian measures was still sub judice at the Subgroup 8 to harmonize and coordinate agricultural poli- time of writing. cies. These discussions are intended to lead to a scheme In this respect, it may be noted that Mercosur provides based on the WTO Agreement on Agriculture. National a framework for production-sharing arrangements, under governments would continue to have complete autonomy provisions that are redolent of the various sectoral arrange- on measures falling within the "green box" and "blue box" ments that preceded Mercosur. These are intended to foster measures (permitted subsidies, which are general or are the rationalization of investment and to increase the com- delinked in (different degrees to the amount of production), petitiveness of companies in the region (GATT 1992c). A while there would be some limitations on national auton- production-sharing arrangement for the steel industry was omy in respect of measures that are subject to WTO reduc- in place between 1992 and 1994 among Argentine and tion commitments on the Aggregate Measure of Support Brazilian firms, but since January 1995 Brazil has applied (AMS). Decision 19/94 of the Council of the Common duty-free quotas for steel imports from other Mercosur Market alloiws the sugar sector to take up to the year 2001 members under the adjustment regime. It is feasible, but to adapt to the operation of the customs union. Discussions apparently has not been discussed, that arrangements in are taking place on the export taxes that are applied inde- the automotive sector could come under those provisions, pendently by Argentina, Brazil and Uruguay on exports of but they could not remain as they are because of the raw hides and skins. The implementation of an umbrella TRIMs Agreement. mechanism for agriculture, including a price band (vari- The extended periods of convergence to the CET for the able levy) system was under consideration in 1995 (Henz capital goods and electronics sectors may be seen as part of 1995), but Mercosur members decided instead to the coordinated action on sectoral policies in the area of strengthen and harmonize their trade defense (anti-dump- manufacturing. They are also linked to the asymmetries ing/countervailing) mechanism (WTO 1996a). These dis- that exist in the economies of Mercosur and, to a degree, a cussion are strongly linked to Brazilian concerns about the lack of international competitiveness in those sectors. In vulnerability of its farm sector and food security under a electronics, Brazil was by far the major producer, but it has 143 TRADE: TOWARDS OPEN REGIONALISM essentially renounced its earlier programs to develop the network linkages, user rules and/or rates of commission, but electronics sector that imposed major costs on other parts little work has been done by Mercosur in this area. of its economy (WTO 1996b).22 The convergence periods Transport policy issues have been discussed in Subgroup 5 effectively provide a prolonged adjustment period to a on transport and infrastructure (previously Subgroup 5 on more open regime, especially in Brazil. inland transport and Subgroup 6 on maritime transport). In the national economies of the Mercosur countries, ser- Proposals include the total elimination of standing and vices now constitute the larger part of national production. flexible quotas for road transport, direct transport between To derive greater benefits from international exchange, the rail terminals, a joint shipping register, labor regimes Treaty of Asunci6n provides for free movement of services, (stemming from the original Subgroup 11, now Subgroup capital and workers as well as goods among Mercosur coun- 10), etc. At present, transport of goods and persons across tries. However, while trade in goods has been substantially the frontiers is allowed, as well as returning with cargo or liberalized, much remains to be done in these other areas passengers. However, internal cabotage is not allowed. At where negotiations for a framework were ongoing in an ad the Ouro Preto meeting the members adopted a hoc group at the time of writing. Multi-Modal Transportation Agreement, governing trade The work on services is being carried out by the Ad Hoc operations under a single contract involving more than one Group on Services and its various subcommittees. A frame- transport mode, as well as a Hazardous Merchandise Trans- work agreement was scheduled to be submitted to the Com- portation Agreement, covering the harmonization of regu- mon Market Council before the end of September 1997. lations on the transport of dangerous goods. Technical dis- In Mercosur there is relatively little integration infinan- cussions are continuing in the area of safety standards, etc. cial services, although it is more advanced in banking than Discussion has been taking place on the possibilities for in other areas (Machado 1996).'3 However, there are con- harmonization, coordination and regulation in the areas of siderable asymmetries in Mercosur financial services mar- audiovisual information and data processing and comput- kets, and this is reflected in the different offers of the mem- erized marketing and distribution. bers in the WTO sectoral negotiations, with Brazil's offers being more restrictive than those of other Mercosur coun- Technical Barriers tries, reflecting the earlier stage of sectoral liberalization Mercosur's Standards Committee, comprising the national (Abreu and Bevilaqua 1995) Within the region, it has standards bodies of member states, has already carried out been agreed through the Protocol of Colonia to apply extensive work on the harmonization of a national stan- national treatment to intra-regional investment, with some dards. Resulting from that work, the council has agreed exceptions, with guarantees for investors. A network of (Decision 6/96) on the application of the WTO rules on national banks (Red BANOSUR) is intended to find new sanitary and phytosanitary measures as the set of rules for operational mechanisms to promote intra-regional trade. Mercosur. To the extent necessary, national measures will Subgroup 4 (financial matters) of the Common Market be adjusted to bring them into conformity with that agree- Group has been examining inconsistencies in the capital ment. Where further work is desirable in this area, it market regimes and financial systems, and it has drawn up might be possible to draw on international standards, such some priorities for cooperation, focusing on supervision, as International Standards Organization (ISO), the Codex exchange of information, privatization, money Laundering Alimentarius of the Food and Agriculture Organization and external relations. The work program extends to the (FAO) and World Health Organization (WHO) on food year 2000. products, or the Office Intenationale des Epizooties (OIE) In the area of prudential supervision, the Common Mar- on animal health; failing this, it could be advantageous to ket Council in its decision no. 10/93 adopted standards cov- adopt standards from major external markets or suppliers ered by the Basel Agreement of 1988, but much hinges on such as the United States or the European Union. This surveillance and enforcement. Coordination in financial ser- could facilitate exports, making it easier to gain access to vices could usefully be extended to certain basic manage- foreign markets and expand sales outside Mercosur. On the ment principles for financial intermediaries, including stan- import side, it would also help to ensure the openness of dards of accounting and auditing, compatibility of systems, the customs union. i44 ME RCOSURS BOX 1 U" jnur 1996 brazil anwhetn sgeM iaerl h gement. contains a Sid ,lettralwn agreement for ma sector, rvsin the, text Aretn tocmestetescoal deicifU$5 RignedsL b;twe Thtszdvi tPeoi' mlion, accuatd. withBaizil between 192I n DecEmbr 194. Unde vhAuo rr Agreemen 1994: 'uchcmestoisSak'pcerrlgte moto vehcleswer~~ ~ ~ tralmur,.whil duyfre 'imortaionby Bazilof loseto 8,00 time, Argentinawas granted frfet(er cestZ h raiia titieslandteueo motor veIce mde bycmpne make"ha itwsrqie ogatlbaiiheprs ntld B inteote o Nty Ngotiaton ere rill i 'Ude he aremen indi Janar196 eh- progress in 1996 dles- and parts: may be imported u~fe sln ste' ne thegemn,iprsfo ithr coutries importerbalance its forignpucases w-ith. exot to ar WraeIesfaoalhnitr-einltae Moe k efiedi e C arl 196 Theareetpoie o aif h oa akr The agreement alsosest anavrae ocl onen oIa least 30 percen co Rmptd 'fav rlocton oasebyprtininBrai,a an oer t treeyar eidlecsrlclcnetlvl's fcuesaels likey toconsdeprdcn vhkice'i Each patnr enages to recgnze the li~di of thWsm uniybflaiinr evhices automotive regI'me oif the othe conr ni eebr Bai n rga ranted eah othe anf inctreaseo 31,199. I effet ths fomajizs Arentia'sacep 15 percent in, the import ~quotas for cars underwhich r3L in9995.h Uner ehe4.AgieemenrfotfJauar 196 wr toh loe oxot1,0 units in196 develpmens undr th agrementand t decde onany o 4,50 cas. inf '19 Uuguyddn hailis'nt.o necessry adustmnt. Fro aur ,20,blaea 000uis hihwsetne ni Marc 196 trade will.9 he duty-free. MM S,nWOi996 ' .q On the basis of experience in the European Union, sectors ment are generally similar, having identical intentions where technical standards may constitute serious barriers to although often raking different forms, then there is no tea- trade and, accordingly, where priority needs to be given to son in theory why a product legally manufactured in one work on standardization, include information technology, country could not he sold anywhere in the customs telecommunications, transport, constr-uction, pharmaceuticals union.24 This would obviate the need to meet national and foodstuffs. As far as can be readily ascertained, these are standards, with tests and certification in the importing areas where there has already been extensive work in Merco- member states, allowing any purchaser-wholesaler, sur. A subgroup also appears to he working on an agreement retailer or individual-rn choose freely his supplier from on testing and certification procedures, common conditions anywhere within the customs union. and codes of practice for laboratories and certification bodies. In the longer term, Mercosur countries could move from In the short term, full and immediate mutual recogni- mutual recognition toward greater harmonization that tion of national standards (on quality, food composition, allows for even greater rationalization of production to cap- etc.) should be the goal. Since the main objectives of legis- ture economies of scale and specialization. While harmo- lation for the protection of human health and the environ- nization implies greater regulation, taking longer to 145 TRADE TOWARDS OPEN REGIONALISM implement and perhaps introducing inflexibility and sti- which were developed largely at the insistence of the fling innovation, it is sometimes argued that harmoniza- United States because of perceived weaknesses of the tion is more likely to promote cost reductions and the WIPO framework. Member states are now bound by the achievement of competitiveness at the international level, WTO rules, which should oblige a minimum degree of particularly in high-technology goods and areas where harmonization, including within Mercosur. economies of scale and specialization matter. In the EU, The original Subgroup 7 (industrial and technological harmonization, rather than mutual recognition, was con- policy) of the Common Market Group carried out a survey sidered particularly important in the areas of information of national laws in the area of intellectual property in 1993 technology, telecommunications and transport to ensure and 1994, and made proposals to the Common Market compatibility between users and operators throughout the Group for dealing with intellectual property at the customs union. However, these are also areas where Merco- regional level. This was subsequently adopted in Decision sur does not have any evident comparative advantage, and 8/95 of the Common Market Council, which in its annex it would be undesirable to establish harmonized standards contains the Protocol on the Harmonization of Rules on that effectively operated to restrict access to least-cost Intellectual Property concerning marking and origin, sources or world markets (particularly since these are also which had only been ratified by Paraguay in mid-1997. A important inputs into other sectors of the Mercosur Protocol on Trade Marks, Indications of Source and Appel- economies). lations has been finalized and is also awaiting ratification by Argentina, Brazil and Uruguay. An Agreement on Intellectual Property Copyright and Related Rights was approved by the Coun- In the past, the issue of protection of intellectual property cil, but has been returned to Working Group No. 7 for fur- was an area where there were significant divergences from ther study. overseas practices in the Mercosur area, where intellectual property protection was often based on process rather than Competition Policy, Reform of the State, and product protection and unexploited patents fell into the Government Procurement public domain relatively rapidly. While this has helped Within the market, a technical committee (No. 4) of the some development of domestic industry, it has also been a Trade Commission is identifying public policies that distort source of tension in international relations-and no doubt conditions for competition, including government procure- has also had certain adverse consequences in the developing ment, tax policies and state trading. A common statute on countries, for example by higher royalties to offset risks. competition polices has been proposed by the Trade Com- Rowat (1993) argues that the increased interest in mission to the Common Market Council. Efforts in these strengthening laws for the protection of intellectual prop- areas have been identified as crucial for the consolidation of erty is strongly linked to the revisions of the trade and the internal market by Machado and Markwald (1996). investment regimes, designed to make developing coun- Indeed, the whole area of reform of the state is seen as cre- tries more attractive to foreign investors. He describes this ating a more favorable economic and institutional environ- linkage as a "marriage of convenience. ment for integration (de Almeida 1996). Most international agreements to protect intellectual Privatizarion, one of the key elements of reform of the property are administered by the World Intellectual Prop- state, has a number of goals, including the reduction of for- erty Organization (WIPO), which is responsible, inter alia, eign debt and the control of public-sector deficits. How- for the Paris Convention for the Protection of Industrial ever, a microeconomic objective is the improved allocation Property, first established in 1883; the Berne Convention of resources, deriving from a more competitive environ- for the Protection of Literary and Artistic Works, first ment. While trade reforms can do much to promote this adopted in 1886; and the Madrid Agreement concerning objective, a proactive competition policy is often necessary the International Registration of Marks, adopted in 1891. to overcome market imperfections, especially in the short The most important new development of the international term. This has been noted in GATT (1992a) for Argentina framework is the adoption by the WTO of compromise and WTO (1996b) for Brazil, but this effort needs to be rules on trade-related intellectual property (TRIPs) issues, carried forward across Mercosur in a coordinated manner. 146 MERCOSUR This could pave the way for competition policies to replace also be benefits from improved road and rail connections in anti-dumping on intra-regional trade. the Mercosur area. In the current financial situation of the However, despite the extensive privatization that has member states, it may be opportune to look further at the taken place in recent years, the public sector remains very possibilities for privatization in these areas, as has already important in Mercosur (Baer 1995, Shirley 1994).25 It been done for bulk shipments of grains and minerals. would therefore be consistent with the goal of improving competitive conditions in the region to increase competi- Coordination of Macroeconomic Policy tion and transparency in public procurement for govern- The Treaty of Asunci6n provides for periodic meetings of ments and r2maining state-owned enterprises. This means Ministers of Economy and Central Bank Governors of Mer- eliminating preferences for domestic suppliers, which are cosur countries to coordinate of macroeconomic policy and often granted in public-sector contracts and work very to try to establish joint approaches to common problems much like ad valorem tariffs on imports. (Argentina gener- such as inflation, foreign investment and trade. In practice, ally affords national treatment, whereas there are still sec- there were no meetings between 1995 and 1997 (until a tors where Brazil has to move in this regard.) Apart from meeting in Asunci6n in April 1997 to consider, among explicit preference margins, projects are sometimes specif- other things, Brazilian restrictions on the financing of ically designed to favor local suppliers, working more like imports). In the past the lack of coordination on macroeco- import prohibitions on foreign suppliers. There is often an nomic polices has been most evident with respect to absence of clear-cut guidelines or transparent tendering exchange rates (Abreu and Bevilaqua 1995, Pereira 1996). procedures, opening the way for corruption of public offi- For example, differential movements in exchange rates and cials. Major welfare gains could therefore be realized by interest rates in Argentina and Brazil caused major fluctu- opening up domestic markets to regional, or, better yet, ations in trade and financial flows, giving rise to political extra-regional suppliers. Little work seems to have been tensions and to new trade measures. This has sometimes done in this area. Argentina presented to the Fortaleza led to managed trade solutions, instead of corrective meeting of the Common Market Group a proposal to macroeconomic policies. For example, at one point, to establish an ad hoc group on government procurement, reduce tensions and help to overcome rising trade deficits, and the group decided to consider this later. Brazil made "special efforts" to buy more from Argentina In the WTO, government procurement is one of the few (e.g., through state-owned enterprises). The Argentina- areas where participation is optional, and the Mercosur Brazil automotive arrangements, stepped-up efforts by countries, together with most developing countries, chose Brazil to promote exports and, most recently, Brazilian not to participate. Assurance of the outward orientation of restrictions related to the financing of imports may be seen Mercosur would be enhanced if Mercosur governments in a similar light. chose to sign the WTO Agreement on Government Pro- One approach to convergence in macroeconomic man- curement. This would seem entirely compatible with the agement might be the establishment of targets, such as the need for prudent management of national budgets, and the European Union has set as a condition for participation in disciplines on transparency and openness so introduced the planned monetary union. These cover inflation, long- would also serve as limits on any corruption in letting of term interest rates, exchange rates, budget deficits and procurement contracts. public debt, all consistent with macroeconomic stability, which, as Pereira (1996) points out must be the basic goal Infrastructuvre more than rigid targets.26 However, efforts to restrain fiscal Although infrastructure is not specifically covered by the expenditure are complicated by the ineffectiveness of fed- Treaty of Asunci6n, it has been highlighted in GATT eral controls over the provinces in Argentina (now being (1992a and 1992b) that the trade of Mercosur countries brought under control) and the states in Brazil. Again, would be facilitated by improved physical infrastructure, exchange-rate corrections are complicated because particularly ports and airports. A particularly telling exam- Argentina has pegged its currency to the U.S. dollar, and ple is the degraded port and transport situation in Argentina's use of monetary policy is also effectively Paraguay, described in World Bank (1993). There would excluded by its Ley de Convertibilidad, requiring all cur- 147 TRADE TOWARDS OPEN REGIONALISMI rency issues to be fully backed by U.S. dollars (in essence a frame. Concerns about possible trade diversion should be currency board system); this leaves economic management allayed by the extensive unilateral reforms, the major essentially to fiscal policy, for which financing is also lim- increase in tariff bindings in the Uruguay Round, the ited by the outlawing of the monetization of government removal of many non-tariff measures and the additional dis- deficits. It is constraints such as these that have led to the ciplines of the WTO, which also add a degree of security to use of special arrangements, lists of exceptions to the CET trading conditions for third countries. So far the rapid and safeguard actions to manage trade in recent years, increase in trade flows between members has been accom- going against the spirit and intent of a common market, panied by a solid growth in trade with third countries. not to mention the WTO. To complete internal free trade in goods requires the The original Subgroup 10 on coordination of macroeco- elimination of the remaining tariffs on intra-Mercosur nomic policies carried out an examination of taxation trade, and a number of non-tariff measures have yet to be issues, particularly in the area of consumption taxes. In eliminated or harmonized; in most areas, this work is principle, this work was handed over to the Ministers of either in process or is being discussed. While the intention Economy and Central Bank Governors, but, as noted, there is to allow free circulation of goods, the weakness in were no meetings between 1995 and 1997, and no specific administrative arrangements for free circulation represents actions have yet been taken along these lines, despite the a serious gap in tbe completion of a customs union; there range and complexity of internal taxes, including those at is scope for improved implementation of agreed stream- the provincial and state levels in Argentina and Brazil.27 lined customs operations. The provisions for sectoral policy This divergence in rates has encouraged "shopping coordination, including production-sharing arrangements, tourism" for which Paraguay, where indirect taxes are low, need to be monitored to ensure that they do not revert to is a regional center. However, little progress has been earlier import substitution policies. No agreement has yet achieved on tax harmonization, and the possibilities for been reached on free movement in services, capital and agreement must inevitably be complicated by the current workers, despite the commitment to include these by Jan- process of reform of domestic taxes in each member, uary 1995. Mercosur needs a strong and coherent competi- intended to overcome tax evasion and lift tax receipts. tion policy to improve the competitiveness of domestic The case for coordination or even harmonization of industry, and this could also replace anti-dumping proce- indirect taxation is to avoid resources flowing from one dures in intra-Mercosur trade. Opening up public procure- member to another, not on the basis of resource endow- ment regionally and beyond could bring both welfare and ment but rather on avoidance of potential tax liability.28 direct fiscal benefits. A supranational framework may be Harmonization would also reduce the risks of fraud and needed to advance and deepen integration in investment smuggling that occurs today. Pita (1996) argues that it is and services, and some steps have been taken in this direc- necessary to accelerate the institutionalization of the tion. In the meantime it remains to be seen if formal dis- processes of tax harmonization to gain the full benefits of pute settlement procedures are sufficient protection for the integration. There might also be some harmonization of smaller countries. tax incentives for investment, domestic or foreign. How- Given the history of macroeconomic instability in the ever, in the present fiscal situation, it is difficult to foresee region, often linked to reversals of previous trade policy any Mercosur agreement on the establishment of an EU- reforms (Papageorgiou, Michaely and Choksi 1991), one of type mechanism whereby goods would only be taxed at the highest priorities for the authorities must remain coor- the point of production, passing across frontiers without dination and cohesion in the pursuit of responsible fiscal, border tax adjustments. monetary and exchange-rate policies to achieve economic stability with steady growth, perhaps with specific targets Conclusion for key economic variables. As we have seen, the failure to The Mercosur countries have taken important steps toward achieve such coordination can lead to disruptive imbalances the construction of a customs union, and beyond that in trade, reversals of trade or exchange-rate policies, and toward a common (but not EU-style single) market. This deteriorating political as well as economic relations. The work has been accomplished in a relatively short time- resort to managed trade solutions, particularly in automo- 148 MERCOSUX K biles, run counter to the spirit of recent unilateral trade Behar, Jaime (1995). "Measuring the Effects of Economic Integration reforms and The common market, not to mention the WTO. for the Southern Cone Countries: Industry Simulations of Trade On external policy, there are still a number of important Liberalization," The Developing Economies, Vol.XXXIII, No. 1 exceptions to the common external tariff, including those (March), pp. 3-31. Bouzas, Roberto (1996). "Mercosur y Liberalizaci6n Comercial Pref- that come through flexible additional lists that have pro- erencial en America del Sur: Resultados, Temas y Proyecciones," vided legal cover for recent policy reversals, but the pro- in NAFTA y Mercosur: Un Didlogo Canadiense-Latinoamericano. grammed elimination of these measures is still within the Santiago: CIEPLAN/Dolmen Ediciones. agreed timetable. In the longer term, greater welfare ben- Brandao, Antonio Salazar P and Lia Valls Pereira (eds.) (1996). efits could be realized by scaling back escalation in the 2MERCOSUL-Perspectivas da Integragd7o. Rio de Janeiro: Fundalao Gerulio Vargas. CET and the elimination of TRIMs exceptions in the auto- GeuiVags da Silva, Carlos Roberto Lavalle (1996). "Harmonizacdo tribatdria no motive sector. Between members, there seems to be more MERCOSUL," in Brandao and Pereira (1996). competitive than coherence of approach in attracting for- de Almeida, Fernando Galvao (1996). "Reforma do Estado como eign investmrent, and this also seems to have linkages to fator favoravel ao proceso de integracao," in Brandao and Pereira the outmoded import-substitution policy. While efforts (1996). mre being niade to agree on a common anti-dumping de Melo, Jaime and Sumana Dhar (1992). "Lessons of Trade Liberal- arechbeing consideration might be also given to using ization in Latin America for Economies in Transition," Policy mechanism, consideration might be also given to using Research WPS No. 1040. Washington D.C.: World Bank. competitive policy standards on external trade to avoid Edwards, Sebastian (1994). "Trade and Industrial Policy Reform in possible abuse of anti-dumping actions whose use has Latin America." Cambridge, Mass.: NBER Working Paper Series become important, especially in Brazil and Argentina. No. 4772. External trade relations are still centered on other LAIA GATT (1992a). Trade Policy Review: Argentina. Geneva. countries and talks with Andean Group members, which (1992b). Trade Policy Review: Urugaay. Geneva. (1992c). Trade Policy Review: Brazil. Geneva. could lead to a virtual South American FTA. There seems (1994). Working Party on the Southern Common Market to be a more ambivalent interest in the proposal for a Free- (Mercosur) Agreement, Document No. L/7540, October 26. Trade Area for the Americas, which would have the advan- Geneva. tage of extended market opportunities, greater economies Henz, Renato A. (1995). "Condicionantes externos a politica agricola," of scale in production and distribution, the dilution of the Ensaios FEE, Porto Alegre, No. 16 Vol. 1, pp. 52-65. power of national industrial lobbies and greater discipline Krueger, Anne (1995). "Free-Trade Agreements vs. Customs Unions." Cambridge, Mass.: NBER Working Paper No.5084. on potential policy slippages. Kume, Honorario (1996). A Politica de Importa,co no Plano Real e a The potential welfare gains from completing efforts on Estrectura de Prote,co Efecciva, Texto para Discuss&o No. 118, Fun- internal and external barriers, ensuring the smooth func- da,co Centro de Estudos do Comercio Exterior, February. tioning of trade and looking beyond existing plans to a Laird, Sam (1995). "Trade Liberalization in Latin America," Min- lower and more uniform structure of protection are very nesotaJosernal of Global Trade, Vol. 4, pp. 101-27. large, alrhough their attainment will entail adjustment for Langhammer, Rolf and Ulrich Hiemenz (1991). "Regional Integra- 71 tion among Developing Countries: Survey of Past Performance these sectors, which hitherto have flourished under pro- and Agenda for Future Policy Action," UNDP-World Bank tected, uncompetitive markets. The willingness of the Trade Expansion Program Occasional Paper No. 7. Washington member countries to confront the "hard cases" will consti- D.C.: Trade Policy Division, World Bank. tute a test of the depth of change in the economic philoso- Lopes, Mauro de Renende (1996). "Mercados agricolas e o proceso de inte- phies of the member states. gracao no MERCOSUL," in Brandao and Pereira (1996). Machado, Marcos Fernandes (1996). "A integracao financiera no MER- COSUL," in Brandao and Pereira (1996). References Machado, Joao Bosco M. and Ricardo Markwald (1996). Abreu, Marcelo de Paiva and Afonso S. Bevilaqua (1995). Macroeco- "MERCOSUL: Padrao de Come'rcio e Estrategia de lntegra,io," Texto nomic Coordination and Economic Integration: Lesson for a Western para Discussao No. 117, FUNCEX, Rio De Janeiro (February). Hemisphere Free Trade Area, Texto para Discussao No. 340. Rio de Nogues, Julio (1983). "Proteccio'n comercial y cambiaria: nna inter- Janeiro: Departamento de Economia, PUC-Rio. pretacion de la experiencia argentina darante 1976-77," Banco Cen- Baer, Werner (1995). "Privatization in Latin America," The World tral de la Republica Argentina, Serie de Estudios Tecnicos Economy, Vol. 17, No. 4, pp. 509-28. Numero 52, 1983. 149 TRADE: TOWARDS OPEN REGIONALISM PapageorgiooL, Demetrios, Michael Michaely and Armeane Choksi 7. Legislative rarification and national implementation some- (eds.) (1991). Liberalizing Foreign Trade. Oxford and Cambridge times lags behind the commonly agreed measures. For example, ir. Mass.: Basil Blackwell. early 1997, the author was informed that, out of the total of 97 deci- Pereira, LiaValls (1996). "TratadodeAssan;do: resaltadoseperspectivas, sions, 501 resolutions and 43 directives approved in the period in Brandao and Pereira (1996). 1991-96. Argentina and Brazil had ratified somewhat more than Pita, Claudino (1996). 'Harmonizaado tribeetdria no M4ERCOSUL.' in 50 percent. This is less serious than it appears, because some of these Brandao and Pereira (1996). legal measures merely concern the undertaking of studies rather thar. Rajapatirana, Sarath (1994). "The Evolution of Trade Treaties and mandating behavior. Trade Creation: Lessons for Latin America," Policy Research 8. The case involved a Uruguayan paper firm that complained Working Paper 1371. Washington, D.C.: World Bank. that Argentine levies were not in the exception lists established at Rowat, Malcolm D. (1993). "An Assessment of Intellectual Property Ouro Preto. Protection in LDCs from both a Legal and Economic Perspec- 9. The number of items is computed on the basis of the Merco- tive Case Studies of Mexico, Chile and Argentina," Denverjoar- sur tariff classification. The number of tariff items would be some- nal of International Lauo and Polig, Vol. 21, No.2 (Winter), pp. what higher if measured in terms of the pre-existing national tariff 401-29. classifications. Shirley, Mary M. (1994). "Privatization in Latin America: Lessons for 10. Imports from export-processing zones are normally treated as Transitional Europe," World Dev'elopment, Vol. 22, No. 9, pp imports sub ject to the CET. 1313-23. 11. Unusually, in the GATT/WTO context, this requirement World Bank (1993). Paragueay: Mercosair and Paragzeays Grouth refers to applied tariffs, rather than to tariff bindings. Prospects, Report No. 10902-PA. Washington, D.C. 12. Nogues (1983) notes that in Argentina the average nominal WTO (I 995a). "Sotuthern Common Market (Mercosolr) Agreement,' tariff was reduced from 98 percent to 49 percent in the Martinez de Document No. WT/COMTD/1 and Add. 1, May 2. Geneva. Hoz period (1976-8 1), and stood at more than 30 percent in 1988, _ (1995 b). Regionalism and the inorld Trading System. Geneva. but the ratio of duty collected to total imports was close to 2 percent _ (1996a). Southern Common Miarket (Mercosar) Agreerrent: QOestions for all the years between 1970 and 1988. This was a result of con- and Replies, Addendiam. WT/COMTD/1/Add. 4, August 14. Geneva. cessional rates as well as preferences. _ (1996b). Trade Po/icy Review of Brazil. Geneva. 13. Paraguay acceded to the GATT during the Uruguay Round _ (1997). Trade Policy Rev-iew of Paragieay. Geneva (forthcoming). with a comprehensive binding coverage. 14. Estimated protection would be even higher if the tariff reduc- rions nn inputs annnsinced in late 1995 were included. The industry Notes is further assisted by non-tariff measures-e.g., export subsidies and 1. Of course, there is trade diversion if the counrterfactual is free investment incentives. trade. 15. The level of the Argentine statistical tax has varied between 2. Mercosur stems from the earlier Integration, Cooperation and zero and 6 percent in recent years. Development Treaty of August 1989 between Argentina and Brazil, 16. It may be noted that, since 1992, no anti-dumping actions which in turn was based on the Iguazu Declaration of November have been allowed on intra-regional trade in the EU and the Euro- 1985. See World Bank (1993) for details. pean Economic Area, with reliance being placed solely on common 3. Mercosur is formally a subregional agreement wirhin the LAIA competition policy enforcement. The same is the case in the Aus- framework. tralia-New Zealand Closer Economic Relations Trade Agreement 4. However, Article XXIV requires that a customs union or FTA (ANZCERTA). cover substantially all trade. For those items that are included, 17. During the period of implementation when national excep- "duties and other restrictive regulations are (to be] eliminated"- tions to the CET are allowed, the difference between the CET and i.e., that there is to be a 100 percent reduction in tariffs and other the national level would be paid (or, in principle, rebated where the barriers. A preferential rate other than zero is not contemplated. national rate is lower) on entering the market of final destination. 5. Disagreement about GATT consistency is the norm; confor- 18. Member countries of LAIA are Argentina, Bolivia, Brazil, mity with GATT Article XXIV has been explicitly acknowledged in Chile, Colombia, Ecuador, Mexico, Paraguay, Peru, Uruguay and only six our of 69 cases where reports of working parties were sub- Venezuela. mitted to the GATT Council. See WTO (1995b) for further derails. 19. Since Chile now has an FTA with Mercosur, it is not obliged 6. A headquarters agreement (Acordo de Sede) with the Uruguay to change its uniform 11 percent tariff to the Mercosur CET. This government was agreed under Decision 4/96 of the Common Market also will allow Chile to maintain its FTAs with Canada and Mexico Council at its December 16-17, 1996, meeting in Fortaleza. Financ- as well as pursuing its possible NAFTA accession. ing of the secretariat is on the basis of equal cost-sharing among 20. Brazil's Ambassador to London, Rubens Barbosa, quoted ir members; there is no independent source of finance. Financial Timnes, December 8, 1994. 1 5 0 MERCOSUR 21. Article III of the GATT prohibits discrimination against 25. Baer and Shirley, taking different views of the lessons from imports in the application of domestic taxes. privatization, provide a useful overview of the importance of the state 22. In developed economies, these sectors are often supported under sector in Latin America. Baer mentions that there were some 353 research and development programs as well as defense contracting. state-owned enterprises in Argentina in the 1980s, 268 in Brazil and 23. There hlas been a substantial restructuring of Latin American more than 1,000 in Mexico. Shirley notes that Latin America is the banks, linked in some instances to inefficient and corrupt manage- major privatization area in the world, generating revenues of some ment (see "Banks in Latin America: An Urgent Case for Repair" and US$40 billion in the period 1988-92. "Miami or Bust" in The Econornist, July 1, 1995, pp. 79-80). The sit- 26. These targets have prompted the comment that the condi- uation deteriorated because of the Mexican financial crisis, but this tions of membership are like a slimming club's requiring weight loss was not the ptime source of the problems. prior to admission ("From here to EMU," The Economist, August 5, 24. This was also a matter of extensive debate and legal action in 1995). the European Union. See, for example, the Cassis de Dijon case 27. Da Silva (1996) points out that Brazil would probably have (Rewe-Zentsal AG v Bundesmsonopolverwaltungfvr Branntweil, ECJ case to make the major effort since its tax system contains elements that 120/78) that cetermined, inter alia, that differences in national mar- significantly distort the efficient allocation of resources. keting laws should only be allowed as a restriction on free movement 28. Tax harmonization does not necessarily imply absolutely of goods to the extent necessary for the enforcement of fiscal super- identical rates of taxation. For example, it is possible to envisage a vision, protection of public health, fairness in commercial transac- fixed percentage variation from an agreed norm or target across tions and defe:nse of the consumer. member states. 151 VI. Roundtable: The nmall Economies The Challenges to Small Economles L U IS MO SC A T ~ HERE IS SO FAR NO CONSENSUS ON WHAT CHALLENGES THE SO-CALLED SMALL ECONOMIES are actually confronted with in a globalized world economy, and particularly not when they belong to large regional groupings. The difficulty here may lie in the very definition-perhaps too simplistic, or mistaken- of what constitutes a small economy. Undoubtedly, there is a confusion in terms, because, with very few exceptions, the world's economies are all small, in the sense that they are unable to isolate themselves or determine the course of external events. In addition, many large-scale indus- trial or service activities with worldwide presence and impact are located in geographically small countries. Good examples are the chemicals industry in Switzer- least three issues emerge clearly as challenges, and it is on land and electronics industry in the Netherlands. So, from these that I want to focus here. The first issue is whether it the outset, let us be clear that the relative nature of the def- is possible to mitigate the impact of external shocks, inition limits the possibilities for analysis considerably. mainly regional, so as to enhance the effectiveness of pub- At the present time, there is not one set of problems lic policies and prevent social conditions from deteriorat- specific to large economies and another to small economies. ing when the regional economic cycle is adverse. The sec- We are simply dealing with new issues, and some not so ond is the importance of a better understanding of the new, arising out of the growing interrelationship linking investment dynamic within areas such as Mercosur, a key nations for economic, technological and cultural reasons. consideration for economies such as Uruguay's, where Each country perceives these new issues as new challenges, present investment levels are too low to support higher which it tries to resolve from its own specific position. growth rates. And the third has to do with exploring the I do not believe there exists a typology of challenges right stance for economies in situations similar to that parallels the size of countries. What does exist is a Uruguay's to adopt toward the new generation of agree- generalized set of conditions in which all factors external to ments that are expanding the areas of economic interrela- a country have significant internal connotations for it. In tionship. The FTAA is a good example. my view, this precisely is the correct stance for determin- ing which challenges to resolve in the prevailing circum- Reducing External Volatility stances, and for formulating a strategy by which to over- Greater interdependency and integration mean increasing come them. the level of domestic accommodation to external events. In the particular case of an economy that operates under The apparent loss of independence that this entails in the the conditions and in the situation of that of Uruguay, at design of domestic policies is obvious, but at most the Luis Mosca is the Minister of Finance of Uruguay. 155 TRADE: TOWARDS OPEN REGIONALISM question is one of degree and not one of substance, since likely pattern of specialization in production and services experience demonstrates in the end that external occur- among the partners involved. This latter point is of singu- rences tinge all countries' internal events in one way or lar importance, given that the invariably limited degree of another. public-sector investment is merely a complement to pri- Nevertheless, recognizing this reality does not mean vate investment. Thus, prior identification of likely invest- ignoring the importance of instituting strategies or poli- ment areas reinforces the process of external investment cies designed to reduce the domestic volatility it causes. and the effectiveness of public-sector capital spending. The aim is to enhance the effectiveness of the policies I would like to draw attention to three considerations adopted and lower the concurrent social costs. I believe this that I regard as looming large in this issue: First, services to be an area of challenge where we cannot yet claim are not a field of activity where investors need to assure enough proven experience, except in seeking to diversify themselves of minimum market dimensions independently our external markets as much as possible and reduce the of the future course of integration agreements, or where distortions affecting domestic factor markets. In any case, I choice of location is dictated by the availability of specific have the impression that what has been done has not been resources. Second, decisions on site choices and develop- sufficient-and neither is the number of proposals put for- ment plans where services are concerned are very sensitive ward. I am, of course, not providing an answer here. What to regulatory environments, to the point that it is differ- I am doing is raising a question to help catalyze opinions, ences between these that are the real persuaders when it although I suspect that we are faced with a problem in comes to selection of this or that country as an investment which the number of unknowns exceeds the number of site. Third and last-and as an adjunct to the second con- available tools. sideration-the regulations governing factor markets, As a final warning against falling into temptation, I including the labor market, are the reality that underlies must add that the facts demonstrate that in integration basic entrepreneurial decisions. Tariff structure is merely systems, particularly on a regional scale, effective coordina- one dimension in the universe of inter-country integration, tion among countries on such concerns as macroeconomic and one component more-although a major one-in the policy still languishes in the "topics pending" category. taking of a final investment decision. To dwell just on this question, and attempt to come up with the perfect tariff Investment Dynamics in Economic Integration system, is to stay on the margins of the subject of invest- Areas ment site selection. It is indisputable that investment is a key factor in sus- All the same, countries need to ask themselves what tar- taining high levels of growth and improving social condi- iff-like measures affect foreign investors' decisions on the tions, the ultimate objective of all government policy. In geographic siting and the development of new activities. this equation, private-sector investment is pivotal: It has Not to go into this question thoroughly is to risk losing out been demonstrated that current levels of global investment seriously in the distribution of the results of integration, are too low, and that public-sector investment, apart from with the smaller-scale economies being left furthest behind. being limited for fiscal reasons, is inefficient in the major- In my view, discussion of this aspect of the investment ity of cases. In countries like Uruguay, with a low level of dynamics at work in the integration process is essential. domestic saving and a correspondingly low level of invest- ment, the crucial question is how to raise the latter. The Small Economies and New Integration Systems next question concerns the likely impact of integration Developments in the region show that the present integra- processes on the choice of foreign investment sites. One tion systems are expanding-Mercosur in particular. The intuitively tends to assume that expanded areas like Mer- addition of new partner countries to this system, the sig- cosur are focal points for foreign investment. This means nature of a framework agreement with the European Union that for the so-called small economies the challenge then in 1995, and the commencement of conversations on becomes that of orchestrating an appropriate set of condi- adjustments to the FTAA are the most recent proofs of this tions that will facilitate the process of attracting invest- expansionary trend. Its impact on the region's small-scale ment to themselves, and of subsequently plotting out the economies, and particularly on countries such as Uruguay, 156 ROUNDTABLE: THE SMALL ECONOMIES definitely will be major. One of its clearly positive effects is eralization are still few and far between in such sectors. that it is accompanied by market growth and extension, This is one of the challenges facing small countries like which help to moderate the fluctuations that external Uruguay, since continuation of this state of affairs adversely events provolce in the domestic activity levels of the coun- skews the natural costs of every integration process, pre- tries in the region. venting expansion of those activities in which such coun- However, the expansion of these agreements makes for- tries might enjoy major comparative advantages, and mulation of member states' participation strategies that therefore also the improvement of their welfare status. much more of a challenge. Two examples of the kinds of problems that have to be resolved are correct sequencing of Conclusion issues for consideration during the course of negotiations The dichotomy between large and small economies has no and configuration of the new economic spaces. practical consequences, with rare exceptions. There is thus Another problem will be how to rreat very sensitive sec- no specific rypology of challenges to be dealt with by the tors like temperate-zone agriculture, an activity that enjoys countries in each group. The differences between them are strong protection in new partner states while at the same more a matter of degree than an absolute linked to their time constituting one of the production mainstays of many size. Nevertheless, problems do exist, and we are con- small and already very open economies. Is there a possible stantly trying to resolve them, despite being aware there is compromise course that would free us from long transition no single solution to apply. periods sure to dilute the expected beneficial impact on the The issues reviewed here probably belong in that cate- general welfare? Our concern stems from our having been gory, too. However, given their significance where our witnesses over the last decade to unprecedented openness countries and progress with integration are concerned, I and growth in international trade. However, there were believe it is important to put them up for debate, and for areas that fell. behind as the process went ahead, with many us to be aware that in the course of our search for solutions countries still given to excessive protection of key sectors other new issues will appear that are going to present us like agriculture; the result is that the effects of global lib- with other challenges. 157 The Challenges to Small Economies FERN A N D O N A RA NJ O T PRECISELY DEFINE THE CONCEPT OF SMALLER ECONOMIES IS NO EASY TASK. THERE ARE different ways to determine the economic dimensions of a country. There are economies that may be considered small if we take into account variables such as land area and population, although some of these countries have achieved great economic development. In this sense, we could mention economies in different parts of the world that have achieved a high degree of development, being small in one or both of the above-mentioned variables-for example, Singapore and Taiwan, China in Southeast Asia or Switzerland and Denmark in Europe, which, in spite of their geographical dimensions and the size of their populations, have achieved very high economic devel- opment standards. The economic dimension also could be measured by taking into consideration the GNP, foreign trade or per capita income. All of these criteria are imprecise. In spite of the difficulty to define what the "smaller been recognized by the agreements of the World Trade economies" are, in general they present certain special Organization (WTO), where it is stated that the small characteristics, such as: economies may delay the application of some of the clauses * Reduced internal markets, resulting in the lack of ade- of the agreement and the developed countries should pro- quate infrastructure; less ability to take advantage of vidc technical assistance in different areas to the develop- the economies of scale, causing low productivity rates; ing countries as may be required. Likewise, in the North and, in general, higher costs per unit of production. American Free Trade Agreement (NAFTA) and Mercosur, * High depcndency, especially in the dcveloping world, schcdules for the reduction of tariffs arc stipulated for Mex- on just a few export products, which generates a ico (in the case of NAFTA), and for Uruguay and Paraguay greater economic vulnerability due to price variations (in the case of Mercosur). or to seasonal factors. The element of a special and differentiated treatment But even after considering these characteristics, to say was also recognized by the presidents and the heads of state that size is an insurmountable obstacle for development is of the hemisphere when they included the following para- not a valid argument, although some insinuate that it is. graph in the Declaration of the Miami Summit in Decem- There are some highly developed small economies and oth- ber, 1994: ers with a very limited degree of development. In the context of the globalization and economic inte- We recognize that the economic integration and the gration processes, the theme of "special and differentiated establishment of a Free-Trade Area are complex treatment" for the small economies arises. This element has objectives, particularly because of the difference in Fernando Naranjo was Foreign Minister of Costa Rica. 158 ROtUNDTABLE: THE SMALL ECONOMIES the levels of development and the sizes of the Guacimo, Costa Rica, the foundation for the new Central economies of the hemisphere. We will be aware of American process of development and integration was set. these differences while we work toward achieving the I refer to the Alliance for Sustainable Development, or economic integration of the hemisphere. ASD, which is the conceptual framework that the presi- . . ~~~~dents defined for the region.1 We will use our own resources, creativity and mdi- vidual caactesaelahoeofthntrntoIn the middle of the profound political transformation vidmua aiti, as well eash othoe of tchieint ton experienced by Central America during the last decade, communii:y to help each other in achieving our from war to peace, free elections and the strengthening of democratic institutions, the region has taken far-sighted And in paragraph 5 of the principles of the Plan of Action steps to ensure its insertion in the world economy. The it says: degree of liberalization of the economy has accelerated, lib- eralizing in good measure the trade regimes and the finan- While we work to achieve the Free-Trade Area of the g g g cial systems. All these changes were undertaken unilater- Americas, opportunities, such as the provision of technical assistance, have been foreseen to facilitate ally, as well as under the multilateral processes within the the interatinothesmalerconmiesandincras framework of the World Trade Organization. Additionally, the integration of the smaller economies and increase exports have been diversified toward light manufacturing, nontraditional agricultural products and maquila. More It is clear that the smaller economies have their very own recently, in Costa Rica, a small high-technology cluster is and special characteristics. To request longer time frames for being developed. trade liberalization is not, by any means, an evasion of the Sound stabilization policies, trade liberalization and responsibility assumed in the processes of integration and structural adjustment measures put in place by the gov- free trade. Moreover, it implies the adjustment of responsi- ernments in the '90s have induced important capital flows, bility and reciprocity to guarantee the sustenance of regional particularly direct investments. Nonetheless, in spite of accords. The first conclusion I would like to emphasize is these significant advances, Central American countries that smaller economies may not and should by no means must increase their internal savings capacity and remain isolated from the benefits of globalization. strengthen their macroeconomic stability, so that they may Central America has been participating in the transfor- generate productive investment processes that sustain eco- mations that have been occurring throughout the world nomic development over time, improve infrastructure and and that represent for the area a series of positive develop- provide better services and carry out activities with a ments. Nevertheless, the potential offered by the world higher added value and technological content. environment has not been sufficiently taken benefit of, due Central America has a population of 32 million inhabi- to some persistent structural and institutional weaknesses tants. In 1996, the GNP reached almost US$40 billion, and distortions that have been present in the region for with a per capita income of US$1200, total exports of the decades. region were almost US$8 billion (20 percent of the GNP), With the deactivation of the armed conflicts and the of which intra-regional trade accounts for US$1.6 billion establishmert of peace in Central America, the challenges (20 percent of the total). The main destination of the extra- of reconstruction, integration, structural adjustment and regional exports is North America, where more than 50 development have emerged as the main axis of the Central percent of the exports were sold. While the European American economic agenda. It is important to emphasize Union is the second most important market, trade with the that today t.lhere is an awareness of the close relationship rest of Latin American is only just beginning. between the urgency to advance in the economic and social Central America is a medium-size economy. Its market spheres and she capacity to sustain political achievements. is attractive for the world's industrialized countries and for Likewise, an "outward bound" vision of Central Amer- many Latin American countries even more, if we take into ica based on the consolidation of the regional market and account that the regional free-trade zone has been consoli- its insertion in the world economy is now in place. Almost dated and considerably improved in recent years. These three years ago, at the Central American Summit of achievements have been successful in spite of the disparity 159 TRADE: TOWARDS OPEN REGIONALISM in income levels and the degree of trade liberalization, The smaller economies have great opportunities to benefit which are very significant among these countries. from the formation of a free-trade area of the Americas, but, The Central American nations, and small countries in at the same time, they are the ones that have greater possi- general, must advance in regional integration schemes that bilities of incurring relatively higher costs if they do not pre- allow them to achieve more substantial benefits than those pare adequately to reap the most benefits from their interna- they may obtain by acting on their own. In unity there is tional economic insertion. Possibly, the smaller economies are strength, and on this subject our experience is valuable. the ones with the most opportunities to benefit from the Since August 1994 our presidents agreed that trade nego- hemispheric integration, although as a paradox, the risks and tiations would be made jointly as a region. the preparation process itself are greater for them. Before this decision was taken, Mexico and Costa The factors that determine the possible costs and bene- Rica signed, at the beginning of 1994, a free-trade fits are not precisely the most favorable in the great major- agreement, which has encouraged trade between both ity of the smaller economies. Factors such as the macroeco- nations, especially during 1996 and 1997. The rest of nomic environment, capacity to face competition, the the Central American countries have advanced signifi- investment climate, the attitude of the entrepreneurial sec- cantly in their bilateral negotiations for a free-trade tors, physical infrastructure, and education of human agreement. At the end of this phase the objective is to resources are essential to successfully take advantage of the reach a Mexico/Central America Free-Trade Agreement international trade. to establish a Middle America free-trade area, a market To incorporate efficiently into the processes of global- of 120 million inhabitants.2 ization it is important that the smaller economies be pre- During the next few months, a free-trade agreement is pared. This refers to the necessary factors a country requires expected to be signed between Central America and to achieve successful access into international trade. Panama, and initial discussions are under way to begin Among these factors one should mention the macroeco- trade negotiations with the Dominican Republic and with nomic conditions of price stability, healthy management of CARICOM. the internal and external debt, exchange-rate stability, and In the second half of 1997, a political, economic and non-macroeconomic conditions such as international com- cooperation framework agreement between Chile and Cen- mitments in matters pertaining to trade, human rights, the tral America is expected to be signed. This instrument will environment, intellectual property and labor regulations. establish a strategic and privileged relationship between To obtain maximum benefits from the process of eco- our region and Chile. Likewise, in the following months a nomic liberalization and to be able to compete efficiently, political rapprochement is foreseen with the countries of programs and specific actions are required. The liberaliza- Mercosur.3 tion does not generate automatic increases in the degree of Last May 8, on the occasion of the visit of President Bill competitiveness. There are structural factors that must be Clinton to San Jos6, the presidents of Central America changed to take advantage efficiently of the trade and indicated to the American president their satisfaction with investment opportunities that a process of economic inte- his commitment to extend the benefits of the Caribbean gration offers. In the case of the smaller countries, the Basin Initiative, as they expressed emphatically and plainly infrastructure limitations, inefficient capital markets, Central America's wish to negotiate, in the near future, a untrained workforce, and other deficiencies limit the abil- free-trade agreement with the United States. ity of these countries to take full advantage of these oppor- My second conclusion is that Central America expects tunities. Besides, there are a series of key microeconomic to be a key player in the integration process of the hemi- factors that are essential to increase productivity and com- sphere. We are fully committed to take Central America petitiveness of the economies.5 beyond the region. We feel that through the consolidation My third conclusion is that although the cost may be of the regional free-trade area, jointly with the integration high, the smaller economies must fully and decisively pre- process in different fields, and the joint trade negotiations, pare to face the challenges of globalization. Even though we will be in a better position to take advantage of the the cost of integration could be higher for the smaller opportunities of globalization.4 economies, the question we should ask ourselves is, What 160 ROUNDTABLE: THE SMALL ECONOMIES is the cost of not timely integrating and of not taking the alliance with nature, so that the 12 percent of the biodiversity of advantage of the potential benefits? the planet, which is located in Central America in less than 2 percent One last word on technical and financial assistance: The of the world area, may serve as an instrument for the promotion of sustainable development in the future. larger count:ries and the multilateral organizations must 2. The Presidents of Central America and Panama and the Prime play a key and decisive role in this field. Without this sup- Minister of Belize met in San Jose Costa Rica in February 1996 with port the way may be steeper for the smaller economies. President Ernesto Zedillo of Mexico. This was the second meeting of Shouldn't the hemispheric integration process precisely the eight heads of state. This last meeting, known as Tuxtla 11, repre- serve to help the smallest and the least developed? sents a historic landmark in the integration process of the region, We then can conclude in fourth place that technical and where the eight countries dccided to move on a preferential relation- ship on the basis of a joint agenda, which included political, economic, financial assistance, bilateral and m ultilateral, is decsocial, cultural and environm ental issues. In the second half of 1997 for the smaller economies. the first follow-up meeting of Tuxtla 11 will be held. 3. In principle, the first meeting of the Ministers of Foreign Notes Affairs of both regions will take place in September 1997. 1. ASD is based on four important elements: (i) Democracy as the 4. Very important steps are being taken to integrate the electri- basic form of human coexistence and only way to achieve well-being cal systems, passenger transportation, tourism promotion, conserva- and justice; (ii) The transformation, modernization and the liberal- tion of the environment and stock markets, among others. ization of the economies; (iii) The dependency on the market mech- 5. In the Central American case, Harvard University, through anisms and the harmonization of economic policies among the Mem- Professors Jeffrey Sachs and Michael Porter, with the support of ber States of CACM; and (iv) Investment in the human being, INCAE and the Central American Integration Bank, and with the particularly in education and health as a key element to give suste- active participation of the entrepreneurial sectors, are working on an nance to the political and economic transformation of the region and intense program to increase efficiency and productivity. 161 V7II. Roundtable: Agricultural Policy and Integration Regional Trade Agreements: The Case of Agriculture B RU C E L . GARDNER N RECENT YEARS THERE HAS BEEN ENLIGHTENING EXPERIMENTATION WITH COMBINED UNI- lateral and regionally integrated agricultural and trade policy reform in Canada, the United States and Mexico. Major steps in formal regional integration were the Canada-U.S. Trade Agreement (CUSTA) and the North American Free Trade Agreement (NAFTA), which came into effect in 1989 and 1994, respectively. During this same period, substantial domestic policy reforms occurred in all three countries, and the Uruguay Round agreement on agriculture in the GATT was concluded. These events provide an opportunity to investigate several issues and questions that arise concerning the prospects and pitfalls of regional trade agreements. The main conclusion I arrive at is that the agricultural Some commodities received further special treatment. provisions of NAFTA/CUSTA have been a surprising suc- Consider the case of sugar. Sugar in the United States is cess in opening agricultural markets. Success is surprising protected by import restrictions that keep the U.S. domes- because of obstacles that faced the negotiators due to the tic price of raw sugar at about double the world price (cur- dogged opposition of interest groups (commodity produc- rently 22 cents per pound for domestic raw sugar traded at ers) who expected to lose from liberalization, and obstacles New York, compared with a world price of 11 cents per those same political forces placed in the way of meaningful pound). U.S. sugar cane growers were concerned that implementation. The reasons for this success, and the NAFTA would seriously damage the U.S. sugar pricing implications for the broader issue of assessing regional trade regime, notwithstanding the fact that Mexico is a sugar agreements on an economy-wide basis, remain unclear, importer. Mexico, however, did ask for access to the U.S. market for 1.5 million tons annually in the NAFTA nego- Politics and Economics of NAFFA tiations (an amount only slightly below the total of current The chief political problem facing the U.S. negotiators was U.S. sugar imports). U.S. sugar interests worried that Mex- that the opponents of NAFTA had reasonable prospects of ico would import U.S. corn under NAFTA and make high stopping the U.S. Congress from ratifying and implement- fructose corn syrup for use in Mexican soft drinks to such ing any agreement. Therefore, they had to win the support an extent it would cause an excess supply of sugar in Mex- of these interests, or at least soften their opposition, by ico, which would then be exported to the United States building in protections for these industries-i.e., by not and ruin the U.S. sugar price-support system. The upshot liberalizing. So, for example, tariffs were reduced in easy was that Mexico received only very limited access to the stages, and "safeguards" were included to permit tariffs to U.S. market, for up to 0.25 million tons annually in years be restored if imports increased substantially. seven through 14 of the agreement. In the end, U.S. sugar's Bruce L. Gardner is a Professor in the Department of Agriculture and Resource Economics at the University of Maryland. 165 TRADE TOWARDS OPEN REGfOINALISM FIGURE IA U.S. Agncultural Imports: NAFTA Countries U.S. Agricultural Exports: NAFrA Countries 8,000 7,000 - _ . 7,000 - 6,000 - 6,000 - 5,000- 5,000- 5 5 ~~~~~~~~~~~~~4,000- 4.000- 3,000 3,000 2,000 1,000 1,000 0- 0- _ _ N , t' o C£ o' _E m v' from Mexico :: from Canada to Canada to Mexico economic situation was unchanged by NAFTA. In addi- negotiations on one make it easier to achieve the tion, sugar also largely withstood challenges to its pro- other. tected position in the Uruguay Round of GATT and in the 2. Regional agreements make it more difficult to U.S. farm policy reforms of 1990 and 1996. Any increased achieve global agreements such as GATT. U.S. imports of Mexican sugar will be offset by reduced Second, there are hypotheses about the fate and conse- imports of sugar from some other country (as also hap- quences of regional trade agreements, once reached: pened when the Reagan Administration in the 1980s 3. Regional agreements tend to be loosely specified increased selected countries' import quotas as part of its statements of intentions or more precisely specified Caribbean Basin Initiative). targets that are either easily evaded or already These and other anecdotes too intricate to be detailed achieved; therefore, the agreements tend to have lit- here provide observations relevant to the political econom- tle effect. ics of regional trade agreements. (For an illuminating 4. To the extent that regional agreements have real eco- description, see Orden 1996.) To organize discussion of the nomic effects, they tend to result in less trade implications more systematically, I will list four hypothe- between the region and countries outside the region ses, then will discuss the implications of further observa- (trade diversion). tions about CUSTA/NAFTA for the truth of these What were the political forces that permitted NAFTA? hypotheses. I will speak only to agriculture in the U.S. context. The First, there are hypotheses concerning the possibilities key underlying element of the situation, and of the truth of negotiating trade agreements. In particular: in hypothesis 1, is that the toughest negotiations involve 1. Regional agreements and unilateral domestic reforms each country's negotiating with itself, i.e., among its inter- tend to be complementary activities, in the sense that est groups. What made NAFTA possible politically was 166 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATION FIGURE lB U.S. Agricultural Imports: ROW and NAFTA U.S. Agricultural Exports: ROW and NAFrA 25,000 - 60,000 50,000- 20,000- 40,000- 15,000 7-- '-E ''30,000 - 1D 10,000 .... 20,000 5,000 10,000 ON ON ON ON ONC ON ON ON ON ON ON ON O ON ON ON a GoN N ON N ON ON ON ON GoN ON - - ,_ Ti-T | from NAFTA C from non-NAFTA to NAFTA H to non-NAFTA the existence of winners as well as losers among domestic and gradual elimination are coming into effect slowly under producers. U.S. feed grain, livestock and food processing NAFTA, beginning in 1994. However, the events that have industries expected to gain from NAFTA. They wanted received most attention under both CUSTA and NAFTA NAFTA to achieve substantive increases in their export are the many disputes that have arisen over trade expansion. opportunities. Some fruit and vegetable producers, espe- I will focus on protectionist impulses that have developed cially those in Florida, expected to lose, as did sugar and in the United States. Two cases in particular involved for- peanut growers. The losers mounted strong campaigns of mal complaints to the U.S. International Trade Commission opposition to NAFTA. But given the division of interests, (ITC) by U.S. producers and ended in trade-restricting bi- and a strong push for an agreement by both the Bush and national agreements: winter tomatoes and bell peppers from Clinton administrations, the liberalizing forces were able Mexico, and durum and spring wheat from Canada. to gain the day. After tomato imports from Mexico increased sharply in Nonetheless, the potentially losing interests were able the winter of 1996, growers led by those in Florida filed a to build protections for themselves into NAFTA and in petition for relief with the ITC, alleging dumping of toma- domestic leg,islation. Sugar, peanuts and other protected toes by Mexico. The ITC in July 1996 issued a preliminary commodities have been able to keep their price supports determination that the tomato industry had not suffered largely intact. In this respect the U.S. agricultural policy serious injury, but it left the dumping allegation open for debates and limited reforms in 1990 and 1996 fit in well a fuller investigation. In October, the United States and with NAFTA and helped it succeed. Mexico reached an agreement on minimum prices for Mex- Because CUSTA and NAFTA have in fact gone into ican tomatoes sold in the UJnited States. effect and we can look at the preliminary results, some evi- In the Canadian wheat case, U.S. wheat growers sought dence on hypotheses 3 and 4 is available. Tariff reductions protection based on a U.S. law that calls for limitations on 167 TRADE. TOWARDS OPEN REGIONALISM imports when such imports interfere with a price-support April 1994 the United States notified the GATT that it program. In November 1993 President Clinton, seeking to intended to modify wheat tariffs under GATT Article 28. assure wheat growers that their interests would be pro- The article's provisions require consultations with affected tected under NAFTA at a crucial point in the debate on countries. In August 1994 the United States and Canada U.S. acceptance and implementation of the free-trade announced an agreement that U.S. imports of wheat from agreement, wrote to the ITC stating his belief that wheat Canada would be limited through tariff-rate quotas to 1.5 was being imported "under such conditions or in such million metric tons for the next year. In September 1995 quantities as to render or tend to render ineffective, or the agreement lapsed, and there have been no further materially interfere with the price support, payment, and restrictions. Supply-demand forces in 1995-97 kept Cana- production adjustment program for wheat" (USITC 1994, dian exports of wheat to the United States below the 1.5 p. A-3). He therefore asked the ITC to investigate the million ton level. truth of this claim. In July 1994 the ITC, after hearing an Despite the expense and disruption caused by these and interesting array of evidence (reviewed in Babula, Jabara other trade disputes under CUSTA/NAFTA, it is worth and Reeder 1996), split 3-3 on the key issue of material noting that even in the disputed commodities, trade interference. Three commissioners concluded that material remains substantially more open than it was before those interference occurred, and three concluded that the evi- agreements. (In the period before 1989 wheat imports dence "could support the President finding either material from Canada, for example, did not exceed 0.3 million met- interference or no material interference" (USITC, p. 3). All ric tons.) In general, U.S. agricultural trade with Canada six recommended that the president impose some addi- and Mexico has increased. The table below shows some rel- tional barriers to U.S. imports of wheat from Canada. evant data. Both U.S. imports from and exports to Mexico CUSTA requires that a Section 22 import restriction have increased, notwithstanding the peso crisis at the end must be a response to imports caused by a substantial of 1994 when the peso depreciated by roughly one-half rel- change in either U.S. or Canadian policy. In addition, the ative to the dollar. Similarly, trade in agricultural products agriculture provisions of the Uruguay Round of GATT increased in both directions between Canada and the prohibit imposition of new quantitative restrictions on United States. Some fairly crude econometric evidence, agricultural imports, limiting the president's options dur- holding currency exchange rates and trend growth rates ing the 1995-2000 implementation period. However, in constant, suggests the increases in trade are significant and TABLE I U.S. Agricultural Imports and Exports under NAFTA (US$ millions, except for NAFTA share in percent) IMPORTS FROM MEXICO FROM CANADA FROM NAFTA FROM NON-NAFTA NAFTA SHARE WORLD 1990 2,610 3,152 5.762 17,008 25.31 22,770 1991 2,527 3,306 5,833 16,886 25.67 22,719 1992 2,372 4,102 6,474 lS,150 26.29 24,624 1993 2,708 4.620 7,328 17,653 29.33 24,981 1994 2,855 5,231 8,086 18,732 30.15 26,818 1995 3,780 5,559 9,339 20,654 31.14 29.993 1996 3,763 6,782 10,545 21,885 32.52 32,430 EXPORTS TO CANADA TO MEXICO TO NAFTA TO NON-NAFTA NAFTA SHARE WORLD 1990 4,197 2,553 6,751 32,612 17.15 39,363 1991 4,554 2,999 7,553 31,651 19.27 39,204 1992 4,902 3,791 8,693 34,237 20.25 42,930 1993 5,271 3,603 8,874 33,734 20.83 42,608 1994 5,504 4,5t3 10,017 35,687 21.92 45,704 1995 5,738 3,519 9,258 46,556 16.59 55,814 1996 6,145 5,446 11,591 49,840 18.87 61,431 168 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATION are attributable to NAFTA (Tweeten, Sharples and Evers- eralizing trade globally. It is the opportunities for tradeoffs Smith 1997; Gardner 1997). among interest groups within the negotiating countries The increase in North American market integration that make this so, and I believe that, in the case of U.S. appears to have increased Canada's and Mexico's share of agriculture at least, CUSTA/NAFTA and the Uruguay U.S. agriculcural trade. The two countries' share of the Round of GATT were mutually complementary rather value of U.S. agricultural imports increased from 25 per- than competitive; and the evidence available gives no indi- cent in 1990 to 32 percent in 1996, and their share of U.S. cation that CUSTA/NAFTA has resulted in less trade agricultural export value rose from 17 percent to 23 per- between the United States and countries outside NAFTA. cent over the same period. However, there has not been suf- ficient trade diversion to reduce U.S. agricultural trade References outside of NAFTA, which remains the source of 75 percent Babula, Ronald A., Cathy L. Jabara and John Reeder (1996). "Role of U.S. agricultural imports and the destination for 67 per- of Empirical Evidence in U.S./Canadian Dispute on U.S. Imports of Wheat, Wheat Flour, and Semolina," Agribusiness 12, pp. cent of U.S. exports. As Table 1 indicates (right-hand pan- 183-99. els) trade outside NAFTA has increased even more in dol- Gardner, Bruce L. (1997). "Canada/U.S. Farm Policies and the Cre- lar value than NAFTA trade since 1990. ation of a Single North American Grain Market," Research Sym- These data thus cast doubt on both hypotheses 3 and 4 posium on the World Wheat Market, Montana State University, above. NAFTA (which incorporates CUSTA) has led to real May. liberalization that has not been evaded. And intra-NAFTA Orden, David (1996). "Agricultural Interest Groups and the North American Free Trade Agreement," in A. Krueger, ed. The Politi- trade creaees do not seem to have occurred at the expense ca Economy of American Tra Picy, Chicago: University of of trade between NAFTA and the rest of the world. CiaoPes p 3-4 Chicago Press, pp. 335-84. However, the way of CUSTA/NAFTA has not been Tweeten, Luther, Jerry Sharples and Linda Evers-Smith (1997). smooth. If hypothesis 3 has failed, it is not for lack of try- "Impact of CFTA/NAFTA on U.S. and Canadian Agriculture," ing by domestic interests in all three countries. lnt. Agr. Trade Research Consortium, Working Paper #97-3, Finally, zL word about hypothesis 2, which is perhaps March. U.S. Department of Agriculture (1996). "NAFTA: Year Three," mostpenrtginentlismto thet o rever tsu ofen tgisonalimferen NAFTA Economic Monitoring Task Force, NAFTA-6, October. Open regionalism." What prevents open regionalism from U.S. International Trade Commission (1994). "Wheat, Wheat being an oxymoron is that regional trade agreements can in Flour, and Semolina," Investigation No. 22-54, Publication fact improve rather than detract from the prospects for lib- 2794, July. 169 Regional Trade Agreements: The Case of Agriculture ROBERTO J UN GUITO T| _ HE PURPOSE OF THIS PRESENTATION IS TO DISCUSS THE PROBLEMS POSED BY MACROECO- nomic events and mainly by the recent but persistent exchange-rate appreciation on the efforts made by Latin American countries to liberalize agricultural trade and to engage in regional trade agreements. From that perspective, I start with a brief retrospective view of the expectations held in the 1980s regarding the impact of import liberalization and overall structural reforms that the Latin American countries were beginning to adopt on agriculture. The second section deals with the evidence on what happened and what went right and wrong, emphasizing the impact of the exchange-rate appreciation on agriculture. The last section takes a look at what is taking place in the areas of the agri- culture commitments in the Uruguay trade negotiations and the Latin American regional trade agreements, and the possible conflicts that may arise as a result of the exchange-rate appreciation process. The Agriculture Consensus of the 1980s torted price incentives against agriculture." As a result, Coinciding with the economic literature of the 1980s agricultural output was lower than would have been the regarding the benefits of stabilization policies, trade lib- case under more neutral incentive structure (Bautista and eralization and structural reforms, a group of outstanding Valdes 1993). In essence, they argued that there had been economists linked to the World Bank and the Interna- discrimination against agriculture, that industrial protec- tional Food Policy Research Institute (IFPRI), who had tion acted as a tax on agriculture, that existing policies led been working on agricultural policies, developed what, in to overvalued exchange rates, and that policies had favored the words of John Williamson, could be termed their own the production of agricultural importables rather than "Washington agricultural consensus." This term alluded exportables (Krueger, Schiff and Valdes 1990). What is to what they considered should be the agricultural poli- most significant from the point of view of this presentation cies to be adopted mainly by Latin American countries, is their hypothesis, supported by the Argentine case, that The consensus, supported in well-documented country the agricultural sector would be the most benefited by the and cross-section papers, published among other places, in macroeconomic policies, among other factors, because the books edited and co-authored by our roundtable colleague, sector is particularly vulnerable to exchange-rate distor- Alberto Valdes, parted from the view that, in their efforts tions, given that its output is highly tradable (Mundlak, to promote industrialization, developing countries "dis- Cavallo and Domenech 1988). Roberto Junguito is an economist, working as a full-time member of the board of the Colombian Central Bank, Banco de la Republica. The author acknowledges the contribution of Felipe Barrera of the IDB in the regression analysis linking structural reforms with agriculture per- formance in Latin America. 170 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATiON Incidence of Structural Reforms in the Latin the reforms have provided during the last five-year period American Economies is not of a transitory nature. Recently, there has been an increasing academic effort, The topic of the particular impact of import liberaliza- including the World Bank (1993), to assess the overall tion and structural reforms on the agricultural sector has impact of the structural reforms and stabilization policies not been specifically addressed. In that regard, with the adopted by the Latin American countries in the late 1980s purpose of this round-table, Felipe Barrera, one of the co- and early 1990s. A paper by Lora and Barrera (1997) of the authors of the IDB study referred to above, ran a regression IDB measured the incidence of reforms on economic growth, model for the cross-section of Latin American countries through a cross-sectional analysis linking growth perfor- linking agricultural sector performance measured by its mance to an index of structural reform policies including GDP growth in the period 1993-95 (labeled as ZCAGRIC trade, fiscal. financial, labor and privatization measures. and WCAGRIC for the cases, respectively, of fixed and ran- They attribute to reforms an additional 1.7 percent to 3.0 dom effects) with the IDB-index of structural reforms percent annual growth of GDP, more than half of it of a tran- (ZDIP5 on WDIP5), and the real exchange-rate (ZCITCR sitory nature. In terms of particular reforms, they find that and WCITCR). The regression equation was also corrected trade liberalization has been the most important. by two variables common in the economic growth litera- On the o.her hand, they also find that stabilization poli- ture, namely the initial period GDP (noted as cies have helped to accelerate economic growth and argue WLPERC83) and education (WSCH83). that the joint effect of lower and less volatile inflation rates Figure 1 illustrates a positive association between agri- have added an additional 0.7 percent annual GDP growth. cultural GDP growth and the country specific index of Using a larger sample of countries, Fernandez and Montiel structural reforms. The regression analysis, shown in the (1997) conclude that the positive impact of reforms in Appendix, which assumes both fixed and random impacts Latin America conform to the worldwide experience and on agricultural growth, confirms that there has been a pos- attribute to stabilization policies an even higher impact on itive and significant relationship between agricultural growth. They also consider that the boost on growth that growth and the reform index. Such results back up the FIGURE 1 Association hetween Agricultural GDP Growth and the Country-Specific Index of Structural Reforms 0.12 Peru 0.1(1 U 0.081 Jamaica Nicaragua 0.06 = ~~~~~~~~~~~~~~~~~~~~~~~~~~~~Chite * Bolivia Brazil .04-Bi Paraguayfl Trinidad & Tobago o Costa Rica Colombia Guatemala 0~ ~~~~~~0U Honduras U ruguay 0.02_ Venezuela Ecuador * Argentina * Dominican Republic 0.00 _ iMexico * El Salvador -0.02 _ -0.02_ l ~ ~ ~ ~~I I I I 0.40 0.45 0.50 0.55 o.60 o.65 0.70 Structural Reform Index 171 TRADE- TOWARDS OPEN REGIONALISM "Washington agricultural consensus" regarding the exchange-rate appreciation. Undoubtedly, industrial pro- expected incidence of import liberalization, structural tection and trade barriers were negatively affecting the reforms and stabilization on agriculture. Latin American agricultural sectors. Another very important result that stems from the regression analysis regards the positive and significant rela- Agriculture Trade Policies tionship between the real exchange rate (measured in terms A final item I would like to discuss is the potential impact of local currency units per U.S. dollar) and agricultural of trade policies on agriculture. Such policies include the output growth. As Table 1 shows, however, in the period independent trade liberalization process discussed above, under analysis, the general trend in Latin America was one the 1994 GATT Agreement on Agriculture, and the trends of appreciation of the real exchange rate due mainly to the toward regional trade agreements. In this context, I ask the capital inflows that occurred in the first half of the 1990s. following question: To what extent is the current trend in Therefore, it appears that in the 1990s there were conflict- exchange-rate appreciation in Latin America an impedi- ing signals in terms of a positive stimulus of reform poli- ment toward agriculture trade liberalization, and do cies on agricultural output growth and a negative one from regional trade agreements represent a step toward a more the real exchange rate, and the international price trends. protectionist environment? From that perspective, there are two major points to be As an answer to the first part of the question, there is made. The first is that the direction of the real exchange wide recognition that Latin American nations undertook, rates in the Latin American countries was, ex post, the in an independent and unilateral fashion, a process of trade opposite of what was expected, despite the depreciation liberalization, including agriculture, that preceded the that took place in the second half of the 1980s; moreover, 1994 GATT Uruguay Round Agreement (Valdes 1996). it acted against agriculture's GDP growth. The second and On the other hand, the vast literature on the subject qual- mnore significant result is that the positive impact of the ifies the Agreement on Agriculture as a "historic interna- structural reforms overrode the negative impact due to tional milestone" and as the "most sweeping attempt to TABLE I Real Exchange Rate to the U.S. Dollar* (1986 = 100) ANNUAL AVERAGE YEAR VENEZUELA PERU ECUADOR CHILE BRAZIL ARGENTINA URUGUAY MEXICO COLOMBIA 1980 76.41 93.01 64.18 47.73 68.53 37.73 50.39 57.14 59.72 1981 72.59 85.40 60.81 43.89 64.38 45.74 49.23 52.62 59.57 1982 70.28 90.25 66.24 53.78 66.55 101.03 54.54 77.51 59.70 1983 68.29 102.91 67.93 69.48 89.15 109.83 96.73 88.38 63.20 1984 102.74 108.27 76.64 75.03 102.35 99.47 105.86 77.74 72.51 1985 103.55 135.91 68.74 97.22 109.62 116.18 115.30 76.74 85.05 1986 100.00 100.00 100.00 100.O0 100.00 100.00 100.00 100.00 100.0( 1987 148.72 67.05 111.37 98.34 90.92 99.91 94.00 103.40 105.14 1988 119.56 58.09 128.64 99.65 81.05 101.09 95.90 82.09 105.27 1989 155.59 51.07 135.85 97.19 69.22 14O.17 93.68 77.44 L12.09 1990 164.19 33.25 140.98 93.03 54.37 86.66 90.86 73.74 120.15 1991 154.37 35.48 134.59 91.13 62.44 58.57 80.34 67.20 117.68 1992 145.49 33.54 131.06 84.37 66.75 50.02 73.47 61.44 105.66 1993 144.35 37.51 117.01 85.92 60.20 46.92 64.08 58.01 102.70 1994 149.43 34.33 107.78 82.21 46.03 46.21 57.99 60.18 90.19 1995 115.14 32.59 105.07 73.70 34.37 45.98 52.61 87.45 84.55 1996 139.65 32.73 108.28 73.43 32.74 47.25 52.91 79.29 81.95 1997 ** 113.32 33.52 106.30 72.03 33.24 48.07 53.48 69.97 77.67 * The CPI is used domestically and internacionaLly. * Foe Venezuela, Aegeneina and Uruguay, refers to the average up to November, Brazil up to October Source: IMF Ince-nactonal Financial Statistics. 172 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATION liberalize agriculture" (FAO-World Bank 1996). More domestic economy. In the same vein it is argued, from the recently, the concerns have turned toward regional trade Argentine perspective, that Mercosur represents a transi- agreements and open regionalism, the subject matter of tory phase toward a complete insertion into the interna- this World Bank Annual Conference. tional markets (SAGYP 1995). It appears, at least formally, However, even though the surveillance of price and that the agricultural trade component of Mercosur is not trade policies (Valdes 1996) signals that trade liberaliza- directed to introduce protectionism as a result of the real tion in the agricultural sector in Latin America led to the exchange-rate events. elimination of most quota restrictions, the removal of Such a situation represents a contrast to the tendency export taxes, the reduction of the role of state trading agen- shown by the Andean countries. As Valdes (1997) argues, cies and a unilateral process of tariffication, the results also countries such as Colombia and Venezuela could be show that the trade liberalization process also coincided tempted to adopt protectionist policies toward agriculture. with a period of decline in real domestic farm prices due In fact, it has been shown that Colombia bounded tariffs at mainly to the exchange-rate appreciation that occurred "excessively high levels" representing a case of dirty tariffi- during the early 1990s. cation at the GATT level (Valdes and McCalla 1996). By the same token, in a seminar with wider geographi- Besides, it has adopted a prior license system-clearances by cal spectrum sponsored by the IMF, Wong and Kirmani the Ministry of Agriculture. The clearest expression of the (1997), dealing with trade policy issues, specifically Andean Group's trends toward the protection of agriculture pointed out that "the evidence from a number of develop- is the existence of a price band system, originally adopted ing countries and transition economies was that the real by Colombia and extended to the group, and the increase in exchange-rate appreciation, brought about by large capital measures of aggregate support relative to the value of pro- inflows or other reasons, could adversely affect the author- duction by its member countries during the first half of the ities' efforts to pursue further trade liberalization or might 1990s (GRAN 1997). Such policies at least coincided with even cause partial reversal in such liberalization." In that the real exchange-rate appreciation and with the increasing sense, an article by Kirmani, Papageorgiou and Michaely complaints by agricultural sector representatives regarding (1997) argued that exchange and trade restrictions often the negative incidence of macroeconomic policies on the act as substitutes, although he rightly indicated that agricultural sectors of the member countries. exchange-rate appreciation should be tackled by monetary and fiscal policies. The point that should be explained is Conclusions whether sorne of the trends in agricultural regional inte- This comment has aimed at discussing the importance of gration efforts could evidence elemenrs of protectionist trade liberalization and structural reforms on the agricul- nature used as a compensation to the exchange-rate appre- tural sector performance of Latin American countries. The ciation process. comments also point out the risks imposed by the process The growing literature on the subject suggests that of real exchange-rate appreciation on the efforts to pursue regionalism could be looked at as a supplement to multila- further liberalization. It formulates the hypothesis that terism, as an alternative to it, or as a path toward it (Cor- regional trade agreements could be used as vehicles for pro- den 1997). In that regard, pointing to the agricultural sec- tectionism, as the Andean Group decisions seem to suggest. tor, Robert and Schiff (1997) argue that the impact of regional agreements on welfare could be ambiguous, while References the impact of the Uruguay Agreement on Agriculture is Bautista, R. and A. Vald6s (1993). The Bias against Agriculture, Inter- positive, thus suggesting that regional agreemcnts should national Center for Economic Growth, and International Food positive, thus suggesting that regional agreements should Policy Research Institute. conform to GATT. Analyzing from the Brazilian perspec- Corden, M. (1997). "Regionalism in World Trade," Trade Policy rive the case of agriculture in Mercosur, Salazar-Brandao, Issues. Washington, D.C.: International Monetary Fund. Rezende-Lcpez and Pereira (1997) emphasize the impor- FAO-World Bank (1996). Implementing The Uruguay Round Agreement tance of regional trade agreements and argue that Merco- in Latin America: The Case of Agriculture. Santiago, Chile. Fernandez, E. and P. Montiel (1997). "Reform and Growth in Latin sur was built on the principle that less protection and America: All Pain, No Gain?" Banco Interamericano de Desar- increased competition have favorable welfare effects on the rollo, February. 173 TRADE: TOWARDS OPEN REGIONALISM GRAN (1997). 'Subsidios y Ayudas al Sector Agroalimentario en el SAGYP (1995). "Mercosur Agropecuario: Actualidad y Perspecti- Grupo Andino," Seminario sobre Subsidios y Ayudas al Sector Agroal- vas," Subsecretarfa de Economfa Agropecuaria, Ministerio de imentario en el Hemisferio Americano: Integracion con Competitividad. Economia y Obras y Servicios Publicos, Buenos Aires, Argentina, Bogotah, Colombia: Ministerio de Comercio Exterior. January. Kirmani, N., D. Papageorgiou and M. Michaely (1997). "Issues in Salazar-Brandao, A., M. Rezende-L6pez and L. Pereira (1996). "Trade the Design and Implementation of Trade Reforms in Developing Liberalization in Brazilian Agriculture: Qualitative and Quanti- and Transition Economies," Trade Policy IssJoes, Washington, D.C.: tative Analysis," Implementing the Uregicay Round. International Monetary Fund. Valdes, A. (1996). "Surveillance of Agricultural Price and Trade Pol- Krueger, A., M. Schiff and A. Valdes (eds.) (1990). Economia Pol/'tica icy in Latin America during Major Policy Reforms," World Bank de las Intervenciones de Precios Agricaolas en Ame&'ica Latina. Banco Paper 349. Washington, D.C.: World Bank. Mundial y Centro Internacional para el Desarrollo Econ6mico. . (1997). "Brief Overview of the Global Impact of the URA Lora, E and E Barrera (1997). "Una Decada de Reformas Estruc- and Lessons from Early Reformers," Implementing the Uruguaa turales en America Latina: El Crecimiento, la Productividad y la Round. Inversi6n ya no son corrio antes." Banco Interamericano de Desar- Valdes, A. and A. McCalla (1996). "The Uruguay Round and Agri- rollo, February. cultural Pnlicies in Developing Couintries and F.conomies in Mundlak, Y, D. Cavallo and R. Domenech (1988). "Agriculture and Transition." Food Policy. Economic Growth: Argentina: 1913-1984," Research Report 76. Wong, C. H. and Kirmani, N. (1997). Trade Policy Issues. Washing- International Food Policy Research Institute. ton, D.C.: International Monetary Fund. Robert, C. and M. Schiff (1997). "Regional Agreements and the World Bank (1993). Am6iica Latina y el Ca,'ibe: Diez Anos Despoies de GATT: Implementation Issues for Agriculture in Latin America," la Crisis de la Decada. Office of Latin America and the Caribbean, Implementing the Urlgeeay Roend Washington, D.C.: World Bank. 174 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATION Appendix DEPENDENT VARIABLE ZCAGRIC - ESTIMATION BY LEAST SQUARES Panel (4) of Annual Data From 1//1992:01 To 19//1995:01 Usable Observations 57 Degrees of Freedom 55 Total Observations 76 Skipped/Missing 19 Centered R**2 0.182838 R Bar **2 0.167981 Uncentered R**2 0.187241 T x R**2 10.673 Mean of Dependent Variable 0.0018725903 Std Error of Dependent Variable 0.0256694848 Standard Error of Estimate 0.0234144412 Sum of Squared Residuals 0.0301529831 Durbin-Watson Statistic 2.428761 VARIABLE COEFF STD ERROR T-STAT SIGNIF 1. ZIP5 0.1081640308 0.0356060620 3.03780 0.00364077 2. ZCITCR 0.045795412 0.0168476611 2.82410 0.00659124 LINREG WCAGRIC #WIP5 WCITC.R WSCH83 WLPERC83 DEPENDENT VARIABLE WCAGRIC- ESTIMATION BY LEAST SQUARES Panel (4j of Annual Data From 1//1992:01 To 19//1995:01 Useable Observations 57 Degrees of Freedom 53 Total Observations 76 Skipped/Missing 19 Centered R**2 0.243584 R Bar**2 0.200768 Uncenrered R**2 0.709206 T x R**2 40.425 Mean of Dependent Variable 0.0388303014 Srd Error of Dependent Variable 0.0309592561 Standard Error c,f Estimate 0.0276774980 Sum of Squared Residuals 0.0406003266 Durbin-Watson Statistic 2.220222 VARIABLE COEFF STD ERROR T-STAT SIGNIF WIP5 0.102841513 0.031926022 3.22124 0.00218321 WCITCR 0.034353866 0.013566549 2.53225 0.01433194 WSCH83 0.005898420 0.002808824 2.09996 0.04050865 WLPERC83 -0.006632666 0.003029140 -2.18962 0.03297551 175 Regional Trade Agreements: The Case of Agriculture ALBERTO VALDES _T _ HE PROLIFERATION OF FREE-TRADE AGREEMENTS (FTAs) IN LATIN AMERICA AND THE Caribbean during the 1990s is remarkable. Unilateral trade liberalization processes within individual Latin American countries created a new economic environment that fostered the emergence of Mercosur, NAFTA and a number of bilateral and trilateral agreements, while bringing life to otherwise sleepy legal agreements such as the Andean Group and the Cen- tral American Common Market. These new FTAs have proven far more significant than pre- vious arrangements. The perceived costs of remaining excluded from these free-trade agreements has sus- tained the trend's momentum, influencing more countries to join. Unlike previous trade-agreement initiatives in the reflecting the understanding among policymakers that the 1960s and 1970s such as LAFTA, EFTA, ASEAN in Asia, long-term growth potential of agricultural export markets and a number of others, agriculture is an integral part of abroad far exceeds that of markets within the region. these new FTAs. Two factors facilitated incorporating agri- Inward-looking FTAs would impose high domestic welfare culture as part of the agreements in Latin America: costs on their member countries by diverting trade, signif- * Agricultural trade reform took place within a larger icantly reducing potential agricultural growth. context of unilateral economic reform that included An optimistic view of trade liberalization holds that rather trade liberalization, deregulation and privatiza- than limiting the openness of FTAs to non-members, these tion-with no particular emphasis placed on any free-trade agreements will accelerate Latin America's integra- given sector. tion into the rest of the world, preparing member countries * Uruguay Round commitments involving the bind- for competition from NAFTA and elsewhere through region- ings of tariffs, restrictions on export subsidies, the ally-and subregionally-coordinated improvements in removal of widely used quota restrictions, and the vir- infrastructure, telecommunications, technical standards, san- tual elimination of state trading throughout the itary and phytosanitary measures, customs administration region served to circumscribe policy options and and financial services, in addition to the tariff reductions that reduce the likelihood that agricultural price support have already taken place unilaterally. measures would be reintroduced. The experience of Mercosur's associate members, which Compared with the 50 years that preceded the current are not subject to Mercosur's common external tariff- period, the prevailing economic environment in Latin namely, Chile and Bolivia (and Peru, which applied for America today is much more strongly oriented toward an membership in 1997)-and the Canada-Chile free-trade efficient regional integration vis-a-vis the rest of the world, agreement are recent developments well worth monitoring Alberto Valdes is Agricultural Adviser of the Rural Development Department of the World Bank. 1 76 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATION as they incorporate innovative approaches and may come to guidelines and the Codex Alimentaris for sanitary and phy- be extended to more countries in the region. tosanitary regulations would serve to promote the harmo- nization of national standards. Selected Issues of Implementation There are several sensitive issues of implementation that Rules of Origin require coordinated action among members of FTAs. For The monitoring of how rules of origin are applied is a this panel, I have elected to highlight six of them: dispute complex administrative issue for homogeneous products resolution, harmonization of national standards, rules of (commodities) in general, which is the case of many agri- origin, dealing with exceptions and tariff escalation, the cultural products and which requires some administration exchange rate policies, and the non-tariff agenda. in common. Under a customs union such as the one within Mercosur, rules of origin, in principle, is not an issue and Dispute Resolution thus it does not require an administration.' The incorpora- How expeditiously and objectively do Mercosur, the tion of "associated states" that are not subject to the com- Andean Group, NAFTA and the Central American market mon external tariff (such as Chile and Bolivia in Mercosur, handle disputes among their members? Are most disputes and perhaps Peru in the near future) does require monitor- resolved on a technical basis, or do they resort instead to ing, which can be done at lower administrative cost by a political negotiations? Do these FTAs have a politically central agency, but this is not necessarily specific to neutral and regional (rather than bilateral) body to deal agriculture. with trade disputes? The absence of an effective and credi- ble system for settlement of disputes would hinder the Dealing with Exceptions and Tariff Escalation process of integration. The current FTAs include several exceptions, such as sugar Disputes among member countries of Mercosur so far in Mercosur, maize and beans in Central America, products appear to be resolved at the political level, which, one under the price-band scheme in the Colombia-Chile would reasonably expect, works to the disadvantage of Agreement, and some others. There is furthermore a sig- smaller economies. The same tendency is evident within nificant degree of tariff escalation at higher levels of pro- the Andean Group. NAFTA, which is reputed to have the cessing, with the end result that effective rates of protec- most developed legal framework for dispute resolution in tion on semi-processed and processed goods, both with the hemisphere, and which allows member countries to regard to member and non-member countries, are higher choose between regional and multilateral forums, should than the nominal tariffs on those products. eventually provide valuable insights. Exchange-Rate Policies Harmonization of National Standards In the past, developments on the foreign exchange markets FTA trade in primary and processed agricultural products, in this region have been an influential factor on the pres- and on perishables in particular, is very sensitive to national sures for higher agricultural protection. Continuous appre- differences in technical standards, sanitary and phytosani- ciation (i.e., fall of the real exchange rate) is already a tary measures, and customs formalities. And red tape on source of tension among members of some FTAs. The sig- imports frorn other member countries, such as those within nificant appreciation of the local currency observed during Mercosur, remains heavy in a number of countries. these last three to five years in Colombia, El Salvador, The question of how to facilitate progress in harmoniz- Brazil and Chile (and until a few years ago in Peru and ing nationaJ regulations leads to a number of possible mea- Argentina) has induced a reduction in "real" domestic farm sures. One is the establishment of a secretariat to design prices, generating strong resistance to further agricultural and implernent common rules and standards between trade liberalization in these countries (Valdes 1996). members. Another would be a system of mutual recogni- Within Mercosur, changes in the Brazilian-Argentine tion of one another's regulations and standards, which is bilateral exchange rate have been recognized as a strong less threatening to members' policymaking autonomy. force in redirecting trade flows of farm products between Whichever course is followed, World Trade Organization these two countries. 177 TRADE: TOWARDS OPEN REGIONALISM The Non-Tariff Agenda tries, with most of the tariff quotas introduced in only two Under the new regime based on tariffs prevailing in agri- countries. culture in this region, countries may be expected to turn An important implication of this is that under the increasingly to a nontariff agenda, such as sanitary and URA, less than 20 percent of agricultural products were phytosanitary measures and technical standards. This could tariffed (the rest are covered by ceiling bindings) and thus be a particularly complex challenge in the integration have no recourse to the "special safeguard" provision on process. The evidence presently available is too sketchy to agricultural imports. These countries received no credit for enable assessment of whether, where, or to what extent this having converted to tariffs before the URA. is happening, but it warrants careful monitoring. Limited to applying quantitative restrictions and unable to use the special safeguard provision, one indirect Is a common agricultural policy necessary for an efficient method of protection becoming common in the region is customs union? Unlike the European Union, Mercosur and the fluctuating tariff that is applied as part of price-stabi- the Andean Group do not have a common agricultural pol- lization schemes, such as price bands in the Andean Group icy, and in this author's opinion, they should not. A "good" and in Chile, and surcharges in Peru. Under the new rules common policy is of course a good cause. However, consid- and commitments accepted in the URA, price bands are erations of political economy are always important, and to legal only when the sum of the basic tariff and the sur- give agriculture, or any other sector, a special standing and charge does not exceed the bound tariff level, and it is not its own hureaucracy is likely to make subseql]ent reforms linked to any internal reference price. However, given that more difficult even when circumstances change. Adminis- that the specific duty is charged to the CIF (cost-insurance- trators of such bureaucracies are likely to become captives freight) price of each shipment, it is likely that imports of producers' lobbies, pressured to introduce or maintain from different countries will be charged different tariffs, subsidies and protection. Hopefully Montevideo will not amounting to discriminatory treatment among different become the Brussels of Mercosur. foreign suppliers (Cordeu, Valdes and Silva 1997). However, while not necessarily requiring a common This variable import-tariff scheme is a contentious agricultural policy, the six issues highlighted above do international trade issue, and potential legal problems require a coordinated action and close monitoring. How is loom on the horizon. Furthermore, within Latin America, the current coordination system working, and which insti- products covered under such schemes have generally been tutional set-up is most effective in guarding the interests given special treatment, including exempted, or slower of consumers as well as producers? These are increasingly tariff reductions under bilateral agreements (such as those complex issues that continue to unfold over time. between Colombia, Venezuela and Mexico, and between Colombia and Chile). Hopefully, the next round of multi- Price Stabilization Policy: A Contentious Issue lateral negotiations will consider extending the special By 1993 most countries in Latin America had imple- safeguard provision to these countries in exchange for the mented bold programs of unilateral trade liberalization, removal of fluctuating tariff schemes. eliminating most quantitative restrictions on agricultural imports, removing taxes, quotas and licenses on exports The Importance of Infrastructure Development and dramatically reducing or eliminating state trading in for Regional Integration of Agriculture agricultural trade. Particularly in South America, transport costs usually rep- In their commitments under the Uruguay Round resent a high share of the CIF value of agricultural prod- Agreement (URA), the overwhelming majority of Latin ucts. Mercosur is giving a great impetus to coordinated American countries adopted the "ceiling bindings" efforts by both public and private investment in roads, modality, in which a ceiling binding was proposed for pre- ports and railways. These efforts should significantly viously unbound tariffs. By 1995 all agricultural tariffs in reduce transport costs and thus contribute to further inte- the 17 Latin American countries examined by Carson gration in agricultural markets among Mercosur members. (1997) were bound under the URA, while proper tariffing Elsewhere in South America, particularly in the Andean under the URA took place in only seven of those coun- region, the magnitude of the resources needed to improve 178 ROUNDTABLE: AGRICULTURAL POLICY AND INTEGRATION transport ancl communications is enormous and likely to backsliding is evident, such as the effect of "convenios de remain a substantial bottleneck for further integration. absorcion" in Colombia, and the recent increase in surcharges in Peru. Concluding! Remarks: The Slower Pace of Trade Liberalization in Agriculture References As elsewhere in the world, trade liberalization and integra- Carson, Chris (1997). "The Removal of Quantitative Restrictions, tion of agriculture in Latin America advances at a slower Tariffication and Minimum Access Provisions as They Affect paceof thagriculturen that sti o erictradable nsetor lowever, Importables," in Cordeu, Valdes and Silva (1997), cited below. pace than that of most other tradable sectors. However, M Cordeu, J. L., A. Vald6s and E Silva. (1997). Implementi&g the Uragceay Mercosur has not applied high levels of protection against Round Agreemient in Latin America: The Case of Agricultere. Santi- agricultural imports from non-member countries, main- ago: FAO and World Bank. taining with few cxceptions a tariff range of 10 to 20 pcr- Valdes, Alberto(1996). SurveillainceofAgriculturalPriceandTradePol- cent. Unfortunately, the same cannot be said of the three icy in Latin America diring Major Policy Reforms. World Bank Dis- remaining rnembers of the Andean Group (Colombia, cussion Paper No. 349, World Bank, Washington, D.C. Ecuador and Venezuela), which maintain a controversial price stabilization scheme with a strong protection effect on Note eight agricultural products (and on most of its substitutes 1. Starting in 1994, the Andean Group adopted a common exter- and complernents, covering more than 100 items). Some nal tariff but limited it to Colombia, Ecuador and Venezuela. 179 7III. Infrastructure and Integration Promoting Latin American Trade: The Role of Infrastructure S RI -RA M A I YER H AN S J . P ET ER S AT THE OUTSET OF THE 1980s THE VALUE OF LATIN AMERICA'S FOREIGN TRADE REACHED a level just above US$200 billion, with no major increments during most of that decade. The scenario changed dramatically in the early 1990s when substantial advances in the region's annual trade performance materialized. Between 1992 and 1996 Latin America's annual foreign trade transactions increased 58 percent from US$315 billion to almost S$495 billion. While exports exceeded imports significantly in 1992, the 1996 trade data reveal that, for the region as a whole, the value of exports was only marginally higher than the value of imports during that year. Most of Latin America's countries experienced trade deficits. In 1996 Brazil had the highest deficit-US$4.5 billion-but other countries, like Colombia and Peru, also showed sig- nificant shortfalls. Yet due to huge surpluses in Mexico and Venezuela, Latin America's total trade ended on the plus side. Latin Ameirica's Trade within a Changing raw materials and semi-processed goods. But the interna- Competitive Environment tional markets for these commodities have been volatile as The region's main sources of imports and destinations for demand patterns have changed. Only in a limited number exports are the United States (46 percent of total foreign of Latin America's economies has trade growth been trade in 1996), the markets of the European Union (19 per- spurred by export diversification. In many instances the cent), Japan, and the "Asian Tiger" countries (13 percent). share of manufactured goods in total annual merchandise The most noteworthy development relates to the explosive exports has remained modest (see Table 1). The World growth of intra-regional trade among the Latin American Bank's 1997 assessment of global economic prospects con- nations, which increased from US$60 billion in 1992 to cluded that the share of technologically advanced goods in US$110 billion in 1996-almost 17 percent per year on Latin America's total exports is still insignificant when average. By comparison, the overall annual growth rate of the compared, for instance, with the performance of Asia's region's foreign trade was 12 percent during the same period. industrializing economies. In spite of these impressive growth trends, Latin Amer- With growing incomes in the regional economies, it can ica's share in world trade was low-barely 5 percent in be expected that the demand for imported goods will con- 1996. With a few exceptions most regional economies still tinue the growth trend that has been apparent since the rely on their traditional export commodities-essentially early 1990s. By implication there is an inherent danger of Sri-Ram Aiyer was Director of the Latin American and Caribbean Region Technical Department of the World Bank. He now is Director of the Republic of Korea Country Unit in the East Asia and Pacific Regional Office. Hans J. Peters is Principal Transport Specialist in the Finance, Private Sector and Infrastructure Unit of the Latin America and Caribbean Regional Office of the World Bank. 183 TRADE TOWARDS OPEN REGIONALISM TABLE I Latin America's Foreign Trade-Perfornance Indicators MERCHANDISE EXPORTS MANUFACTURED GOODS NOMINAL EXPORT GROWTH US$, MILLIONS (FOB) % OF FOREIGN TRADE 1988-89 TO 1993-94 ANNUAL FROM EXPORT AVERAGE DIVERSIFICATION 1980 1995 1980 1993 (%) (%§) Argentina 8,020 20,967 23 32 8.6 1.9 Bolivia 942 1,101 3 19 3.8 n.a. Brazil 20,100 46,506 39 60 3.1 1.3 Chile 4,710 16,039 10 19 7.7 2.8 Colombia 3,920 9,764 20 40 7.4 3.0 Costa Rica 1,000 2,611 34 33 12.2 1.3 Ecuador 2,480 4,307 3 8 9.2 1.5 El Salvador 967 998 36 48 11.7 1.2 Guatemala 1,520 2,156 46 48 12.7 6.6 Honduras 830 1,061 12 13 10.6 0.8 Mexico 15,600 79,543 12 74 10.6 0.2 Nicaragua 451 520 13 7 3.3 2.7 Panama 358 625 9 16 2.9 n.a. Paraguay 310 817 12 17 (2.1) n.a. Peru 3,900 5,575 18 17 3.4 1.4 Uruguay 1,060 2,106 38 43 8.4 5.7 Venezuela 19,221 18,547 2 13 7.2 1.7 Sozirce: World Development Indicarors 1997 (The World Bank) widening gaps between annual imports and exports. To practices that are still prevalent throughout Latin America. avoid the build-up of sizable trade deficits, focus will have It is therefore of prime importance for regional govern- to be on efforts to increase exports. But continued reliance ments and the local business communities to take decisive by the Latin American countries on traditional export com- steps to establish infrastructure systems and service provi- modities is unlikely to achieve such objectives-even if the sions that are conducive to improving trade performance. quality and prices of locally produced goods are compati- ble with or superior to those of the international competi- The Elements of Changing Competitive Practice tion. Although quality and price are still important crite- There has been growing evidence of fundamental ongoing ria for success in trading, competitive practice in today's changes in the organization and management of industrial markets has established even more stringent determinants and marketing processes throughout the world economy. of competitiveness (see Box 1). These developments gained momentum during the 1970s. The challenge for the governments and business com- Since then major breakthroughs in communication and munities all over Latin America will be to adjust to the transport technologies materialized. Much of the innova- changing practices in international trading. This observa- tion of the past 20 years has been directed, implicitly or tion gains special importance in view of the fact that more explicitly, at the very task of shrinking distance. The con- than 70 percent of regionally produced export commodi- tinuing transport and communications revolution has ties are targeted at markets in the OECD countries. It is in brought about an enormous reduction in distance as a nat- these countries where evolving commercial practices in ural barrier to trade and investment. Technological change manufacturing and trading have had profound effects on has intensified the pressures of international competition. industrial relations and purchasing behavior. Combined with deregulation in the transport and commu- These new commercial practices are, above all, critically nications sectors, it has led to substantially lower costs and dependent on the availability of efficient trade-supporting improved quality of these services. In virtually all cases, infrastructure and services. The physical characteristics of competition and deregulation of entry barriers have been a these support networks and their organization are in many key factor driving down prices, raising the quality of trans- ways radically different from traditional structures and port and communication services internationally and 184 INFRASTRUCTURE AND [NTEGRAT[ON materials acquisition, production and distribution to the The ChaugIDg Determinants 4~ ~end markets. Particularly evident were the high costs of associated inventory holdings. Parallel to these develop- Newinfrmaiontecnolgis ae tansormng he ay ments, market demand had starred to shift much more selerssel ad byer b~y.As reult tadiionl mr- rapidly than during any time before. Better information kerig sraigie ar stadiy bconiin les pducive available to the consumer through improved information and cosi-effective. ~~~~networks was a key reason. As consumers move faster Incrasig uso e' soh iati is haning he ery between suppliers and product lines, there is more need for natue ofcusomervendr rlatinshis.~Custmersare businesses to anticipate quickly and correctly what will be demadin~ *~~posiv, cnsulatie, alueaddd prt- demanded, and to put into place a foundation for manufac- irerhips whle echeing onvetioal, ranactinal ture and delivery of products that can be shifted economi- relanonships. ~~~~~~~~~cally and quickly as tastes change (see Box 2). Innotis annfctuing nd dstriutik rtazeges,Manufacturers and traders in the industrialized coun- elecronc daa itercang andsysem slutons ill tries were confronted with the high costs inherent in their contnueto volv, aterng ~e wy gods re ar- traditional business organizations and at the same time chasd. Byer wil reuce he umbe ofapprved with increasingly volatile market demand. The need to natonater, ornde srccesadnd mores resonsv services-nmn intanes playing large r rolnethn t echolog incrasnly advances have extuidd the reapons glba productionity laessres,ntew glvomnualctueringr aldleoving thcosprpiona psroducti Crenond rremret il civecmettv staes vermuc log. rditanes, aeve foring.sales wraith ifrni.in yadn au r uhsrie ations to mrefou reshoourcesdaallycope, infrmtnwpior-ifntedencieig-stmrgal utrni r sivies snd tirew.stieproducts and servithe dmns of uemitnfag ohedueeadidstilcins evc a betr iniarmationets ow moiorn becmeaneseniordtoinatiferotnt.g gpaie beamke erossibetmon istor. sallpaes orga ovizangs trodfusc nttmtepae.Sric a eoea.motn thfro it ever mcarceriaresourceso mhoreh selevely chose.rmsendiioatidutieevn oe mora hrcsingh p staenta cutomter consmert. tha-ch mincinoofemonitmdeandosld,Th rvaloe, theorcompeticorncies in ill headitionalosetaup tfis Amngth mjo cnsquncs,loertrnsor cst kndofpefoinin& rdiy;bishes mnaer hv TRADE: TOWARDS OPEN REGIONALISM reduce costs and to become more responsive to changing These developmeats seemed to open prospective market customer preferences forced these enterprises to engage in opportunities for aspiring industries in developing coun- substantial restructuring of their corporate practices. Ris- tries and held the promise of accelerating local economic ing factor costs in their domestic markets induced out- growth. Businesses in a number of developing econ- sourcing of intermediate production processes to offshore omies-notably those along Asia's Pacific rim-have been industries in countries with abundant low-wage labor and able to capture these new opportunities and are now active other conditions of comparative advantage (see Figure 1). participants in the global manufacturing and trading net- Traders also started to explore less costly supply sources. works (see Box 3). A key element behind such success was Growth in global production is deepening the economic competition. As it turned out, two main forces underlie the integration of countries. International outsourcing of pro- increased intensity of competition: greatly reduced barriers duction inputs is gaining strength. Because of greater com- to trade and investment in these countries, and the falling petition, firms in industrialized countries are being forced costs of international communications and transport. to take advantage of lower-cost production opportunities Global sourcing is not an exercise in finding cheap around the world, to disaggregate production processes, sources of supply or suppliers of questionable quality in the especially in manufacturing, into stages that are out- backwaters of the world, however. It is a process that com- sourced to different countries according to their compara- panies use to select the suppliers that offer the best value. tive advantage. This process which has been described as a The promise of these prospects is dampened for many coun- "slicing up of the value change," is an increasingly impor- tries in Latin America by the apparent inability of their ranr aspect of global production today, public administrations and commercial communities to effectively address the ever more stringent requirements of FIGURE 1 foreign businesses for reliable deliveries in increasingly Trends in Outsourcing Production Inputs shorter time intervals, and with low reject rates. The diffi- Sample: 628 North American and 240 European Industry culties encountered in these countries are a reflection of not Groupings being able to grasp and thus to internalize the consequences 140- of technological progress and increasingly volatile market behavior. What appears to be particularly cumbersome to adjust to are the strategies adopted by the international ser- 130- vice industries in reaction to changing global practices in Offshore P,urchases manufacturing and trading. A greater traditional inward of Processed Goods orientation in Latin America until recently provided foreign 120- affiliates with strong incentives for domestic production and consumption, whereas in more output-oriented East Asia, export propensities were high and rising. As considerations of economic efficiency and interna- tional competitiveness have gained importance, the role Z and priorities of governments have been called into ques- _______________________________________ ntion, and the need for new forms of regulations has become 100- apparent. In most industrialized countries, technological progress and greater competitiveness have been the cause and effect of a trend toward deregulation and simplifica- 90- tion, as well as updating the regulatory machinery, in turn .11 _Ui enabling a more flexible and adaptable integration of ser- vices into productive functions. With the growing com- 80 - plexiry of service activities, the role played by governments 1987 1990 1995 as providers of basic services is shrinking. Their place is So0e-: The Wortd Bank being taken by private agents. In this new context govern- 186 INFRASTRUCTURE AND INTEGRATION Globnoeal utCnetionofmres He sand MatetinuofacturesSucceeded eta ffciees hs e nrsrcurssinf Tis aevertcllymintegsratedormtextilean grmelainsbtwmanfc plant tnfanriands buysha Fortumnes plaustricsan toyseFOB ero- thesuplers and exorer. o Thervcmanes.akysple ffn Jifou e nd ipcarrongs for e tiracnsortes Athrog Hoingin ishe garmentslcil tohMden FsionsthembHt indusseldorfuc Kbong mandong beach,t Caiforneiaandusrhliencunviaeth U.S. (Germany)tact Whiletfasioenes designs aroaden the rail-basi- ed leeatindo b nridgetouitsuhoe base iond ho hallha Taietpeoiionheadquares,tiorting srakces pace infrastrecently tmprtansatin aefet donte onrad pejusrmanc-time Easis whichs estab ishe offsoreatt mnuftactrn base dimnsin Tonbuervi,cjst allnow is.h n detooraniz its cretailngmrkt activitiesbwith anort onfrBangkokur (Tehaingd) Thegnewfiactory. imanports- minimgale invntories.ntwrs reicrainl (ionda and conm icttonsfo h infrstrGulfureghvtaiion(e as llyn enequrpieingh up-catountry Talan theroratihas tobee sucesdinto bentheicyar is asuppl siedfrmal chsemictal industry d compe ine expbandingetwrsbusiness vouew inflessthantu10 y aear dfro infJavato Thoo e ponttoaodutse.B airte bfbek erning ansuppleier usof gmportnalomatocrkteriasegmen to fhromgh Bango tehologysseldrfvine all this,erneat Ftreguhas baeomingma prelatipely plesimpfthorthn Amdeermiic ngmar tions bevyr gnresponsv tof inorertructu iehs "Qguicne e to thitry s cmeitiveness.'y&imMls srtg oasm PLAge sTeCS4) T whDE: snfanlad SethoresInc of Cntonyinnai cAis inthe UsinitedxStates. siam, Milsopackos rice readynfsr Ohimsof (Unsited statems) purcaet ag a mounter s of ph vlas paticerti shelvesinacordan oe wihtervspveciald reunivreme,nts mdentseretinctes imotatroeo regulatoryswls, enstiuriongath requirmetstatenbe thesetntalo systtengmsc tof ahefievre thei sondopraio o mrkt ad roecio o cnumrs ptetil ffcivnes.Ths ewinrstucur i s1sgnf TRADE TOWARDS OPEN REGIONALISM BOX 4 The New Dimensions of Trade-Supporting Infrastructure Changing international practices in manufacturing and The new generation of AI has had profound effects on trading have entailed a different demand for supporting the economics of manufacturing industries in industrial- infrastructure. The new generation of infrastructure ized countries. Increasingly, Al has important conse- covers a wide spectrum of technology enhancements quences for the Latin American countries as well. While and specific applications related tO transport and traditional considerations such as costs of labor, utilities communications. These breakthroughs have been and space remain important, regional manufacturers referred to as "Advanced Infrastructure," or Al (Mody, must also have access to global networks in comumunica- Ashoka, Reinfeld 1995). To be specific, Al combines tions and transportation in order to meet the competitive basic transportation and communications technology demands in modern business environments. The premise with information technology, thereby creating an is that if the regional economies do not develop Al, they enhanced service capability. will gradually lose their competitive edge and run the Other than technology enhancements, a key distinc- danger of experiencing declines in their manufacturing tion between Al and basic infrastructure is the increased industries. importance of value-added services associated with AL. Merely having adequate physical infrastructure, such as The idea is that service providers, as well as the services ports, roads and telephones will not be sufficient to pro- themselves, are actually a part of infrastructure, which is mote the activities that growing economies need. This very different from the physical way infrastructure has physical infrastructure will have to be enhanced with tech- ordinarily been characterized. Whereas basic infrastruc- nology and applications (i.e., Al) that allow firms to meet ture is primarily supply-oriented, Al must be examined the competitive demands for shortened product cycle from the demand side by including considerations of times and improved customer service. design and applications that are tailored to meet the The inclusion of value-added services as part of Al needs of specific users. means that governments need to change how they plan This new dimension of service-the value that it and administer this sector. To create an effective Al, they adds-not only expands the definition of infrastructure, must not only plan the physical infrastructure, but must but also makes it more important to create an appropri- also create the appropriate environment in which the ate environment in terms of policies, standards and facilities and the service providers will operate. Govern- resources, in which to develop Al, Without this kind of ments must recognize that availability of Al is becoming complementary support, the effectiveness of Al is simply extremely critical to firms if they plan to compete in the lost. In fact, in industrialized countries, it wag only after global marketplace. This new generation of infrastruc- major policy reforms were introduced and the service sec- ture depends to a great deal on the freedom and fiexibil- tor became more developed and prominent that these Al ity of service provicders and users of this infrastructure. systems first occurred. of economic progress in the Latin American countries. This drivers that affect market behavior and determine interna- creates the need to monitor more closely the causes and tional competitiveness in the world economy. effects of changing international trends in manufacturing Because a large share of exports from developing coun- and trading. Such assessments will facilitate the preparation tries is targeted at markets in the OECD economies, a of practical guidelines that can enable governments and principal World Bank investigation concentrated on business communities in the region to adjust more effec- industrial and trading practices, as well as consumer tively to the continuous restructuring in the international behavior in the OECD markets. Through an extended market environment. The World Bank has conducted sev- cooperative network with professional and commercial eral surveys to ascertain the changing nature of the principal organizations in several countries, it was possible to survey 188 INFRASTRUCTURE AND INTEGRATION almost 1,500 businesses involved in manufacturing, trad- ing and the provision of supporting services-such as Wi4is LogsticsManagemet,? informatics, transport and warehousing-in three areas that are impJrtant in this context: Japan, North America TeUS oni fLgsisMngmn eie and Western Europe. The key finding was that-ubiquitously-the strategic an c tools used to become more responsive to volatile market se dlare4i atn inorin t demand and to cut business costs implied a new manage- point of consum for the purpose of conforming to ment approach. This approach aims to orchestrate the customr reui emens." functions of materials acquisition, production and market- _____a____________________________________________ ing. It is called logistics management (Box 5). The princi- pal task of eFfective logistics-management organizations is infrastructure refers to that segment of public infrastruc- to reduce inventories to the lowest possible levels through ture-in particular transport and telecommunications- streamlined supply and distribution channels. To be effec- that encompasses hardware, software and services that tive, logistics management critically depends upon allow users to move goods and information more rapidly advanced inIfrastructure and information technology. and reliably. Information technology refers to that set of Advanced infrastructure, information technology and technologies (hardware and software) that relate to the logistics management may be thought of as three perspec- storage, processing and transfer of information. Logistics tives of an enhanced business environment that is enabling management refers to business administration techniques, a modern economy to operate more efficiently. Advanced in terms of systems and approaches, that allow a firm to FIGURE 2 Logistics Cost Components OECD Industry Averages Represeiting Almost Half ofl,ogi tics Expenditures Inventory Holding and Processing is Storage the Highest Cost Burden. S2oag Inventory Financing 16% Packaging Transport 10%25 Administrto Soua:rce Bundesveretnigung Logistik (Germany) 189 TRADE: TOWARDS OPEN REGIONALISM optimize the flows of goods and information that apply to inventory holdings-in product final prices could be their operations. reduced to around 20 percent on average (see Figure 4). The survey confirmed that the costs of excessive inven- Companies that were successful in streamlining their logis- tory holdings have been identified as the heaviest burden tics-management organizations reported that a 1 percent on corporate performance (see Figure 2). The application of reduction in their logistics costs had the same effect on cor- logistics-management principles has enabled many indus- porate performance as a 10 percent increase in annual sales. tries and trading organizations to conduct their business In the countries of Latin America the incidence of logistics with inventories that now often represent less than one costs in the market price of domestically manufactured week of supplies required for manufacturing or retailing. products typically ranges between 30 percent and 40 per- Purchasing and selling under "just-in-time," "constant cent. A recent investigation by the University of Costa replenishment" or 'quick response" arrangements are gain- Rica provides telling evidence. ing decisive importance for efficient inventory manage- The World Bank's international survey revealed that ment (see Figure 3), and-more importantly-for compet- success did materialize through far-reaching reorganization itive success. of corporate structures, and a substantial development of These achievements have had remarkable effects on strategic alliances with service providers (see Figure 5). product costs. Expenditures incurred in managing logistics Many industries and trading enterprises have externalized organizations amounted on average to 23 percent of value all activities that are not core to their basic mission, which added and 70 percent of the operating margins in all is production or selling. As a result, transport, warehous- industries that were surveyed (see Box 6). Through dedi- ing and communication services are increasingly provided cared efforts in many industries, the logistics cost share- by third parties. Progressive companies also consider out- essentially reflecting transport intensity and the level of sourcing equipment and private fleets because of their cap- FIGURE 3 Application of "Just-in-Time" Concepts in Purchasing and Distribution Observed Averages by Category of Industry in North America and Western Europe 60 _ 1987 50- F 1990 So: The World Bank1995 40 - 30- 20 10 0 Automotive Chemical Electrical Food Paper Pharmaceutical Textile Type of Industry Sroo-: The World B.nk 190 INIFRASTRUCTURE AND INTEGRATION ithe levesiand structure ofa wllogiticscost vary byinutsturc- Ant resrsentedl pointo av tee 23 phercentex of vauesoaddedg Althoug imanpertfuections its mnceasuint, the cilztonceput of thPerhpsc morte i prortantly the averagvogsic soss weree buas valuer added isntrouighl peqsuivaen to catontrollablhe fuctonsts As quivalens toh7 prcente oftypical toperadutiong margins. By af pronportiond valueaddmed,th logisvetoies efotin trnorthwl ipicahtedtionar10 percentredutiong in aloisitics costsr woul Americans induestses.Ti rnged betw ieen muperfenther yielde an im rporove ensrto ercgins thtpenerating m taregins perentliingo selectedindustrvies i the relarly ed inogistiuc- ogtr line ihol e aeul eetdpo turewithemphsis n sytemsinteratin (se Fiure ). dcersoriiservce O suAppSlerTAs OneF.: heahpate Much f theattetion hat sanwbigdvtdt evc asmscerydfndrsosiiiis hssaigrss stemsfrorr the usines co Dunit'stoeettwr h TYC imprac fi l sta dpnaiiyadaiiyt inegaedmnaeen ratce.Bycosl lnin uy povdVhglyeficentconribtios oPte joit udr andisel OurIcinUs,thes pracicsplaeincesndead taighsbomacresoeofucs-nddvcd upponiantrcnes ir rnprainadlgsisfntos ifatutr ly rca oeihs Apparel~~~~~~~~~~~~~~g TRADE: TOWARDS OPEN REGIONALISM FIGURE 4 The Incidence of Logistics Costs Typical Indicators among Industries in OECD Countries Industry Food * :: > > > e . e s go Chemical s L Metal -3 Paper > ,, 6> , . ; Textile J Electrical Automotive s > v ,- AVERAGE 0 5 10 15 20 25 30 35 Logistics Costs in Relation to Annual Turnover (percent) Source: European Logistics Association The survey findings largely confirmed that the provision processing and assembly are other value-added services in of efficient supporting services had become possible these two industries that are attractive to shippers. through unprecedented advances in transport and cargo In some countries, liberalization of government regula- handling technologies. The focus within the transportation tions, basically by removing infrastructure and service bar- function has shifted from efficiency and cost-effectiveness riers to efficient competition in the transport and telecom- to measurement of on-time delivery, order fill rates and munications sectors, provided facilitation. The net result of order cycle times. Furthermore, there was clear evidence a more open political environment was the creation of that effective communication links are prerequisite for the greater opportunity for logistics to take an active role in success of any logistics-management system. Recent World competitive practice. In short, establishing such conditions Bank investigations revealed how new information tech- is the task of economic policy. Policy is what matters over- nologies have rendered many services increasingly tradable. all. The key to growth of trade through improved compet- These developments have been instrumental in enhancing itiveness is granting producers and consumers the eco- two new dimensions of services: first, the proper integration nomic freedom to face market challenges and to respond to and utilization of modern and efficient logistics organiza- incentives. tions in the production process, and, second, logistics ser- Perceptions of what constitutes quality have evolved as vices as an important source of value added. Shippers well., Total quality management programs-promoted by increasingly select service providers that offer most value the International Standards Organization under the ISO added over and above their traditional activities. An impor- 9000 norms, which relate to quality restrictions-that took tant value-added service in transport and warehousing is the root in the manufacturing environment have grown in facility to electronically track shipments real time. Order influence and are now being applied to every business func- 192 INFRASTRUCTURE AND INTEGRATION FIGURE 5 created opportunities for lowering cycle costs. However, by Trends in Outsourcing Logistics Services the late 1980s the usefulness of the traditional structure Sample: 625 North American and 238 European Industry significantly declined for two reasons. First, a pattern of Groupings diminishing returns set in as freight rates and systems 160- redesign reached near optimum levels after nearly a decade of assault. Second, the demands and opportunities for 150- logistics shifted from costs to market-to the strategy for supplying it and to information about how well the supply 140_ strategy was being implemented. Increasingly, logistics became attached to other front-line functions, such as cor- _1_0_ porate strategy, marketing, information, control and pro- 120 curement, as well as to manufacturing. The survey discov- o^ ered that logistics is inseparable from these functions, and 110 that formal interface with each creates new opportunities co f for profitability. The interface with marketing focuses on 100- executing those aspects of the logistics process that the cus- tomer perceives as most critical to buying a product. Suc- 90 cessful companies have recognized that time is a strategic variable that affects competitive success in the market- 80- place. Increasingly they are looking at logistics for ways to offset market erosion. 70 The Macroeconomic Effects of a Deregulated 60 1 Logistics Service Industry 1987 1990 1995 Creating a policy and regulatory environment that enables Source: The World Bank national producers and traders to adjust effectively to changing practices in the international marketplace is not tion. ISO 9000 certification is also becoming a key compet- only beneficial to corporate performance, but also leads to itive factor in the infrastructure sector. The term "quality," improved fundamentals for the economy as a whole. Of the traditionally associated with product defects, has quickly few countries that maintain sufficiently disaggregated become synonymous with good service. Likewise, customer national statistics to allow conclusive assessments, the service is no longer used exclusively to refer to order entry United States' performance indicators for the national departments, but has taken on much broader meaning and economy provide evidence of the effects of service industry now refers to every process that ultimately affects the cus- deregulation on logistics expenditures incurred in manu- tomer. Mcre and more companies are evaluating perfor- facturing and trading. Particularly noteworthy is the mance by monitoring order fill rates, on-time delivery and decreasing trend over the years in nationally aggregated out-of-stocks, all of which involve transportation and other freight payments (see Figure 7). After successive deregula- logistics ftnctions. It is apparent that quality service has tion of the different transport modes, from the late 1970s become key in gaining competitive advantage. onward, the annual logistics expenditures in the national In the countries surveyed by the World Bank, the tradi- economy dropped from an equivalent of 17.2 percent of tional hierarchical organization structure effectively met GDP in 1980 to 10.8 percent in 1994. The average annual the overriding logistics goals of reducing the costs of the cost savings were in excess of US$60 billion. A further logistics C.eements-transport, warehousing and inventory. acceleration in the decline of the annual logistics expendi- This struccure reached its greatest effectiveness in the early tures/GDP ratio started at the end of the 1980s and can be 1980s when the convergence of freight deregulation, new largely ascribed to the effects of deregulating the telecom- information technology and global supply and competition munications industry. Competition created better telecom- 193 TRADE TOWARDS OPEN REGIONALISM FIGURE 6 Logistics Performance Trends 500 Europe Enterprises Customer Service a Transport Warehousing Materials/Product Planning & Control Achieved between 1990 and 1995 SYSTEMS INTEGRATION C Targeted through 2000 Overall Logistics : v a > 5 a > Performance 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Productivity Improvements (percent) Sasa-c: The NVorld Bank munication service offerings, which enabled manufacturers part of industrial and marketing processes is becoming and traders to reduce order cycles further (see Figure 8) and more pronounced. More than ever before, efficient trans- to turn over inventories more frequently (see Figure 9), port networks have become prime prerequisites for estab- thanks to improved linkages with transport, warehousing lishing market-responsive trading arrangements. This and other trade-supporting service providers. Higher fre- observation applies to both international and domestic quencies of inventory turnover induced major increases in trades. In the case of international trade the availability of the annual returns on assets in almost all industries. efficient transport services and their supporting infrastruc- The policy interventions and subsequent developments ture is now a key determinant of competitiveness. On the in North America and, similarly, in several other OECD domestic scene the quality of local distribution networks countries demonstrated the growing importance of trans- exerts much influence on product availability and retail port. As the practice of manufacturing and trading with prices, with obvious effects on affordability and the minimal inventories has spread, the demand for efficient national industries' ability to engage in modern logistics- transport services has increased measurably. For instance, management practices. while the aggregate annual costs of inventory-carrying in the U.S. economy has remained almost stable since 1980, Issues in Managing Latin America's Trade the concomitant annual incidence of freight transport more Trend projections suggest that in most industrialized than doubled. A typical phenomenon of modern manufac- countries there will be further growth in outsourcing of turing and trading practices is that shipment sizes tend to production processes to foreign markets. The search for become smaller, whereas shipment frequencies increase sig- less costly supply sources will continue unabatedly. There nificantly. Hence the intensity of transport as an integral will be many opportunities for the countries in the Latin 194 INFRASTRUCTURE AND INTEGRATION FIGURE 7 Competition ais a Regulator of Freight Transport The U.S. Experience Trend (1971 = 100) 700 Gross National Product 6 0 0 .. ... . - ..- - -...- --..- - - - 600 Nation's Freight Bill .......... 500 - . 0.i: 400- Start of Tra nsport A 300 - ~~~Industry Deregulation 200 - 100- a 0- - c CA In 't v I' NO N 0 rCO 0 - N r' o NO N r oo (N cON ' uV r r- rN rN rN N N N o C rC co x x oo C oc CC ON ON CN CN ON O" soN CA ON N ON CoN ON GoN ON GoN ON ON ON ON ON ON ON ON ON ON ON ON ON ON ~~~~~~~~~~~~~~~~~~o,_ 7, C, _ : c, C, ,_; (,\ _ Source: ENO Transportation Foundation American region to capitalize on these trends-and identified that will be instrumental in bringing about thereby accelerate the growth of national trade. But here effective development and strengthening of competitive as elsewhere the availability of low-cost labor no longer service sectors, including infrastructure related to services. suffices to attract foreign investments into local produc- The contribution that logistics services make to competi- tive sectors. This observation becomiies even more corn- tiveness in the goods-producing sectors and to the overall pelling when one considers the declining share of labor economy in each country in the Latin America region may costs in product-end prices due to increasing automation well be affected by the speed at which these countries will of manufacturing processes. While direct labor costs in be brought into a much more competitive international manufacturing are now less than 10 percent of the value- setting. added chain in many industries, the costs of required Of highest priority in this context is the need to materials frequently represent more than 60 percent. embrace the concept of advanced infrastructure. For all Latin These circumstances make it essential that the entire American countries the ability to compete in global mar- logistics flow is managed efficiently to cut costs, control kets now rests on the successful development not only of inventories, improve fill rates, speed up receivables and basic infrastructure, but more importantly on the estab- reduce tied-up working capital. lishment of effective mechanisms that facilitate the cre- There is a need in almost all countries in LAC to reassess ation of integrated advanced infrastructure. The problem the role played by services in contributing to growth and in many regional economies is that the regulation, plan- sustainable development, including an examination of the ning and management of the different elements of trade- contribution of the logistics-service sector in facilitating supporting infrastructure are disjointed and without effec- the compel:itiveness of goods-producing sectors, and to the tive coordination. There is thus a need to develop policies local economies as a whole. Policy measures have to be and administrative arrangements that bridge institutional 195 TRADE: TOWARDS OPEN REGIONALISM FIGURE 8 Order Cycle lime Development Electrical/Electronics Industries Weeks 25 20 105 5 1995 >>> less 0 Ehan TWO WEEKS 1980 1985 1990 InJformation Processing Physical DisEribuEion MaoufacturiogLatm Source: Philips lnte-na . B.V. as well as organizational disparities and inconsistencies. A key issue in almost all Latin American countries is Beyond creating a level playing field for advanced infra- that, traditionally, businesses have focused primarily on structure service providers, governments must also provide their internal activities and processes, honing their own a policy environment that actively supports the application systems and procedures, withiout much consideration of of advanced infrastructure concepts and arrangements. their impact on customers and suppliers. Usually, Preliminary evidence in many regional countries sug- autonomous business units have developed within compa- gests that the development of advanced infrastructure has nies, organizing around business functions and internal been hampered by incompatibility of existing systems and cost/profit centers. Often, supply and demand forces are available technologies. The lack of strong industry associa- ignored, diminishing the potential of logistics services for tions that would take an active role in promoting such con- greater productivity, greater competitiveness and more cepts is another issue of consequence. Furthermore, there is value added, both within the firm and at the national level. a substantial shortage of qualified human resources that are Slowness in externalization has meant isolating these func- critical for the operation and management of efficient tions from market signals. In contrast, the logistics ser-vice rrade-supporting infrastructure and service systemns, sector in industrialized countries has evolved sponta- According to successful manufacturers and traders Sur- neously, conducted by firms themselves, based on competi- veyed by the World Bank in Asia (Mody and Reinfeld tion-induced demand and technology-driven supply. 1995) a very important consideration for adopting market- Increased competition facilitated through trade liberaliza- responsive advanced infrastructure arrangements is the tion, combined with technological advance, created availability of well-trained labor supply. demand for both efficient provision of infrastructure and 196 INFRASTRUCTURE AND INTEGRATION FIGURE 9 modern logistics-management concepts, is widespread Trends in Mairketing Management within regional governments and the local business com- Sample: 625 North American & 225 European Businesses munities. But logistics services will only be in demand to 170 - the extent that the activities using such services operate in 180 - ~~~~~~~~~~a setting that is vigorous and competitive enough to require them. Preliminary indications point to the exis- 150 - tence of a dual development of logistics services in Latin American countries. In a number of these countries, several 140- _ _ ___ modern economic activities that rely on technological change and that normally compete in the international o 3 markets coexist with other activities that are technologi- r- 120- cally backward and traditional in their managerial and hierarchical structures. The latter have usually benefited o 110- from protection and have not undergone the productive transformations necessary for competitive success in inter- 100 -, 100 national markets. 90 - Thus, many regional enterprises, which could be poten- tial users of logistics services, resist change and maintain a 80 - high degree of internalization of services. There are various reasons for this, including the low level of competition at 70 - 0 which they operate, the fact that these firms attach higher 60 -- _ importance to needs assessments by their own employees than that by customers, entrepreneurial failure to perceive 198-7 1990 1995 the economic advantages of externalizing, the low level of Source: The World Bank assimilation of new technologies, concerns over confiden- tiality, lack of confidence in the reliability of outside sup- logistics services. Demand has thus stimulated the need to pliers, high costs and/or inadequate quality of available adapt to ccmpetition in the domestic market and to take logistics services and, finally, the absence of policies advantage of export opportunities. encouraging the entrepreneur to externalize services. The supply of logistics services came from knowledge- The fragility of the domestic supply of logistics services able individuals who were capable of adopting technolo- in Latin American countries and the widespread inability gies and efficient use of infrastructure. To a large extent the to express a demand for such services restrict the chances supply of services in industrialized countries has emerged for adapting domestic economic activity to the require- from the manufacturing and agro-processing sectors ments of efficiency and competitiveness in a vigorous and through the process of externalization. Competition, with increasingly globalized economy. Local enterprises should increasing focus on access to and adaptation of technology, be enabled to develop tailor-made competitive strategies. has createcd a growing demand for logistics services. In These strategies will have to be based on such factors as the short, a dynamic aspect of the service sector in modern competitive structure of the industry in which they oper- economies concerns major restructuring of the production ate, their position in that industry and clear identification process, of corporate straregy and of labor markets. It fol- of their particular sources of competitive advantages. In lows that this has given services a new strategic role that doing so they need to explore the possibilities of creating transforms them into decisive factors in the generation of advantages by seizing the opportunities created by new value added in an economy and in the creation of compet- technologies, the new and changing requirements of their itive advantage. customers, the emergence of new segments or niches in the There is much evidence that unawareness of changing industry and changes in the costs or availability of inputs. competitive practice in international markets, based on These strategies have to be oriented toward gaining a posi- 197 TRADE TOWARDS OPEN REGIONALISM tion in good time to take advantage of structural changes, to include specific measures focused on stimulating both to be followed by a clear perception and pursuit of innova- the demand for and supply of such services. Because this tion in order to sustain advantages over time. In some LAC demand is derived from the requirements of other sectors, countries steps have been taken along these lines, but are the situation faced by those other sectors with regard to still at the preliminary stage (see Box 7). regulation and competitiveness will determine the vigor of There are also instances where structural adjustment demand for logistics services, particularly from those seg- programs have been launched in order to create a setting ments of the regional economies that are exposed to inter- that is more conducive to creating greater competitiveness national competition. and efficiency in their enterprises. However many of these The development of efficient logistics services may be adjustment programs do not pay sufficient attention to the assisted by better utilization of market mechanisms with link between logistics services and planned or actual respect to the allocation of resources, price determination restructuring in the manufacturing and agricultural sec- and the appropriate use of supply and demand forces in the tors, whereas the role of logistics services is precisely that clearing of markets. Frequently, the still limited use of of assisting enterprises to attain the necessary flexibility to market mechanisms in several Latin American countries is adjust to more competitive and uncertain situations. reflected in the concentration of demand in a small number Strategies for strengthening the development and compet- of large-scale, generally monopsonic users. But a defective itiveness of indigenous logistics services would thus have market structure is not necessarily connected with state BOX 7 Regional Initlafties To Facilitate Trade Transactions In several countries national trade facilitation bodies have logistics-management practices are not well attuned to been established. In most cases their organization is mod- the need of modern markets. It is now vitally important eled after the British Simplification of International to overcome these deficiencies in the regional economies Trade Procedures Board (SITPRO). SITPRO is a joint in order to enhance the international competitiveness of initiative by government and the local business commu- local industries and national trade organizations. nity. Its mandate is to disseminate modern trade logis- To date, specific initiatives have been launched in four tics-management concepts, to foster interaction between LAC countries-Argentina, Brazil, Chile and Colom- related public institutions and the national industry, to bia-with assistance from the Pan American EDIFACT identify issues and prepare proposals for remedial action Board. EDIFACT stands for Electronic Data Interchange and to organize training on a wide scale that is open to for Administration, Commerce and Transport-a system all sectors of the national economy. These activities have that has been brought about by the U.N. Economic had measurable effects on the ability of British industry Commission for Europe. Most advanced is the system and traders to produce and market their products developed in Brazil under the name SIMPRO-BRASIL. competitively. This body has been organized by local entrepreneurs and The problem in LAC countries has been that the local is the only private participant in a federal government service industries with relevance for trade management program oriented to produce positions and actions were and often remain poorly organized. Little, if any, toward a Brazilian information infrastructure in line with coordination with trade and industry exists, and even ongoing international developments. The membership in within the service sector there is only limited profes- SIMPRO-BRASIL is growing, including businesses sional coherence. For a long time the regional govern- related to production, trading and the provision of ser- ments have neglected this segment of the national vices. The organization is providing a broad array of ser- economtes. As a consequence legal frameworks that gov- vices to its members through electronic communication, ern the service sector are inadequate, the trade-related professional advice and training. (See SIMPRO-BRASIL:s service industry is underdeveloped and current trade web page at http:/Iwww.ibase.br/-simpro). 198 INFRASTRUCTURE AND INTEGRATION ownership of particular services or with the monopsonic appropriate technologies. Other observed tactical applica- nature of demand for them. Many services rendered by the tions include fiscal and credit incentives, and modification private sector exhibit clear examples of market concentra- of the regulatory systems to reduce or remove the bound- tion on the supply side, which does not always contribute aries between different services. There also have been to a better and more efficient integration of services into attempts to promote the supply of logistics services on the production or trading processes. Better utilization of mar- basis of incentives for the externalization of locked-in ket mechanisms can be achieved through changes in the knowledge, the establishment of service centers, and train- regulatory framework that affects market entry and the ing facilities aimed at teaching the essentials of logistics behavior of firms. services and their role in modern-day competitive practice. But success with these measures has been limited. Strategic Options to Enhance Latin America's The outlook is not promising without substantial Trade through Better Infrastructure and Services improvement of regional infrastructure, a liberal regula- Latin America's governments have at their disposal a set of tory framework that governs the service sector and busi- instruments with which to stimulate the development of nesses that are willing and able to apply modern logistics- competitive service sectors. Some of these, such as trade and management techniques. In effect, regional countries and fiscal and monetary policies, fall within the scope of macro- their business communities that are unable to adjust to economic policy. Others may necessitate changes in the these new market practices will face the risk of becoming institutional and regulatory framework, and the achieve- marginalized in the international trade markets. Against ment of more specific objectives may call for particular the international experience record to date, it appears strategies aimed at positioning national enterprises in the unequivocal that a key criterion of the required adjustment international market. Overall, attempts to strengthen the agenda relates to infrastructure and services in support of organization and provision of services will entail devising industrial and trading processes. More specifically, the and applying specific strategies to overcome the weaknesses establishment of effective transport and telecommunica- characteristic of existing trade-supporting infrastructure tions arrangements-both from an infrastructure and ser- and logistics services in the LAC region. The incorporation vice-organization point of view-should be given prime of foreign service providers to assist in the attainment of consideration. sustainable development, whether by facilitating access to Distance is critical in logistics management and equates new technologies, information networks and distribution to transport dependency. Since many Latin American coun- channels, or improving capacity for training and the gener- tries are remote from the principal consumer markets, they ation of domestic know-how, may be an important compo- are crucially dependent on efficient arrangements for nent of industrial and trade-development policy. As can be freight transport. It is also important for policymakers to observed elsewhere, policies with regard to trade in goods appreciate that demand-responsive production with lower have had a marked effect on the evolution and development average inventories translates into an increased require- of services. Imports of equipment and inputs for various ser- ment for small shipments in effectively controlled logisti- vice activities, especially high-technology products and cal environments, which signifies high transport intensity. those that support telecommunications infrastructure, need As the experience in the OECD community exemplifies, a sufficiently liberal regime not to thwart the development the implication is that rectifying transport system short- of the serv.ces that depend on such inputs. comings should be high on the reform agenda of Latin Experience is still limited but shows that in order to American countries in their drive to increase the competi- stimulate the supply of domestic logistics services some tiveness of their national industries. But the OECD expe- regional governments have experimented with different rience also points to a number of significant consequences instruments, including direct financial support. For that a liberalized markets entails, about which the regional instance, in Colombia the government established the policymakers should be aware (see Box 8). A special chal- Vallejo Plan, which provides for the development of service lenge will be to design regulatory provisions and organiza- enterprises through soft financing for the purchase of tional arrangements that will yield optimal use of existing equipment, international promotion and the acquisition of physical network structures. The development of efficient 199 TRADE: TOWARDS OPEN REGIONALISM BOX 8 Shippers' Transpott Preferences In all economies where manufacturers and traders were Attempts have been made in the OECD countries to able to go about setting up efficient physical distribution arrange for the carriage of goods by combining different systems, trucking has been the preferred modal choice (see modes, based on their respective operational and cost Figure 10). Road transport provides flexibility, moving advantages-it is called inrermodal transport. The consignments swiftly from origin to destination without emphasis was on combining freight transport by road trans-shipment. The same service cannot be provided as and rail through different technologies. But the market ubiquitously by any other mode of transportation. Freight share of such intermodal transport arrangements is still transport by air, rail or water requires one or more trans- limited-i5 percent in North America and barely 5 per- shipments for any type of cargo, which adds costs and cent in Western Europe. The economics of truck-only time. Unlike roads, the physical infrastructure for these transport remain compelling. Some OECD governments modes is geographically confined. In addition, their orga- pursue policies aimed at forcing the use of intermodaL nization and management are in many countries cumber- transport arrangements through regulatory intervention. some and inefficient. As a result, reliability is seriously Early indications are, however, that such steps seriously compromised, which curtails shippers' ability to acquire, undermine the competitiveness of the national industries manufacture and selL products within right timeframes, as and trade. required for effective logistics management. Hence their The challenge for LAC governments will be to estab- preference of trucking, which is usually in private hands, lish conditions under which the national transport sys- whereas the other modes are often still managed by ineffi- tem can meet the special logistical requirements of local cient bureaucracies, The result has been rapid congestion industries and trade. Costly mistakes have to be avoided, of national highway networks, especially in urban areas, and the OECD experience shoutld be carefully considered. together with increasing levels of attendant pollution. For a long time trucking will remain the preferred mode Thus, the attainment of economic efficiency in this context of freight transport. has invariably led to escalating social costs. intermodal freight transport systems should be given high regards transport, some affirmative actions have already pro- priority-based on the comparative advantage of different duced beneficial results. Throughout South America, govern- modes-if overly heavy reliance on road transport is to be ments have abolished protectionist policies with regard to avoided. Shippers prefer road transport because of its ocean transport, leading to extensive competition between apparent flexibility, which other modes-in particular the national and foreign-flag carriers. As a result, freight rates for railways-have failed to produce. However, prolonged waterborne transport have tumbled 30 to 50 percent since reliance on road transport for freight movements has led to 1994. Equally decisive in the effects, the admission of private severe economic and social problems in many OECD coun- interests in port financing and management in many of the tries. To counteract these adverse impacts, their govern- regional ports has led to decreases in terminal handling ments are now actively devising measures aimed at increas- charges by up to 70 percent as productivity improved con- ing the share of other modes through dedicated efforts to siderably, driven by fierce intra- and inter-port competition. promote intermodality. Latin American governments While there have been consequential losses of market shares should capture the opportunity to avoid the costly mis- by national entities that formerly enjoyed quasi-monopoly takes of the (initially) misguided regulations and invest- status in shipping and ports, the outcome of opening these ment decisions related to freight transport systems in the sectors to international competition has yielded measurably OECD economies. positive effects on national trade performance. Consequentially, the required adjustment process in the Impressive as these outcomes have been, significant regional countries is diversified and invariably complex. As inefficiencies continue to exist beyond the waterfront. The 200 INFRASTRUCTURE AND INTEGRATION FIGURE 10 The U.S. Expeirience: The Trucking Industry As Manufacturers and Distributors Increase the Trucking Industries Share of the Nation's Freight Bill, Their Investment in Inventory Needed to Support Their Sales Decline 2- 8 80 027- 78z 25- 76 ~3 a 23- \ $ 2]- X -~~~~~~~~~~~~~~~~~~~~74 , 0 . 21 -6 (N 2 Crr I coG O m C , 00 0 00 C C x Y 00 w X G\ GE a C, Sougrce: Federal Reserve Bank of St. Louis, and ENO Transpottation Foundation physical condition and the performance of inland distribu- with modern information technologies, will increase pro- tion networ:ks are still far from adequate. The reasons are ductivity for both goods and other services. They will also manifold. M4ost cumbersome are local regulations that facilitate the application of new management techniques, stymie initiatives within the national transport sector to make relations between the various stages of design, pro- become mcre responsive to changing shipper require- duction and marketing of products and services more expe- ments. The concern of ensuritig quality, competitive prices ditious, allow customized services to appear, help to create and adequate provision of services, while at the same time greater economies of scale and facilitate an efficient global- adapting them to technological advances, ptaces transport ization of the production and distribution functions. service suppliers under pressure to improve management In many Latin American economies the taws and regu- and operational efficiency. It is encouraging that in some lations that govern the conduct of trade and the organiza- countries governments have come to appreciate the appar- tion of trade-supporting services and infrastructure have ent inability of state monopolies in transport and other ser- barely been changed since the time of their promutgation, vices to meet this challenge, giving rise to policies aimed decades back. Hence, their provisions are substantially out at privatizing these monopolies. It is hoped that the of line with the exigencies of modern markets. Such situa- demands of a more competitive environment will provide tions are all the more critical as most elements of the ser- the necessary stimuli for technological progress to focus its vice sector relevant to supporting trade-transport, freight creativity o:n developing new and better services, tailored forwarding, warehousing, telematic information systems, more closelv to the requirements of the production and dis- banking and insurance-have become increasingly inter- tribution pr-ocesses. New and enhanced ser-vices, associated dependent to yield efficiency. Failure of one element in this 201 TRADE TOWARDS OPEN REGIONALISM interdependent setting to respond efficiently to market of local products are when they leave the assembly lines or requirements usually undermines the performance of all the farmers' fields. other elements-the weak link syndrome. Trade-supporting infrastructure needs, above all, to be One of the most important aspects of modern logistics well maintained. Without doubt, such infrastructure will management is internationally harmonized documentary also have to be modernized as well, so that the productiv- procedures which enable speedy electronic transmission of ity potentials of new transport and freight-handling tech- freight bills, payment orders, insurance contracts and nologies can be fully exploited. Budget constraints in transactions relevant to the movement of goods. A modern many Latin American countries will limit the scope of such and efficient telecommunications infrastructure capable of required interventions. Hence, ways have to be explored to supporting, at low cost, the domestic and international tap nontraditional sources of finance, which largely points development of such varied services as banking, transport to attracting private capital. The experience in a number of and other logistics services that are direct users of such net- Latin American ports suggests that attracting private cap- works is indispensable. In countries where major advances ital goes hand-in-hand with involving private expertise in in streamlining logistics-management organizations have asset management. As successfully demonstrated in many been achieved, customs or excise clearance of traded goods countries throughout the world, there is ample scope for has also been integrated into the electronic communication creating synergies through public-private partnerships in systems. As a result, such clearance is usually swift. In con- financing and managing trade-supporting infrastructure trast, complicated and hence tedious customs practices in and services. But to succeed there has to be unequivocal many Latin American countries remain one of the princi- commitment within governments to such partnerships, pal obstacles to better trade performance. Ways have to be accompanied by transparent legal frameworks that will found-on a priority basis-to remove this bottleneck. govern such alliances. Governments are rightfully concerned about ensuring Other than ports, there are now also developments in proper revenue collection and fighting contraband. But several Latin American countries to involve private capital there are successful ways of safeguarding against such con- and management expertise in making national railways cerns while facilitating trade transactions at the same time. more responsive to evolving market requirements. In addi- The reforms in customs regulations, organization and pro- tion, some regional governments have decided to engage cedures in Mexico provide a good example for possible private contractors in highway maintenance and to build replication elsewhere, if adjusted to local conditions (see new trunk roads as toll facilities through concession Box 9). arrangements. Yet more needs to be done to make local Adjusting national infrastructure to the special require- industries more competitive internationalLy and to reduce ments of trade and the supporting services is demanding the costs of domestic trading. Importantly, more effective and considered to be costly. But all LAC countries are dialogue between the public administrations and the busi- endowed with infrastructures that have been developed ness communities is needed. Business requires freedom of over generations. Often, the problem has been not so much action while maintaining the concept of "checks and bal- a lack of resources, but an inability to use existing resources ances." Experience has repeatedly shown that, once given well. To some extent such a state of affairs is, again, due to the freedom to take measures considered to be profitable, overbearing regulations and too much state intervention. private entrepreneurs have invested in market-responsive On the other hand, in arcas where thc state could rightfully infrastructure and service arrangemcnts that turned out to be expected to take a more proactive role, it has largely be beneficial for national trade performance. An outstand- failed to do so. The dismal state of highways and rural ing regional example is represented by the initiatives taken roads throughout Latin America attests to this observation, in Argentina with respect to involving private interests in which has been widely acknowledged by many regional the management of national railways, ports, airports and politicians, public administrators and business managers. parts of the local highway system. Clearly, the state needs to Without well-maintained roads it will be difficult to orga- ensure that abuse of newly acquired rights is avoided, but nize efficient trade and industry logistics-management sys- the dynamics of private management of national trade sup- tems regardless of how good the quality and low the costs porting infrastructure and services cannot be overstated. 202 INFRASTRUCTURE AND INTEGRATION Cusoms Redbtonns: M0Id osSceshiBekn Ilo Tae maret In te erly198s, exio.secoomy xpeiened an soe rplaed t olpecluigciusad to .ance-f-ayens iffiulties Theover n re fan overal fth tasyemancutshsno sponWdedita uber of refoM intativestaoe bee ' ihgnrltxcleto.The riht merchadise eports,which Represete less,ha oneV enrurca bak, hc oeebachofcei cutos thir ofttleprsIfl8 haE douie thershaoadieis sinc the utther reaieda ume of mpdient Wdepradin ofputeizainadeecrncdt tOMo efieitrad flow - Aky newsPMs5i diii whi I esme. I likge t tebckheofted eome npeto CusetomsBu prceues were highly cetrlie ard' ani npce.Teitnini oeoedsrto n anffd nonnianpren tpsdrdesfce ' flpv rofe g prcs as benrdCedfOm. 12 tofuMhenwss orinnes plcbet utoshdpoieae over Mustenow b don at Anitrostewhnteri- theyeas, anywer neerpubishd ad uidewen dctin o a radfs oca fsca.ofice.Ths...asi. constant changes. No uniform standards for application removed the pasi need for lengthy detours to clear cusL~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~t , T t proces is. h ietrt Genealnfn usomsI (I) part th reultes feI tutr a hse uBoeshv h aaard y tihdiitton fauson bo . gusm. pronl benefi Ifitifom. hlialicntv alyconsidered t be6 mjracmlcsicuosirg. eeind pt4' tih: evls Zfthlrte.Tesrue of brokereeiis6Ah feesiwas, such that. 70AThr ist jijdite widc thatteimdaeefcso pere.Inr otpa ho .cmeaP .o unoueted tinheseom, mwteasures ha~E vt bee sustntal A World In 199, th govrmen stepe inand ntrodced alueof merchndier t~rade in 19, orcoe o1pecn roatve, hihwee.ssgnd ooterhgnce witin tinsIn th cot ofk te t, storag an fd trnpota theqs Miiyo iac,adwslf wt h oemn ela oe te 'okeCr fBadi eiiae kudou dat offailiatng he hyica.poc.s fcstmcer ientdepne. tilontwrh tha the& dail ance an prvnin of smgln.Te'hom 4m- ciel.fcun Ii uish dfeetfclte hsraMo as. eetaie.Tpln safwr ehfld icrae 2t 5pret d d I' 1 $10a Thi§ A~~~~20 TRADE: TOWARDS OPEN REGIONALISM The governments' role should gradually shift to being a alone if they or their managers are ignorant of the essential promoter rather than manager of trade-supporting infra- ingredients of modern trade and industry logistics-man- structure and services. The regional governments should be agement organizations? Successful initiatives to overcome facilitators and create an enabling environment in which such issues can be observed in some of the Central and East local industries and the trading community can adopt European transitional economies, which were confronted modern logistics-management concepts with all their with the same problems (see Box 10). The process of proven enhancement potential as regards productivity and change has been assisted by increasing levels of technical competitiveness. Very importantly, governments must education, which has facilitated the international transfer- actively endorse and, where necessary, supply advanced ability of technology. One of the most striking features of infrastructure. For the forward-looking government, estab- the fastest growing among the newly industrialized lishing a regulatory framework that supports advanced economies of East Asia is the very high commitment to infrastructure development is a must. In this general con- basic education. But one also needs to ask under what sort text, government policy should not restrict entry or opera- of conditions will the right sort of action happen? The sim- tions of qualified international service providers. plest answer is, under conditions that grant the investor- There are other important situations where government a firm buying a new machine or sending an employee to be involvement in the development of services and applica- trained-an adequate return. In many regional countries tions may be not only appropriate, but essential. Govern- such conditions are still imperfect, which is why the ment assistance is particularly crucial if the local industry adjustment process has been slow and ineffectual. primarily comprises small companies or if they do not have The bottom line is that the challenge facing policymak- strong industrial associations. In the case of Taiwan, China ers in the countries of Latin America is to establish condi- the automotive industry lacked the kind of competitiveness tions that would enable the local economies to attract more that would drive manufacturers to develop electronic data global production, and to realize more of its benefits through interchange (EDI) systems in support of trade management enhanced trade transactions. Such conditions include on their own. But by having automobile suppliers linked to enhanced political and macroeconomic stability, improve- such a system, manufacturers could be expected to benefit, ments in transport and communications infrastructure and and the industry itself become more appealing to foreign better property rights for investors. A crucial condition for markets for higher value-added products. Since the local capturing more benefits is more competition. In general, value-added service providers did not have strong incentives policy is most effective when it is harnessed to a realistic in this fragmented market to develop EDI, it was appropri- appreciation of where a country's competitive strengths lie. ate and useful for the government to develop such service. Policies that encourage greater competitive pressures on all The technology used for applications associated with firms from all sources are likely to maximize the benefits that advanced infrastructure often requires large research and the regional countries could derive from global production, development expenditures that the private sector, by itself, because competition forces firms to be more efficient. More open might not be willing to absorb. In such cases it may be trade and investment frameworks are among the best chan- appropriate for the regional governments to become nels for introducing such competitive pressures in traded- involved through general funding support and initial goods sectors. In the service sectors, they need to be comple- development at one of its own institutes or local polytech- mented by the encouragement of foreign investment and nics. Governments may also be appropriately involved in well-designed regulatory policies that enhance competition. service or application development in situations where In all this, there should be concerted efforts, involving the there are so many participants that it requires the leader- public administrations and local business communities, to ship and authority of public agencies to garner the cooper- promote and develop advanced infrastructure in support of ation needed to develop a system. national production and trading. A task for which Latin American governments should arguably take a leading role is human resource develop- References ment. What use is modern technology if a country's work- Krugman, Paul (1995). Growing World Trade: Caloses and Consequence. ers cannot read the instructions on a bag of fertilizer, let Brookings Papers on Economic Activity, Vol. 1. Washington, D.C. 204 INFRASTRUCTURE AND INTEGRATION Moidy,shoai and tiading cRcesinfl Hungary. ApplicdIfastiontuof oisdNticsa Treated Fcltration:aing andtforma educatio inde moders,Hns loI93Stics c-manaemNent prusincIpesnaionalManufactu Tehniclgapry 37 ahngo,DC tuing and Trarke.in porocesses wasic therefrchlimited. Mostkingt ateairs Aadvierl eto anyody6 whoist Rincnee Loftcad assi- capritical Wa rldthatHng,arhiamnuctrsandttraers.tacie to inaetrodtuce ornstrtteamInvestrgadioesand Cindu ctay weroeelargsSelyonawareetn of the g erowinngtiConfrtance of tefe lcogisticas Unvrsidaddes Corsto RiarrSang for6 Cogsta ics . Mivenoistics ofrganizartiommnsincomptitivean practice eorseo Wreldated pe99sonTel EatrAsiaing Miand :EooicGotn peiorsto grieian (19heir marklld enzetlsares LPulc Policy Washa.ingto s n, iD.pC. ra.~ao Asi and Wintgal G part. ThNf.rh.Wnrator ld Bantsuptortead its publdc Bankr under lba cnoi Popct n the umrelaolteouna iang Colnm Roleodn Asfa'se Economi mes, Po,antinlLgak daries. fosiindutrn, D.MCA. Asmnaeito"sp Schare toberthe awdarenimesslwthi the5) Infrationa iecndultry adaeatvl novdi P ciiis te nld an tadngcomuitesad hesevie ecor th unaranAsocatonofLgitis,Puchsig5n Comment ERNE S T O T I RO NI I MUST CONFESS SOME CONFUSION ON HOW TO ADDRESS THIS COMMENT. I COULD NOT FIGURE out what is the main question that the authors are attempting to answer. Is it: What is the role of infrastructure availability in promoting foreign trade in Latin America? Is it on promoting total trade or only regional trade and, thus, on commercial integration? Or the question could be more specific or quantitative: To what extent is the lack of adequate infrastructure stopping the growth of trade or exports? If this had been the concern of the authors, one could have moved into analyzing what type of infrastructure is missing most in Latin America. Is it roads, ports, telecommunications, services of a particular nature? If the main interest were in policy decisions, the key sectors within countries. It would be interesting for Latin questions might have been: In which type of infrastructure American policymakers to have empirical evidence about is the rate of exports growth higher per dollar invested? Is the correlation between trade development and the it higher per dollar invested in railways or highways? improvement of infrastructure. This could be done either All these are very interesting and relevant questions. At through case studies or cross-section analysis. least in Chile there is a very strong political and economic Common wisdom would say that there is a very high debate on what the government should do to maintain the positive correlation between investment in infrastructure rate of growth of exports in view of a declining exchange and trade growth. But life is not that simple. Few people rate. Today an exporter gets 420 pesos per dollar, when have addressed this question without prejudices. Engineers seven years ago the exporter got 650 pesos with the 1997 writing about infrastructure development take it for purchasing power. To face this situation some economists, granted that there would be no trade development without businessmen and politicians say that the government adequate infrastructure. But I browsed through several should reduce import-tariff rates, others say it should sign papers and books about trade (like Meller's "El auge expor- more free-trade agreements, others say it should increase tador Chileno"), which attempt to explain the success of the exchange rate by reducing public expenditures, and Chile in expanding its exports, and found almost no men- still others say it should mainly improve the country's tion of the availability of infrastructure as a key explana- infrastructure through the privatization of ports, highways tory variable. aind railroads. With this I do not mean to imply that infrastructure is Unfortunately, the paper we are commenting on did not not relevant at all. What I mean is to call for more analy- address these questions, nor did it present many specific sis of the matter, which would draw more fine distinctions examples of successes or failures of individual countries or about export sectors, kinds of infrastructure, countries and Ernesto Tironi is Director of GasAndes in Chile. 206 INFRASTRUCTURE AND INTEGRATION other conditions, like timing. A key question, in my view, mainly to generate electricity, so prices of electricity were is whether the building of the appropriate infrastructure more than 20 percent higher in Chile than in Argentina. should precede the development of exports. In other words, The elimination of the gas trade prohibition attracted, is it a precondition in order to expand certain exports? If this therefore, a number of investors trying to get into the busi- is the case, should the public sector always run the risk of ness of shipping gas between the two countries. But fol- making thcse investments? What assurance can there be lowing Argentine traditions, all investors sought an associ- that exports will follow, thus justifying the initial invest- ation with the state or the state-owned company (YPF), ment in infrastructure? The world seems to be full of and they wanted government financing or guarantees to examples of "white elephants" (oversized ports, for exam- invest. It came as no surprise that almost five years had ple) built by countries on the expectation of exports or passed and the construction still had not started, notwith- other activities that never developed. standing dozens of press statements about the progress of What I find most interesting and encouraging in Latin the project, state visits and ceremonies. Only when the America in relation to its infrastructure is how it is being countries agreed to open up the market and established an positively affected by regional integration and by competition. open access system (no exclusive state concessions) did Traditionally the economic literature in the region fierce competition start, which resulted in the first pipeline (especially t-he more "structuralistic" literature) has stated being built in less than two years. that we need more investment in infrastructure in order to Five conditions have been crucial for the development of foster integration among our countries. This is the base for an integrated gas and electric infrastructure for Chile: the calls for the construction of "bi-oceanic corridors," 1. Similar and open regulatory frameworks in Argentina cross-ocean railroads, etc. Paradoxically, we are now finding and Chile; a different inverse process: that integration is fostering 2. Companies in the sector that are either private or in investment in infrastructure. I postulate, in addition, that the process of being transferred to the private sector; the latter will contribute strongly to expand Latin Ameri- 3. A competitive free market; can trade, both globally and regionally. 4. Resources that are not defined as "strategic" and thus Look at what is happening in the natural gas and elec- subject to special state regulation; and tric generation sectors. I will refer mainly to the Chile- 5. Openness toward foreign investment. Argentina case, because I know it better, but the same is The result has been a substantial reduction of the price happening with Argentina and Uruguay, Argentina and of electricity in Chile. The prices have already fallen more Brazil, Bolivia and Brazil and is starting (or will start very than 10 percent, even before the gas starts flowing this soon) with Venezuela and Brazil. month. In addition, it has generated a $1.1 billion invest- The return to democracy in Argentina and Chile inau- ment, considering only the pipeline construction in Chile gurated a new era of cooperation between the two countries and Argentina, three power plants in Chile and the gas dis- that, among other things, led to the signing, in 1991, of a tribution system in Santiago. It will also mean about $80 "Gas Protocol" that permitted the export of natural gas million in annual exports from Argentina, more activity from Argentina to Chile. Initially it was quite restrictive. for its gas sector, and demand for the Argentine construc- Later, with the privatization of YPF and further deregula- tion company that built the pipeline, for the steel mill that tion in Argentina, the protocol was amended in June 1995, produced the pipe, and for many other engineering and and a true opening of the gas border took place. service companies. More than 1,000 people have worked in Elemenmary economics teaches us that international the pipeline construction for one-and-a-half years on both trade occur-s as a result of price or cost differentials arising sides of the border. Last but not least, Santiago now has a from differences in natural resources availability. This was hope of eliminating the cloud of gray smoke that is suffo- the case of natural gas between Argentina and Chile. The cating the city. latter has no gas, except in Magallanes, 2,000 kilometers Because of this success, construction of two additional south of its main consumption centers. On the other hand, gas pipelines from Argentina to Chile has started. One is in Argentina has the second-largest gas reserves in the region the south, to Concepcion, called Gas Sur, which will and a welE,-developed gas-pipeline network. Gas is used involve at least $500 million in the next three years. The 207 TRADE: TOWARDS OPEN REGIONALISM other is likely to be built in the north for the copper mines ing large dams, and second, because of technological break- in the Antofagasta Region, and would imply a 900- throughs that have increased the efficiency and reduced the kilometer pipeline, with an investment of at least $650 price of gas-fired power plants. million. More than half the consumption of electricity in Latin This will, in fact, integrate not only the energy sectors America comes from Brazil, but it has less than 3 percent of Chile and Argentina, but the north and central-south of the gas reserves. Argentina and Venezuela, on the other electric grids of Chile, which were not interconnected until hand, have 12 and 70 percent of the proven natural gas now. It will reduce electricity prices in the north by about reserves of the region. Thus, it is no wonder that pipelines 30 percent. In addition, the quality of service will be will start to be constructed soon both northbound and improved both in the north and the south because there southbound. will be a more balanced supply of electricity coming from Bear in mind that cheap and reliable electric power is a hydroelectric and thermoelectric sources. Prices will also necessary condition for economic development, and espe- be much more stable. cially to expand exports in world and regional markets. I see no reason why similar experiences cannot be repli- Take the case of mining. People believe that producing cated in other infrastructure sectors. There is news that a copper is only a matter of being lucky enough to have the private railway company operating in freight transport in deposits. They seldom realize that around 30 percent of the Chile is thinking about joint operations with a private cost of mining copper comes from electricity. Thus, its Argentine railway company. This will probably mean availability and price is critical. The energy integration of building new tracks. With the private concessions of high- Chile and Argentina, with its lower electricity prices, way construction and operation both in Argentina and means a nearly 10 percent cost reduction in mining. Chile, I see no reason why there could not be a joint We shall soon hear more about Argentina's becoming an project soon to improve road transportation between the important copper-mining country. This is also possible as a two countries. And the day that ports are privatized should result of the integration between Argentina and Chile, as bring an important competition in that sector, which is well as from the deregulation that is allowing the two vital for export expansion. countries to build and use the required infrastructure In the case of the electricity sector, the next regional jointly. One of the new mines will be Pachon, about 400 integration projects are transmission lines: First the Elec- kilometers north of Santiago. It will most likely be sup- troAndes Project with a 400-kilometer line between Salta plied with electricity generated in Santiago from gas that in Argentina and the Chilean northern grid close to the has come from Argentina. If not, it will be supplied from Zaldivar copper mine. This involves constructing a com- San Juan, from a new gas power plant constructed by a bined-cycle gas power plant in Salta for a total investment Chilean company there. Anyway, copper is most likely to of around $300 million. Then there will be a new line be shipped from the Chilean port of Los Vilos. Similarly between Argentina and the south of Brazil. It is also very integrated projects will come soon. likely that there will be an electric interconnection Latin America still has a long way to go in developing its between the Comahue area in Neuquen, Argentina, and infrastructure. Take again the example of electricity. Chile the Lakes Region in the south of Chile within the next five is the country that has more electricity available per capita, years. at 2,000 kilowatt-hours per person. Brazil and Argentina As for the rest of Latin America, the integration of the have about 1,600. Peru has 500. These are a tiny fraction of energy infrastructure is only beginning. Aside from the the consumption in the United States (more than 12,000), GasAndes and Gas Sur pipelines, there are at least seven on Japan (8,000) and Germany (7,000). This gives us some the drawing boards. This is especially true in Brazil, but all idea of the needs in infrastructure if we expect to become the region will be requiring large amounts of electricity for developed countries within our generation. its development. Expected investment for the next 10 years Therefore, the concern of the World Bank about the role is more than $60 billion. Electricity will not come mainly of infrastructure development on Latin America's economy from hydroelectric sources, as in the past, for two main rea- and trade is most relevant and timely. And if the Chile- sons-first, because of environmental restrictions on build- Argentina experience with the development of their energy 208 INFRASTRUCTURE AND INTEGRATION infrastructure is considered successful, the key conditions vate companies, free markets, no special "strategic" regula- for it are now known: similar open regulations, strong pri- tions and openness toward foreign investment. 209 IX. Integration of Financial Services The Integration of Financial Services in Latin America RO B ERT Z A H LE R CARLOS B U DNEVI CH ECENT YEARS HAVE WITNESSED A PROFOUND PROCESS OF GLOBALIZATION OF ECONOMIES | all over the world, reflecting both the spectacular progress that has been made in commu- nications and technology and the deliberate strategies of economic openness that most coun- tries have adopted. These factors have promoted the internationalization of the world econ- . omy. At the same time, three great economic blocs, centering on the United States, Europe and Japan, have been emerging, and they pose several major challenges to Latin America. In the first place, in terms of strategy and negotiations, Europe since the Second World War. What Europe did, in the region will need to define and adopt a position with effect, was to launch itself into an experiment of trade and respect to the new realities of international coordination financial integration that, while not without its problems, and cooperation. Second, unless there is a deliberate is now about to reach a level of development that is more attempt at coordination among the three great blocs, the advanced, more sophisticated, deeper and broader than economic cycles in the various parts of the industrialized ever before-a level of integration that has profound impli- world are very likely to fall even further "out of synch" cations for political and economic systems at the national, with one another than they are already. This could have regional and international level. implications for the entire international system and could It is important to note, nevertheless, that Europe's finan- induce the three blocks to turn inward-after all, they cial integration has begun to materialize only after 40 years already see themselves increasingly as equals in terms of of successful commercial integration, which has seen a long power and economic influence, and as less dependent on process of liberalization, sound regulation and coordination, developments in the international economy than they once and several years of uninterrupted functioning of a single were.' Third, we need to ask ourselves whether it is time to market. Our region, in contrast, is far from matching return to a rnore "integrationist" approach to development Europe's integration achievements in terms of their pace, within Latir America. depth and freedom of movement. In particular, the integra- In fact, as globalization has advanced, Latin America tion agreements that have been implemented in Latin has, with varying degrees of commitment and intensity, America have never produced any schemes for financial or been making efforts, inspired largely by Europe's example, monetary integration.2 We have agreements in the field of to promote r egional and subregional economic integration. international payments mechanisms and reciprocal credits, It goes without saying that these efforts to date have not but they fall far short of anything resembling financial inte- enjoyed the same degree of success as those undertaken in gration, much less monetary integration. Moreover, it is Robert Zahler is President of Siemens-Chile and of Zahler & Co., Santiago, Chile. Carlos Budnevich is Manager of Financial Analysis, Stud- ies Division, Banco Central de Chile, Santiago, Chile. 213 TRADE TOWARDS OPEN REGIONALISM clear that much remains to be done in Latin America in TABLE I terms of integrating trade, the environment, transportation Banking System Indicators in Latin America and communications infrastructure, labor mobility and tax (percent) complementarity, areas in which progress has usually pro- LOANS ceeded ahead of or in step with financial integration. INTER MEEDIATED In light of the foregoing, the purpose of this paper is to BANK BY FOREIGN attempt a first approach to the topic of integrating filoan- PARTICIPATION BANK BANKSi cial services in Latin America. The analysis starts from a IN INTER- CONCEN- SYSTEM MEDIATION3 TRATION' TOTALc perspective of autonomy in monetary, regulatory and bank- Argentina 98 37.6 30.4 ing supervision policy. It deals with the concept of finan- Brazil 97 54.9 7.3 cial integration and with the preconditions for such inte- Chile 62 62.1 22.6 gration, and it proposes the kind of financial integration to Mexico 86 2465 23.0 which the region might best aspire, setting out the macro- Venezuela 92 57.2 29.7 economic conditions and the principles of financial and BIS(1996) Datafor1994. banking regulation that the different countries of the b. BIS (1996), percentage of loans of the five Largest banks. Data for 1994. c. Salomon Brothers 1996 data region would need to have in common if financial integra- Soarc: BIS and Salomon Brothers tion were to proceed. And it concludes that financial inte- gration in Latin America has been and remains on a slower percent of intermediation, in all other major countries of track than commercial integration, and that what progress the region at least 86 percent of financial intermediation is it has made is due more to de facto adaptations to reality done by the banking system. than to any top-down programming approach. It also con- It should be noted, however, that pension funds are cludes that for the foreseeable future it will not be possible, acquiring a growing importance. As the trend toward pri- nor even necessarily desirable, to achieve full coordination vatization, individual capitalization and internationaliza- in macroeconomic policies, nor is it realistic to hope for tion of these funds continues, and as the banks feel the complete harmonization in financial regulation. Neverthe- effects of growing disintermediation, nationally and inter- less, the paper suggests some principles and measures that nationally, with the rising competition from non-banking might contribute to a process of financial integration-one activities and from technological progress in informatics that would initially be limited, but effective, solid and and communications, pension funds will be playing an durable, and that might serve as a starting point for more increasingly active role in cross-border trade and invest- ambitious goals down the road. ment and the flow of financial services. Experience shows that the most successful development Preconditions for Financial and Banking strategies based on external economic openness are those Integration where trade liberalization has preceded financial deregula- By financial integration, we mean the movement of capital tion. Moreover, if a country's financial integration into the among member countries of an integration scheme, per- international economy is to be sustainable over time, there formed via financial institutions, without any barriers to has to be a competitive exchange rate, together with a low- impede the free transfer of resources across frontiers. In ering of tariffs and the dismantling of non-tariff barriers, to general terms, it includes all the institutional players in encourage greater development in the export sector. Again, capital markets, of which banks, finance companies, pen- if financial opening occurs ahead of or at the same time as sion funds, investment funds, mutual funds, insurance commercial opening, this may put the country's export sec- companies and brokerage houses are the most important in tor at risk, by inducing strong upward pressures on the Latin America, and the ones that have been subject to the domestic currency. This will not only act as a dampener on greatest degree of financial integration. Table 1 shows that, exports, but will call into question the sustainability of the in Latin America, a significant proportion of financial current account deficit in the balance of payments and can intermediation is performed by the banks. In fact, with the even jeopardize the success of the openness strategy itself. sole exception of Chile, where banks are responsible for 62 The ideal sequence, then, is to go about liberalizing the 214 INTEGRATION OF FINANCIAL SERVICES capital account gradually and selectively, and only after and one lonely country had single-digit inflation; by 1996, trade-opening measures have been completed.3 in contrast, eight countries were in the single-digit range, and only one had annual inflation higher than 30 percent. Macroeconoomic Stability At the same time, real interest rates were between 6 per- If the process of financial integration is to be effective and cent and 15 percent, fiscal balances ranged between -6 per- durable, the countries of Latin America first will have to cent and +3 percent of GDP, public debt stood between 28 achieve macroeconomic stability, or to be well on their way percent and 106 percent of GDP, and the current account toward it. Capital flows can be very volatile, and exchange balance was between -7 percent and + 1 percent, while risks and sovereign risks very high if a country's macroeco- unemployment varied from 3 percent to 17 percent. nomic fundamentals are not sound, and those risks will for The European Union is clearly far ahead of Latin Amer- all practical purposes impede the smooth exchange of ica in Eerms of the convergence of such variables as inflation financial services.4 and real interest rates, and only slightly ahead when it The process of financial integration does not demand comes to the current account deficit on the balance of pay- absolute coerdination among different countries' macro- ments. Yet the fact is that Latin America has better indica- economic policies, but it does require a strong and demon- tors and greater convergence than the European Union in strable commitment to maintain macroeconomic stability. the fiscal sphere, both in terms of deficits and public debt For example, countries that decide to proceed with finan- as a percentage of GDP, as well as in unemployment rates. cial integration might well take it upon themselves to As noted above, it is very difficult in Latin America to comply with criteria similar to those of Maastricht, think of total macroeconomic coordination and a common whereby they agree to set bands or ranges within which the set of overall objectives, given countries' different stages of key macroeconomic variables of the different countries are progress toward reform and stabilization, their varying allowed to move. This does not mean that their macroeco- degrees of external openness, the distinct vulnerabilities of nomic policy tools must be made uniform. Indeed, to the their economic structures, and the disparate sources of extent that they share the same final objectives and produce shocks to which they are exposed. What would seem use- converging mnacroeconomic results, a financial integration ful and desirable would be to create and mobilize an effi- scheme could embrace, for example, countries that follow a cient and regular mechanism for exchanging information fixed nominal exchange rate with full convertibility among national economic authorities, and to reach a sig- (Argentina) and others with a discretionary monetary pol- nificant degree of consensus on the macroeconomic goals icy that is firmly anchored in explicitly stated inflation that they should be pursuing.5 goals (Chile), and still other countries where monetary pol- A regional or subregional agency, as appropriate, might icy is formulated and implemented on the basis of mone- be able to monitor and publish regular information that tary or credit aggregates (Peru and Mexico). countries, and their financial institutions and commercial The economic landscape of Latin America during the banks, could use in making decisions about international- 1990s has been marked by the introduction of structural ization.6 Similarly, the integration process would be given a reforms and stabilization programs aimed at reducing boost if countries could agree to cast their economic policies inflation ancd achieving sustained economic growth. This within certain ranges of current account deficit-to-GDP, has induced a greater degree of convergence in the various inflation and fiscal deficit-to-GDP, and to adopt indicators macroeconornic indicators for countries of the region than of external and internal solvency and liquidity, such as net prevailed during the 1970s and 1980s. Whereas in 1990 foreign-currency external debt, the ratio of public domestic the average inflation rate stood at about 1,200 percent a debt to GDP, the ratio of international reserves to imports, year, by 1996 it had fallen to 19 percent, and spanned a and the level of international reserves in relation to the sum range of 1 percent to 47 percent. Again in 1990, of the 19 of external short-term debt plus domestic short-term finan- Latin American countries reporting statistics to the U.N. cial liabilities in foreign currency, among others. Economic Commission for Latin America and the With respect to this last kind of indicator, the sustain- Caribbean (ECLAC), four were experiencing hyperinfla- ability of the external balance, measured in terms of flows tion, only six had inflation of less than 30 percent a year, and reflected in the current account balance, needs to be 215 TRADE: TOWARDS OPEN REGIONAL[SM supplemented with a proper analysis of the situation in rium interest rate is higher than the international, and terms of stocks, which is reflected in the strength of exter- there is a sudden liberalization of the capital account. On nal solvency and international liquidity indicators. In par- the other hand, if the internal equilibrium allows domestic ticular, countries should have a sound international reserve interest rates to stay roughly in line with those abroad for position, both with respect to the short-term maturity reasonable ranges of country risk and exchange-rate expec- structure of their external debt and their domestic public tations, financial opening can be approached more quickly debt and the foreign-currency liabilities of their banking and completely, without having to mortgage the stability system. This would help to induce a more stable behavior of the external sector, jeopardize inflation goals or put the in net capital flows, and it would above all avoid sudden real exchange rate at risk. interruptions in capital inflows and a stampede to take On the other hand, explicit or implicit guarantees of money out.7 interest rates and exchange rates exacerbate the inflow of capital and give rise to behavior that implies moral hazard Liberalization, Regulation and Financial Supervision and/or risk. The best way to avoid these problems is to fol- Together with stability and greater macroeconomic con- low a combination of monetary and exchange-rate policies vergence, regional financial integration requires the liber- that will allow for greater flexibility and a greater market alization of domestic financial systems, in the sense of role in the setting of interest rates and the exchange rate. eliminating quantitative controls and the selective alloca- If financial flows handled through the banking system tion of credit, deregulating interest rates, regulating the are to be appropriately channeled to the various sectors of granting of banking licenses, etc., so as to encourage the economy, and since loans are not generally secured, greater market competition and setting reserve ratios that there is a need domestically for provisions based on indi- will reconcile the maintenance of adequate bank liquidity vidual credit risk, and internationally for country-risk pro- levels with the development of financial intermediation. It visions, to the extent that there is a risk that the foreign is essential to recognize at the outset, however, that such loan cannot be repaid, and that there may be restriction on liberalization must go hand-in-hand with proper banking the free transfer of foreign exchange to the creditor coun- supervision and regulation. try. Regulations also need to take account of the risk of cur- It must be borne in mind that a simultaneous process rency mismatches. It should be noted that a deeper and of domestic and external financial liberalization can lead more internationalized financial system can digest capital to heightened competition in financial markets and make flows with less volatility in interest rates, the exchange rate it more difficult to digest inflows of capital, particularly in and the prices of assets. the banking sector. Experience shows that this situation On the banking front, what is needed is that the partici- contributes to a temporary and unsustainable rise in asset pating countries' institutions should regulate themselves in prices, especially for non-tradable assets, and it at least light of the ratings they earn for their behavior in terms of allows, if it does not actively encourage, the emergence of solvency, liquidity, profitability, risk and efficiency. At the price "bubbles" in the stock market, as well as the over- initial stage, it is the state that should do the ranking of bank valuation of domestic currencies, all of which stimulates instruments, so as to ensure that a clear and transparent excessive domestic spending and raises the current methodology is established and will prevail. Subsequently, at account deficit of the balance of payments, factors that later stages, this role can be handed over to independent pri- work against exchange-rate and financial stability and vate institutions, including the international rating agencies, overall macroeconomic equilibrium. Moreover, in these particularly in the case where banks make use of international cases there is a tendency to relax both supervision and self- capital markets to fund themselves. Similarly, a key to stabil- regulation of the banks, which can lower the quality of the ity in the banking sector is to ensure compliance with pru- banking system's loans portfolio and increase its risks, dential standards on capital adequacy, consolidated supervi- while the banks lose their effectiveness as transmission sion and the granting of banking licenses. belts for monetary policy.8 In addition, because it is not feasible or credible in most The foregoing will be even more true in cases where, countries to withdraw the implicit or legal guarantee of prior to external financial opening, the domestic equilib- deposits, the problems arising from abuse of such guaran- 216 INTEGRATION OF FINANCIAL SERVICES tees must be minimized by stressing the role of proper known. Much less attention has been paid, however, to banking supervision and regulation. It may be advisable to such aspects as analyzing the current status of financial supplement that system by arranging for substantial integration in the region, and comparing it with the involvement: of foreign banks whose home countries prac- progress in the financial integration of Europe, which is tice serious and high-quality banking supervision and who now in its final process. Nor has there been much explo- have, moreover, open lines of liquidity to their parent insti- ration of the differences between unilateral integration, in tutions, if problems should arise.9 which a single country decides to open its financial markets To the extent that banks from one country set up shop to others, and the "integrationist" approach, in which a in another country, there is a need, as well, for the mutual bloc of countries agree on the reciprocal opening of their exchange of information among the supervisors of banking domestic systems. and financial institutions in those countries. Although When examining the feasibility, speed and depth of the financial integration is easier when countries have liberal- financial integration process in Latin America, we need to ized their banking systems and have equipped themselves consider a number of institutional and structural questions, with a proper regulatory and supervisory system, it is such as the magnitude of the differences in such variables in essential that such regulation and supervision share certain income per capita, country risk, the financial depth of cap- basic principles in common, even if the banking structure ital markets, the degree of banking intermediation, and the in those countries is not necessarily homogenous. For type and quality of supervision in the various countries. example, in Argentina certain authorities are calling for Note, for example, that GDP-per-capita ratios are much the establishment of a "narrow banking" scheme, while more similar, and relatively more stable, among the coun- full-service banks are expanding in Mexico. It is even pos- tries of the European Union than is the case in Latin Amer- sible for financial systems to coexist where development ica. Homogeneous data for 1993, shown in Table 2, show banks play an important role. that the highest GDP per capita in the European Union Nevertheless, despite the fact that member countries of an integration scheme do not share the same approach to TABLE 2 banking regulation in all its details, integration of financial GDP Per Capita in 1993 (US$1 services in Latin America must not countenance, much less LATIN AMERICA EUROPEAN COMMUNITY encourage, regulatory arbitrage-for example, the creation ' ' ~~~~~~~~ ~ ~~~Argentina 7,220 Austria 23,510 of tax havens and regulatory paradises. These tend to intro- Bolivia 760 Belgium 21,650 duce a form of unfair competition and, moreover, to pro- Brazil 2,930 Denmark 26,730 Chile 3,170 Finland 19,300 voke instability for countries where supervision and taxes Colombia 1,400 France 22,490 are taken seriously. Furthermore, financial integration Costa Rica 2,150 Germany 23,560 must strive :o avoid major tax distortions and substantive Dominica Republic 1,230 Greece 7,390 Ecuador 1,200 Ireland 13,000 regulatory differences between domestic and international El Salvador 1,320 Italy 19,840 markets, since otherwise there will be significant financial Guatemala 1,100 Luxemburg 37,320 Guyana 350 Netherlands 20,950 disintermediation in the countries that are trying to inte- Honduras 600 Portugal 9,130 grate themselves financially. It must be stressed that incen- Mexico 3,610 Spain 13,590 tives for local and foreign capital to seek out places with Nicaragua 340 Sweden 24,740 Panama 2,600 UK 18,060 less onerous regulatory and taxation requirements must be Paraguay 1,510 resisted, because it is those places that present the greatest Peru 1,490 Umuguay 3,830 country risk and the least macroeconomic stability. Veneztula 2,840 Financial Integration and Cross-Border LATIN AMERICA EUROPEAN COMMUNITY Competition in Financial Services in Standard Deviation 1645.01 Standard Deviation 7435.40 Average 2086.84 Average 20084.00 Latin Ameirica Coefficient of Variation 0.79 Coefficient of Variation 0.37 The long-term costs and benefits and the transition prob- Max./Min. Ratio 21.24 Max./Min. Ratio 5.05 lems associa-ed with external financial opening are well- Source: World Barnk, World Tables 1995. 217 TRADE. TOWARDS OPEN REGIONALISM was five times greater than the lowest, while in Latin practices,10 also show major discrepancies among the coun- America the gap reflected a factor of more than 21. More- tries of Latin America. For example, as Table 5 shows, the over, variability in GDP per capita in Europe was less than ratio of capital to total assets for the Argentine and Colom- half that of our region, as can be seen from the coefficients bian banking system is about 15 percent, while in Peru of variation shown in Table 2. and Brazil it is around 9 percent; in Mexico it stands at 7 As Table 3 makes clear, Latin America also shows sig- percent, and in Chile it is at slightly more than 5 percent. nificant differences in country risk: In 1996 the spread over Overdue loans as a percentage of the total lending portfo- LIBOR on external dollar-denominated bond issues with lio stand at 1 percent in Chile, and about 7 percent in Peru, comparable maturities from Argentina, Brazil and Mexico Mexico, Brazil and Colombia, while in Argentina they are was more than four times as high as that for companies as high as 12 percent. Gross interest margins (as a percent- operating in Chile. In the first four months of 1997, these age of assets) in Chile are one-half those recorded for Peru spread differentials declined to between 3 and 3.5 times. and Brazil, and less than half of those in Argentina, while There are also sharp discrepancies in indicators of the support costs in Chile, again as a percentage of assets, are scope and characteristics of domestic banking markets in some 50 percent lower than in Argentina, Mexico and Latin America. As can be seen from Tablc 4, intcrnal credit Peru, and lcss than onc-third the level of those in Brazil from the banking system amounted to 58 percent of GDP and Colombia. Paradoxically, as can be seen from Tables 1 in Chile, while the figure for Mexico was 48 percent; and 5, the Chilean banking system is the one with the low- Brazil, 38 percent; Argentina, 26 percent; and Peru, only est spreads and the highest degree of concentration; in con- 11 percent. Again, the M2/GDP ratio was 29 percent for trast, Colombia, where there is the highest degree of mar- Chile, and around 23 percent in Brazil, Costa Rica and ket concentration, is the country with the lowest spreads. Mexico, while in Argentina and Peru it was half that level. Finally, the dynamism and strength of financial inte- And the ratio of demand deposits to GDP, which indicates gration will also depend on the initial conditions in which the degree of banking intermediation, reached a maximum the various countries find themselves. For example, the in Chile, at about 6 percent, while the lowest value-that nature and the speed of financial integration are sure to be for Argentina-was no more than 2 percent of GDP. affected if one country has a problem of external over- Indicators of soundness, management and competition indebtedness or a banking system that is weak or in col- in the banking industry, which are difficult to compare lapse. Considerations of this kind, as noted in the preced- given the differences in certain accounting and regulatory ing paragraphs, have played and continue to play a major TABLE 3 Spreads in Bonds Issued in US$a (basis prints) COUNTRY-RISK 1996 1997b CLASSIFICATIONSc ARGENTINA Companies 474,30 280,80 BB Banks - 283,00 BRAZIL Companies 403,95 292,50 B+ Banks 332,80 216,00 CHILE Companies 85,80 103,30 A- Banks - - MEXICO Companies 455,40 362,50 BB Banks 414,00 a. Refers to the differential between the yield on the bond issued and the relevant U.S. Treasury Bond. b. Information through Apeil 1997. c. May 1997 long-term international country risk classificarions in foreign currency Losors: Bloomberg, Cmz Blanca and Standard & Poors. 218 INTEGRATION OF FINANCIAL SERVICES TABLE 4 tionalized" financial integration. In effect, the 1990s have Indicators of [Financial Deepening given rise to greater flows of capital within the region, DOMESTIC DEMAND with a particular emphasis on real estate investment. Par- Ml/ M21 CREDITI DEPOSITSI ticularly noteworthy here is the volume of direct invest- GDP GDP GDP GDP ment that has been flowing from and to countries of the ARGENTINA 1994 5.8 13.2 24.2 1.8 region, originating in or destined for capital market com- 1995 5.9 12.6 25.6 2.0 panies and institutions such as banks and pension funds.12 These flows have been stimulated by the structural BRAZIL 6.8 31.5 55.1 reforms, the greater openness and the privatizations that 1995 4.9 24.3 38.0 2.1 most countries of Latin America have embarked upon over the last decade. CHILE 1994 8.6 28.0 58.7 5.6 In particular, there has been a remarkable growth in 1995 8.7 29.2 58.4 5.7 regional banking services networks offered both by Latin COSTA RICA American financial groups and those from outside the 1994 12.4 27.5 23.9 6.6 region. Among the former, the Brazilian banks are promi- 1995 9.2 23.9 ND 4.1 nent in the countries of Mercosur, Colombian and MEXICO Venezuelan groups are active in their own countries as well 1994 10.2 19.4 51.5 6.4 as in Ecuador and Chilean financial groups are engaged in 1995 8.4 23.0 47.5 4.8 banking circles and pension funds in a number of countries PERU of the region. The growing presence of non-Latin Ameri- 1994 5.1 12.5 9.4 2.9 can banks is also striking, especially that of the Spanish 1995 5.7 13.2 11.2 3.3 banks, led by the Bilbao-Vizcaya and the Santander, which Source: IMF Statistics, IES 1996. have been buying up domestic financial institutions, banks role in explaining the slow pace of financial integration in and non-banks alike.'3 Latin America.'1 Note that the more this kind of de facto integration proceeds among the same banks and/or groups in different Financial Integration in Practice countries, the more quickly local or domestic problems, What we have been witnessing in Latin America in recent including runs on banks, are likely to be transmitted years is a de facto and rather hesitant form of "non-institu- regionally. This can only underline the importance of swift TABLE 5 Financial Systems Indicators (percentages) MEXICO BRAZILb COLOMBIA PERU CHILE ARGENTINA' INDICATORS 1994 1995 1996a 1994 1995 1996 1994 1995 1996 1994 1995 1996 1994 1995 1996 1994 1995 1996a Spread 5.3 4.9 3.3 9.2 8.7 6.9 8.0 7.5 7.7 6.2 7.5 6.2 3.9 3.4 3.1 5.7 5.8 5.0 Admin. Expenses/ Assens 5.6 5.4 4.5 9.3 9.2 8.9 8.9 9.0 8.5 9.2 10.1 5.9 3.0 2.6 2.4 7.2 5.7 5.1 Loans due/ Loans 7.3 7.0 6.9 7.0 10.0 6.8 5.5 6.3 7.6 8.0 6.9 6.7 1.1 1.0 1.0 13.1 13.5 12.0 Provisions/Loans 3.5 5.1 6.4 1.7 9.0 7.1 2.1 2.9 3.2 8.9 5.9 5.7 2.2 1.8 1.7 7.6 7.2 6.7 Liquid Assets/ Assets 23.3 23.6 23.7 21.1 15.6 12.5 16.3 12.1 9.2 29.4 29.6 30.1 8.7 8.7 8.9 11.5 6.6 6.9 Capital/Assets 5.4 6.8 6.7 10.2 8.9 9.4 13.5 14.3 15.6 9.6 9.2 8.5 5.7 5.5 5.3 16.5 15.2 14.5 Growth Rate of Assets 43.8 22.6 28.6 -3.6 1.2 9.5 32.2 25.8 23.8 31.5 35.8 45.8 38.4 22.3 19.1 11.0 20.2 14.8 a. Information as of September 1996. b. Indicators for the 17 largest banks. c. Indicators for the 20 largest banks, except for "liquid assetslassets," which is the ratio foe the entire system. Source: Salomon Bro -hers 219 TRADE. TOWARDS OPEN REGIONALISM and effective cooperation and coordination among national Mercosur and NAFTA supervisory powers in Latin America, and the need to As noted above, Latin America has made little progress ensure that national regulations and rules are guided by toward financial integration as such, and is still very much similar principles and standards, so as to avoid arbitrage or in a learning process. Until now, the region's financial "shopping" among regulatory systems. internationalization has proceeded essentially through a Increasingly, we find that a greater number of Latin series of unilateral openings, despite initiatives such as American countries and companies, financial or non-finan- NAFTA and Mercosur, and the worldwide trend toward cial, are placing financial instruments, such as shares the formation of bid trading blocs. (ADRs in the United States, for example), bonds and con- Mercosutr is a trade agreement among Argentina, Brazil, vertible bonds on international capital markets, which are Paraguay and Uruguay, to which Chile has recently much more highly developed than those of the region. This adhered, that has had the essential objective of reducing tar- kind of activity could be attempted within the region iffs and other barriers to foreign trade. That agreement does itself, but this would require national reforms in the insti- not include the services sector as an area of negotiation. tutional, taxation and legal areas, and greater harmoniza- NAFTA is a much more comprehensive agreement, tion of the relevant accounting data, of the kind that will since the treaty includes principles and provisions govern- allow Latin American securities markets, separately and ing investment and financial services. With respect to the independently, but in a complementary manner, to engage latter, the key clauses of NAFTA seek to ensure that for- in local securities transactions.'4 eign providers of services can have access to local markets, More generally, experience suggests that international- while at rhe same time maintaining certain principles of ization of the banking system goes hand-in-hand with prudential regulation. trade liberalization and the growth of trade and investment Generally speaking, NAFTA covers cross-border trade abroad. At first, the banks' international business tends to in financial services, including banking, insurance and concentrate in external financial investments and in the securities intermediation products. That is to say, it allows financing of foreign trade among neighboring countries. foreign providers to offer such services without the need to Subsequently, and still within the foreign trade sphere, the have a physical presence in the country of destination. In a banks will participate in financing linked to trade with manner that is integrally linked to cross-border trade in third countries, and will then start to become involved in services, NAFTA seeks to ensure the free flow of capital. In financing domestic companies that are "going interna- addition, NAFTA protects the right of establishment- tional." The most advanced stage of international banking i.e., the ability of suppliers from one member country to is typified by making loans to non-residents, and for phys- have a commercial presence in another member country- ical facilities located abroad. and guarantees national treatment for the physical installa- To sum up, the integration of banking services in the tion of a financial institntion in another country. To put it countries of Latin America is still at a rudimentary stage,'5 another way, it forbids discrimination on the basis of the as a result of the many different monetary and exchange origin of the investment. In this sense, it applies the prin- regimes, high and differentiated sovereign and exchange ciple of most-favored-nation. risk perceptions, the widely diverse solvency situations of Yet, notwithstanding these principles, which are the banking systems, and a lack of understanding and intended to open the way to trade and participation in each coordination in regulation and supervision among the other's financial markets, NAFTA also provides for safe- region's countries. Even efforts at trade integration are pro- guards, either by exceptions or by reservations, to protect ceeding only slowly. For all of these reasons, it is unrealis- each country's approach to regulation and to macroeco- tic to expect a powerful movement toward Latin American nomic stabilization.'6 financial and banking integration in the foreseeable future, The need to preserve the differences among the various beyond the developments discussed earlier. This is not to regulatory approaches was a key factor in the financial ser- say, however, that we should not press ahead with institu- vices negotiations under NAFTA. The pertinent clause of tional, legal, regulatory and accounting reforms aimed at the treaty ensures that member countries retain their eventually achieving such an objective. autonomy with respect to prudential regulation and to tak- 220 INTEGRATION OF FIN'ANCIAL SERVICES ing stabilization measures that affect the financial sector. nature, have fallen far behind progress on trade in goods.'8 Since the regulatory approaches of the countries involved The Uruguay Round was the first occasion that services were very different, there was no possibility of arriving at became an important feature of efforts to achieve openness full harmonization, as the members of the European Union and internationalization. have been able to do, for example. The most important multilateral accord affecting inter- While the focus of NAFTA is fairly liberal with respect national trade in this area has been the OECD agreement to cross-borcler trade in financial services, it states that the on codes for liberalizing capital movements. These codes right of establishment may be exercised under any desired call for the progressive liberalization of capital flows and legal form, whether as a branch, a subsidiary or an affili- financial services, and national treatment for foreign service ate-but within the legal and regulatory framework of the providers. In recent years, OECD members have made host country. drastic changes to their domestic legislation in order to In general, free-trade agreements in the financial ser- comply with the spirit of these codes. vices sector .mnay involve distinct levels of commitments by The European Community, for its part, has agreed on member countries, ranging from freedom for cross-border the free right of establishment in its members' markets, a trade, the right of establishment and national treatment, to liberalized regime for the delivery of cross-border financial the creation of a common market with fully harmonized services, and the harmonization of the principal rules of regulations. regulation among member countries. To establish a com- Freedom to conduct cross-border trade in financial ser- mercial presence in any country of the Community, a bank vices was not an issue of debate in the NAFTA context, of any member countries needs only one license-i.e., a mainly because both Canada and the United States already home-country authorization to operate regionally. For pur- maintained a liberal approach to capital movements. In the poses of supervision, the regulatory principles of the home absence of exchange controls, residents of one country are country prevail, and there is mutual recognition of regula- at full liberty to make financial and non-financial invest- tory regimes among all EC members. There are also agreed ments in other member countries. What obstacles there are standards relating to capital adequacy and transparency in this area arise essentially from the lack of tax harmo- rules. nization and from certain exchange controls that Mexico More generally, in the developed world, and particularly has maintained, despite the thorough-going process of in Europe, the various ways of conducting international external opening in which it has engaged.'7 trade in banking services are governed by the following Consequently, the negotiations on financial services rules: In the case of cross-border banking, the rules of the focused on the right of establishment subject to national home country prevail; in the case of subsidiaries in another treatment--i.e., on allowing foreign providers of financial country, the rules of the host country prevail; and in the services to have a physical commercial presence, subject to case of branch banks, the rules must be harmonized. the same rules as those governing domestic financial insti- In Latin America, although for the moment, as explained, tutions in the country concerned. there can be no harmonization of regulations as in Europe, Although the elements discussed here are certainly any real progress toward financial integration must presup- encouraging the more direct integration of financial ser- pose certain basic, shared elements of regulation. vices in the region, both Mercosur and NAFTA are treaties In the first place, regulation and supervision must be of limited scope both in terms of their field of application conducted with a view to maintaining a solvent, stable and and their country coverage. If the "Americas Initiative" efficient financial and payments system. Additional official becomes a reality, we hope that the process of financial objectives must be to keep runs on banks to a minimum integratiotl will be deepened and accelerated. and to reduce as far as possible any implicit or explicit guarantees on bank liabilities. Finally, the authority must Regulatory Principles To Promote Greater be constantly on guard to maintain public confidence in Financiall Integration in Latin America the financial system. Worldwide, moves toward multilateral liberalization of Another important element is that the banking system trade in services, and in particular those of a financial must be subject to a series of prudential regulatory mea- 221 TRADE TOWARDS OPEN REGIONALISM sures. Banks must maintain an adequate level of capitaliza- requirement for demand deposits in domestic currency is 9 tion, consistent with the risks they incur. These risks may percent, and for term deposits and other term obligations be of various kinds-for example, credit risk, currency risk 3.6 percent, except for obligations of more than one year, and interest-rate risk. For purposes of quantifying these which are not subject to cash reserves. In addition, any risks, account must be taken of net positions, exchange- deposits or obligations in foreign currency are subject to a rate volatility and asset prices.'9 Even in the case of com- 30 percent cash reserve. In any case, the elements that must plicated derivatives like options, capitalization must be be considered in managing liquidity effectively will required to reflect all the dimensions of risk. include proper information systems, an analysis of net An adequate level of capitalization represents a safety funding requirements under different scenarios and diver- cushion for coping with potential losses and for avoiding sification of funding sources. moral hazard arid risk. At the present time, the Basle cap- The institutional arrangements governing the paymenits ital adequacy requirement for banks of 8 percent is a stan- system must be such as to minimize periods of counter- dard that is frequently applied internationally. Yet that party risk that could have a negative impact on the public standard must be regarded as a minimum. In particular, finances. In particular, an efficient payments system should given the greater volatility of Latin American economies eliminate any credit risk to the central bank. Payments and the vulnerability of their financial systems, banks should be made against instantaneous verification that engaged in the process of financial integration should have funds are available in the current accounts deposited in the a capital adequacy indicator of at least 10 percent. These banks and in the central bank. higher capital requirements must not give rise to differ- Generally speaking, banks keep a combination of mar- ences in the weighting of risk factors for countries of the ketable assets and others, such as placements, for which region, since that would constitute manipulation that there is no liquid market. In order to foster self-regulation would allow capital requirements to be altered indirectly. and transparency, assets should be valued at market prices. Currently, Argentina calculates capital requirements as a Placements should be valued according to the repayment function of asset risk, including market risk, while Brazil prospects of the credit, by means of provisioning require- demands liquidity compatible with the risk structure of ments, structured on the basis of the expected cash flow assets, and Chile requires a maximum debt-to-capital ratio from the borrower and the quality of the guarantee.20 The of 20. frequent valuation of a bank's assets at market prices, reg- Regulation also should limit the concentration of assets ular risk classifications of the liabilities they issue, the pre- and transactions with related parties, so as to enhance sentation of monthly financial statements and verification diversification and minimize conflicts of interest. Loans to of the quality and accuracy of accounting information by related borrowers should be subject to strict limitations, to external auditors are all ways of providing transparent capital deduction rules or to the need to collateralize such information to the market; this is a key aspect of self-reg- loans. In any case, the terms and conditions of such loans ulation. The presence of international rating agencies and must be based not only on current market factors, but also external auditors in different countries should facilitate the on medium- and longer-term fundamentals. standardization of information and thereby stimulate the In addition, banks, for the most part, are exposed to a region's financial integration. term conversion process that makes them relatively illiq- The process of granting licenses needs to strike a deli- uid. Cash-reserve requirements can protect the payments cate balance between being overly generous, which can system and induce banks to act prudently in the face of lead to problems of management and solvency, and being possible withdrawals of deposits. In Argentina a cash too restrictive, which can discourage competition in the reserve of 18 percent is required for demand deposits and industry. In any case, the process must be selective, with for term deposits maturing within 3 months, and this per- objective rules and clearly identified barriers to entry. In centage is reduced gradually to zero for transactions of particular, it is important to avoid bouts of destructive more than one year. In Brazil, demand deposits have a 75 competition and the adverse selection of new bank owners. percent reserve requirement, term deposits 20 percent, and The foregoing presupposes a clear understanding that the savings deposits 15 percent. In Chile, the cash-reserve costs of entry to and exit from the banking system are very 2 22 INTEGRATION OF FINANCIAL SERVICES different. As is well known, a bank's exit from the market bank's belonging to the group. Moreover, if the market can entail negative externalities and can be very costly. regards the conglomerate as a single unit, then it is futile to In conclusion, there can generally be no such thing as try to keep its dealings separate, and consolidated supervi- free access to the banking business. This is true for reasons sion becomes not only necessary but urgent. In these cases, that are both microeconomic (to avoid the adverse selection a functional regulation is not enough, since the risk of the of owners) and macroeconomic (since a banking market in conglomerate as a whole is greater than the sum of the risks which many entities are competing may tend to fall into of its component entities because of the "systemic" risk generalized insolvency, with consequent problems for the involved.2" The failure to exercise consolidated supervision entire systenm). Thus, in Chile, for example, the granting of over offshore banking activities involving the banking sys- new banking licenses has in effect been closed since 1990, tems of Ecuador and Venezuela was one of the prime causes and the authLorities have complete discretion over the issu- of the recent financial crises in those countries. ing of any new ones, while in Uruguay new licenses are In addition to the foregoing, when problems arise with limited to no more than 10 percent of the number of banks certain banks, national regulations should provide for a in existence at the end of the previous year. Some countries, safety net with a set of instruments (carefully limited to such as Brazil and Mexico, still maintain restrictions on reduce moral hazard) such as guarantees for demand foreign ownership. In Brazil, foreign participations require deposits and means of injecting liquidity into the banking prior approval through a presidential decree, and the num- systems, so as to ensure a greater degree of stability in the ber of foreign bank branches is de facto limited to the financial system. Schemes of this kind, however, can count at the present time. In Mexico, foreign investors may increase the risk of imprudent behavior on the part of the own up to 30 percent of a bank's capital, a restriction that banks. To limit this risk, the insurance offered can be for is being gradually relaxed for NAFTA member countries. partial coverage only and can carry a premium that varies The scope of activities in which banks engage should with the risk incurred, while liquidity credits can be made also be kept under review, so as to avoid overextension of the available only against collateral of the lowest risk. In addi- safety net and to limit the risks arising from new activities. tion, regulation must provide for a set of corrective actions With respect to monitoring, experience suggests that in the case of banks that are undercapitalized, whereby supervisors should rely on a combination of reports of the shareholders and/or creditors can be called on to provide "Camel" type, which describe capital adequacy in terms of preventive capitalization and so minimize the risk of bank- the risks assumed, the quality of assets and management, ruptcy. In extreme situations, where a banking entity is no and the structure of revenues and liquidity, with on-site longer financially viable, the supervisor can and should inspections, which will provide for clarity with respect to encourage a merger or find another bank to take it over. In the quality of the portfolio, the institution's management the end, the authority must have the power to declare a capabilities, and the quality of the internal control systems bank insolvent and to wind it up if necessary, so as to pro- in place. tect the stability of the rest of the financial system.22 The As banking groups begin to take on new financing guiding principle should be that losses, just as much as activities, whether in the domestic arena or abroad, it profits, are for the account of the private sector, at least for becomes increasingly important to exercise consolidated the former shareholders, and to some extent for the credi- supervision that will provide a comprehensive overview of tors and the depositors. And when fraud is involved in a the risks assumed. Supervision on a consolidated basis lim- bankruptcy, the owners should be liable with their personal its the risks of "contagion" arising from inter-group posi- assets. This principle will minimize situations of moral tions, significant common positions of the group, psycho- risk. logical conragion and possible double-leveraging of capital. To summarize, it is essential, in our view, that Latin In effect, loans and other related group activities, the use of American countries adopt jointly the regulatory principles capital for purposes of multiplying the group's dealings and we have outlined. Nevertheless, given the great differences its debt, and the risk of over-concentration of loans and to be found in organization and institutional structure of assets in certain activities by various companies within the the financial and banking industry, and in the cyclical eco- group-all can increase the risk and vulnerability of any nomic conditions facing the industry in the various coun- 223 TRADE TOWARDS OPEN RFGIONAT ISM tries of Latin America, we do not see it as feasible to har- sionally. Moreover, and even more important, there should monize specific rules. Indeed, it is very difficult to achieve be supervision agreements between the member countries complete harmonization in the fields of banking activity, of any financial integration system that provide for the the banking structure, the treatment of branches abroad smooth and timely exchange of information, and allow for and the safety of deposits, to cite just a few examples. on-site inspection of branches and subsidiaries by the With respect to the internationalization of the banking home-country authorities, so that they can maintain effec- industry through cross-border credit transactions, this tive supervision.23 demands regulatory features that are different from those Most of these principles have been adopted by the G- 10 that apply to domestic lending. While it is true in theory within the framework of the Basle Committee on Banking that opening up the spectrum of business possibilities for Regulations and Supervisory Practices.24 the banks by authorizing credits abroad should allow them Consideration also needs to be given to the aspects of to diversify risk more effectively, it must be remembered taxation, exchange controls and investment protection. that operations of this kind are exposed to country risk, Particular care should be taken in dealing with "tax exchange risk and differences in legal, regulatory and havens," which generally exercise only lax supervision, and supervisory treatment. where it is common not only to conduct banking activities Pursuing the issue of internationalization, while an that are prohibited at home, but to evade domestic regula- individual debtor's risk can be measured on the basis of the tions altogether. Clearly, it is important to avoid regulatory expected flow of repayments and the quality of the guaran- arbitrage associated with situations of this kind. In this tees offered, an additional degree of caution is needed in respect, international agreements need to be sufficiently light of the increased difficulty of executing those guaran- broad in their geographical coverage that they can avoid tees and the possible divergence of the currency in which regulatory competition with countries that do not share the debtor's revenues are earned from that in which the these principles. Otherwise, a potentially explosive situa- loan is denominated. In addition, limits are required to tion could develop that would threaten the stability and ensure proper diversification by country, and provisions for solvency of the banking systems of the countries that are country risk are needed that will take account of the polit- members of the agreement. ical risk and the transfer risk involved in international transactions. Finally, there are financial risks to be covered, Summary and Conclusions in particular the risks of currency mismatching and inter- Globalization is one of the trends that has shaped the est-rate risks. development of the world's economies in recent times. When it comes to the physical establishment of sub- Latin America is no stranger to this phenomenon, particu- sidiaries, branches and affiliates of financial groups, a series larly in light of the institutional and structural reforms it of regulatory measures are required, such as consolidated has undertaken, to varying degrees, and the steps it has supervision by the home country, which can take an taken to open its economies. In particular, the region has overview of the risk situation of the holding or parent com- moved decisively to integrate itself into world capital mar- pany and all the activities flowing from it. Such regulation kets, by participating more actively in the flow of foreign must also ensure that the consolidated capital is adequate direct investment and by placing its shares and bonds on in light of the risks to be undertaken, and that authoriza- international financial markets. tion has been obtained from both the home and the host In terms of economic integration, despite the progress country. At the same time, it must enforce the principles of that has been made, Latin America is clearly well behind integrity and solvency of the owners. In particular, bank Europe. Europe's successful experience shows that the shareholders must demonstrate that their financial and changes associated with the processes of financial integra- legal conduct are unimpeachable and that they are eco- tion make themselves felt only gradually, and only after nomically sound enough to undertake banking activities, commercial integration has been achieved. More specifi- to minimize the risk of adverse selection of controlling cally, the case of Europe demonstrates that economic intc- shareholders of a bank, and allow a reasonable assurance gration is a process in which countries must advance step that the institutions will be managed seriously and profes- by step, dealing first with trade, then with finance, and 224 INTEGRATION OF FINANCIAL SERVICES finally, at the most advanced stage of integration, with banking systems with the principles cited above are pre- monetary union. conditions for making Latin American financial inte- The imminent creation of the European Monetary gration viable in the first place and for reinforcing it Union, taken together with the various country agree- subsequently. ments and subregional integration schemes between Latin In terms of the feasibility, speed and depth of the finan- America and the European Union, and the growing partic- cial integration process in Latin America, we must take ipation of Latin American countries in the work of the into account the differences in such variables as income per OECD and the Bank for International Settlements (BIS) in capita, country risk, the financial depth of capital markets, Basle, represent an excellent opportunity for the region to the level of bank intermediation and the methods and internalize the institutional, juridical, normative, regula- quality of banking supervision in different countries. tory and economic aspects of financial integration.25 Moreover, the dynamism and durability of financial inte- A key element in any process of financial integration is gration will also depend on the initial conditions prevail- the ability to achieve and sustain basic macroeconomic ing in those countries. For example, the shape and speed at equilibrium. This is a necessary condition for strengthen- which financial integration proceeds cannot be immune to ing domestic financial systems, and to establish that mini- situations such as those where one country has a problem mum degree of stability that is required to conduct effec- of external over-indebtedness or a banking system that is tive international trade in financial assets and services, on the verge of collapse. There is no doubt that considera- particularly with respect to country risk and exchange risk. tions of this kind have been and continue to be a major fac- This does not call for full policy coordination, but rather tor in explaining why financial integration in Latin Amer- a strong ancd real commitment on the part of the members ica has made such little progress. of any integration scheme to achieve and maintain macro- What we are witnessing today is a kind of non-institu- economic stability. As a minimum, it would seem desirable tionalized financial integration in Latin America. The and appropriate for such countries to move forward in the 1990s have seen a surge of capital flows within the region, area of coordinating their macroeconomic information. A with a particular emphasis on real estate investment. Espe- regional or subregional agency could monitor and publish cially noteworthy here is the volume of direct investment regular reports on such information, so that countries and that has been flowing from and to countries of the region, their banks can take it into account in their plans for originating in or destined for capital market companies "going international." and institutions such as banks and pension funds. It is Another prior condition for financial integration is noteworthy that this process has taken place through their domestic finanicial liberalization, which must, of course, go owners, whether individuals or corporations, rather than hand-in-hand with a strict and efficient system of banking through the institutions themselves, given the legal supervision and regulation. restrictions in place and the lack of consolidated supervi- To make progress toward greater financial integration it sion. Moreover, there has been a recent reawakening of is essential, on the one hand, that there be a smooth and interest on the part of non-Latin American banks, particu- timely reciprocal flow of information among the region's larly from Spain, in acquiring domestic capital market national supervisors and, on the other hand, that they players and in taking an active part in the region's financial share the basic prudenrial and regulatory principles and banking business. adopted by the Basle Committee on Banking Regulations On another front, an increasing number of Latin Amer- and Supervisory Practices. These relate in particular to the ican countries and companies, financial or non-financial, adequacy cf capital bases and to consolidated supervision. are placing financial instruments such as shares (ADRs in Yet given the differences in the institutional and organiza- the United States, for example), bonds and convertible tional structure of Latin America's financial sectors-and bonds on international capital markets, which are much in the initial economic conditions facing them-it would more highly developed than those of the region. This kind seem unrealistic to try to harmonize the specific points of of activity is beginning to appear within the region itself, financial and banking rules. The exchange of information but if it is to become deeper and more widespread, it will among authorities and compliance on the part of national require national reforms in the institutional, taxation and 225 TRADE: TOWARDS OPEN RFCIONALISM legal areas, and greater harmonization of the relevant international economy. Those blocs, because of their size and nature, accounting data. should be less sensitive to external economic repercussions on their Until now, the region's financial internationalization has development and their policies. This, in turn, may lead to their eco- nomic cycles being increasingly out of phase, and it could induce proceeded essentially through a series of unilateral openings, greater exchange- and interest-rate volatility among the major cur- despite initiatives like NAFTA and Mercosur, and the rencies, including most certainly the dollar and the euro. Conse- worldwide trend toward the formation of big trading blocs. quently, Latin America will need to prepare itself to face greater The integration of banking services in the countries of Latin instability in coming years, both in the prices of the major curren- America is in fact in its early stages, primarily as a reflection cies and in international interest rates. Of high and differentiated sovereign and exchange risk per- 2. The theory and practice of economic integration have focused almost exclusively on the implications for the allocation of resources. ceptions, the widely diverse solvency situations of the bank- With the sole exception of the analysis and debate that surrounded ing systems, and a lack of understanding and coordination in the question of optimum currency zones in the 1970s-and that financial regulation in general (and banking supervision in never had much of an impact on the integration effort, although it particular) among the region's countries. It would be a clear may have been helpful in refining some individual countries' step forward in the process of financial integration if coun- exchange systems-there has been little analytic or empirical atteen- tries would allow and facilitate reciprocal access to their tion paid to the macroeconomic, financial and monetary effects of markets by guaranteeing the right of establishment to for- integration. 3. There are countries within the region that have introduced policies that are formally very "liberal and open" with respect to the must be recognized that even trade integration schemes in capital account of the balance of payments, and that, nevertheless, Latin America have run into roadblocks, which makes it have recorded higher domestic interest rates and greater differentials seem unrealistic to think that full financial and banking against international rates than other countries that embarked on a integration can be achieved any time in the near future. This more gradual and selective strategy of financial opening. This para- is not to say, however, that we should not move ahead with dox seems to be explained primarily by their greater degree of coun- the kind of institutional, legal and regulatory reforms that try risk, or by the devaluation that they may have practiced. It is by no means clear, therefore, that a swift and wholesale financial open- t[hat objective will demand. Latin America's present eco- ing will allow a country to generate greater confidence among inter- nomic outlook is broadly favorable, since we can expect that national creditors and foreign investors, or to break its way effec- its current business and trading dynamism will persist, that tively and permanently into international capitaL markets. inflation will be brought further under control and that eco- 4. The cases of Mexico in 1994 and the Czech Republic and Thai- nomic reforms will be reinforced. In addition, the relatively land in 1997 bear witness to the importance of having sustainable low degree of banking intermediation in the region, the current account deficits as a precondition for effective and durable financial integration. high returns recorded in the financial sector and the growing 5. Recently, because of the problems surrounding Brazil's adop- opportunities for doing business on a regional scale are all tion of special short-term import financing measures, Mercosur factors that augur well for regional financial integration. agreed to set up a Macroeconomic Coordination Group for the bloc Nevertheless, whatever the degree of progress it may and to work through it on the preparation of a scheme of this kind. make in achieving financial integration, Latin America 6. An initiative in this direction was considered in New Orleans must press ahead in its national efforts toward stabilization in 1996, during the first meeting of hemispheric finance ministers, and structural reform, in order to provide a solid under- where the countries attending agreed to adhere to the information pinning for the long-term sustainabiliny of its economies. disclosure standards of the International Monetary Fund. 7. Guillermo Calvo, among other authors, has dealt with this A responsible approach to domestic macroeconomic man- topic. agement, the encouragement of domestic savings, and 8. This is because, during such episodes, domestic interest rates attention to the growth and proper funding of capital for- will be much higher than normal, which leads to lower effective mation are essential preconditions for Latin America's eco- repayments of credits from the banking system. This means that the nomic development. banks' debtors no longer face, in practice, any budgetary restraints on their spending decisions, and monetary policy loses most of its effectiveness. Notes 9. Table 1 shows that foreign banks in selected Latin Ameri- 1. In strictly economic terms, the European Monetary Union will can countries have a share of between about 22 percent and 30 help to highlight even more clearly the role of economic blocs in the percent of the credit market, except for Brazil, where at present 226 INTEGRATION OF FINANCIAL SERVICES foreign banks account for only 7 percent of total credit from the 18. Recently the Basle Committee on Banking Regulation and banking system. Supervisory Practices published a paper that sets out 28 "Core Prin- 10. For example, while Chile and Mexico apply monetary correc- ciples for Effective Banking Supervision." These principles were tion and have financing indexation mechanisms, this is not the case developed by the countries belonging to the G-10, in collaboration in Argentina or Brazil. There are also different rules regarding credit with the supervisory authorities of 15 emerging economies, includ- arrears for purposes of classifying overdue loans in each of these coun- ing Brazil, Chile and Mexico. The paper provides a useful frame of tries. When it comes to the banks' financial investments, in Chile reference for supervisors and banks worldwide. these are valued at market prices if they are for more than one year, 19. Increasing competition between banks and other types of and at cost of acquisition if they are short-term; in Argentina, on the financial institutions, and the growing sophistication of new banking other hand, short-term instruments are valued at market prices, and products like derivatives, make it essential that banking risks be man- those of more than one year are recorded at their cost of acquisition. aged in a professional way, both by the banks and by the supervisors. In Mexico, they are recorded at their cost of acquisition, while in 20. Nevertheless, under some circumstances (typically in the Brazil they are carried at market value, whatever their term. With growth phase of the business cycle or when some key prices, such as respect to the taxation of credit, Chile imposes a stamp tax, the exchange rare, are out of line with their fundamental determi- Argentina applies the VAT, and in Brazil there is a so-called financial nants), the market may send "distorted" signals, either directly or in operations tax, which varies depending on the maturity of the instru- relation, for example, to the value of the guarantees. One way to alle- ment. In Mexico, loans are not subject to any taxes at all. viate this problem is to require banks to put a minimum percentage 11. Note, however, that institutional and structural differences, of their own capital into the financing of an investment project and discrepancies in economic cycles and starting conditions, can and/or into the acquisition of an asset. also present opportunities and foster trade in the financial area. 21. The growing interrelationship between banks, insurance 12. Note that for most of the region's financial institutions, companies, pension funds and securities dealers requires a consoli- inclu-ding the banks, this process has taken place through their own- dated approach to supervision that will deal squarely with the main ers, whether individuals or corporations, rather than through the institutional challenge, which is to coordinate the financial system's institutions themselves, given the legal restrictions in place and the various supervisory bodies. lack of any properly consolidated supervision. 22. On this point, it is important that national legislation and 13. The Grupo Santander, for example, has interests in 14 Latin legal systems should be equipped to deal promptly and effectively American countries, representing an investment of US$3.5 billion with the particular features of financial activity in order to resolve and accounting for more than 50 percent of the group's entire invest- crises and/or bankruptcies. ment outside Spain; its Latin American assets total some US$45 bil- 23. Such exchanges of information will of course have to over- lion, and has created jobs for more than 43,000 people in the region. come the obstacles presented by the banking-secrecy laws that exist 14. Mexico is about to open its stock market to foreign compa- in every country. nies that quote their shares in the United States. Similarly, Argentina 24. See footnote 18. and Mexico, following the recent example of Brazil, are the two 25. To this end, it would be useful for the region's central banks countries in the region that are planning to introduce the use of cer- to participate more fully and actively in the regular BIS meetings to tificates of deposit for shares. In Chile, the Congress is currently con- which they are invited. It was only recently that Brazil and Mexico sidering legislation for the Bolsa Internaciiosal, an international stock became the first two Latin American countries to join the BIS as exchange that will allow foreign securities to be traded locally. members. The BIS could become a forum for examining questions 15. By way of example, despite the sharp growth in international about exchange rates, monetary and financial issues among Latin trade, electronic transfers among banks in different countries are not American monetary authorities, and for sharing experiences with done directly, but via the United States, much the same as happens other emerging economies. In addition, the BIS offets the chance to in the regions air transport business. compare knowledge and practices with European countries; this 16. While it is no doubt desirable to allow a degree of national could help countries of our region to internalize the lessons of Euro- maneuvering room, it must be recognized that this may slow pean financial integration. Sinsilarly, Latin American countries are progress toward full financial integration. increasingly active in the various workings of the OECD in Paris, 17. Latin America is still a long way from free trade in cross-bor- and are maintaining close ties with the European Union through the der capital movements, because several countries-e.g., Brazil, Commission in Brussels, which has played and continues to play a Colombia and Chile-apply restrictions on capital inflows, major role in Europe's financial and monetary integration. 227 XZ. Macroeconomic Policy Coordination Free Trade and Macroeconomic Policy B ARRY E I CH E NGRE E N HAT ARE THE CONNECTIONS BETWEEN THE ESTABLISHMENT OF A FREE-TRADE area and the conduct of macroeconomic policy? Most economists' instinctual answer would be "none." Trade theory and policy tend to be taught in one part of the international economics curriculum, open economic macroeconomics in another. Trade theory concerns itself with the allocation of resources between sec- ; T ~~tors producing importables, exportables and nontraded goods, open-economy macroeconomics with financial variables like inflation and exchange rates. Rarely do the twain meet. They should meet, of course, as the instructor struggling policies for bilateral real and nominal exchange rates and through the standard syllabus is reminded each semester enthusiastically trade with such a country like any other? by the outspoken student at the rear of the lecture hall. What is the evidence on these questions from existing free- Inflation and exchange rates, not to mention cyclical fluc- trade areas, and what are the implications for Mercosur? tuations, affect resource-allocation. Trade transmits distur- In the first section I consider two channels through bances to inflation and unemployment in a variety of mod- which erratic, uncoordinated policies might prevent the els. Even a moment of reflection will make clear that the full gains from trade from being realized: that they might connections between trade and the macroeconomy run in give rise to exchange-rate and relative-price fluctuations both directions. that discourage producers from allocating resources along In this paper I seek to put some structure on the out- lines of comparative advantage; and that they might give spoken student's intervention, asking questions like the rise to exchange-rate and relative-price fluctuations that following: To what extent must participants in an FTA encourage domestic agents to lobby for protection. Both coordinate their macroeconomic policies in order to fully theory and empirical evidence provide support for the reap the benefits of trade? Are there political or economic operation of these two channels; they suggest that the first mechanisms that, in the presence of inadequate policy effect, while significant, is likely to be small, but that the coordination, prevent the gains from trade from being real- second effect may have major consequences. ized? Should countries hesitate to negotiate a free-trade In the second section I turn the question around and agreement with a potential partner who has yet to put in consider the implications of trade for macroeconomic pol- place sound and stable macroeconomic policies, or should icy. Much recent research has explored the implications of they disregard the consequences of those macroeconomic capital mobility for monetary and fiscal policy. Here I Barry Eichengreen is a Professor at the University of California, Berkeley. Much-appreciated comments were provided by Guillermo Calvo, Claudio Loser and an anonymous referee. Maria Soledad Martinez-Peria and Carlos Arteta provided sterling research assistance. This paper was written during a visit to the Center for Advanced Study in the Behavioral Sciences, whose hospitality is acknowledged with thanks. 231 TRADE- TOWARDS OPEN REGIONALISM revisit the prior question and the older literature on the to currency fluctuations will mount. This will encourage implications of trade openness for the conduct of stabiliza- efforts to coordinate exchange-rate and macroeconomic tion policies. policies among the member states of the FTA. But the The next section considers the relevance of these argu- options for more closely harmonizing the macroeconomic ments to Mercosur. Despite the FTA's short history, experi- policies of the Mercosur countries are relatively limited. ence with Mercosur confirms that real exchange-rate fluc- Feasible policy coordination is likely to focus on harmoniz- tuations, which give rise to sharp shifts in competitiveness ing policy rules and guidelines. This could involve, for and sudden import surges, threaten a protectionist back- example, the adoption of common inflation targets for lash against trade liberalization. At various times these monetary policy. More ambitious policy-coordination exer- pressures have been transmitted from and felt by both of cises are unlikely to be successful, however. the two large partners in the FTA, Argentina and Brazil- The final section concludes by returning once more to the most prominently in trade in motor vehicles and compo- connections between free trade and macroeconomic policy. nents, but in other sectors as well. To date, the protection- ist pressure induced by macroeconomic instability has not Policy Instability and Trade seriously undermined political support for the FTA, but The impact on trade of macroeconomic policy generally the danger remains. and macroeconomic policy instability in particular is likely For those concerned with the connections between free- to depend on the form taken by that instability. Still, for a trade areas and macroeconomic policy, the European Union range of plausible specifications, the uncertainty associated (EU) is an especially pertinent case.i The Treaty of Rome, with policy instability and the protectionist pressure it under which the members of the European Economic Com- evokes may prevent the gains from trade from being fully munity committed to establishing a customs union, sin- realized. Although results vary by model, both theory and gled out the exchange rate as a matter of "common inter- empirical evidence point in this direction. est." Movements toward freer intra-EU trade have consistently been accompanied by efforts to coordinate Theory macroeconomic policies more closely. While this dynamic The earliest partial-equilibrium analysis of the impact of is particularly evident in the current drive for monetary policy-induced exchange-rate uncertainty on trade is unification, concern with macroeconomic policy coordina- Ethier (1973). Using a mean-variance specification of the tion in general and exchange-rate stability in particular has expected utility of a representative firm competing atom- been a constant of EU history. The fact that the EU is the istically in international markets, Ethier showed that the most extensive and successful free-trade area (and, for those volume of trade declines with the level of exchange-rate who insist it is more than a trading arrangement, that its uncertainty. Hooper and Kohlhagen (1978) demonstrated customs union has led to even deeper forms of integration) subsequently that the essential result carries over to suggests that the coordination of macroeconomic policies monopolistic competition. may be essential to the success of a regional trade arrange- Neither of these models took general-equilibrium ment. The fourth section takes a detailed and ultimately repercussions into account. Perhaps the simplest general- somewhat skeptical look at this proposition. equilibrium model that can be used to analyze the impact With this analysis in hand, section five returns to the of policy instability on trade is that of Kemp and Liviatan core question: In how much and in what kind of macro- (1973),V who consider a two-country Ricardian model in economic policy coordination must the Mercosur countries which both countries typically specialize completely, pro- engage to reap the full gains of regional integration? My ducing and exporting one of two traded goods.3 They conclusions are follows: In the short run, the need for assume that the relative price of the two goods (the terms macroeconomic policy coordination is limited. But as Latin of trade) fluctuates randomly. We can think of the terms- American financial markets become still more open and of-trade fluctuations as reflecting macroeconomic policy- integrated, and as competition between the manufacturing instability abroad. Intuitively, in periods when Argentine and service sectors of the Mercosur economies grows even inflation is relatively high and variable, Brazil's terms of more intense, the potential for a protectionist backlash due trade should be relatively unstable. 232 MACROECONOMIC POLICY COORDINATION To provide a channel for relative price fluctuations to plausible conditions it is lower than in the certainty case.6 affect trade, it is assumed that production decisions occur Thus, the simplest model yields strong conclusions. prior to the resolution of uncertainty (on the basis of expec- Relative price uncertainty due to policy instability may tations) and that trade and consumption decisions occur discourage risk-adverse producers from specializing along afterward.4 Consumers are assumed to have standard quasi- lines of comparative advantage. Even if firms continue to concave preferences and to be risk-averse. specialize, consumers may lobby for protection for the It simplifies matters to start with the special case where import-competing sector. Under plausible conditions the producers are risk-neutral.5 Given this assumption, price volume of trade will be lower than otherwise. A macroeco- uncertainty does not affect output: Producers maximize nomic policy that creates relative price uncertainty may expected profits, and our Ricardian economy still special- then be incompatible with fully exploiting the gains from izes completely in production. This is not a happy situa- trade.7 tion for consumers, however. The relative price of the exportable good can turn out to be high or low. If its price Evidence is very low, it will turn out to be even more expensive to There is no body of empirical work directly concerned with purchase the imported good ex post than it would have policy instability and trade. But literature on related sub- been to produce it at home ex ante. In this state of the jects speaks obliquely to the question. I consider the works world the utility of domestic residents is very low. If they concerned with three links in the chain connecting macro are sufficiently risk-averse, they will want the insurance of policy to trade and trade protection: the impact of some domestic production of the importable. This makes exchange-rate variability on trade, the impact of policy them better off in states of the world where the relative instability on the exchange rate and the liking of protec- price of the importable turns out to be high (specifically, tionism to policy instability. higher than the ex ante marginal-and, given the Ricar- The findings of the literature on exchange-rate variabil- dian assumption, average-rate of substitution between ity and trade would seem to be contradictory and incon- exportable and importable goods), and worse off in states of clusive. In fact, however, the most recent wave of (third- the world where the relative price of the importable turns generation) studies yields strong, consistent conclusions. out to be low. Because they are risk-averse, they attach par- The first wave of methodologically rudimentary studies ticular valuz to utility in the state of the world where the reported implausibly large effects. Thus, Cushman's (1983) terms of trade and national income are unfavorable; hence, estimates for OECD countries in the 1970s implied a 7.2 they wish to insure themselves by having some domestic percent decline in the level of intra-EC trade due to production of the imported good. They will lobby, there- exchange-rate variability. The second generation of work fore, for government policies to limit imports and other- (e.g., IMF 1984, Gotur 1985) subjected these specifica- wise encourage the import-competing sector. The protec- tions to sensitivity analysis and established that their tionist pressure induced by price uncertainty will cause the results were not robust; this did much to inform the skep- pattern of domestic production to resemble the pattern of tical view of the relationship. In recent years, a third wave domestic consumption more closely. Specialization along of more methodologically sophisticated studies has docu- lines of (ex ante) comparative advantage will be limited, as mented the existence of small but robust negative effects of will the volume of trade. exchange-rate variability on trade. Bini-Smaghi (1991), If we now add risk-aversion on the part of producers and employing a dynamic specification, finds evidence of a assume that they have no recourse to financial markets or small but consistently significant negative effect. Chowd- opportunity to establish foreign subsidiaries, they will hury (1993) obtains essentially the same result from an respond to uncertainty by producing a diversified portfolio error-correction model suited to the stochastic properties of of both exportables and importables to hedge against rela- the relevant time series. Arize (1996) obtains essentially tive price risk. The degree of specialization will be less the same results, also using the error-correction methodol- than in the absence of uncertainty even if we allow for self- ogy but limiting his analysis to European countries. Kroner insurance by consumers. It is not, in general, possible to and Lastrapes (1993) impose rationality on perceived say what happens to the volume of trade, although under exchange-rate volatility and again find a small, statistically 233 TRADE: TOW'ARDS OPEN REGIONALISM significant impact on their reduced form equation for puted on the basis of the coefficient of variation of quar- export volumes. Holly (1995) estimates the structural terly data).'1 export supply and demand equations corresponding to Edwards' framework provides a logical starting point Kroner and Lastrapes' reduced-form system: He confirms for analyzing the impact of policy volatility on real their finding of a small negative effect and shows that it exchange rates and on political support for open trade in operates mainly by depressing export supply. Frankel and integrating economies. A logical question to ask is not just Wei (1993) analyze this relationship using a gravity model how policy variability affects the real exchange rate, as in framework in which the volume of bilateral trade depends Edwards' paper, but also how increasing openness alters the on countries' distance, population, national income and, in link. In an effort to answer it, I updated Edwards' analysis addition, the variability of the nominal or real exchange to cover the 1979-93 period. I used annual data; thus, the rate. Instrumenting the exchange-rate variability term analysis can be thought of as extending Edwards' analysis using the variability of relative national money supplies to of long-run real exchange-rate variability. Using data for control for endogeneity, they too find a statistically signif- all non-oil developing countries for which the relevant icant but economically small impact of exchange-rate vari- variables were available, I re-estimated his basic specifica- ability on trade.8 tion. This relates the coefficient of variation of the real Thus, recent studies support the hypothesis of a nega- effective exchange rate to the coefficient of variation of the tive relationship between exchange-rate variability and terms of trade, the coefficient of variation of real GDP trade while all but universally finding that the posited growth, the coefficient of variation of the rate of money effect is small.9 growth, the coefficient of variation of the rate of domestic Is the exchange-rate variability considered in these credit growth, the coefficient of variation of inflation, the analyses plausibly linked to policy? A variety of studies average inflation rate and two measures of the volatility of suggest that this is the case. DeGrauwe, Janssens and the nominal exchange rate, one based on the effective Lelianert (1985) report strong evidence that real exchange- exchange rate, one based on the bilateral rate vis-a-vis the rate variability increases with the variability of inflation dollar and finally a measure of openness (the export/GDP and money growth rates. DeGrauwe and Rosiers (1984) ratio).'2 The dependent variable, the price of tradables rel- find that monetary instability is one of the principal deter- ative to nontradables, is constructed as a trade-weighted minants of real exchange-rate variability in a cross-section average of the price level in each country's 10 most impor- of 39 developed and developing countries. Bayoumi and I tant trading partners, converted by the period-average (1996) analyze the determinants of real and nominal exchange rate, relative to the home price level.'3 All vari- exchange rates in a sample of industrial countries. We find ables are expressed in logs, and standard IMF sources are that the relative rate of growth of money supplies is an used throughout. important determinant of bilateral exchange-rate variabil- The results of estimating this model are shown in Table ity; country pairs for which the average rate of growth of 1. While the size and significance of the coefficients differs the money supply differs significantly tend to have more by specification, the variability of the terms of trade, the variable (real and nominal) bilateral rates.10 variability of the nominal effective exchange rate, and the In a study closely related to the issues of concern to this average level of inflation matter for long-run real exchange- paper, Edwards (1987) analyzes the determinants of both rate variability.4 This is consistent with the view that eco- long-run and short-run real exchange-rate variability (the nomic structure and policy (in particular, the policies price of tradables relative to nontradables) for a sample of affecting inflation and the nominal exchange rate) matter developing countries, finding that real rate variability is for the volatility of the real exchange rate.'5 caused by both monetary and real disturbances, with real I next addressed the question of whether any of these variables (notably the external terms of trade) being rela- effects, which presumably work to undermine political tively important in the explanation of long-run instability support for free trade, grow stronger as integration pro- (computed on the basis of the coefficient of variation of ceeds. In addition to including the level of openness, as annual real exchange-rate data) and nominal variables before, I extended the specification by interacting openness being relatively important for short-run volatility (com- with other independent variables; a positive coefficient on 234 MACROECONOMIC POLICY COORDINATION TABLE 1 (1993) analyzes the determinants of non-tariff barriers in Determinants of Real Exchange-Rate Variability, 1979-93 the United States using industry-level data."1 He finds sig- (standard errors in parentheses) nificant upward pressure on non-tariff barriers in industries (1) (21 (3) experiencing increases in import penetration. His result is Constant -3.49 -2.57 -2.78 consistent with the notion that policy shifts which give rise (0.72) (0.64) (0.75) to changes in import penetration affect the endogenously Terms of trade variability 0.62 0.40 0.43 determined level of protection. But for policy variability (0.14) (0.13) (0.14) GDP growth variability 0.34 0.26 0.25 that increased the variability of import penetration ratios (0.07) (0.12) (0.11) to raise the average level of protection, it would be neces- Money growth variability -0.14 -0.15 - (0.16) (0.15) sary to establish the existence of a ratchet effect (for Credit growth variability -0.20 -0.11 -0.13 whether positive changes in import penetration had a (0.15) (0.13) (0.14) Inflation variability -0.28 - -0.16 larger Impact than negative ones or, equivalently, for (0.13) (1.10) whether the variability of the import penetration ratio Bilateral excharnge-rate when added also entered with a positive coefficient). This variability 0.01 0.06 0.08 (0.12) (0. 10) (0.13) has yet to be done. Effective exchanige-rare variability 0.62 0.77 0.81 O (0.17) (0.21) (0.24) Overall Conclusions Average inflation rate 0.52 - - The impact on trade of macroeconomic policy generally (0.1 1) and macroeconomic policy instability in particular is one of Openness 0.37 0.07 0.02 (0.21) (0.27) (0.26) the more contested areas in the literature of international n 30 30 30 trade and finance. A variety of theoretical models suggest R2 .81 .61 .71 F 9.57 7.51 7.72 that uncertainty associated with policy instability (and the protectionist pressure it evokes) may prevent the gains SOrce, See text from trade from being fully realized, although the work of Helpman and Razin, among others, should remind us that an interaction term would suggest that the effect in ques- these results require strong assumptions whose generality tion, linking economic structure or policy instability to may be questioned. Recent empirical work also points in real exchange-rate variability, is stronger in more open this direction: It all but uniformly suggests that policy- economies. The baseline model included all of the regres- induced exchange-rate instability discourages trade, sors, as in t]he first column of Table 1. although the magnitude of the effect is small. More needs The results of these tests for increasing sensitivity of the to be done to establish the channels through which this real exchange rate with growing openness turn out to be effect operates, and in particular whether the relative-price mixed. The only variables for which there is evidence of uncertainty to which policy uncertainty gives rise is a different effects in more open economies are the variability source of protectionist pressure. of credit growth and the variability of GDP growth. But in both cases, the interaction terms enter with positive coeffi- Trade Openness and Macroeconomic Policy cients: that for the variability of credit growth differs from The literature on the implications of openness to trade for zero at the 95 percent level, while that for GDP growth the conduct of macroeconomic policy goes back to differs from zero at the 90 percent level. Thus, there is Machlup (1943), Meade (1951) and Mundell (1963). They some evidence that the variability of both domestic mone- suggested that fiscal policy was likely to be less effective for tary policy and cyclical instability have a greater impact on stabilization in more open economies. If it is costly to the variability of the real exchange rate in more open increase the size of the deficit or to change levels of public economies.'6 But this evidence is not overwhelming. spending and taxation, then we should expect to see fiscal Substantial literature supports the hypothesis that real policy used less actively in economies where multipliers are exchange-rate variability and import penetration are smaller due to the existence of larger import leakages. If sources of protectionist pressure. For example, Trefler most of the benefits of a countercyclical fiscal policy leak 235 TRADE TOWARDS OPEN REGIONALISM abroad in small open economies, then we should expect to Kingdom).'9 It is not feasible to follow Bayoumi and Mas- see them make less use of automatic stabilizers. Bayoumi son's procedure and estimate the 16-equation system using and I (1995) provide evidence consistent with this view, three-stage least squares, since this procedure uses personal showing that individual U.S. states, which are far more income and openness for each of the 16 countries as an open than the country as a whole, do less than a quarter of instrument, introducing serious degrees-of-freedom and the automatic stabilization of the federal government."8 multicolinearity problems. Instead, I estimated the system Working in the other direction is the fact that more open using a two-step procedure. In the first step I estimated the economies tend to have larger public sectors. Rodrik (1996) equation for each country with instrumental variables, has documented this regularity, arguing that the public sec- using as instruments own-country lagged income, the con- tor plays a risk-reducing role in economies exposed to sig- stant and the time trend. In the second step I substituted nificant amounts of external risk (terms-of-trade uncer- the fitted values for the actual values of the independent tainty, concentration of exports, etc.). Countries in which variables, and re-estimated the resulting set of equations the public sector makes up a large share of national income using seemingly unrelated regressions to take into account may find it relatively easy to effect large changes in tax rev- the cross-country correlation of the error terms. enues or government consumption (as percentage of GDP), The result of estimating Bayoumi and Masson's basic facilitating the use of automatic stabilizers. specification (with standard errors in parentheses, and Thus, whether automatic stabilizers are used more or where D denotes the first difference) is: less actively in more open economies is an empirical ques- tion. To compare the use of fiscal policies for countercycli- D Post-Fisc Income = Constant Terms + 0.900 D Pre-Fisc cal stabilization in more and less open economies, we Income (0.009) extended the analysis of Bayoumi and Masson (1995). These authors compare the change in incomes before and Following Bayoumi and Masson, the coefficients on the after taxes in five European countries: Germany, France, the slope coefficient is constrained to be equal across countries. United Kingdom, the Netherlands and Belgium. Using (In contrast, the country-specific constant terms, not time series for each country, they regress the change in per- reported here, are not constrained across equations.) A one sonal income net of taxes and transfers on a constant and dollar change in pre-fisc personal income leads to a 90 cent the change in personal income gross of taxes and transfers. change in post-tax-and-transfer personal income. This (Personal income is expressed in per capita terms, con- effect is smaller than that obtained by Bayoumi and Mas- verted into dollars, and normalized by average per capita son, reflecting the longer sample period, the larger sample income for all five countries.) Since fiscal policies are of countries, and the different estimation strategy. endogenous, the estimates are obtained using instrumental Extending the specification to include the interaction variables (where the instruments include the constant, a term between pre-fisc income and openness (where open- time trend and the first lag of the pre-tax and transfer ness is measured as the ratio of exports to GDP), one income series). To obtain efficient estimates that take into obtains: account the cross-country correlation of error terms, esti- mates are by three-stage least squares. D Post-Fisc Income = Constant Terms + 0.169 D Pre- To analyze differences in the extent of automatic stabi- Fisc Income (0.022) lization in more and less open economies, I re-estimated Bayoumi and Masson's model, adding an interaction term + 1.379 (D Pre-Fisc Income * Openness) between pre-tax income and openness (where the latter is (0.041) defined as exports as a share of GDP). All variables come from the OECD's Main Economic Indicators- The data period The coefficient of 0.1 69 on Pre-Fic InJcome means that a is 1970-92, and the extended sample includes 16 countries closed economy would offset some 83 percent (1 to 0.169) (Austria, Australia, Belgium, Canada, Finland, France, of cyclical fluctuations in pre-tax personal income using its Germany, Greece, Italy, Japan, the Netherlands, Spain, tax and transfer system. The positive coefficient on the Sweden, Switzerland, the United States and the United interaction term confirms that more open economies do 236 MACROECONOMIC POLICY COORDINATION less automatic fiscal stabilization. For a country with an real) exchange rates; this regularity is evident in both the exportlGDP ratio of 0.5, the slope coefficient is 0.169 + 1970s and 1980s. (0.5 * 1.379), or 0.86. Such an economy thus offsets a Thus, there is some evidence that both fiscal and mone- smaller 14 percent of cyclical fluctuations in pre-tax-and- tary policies are used less actively in more open economies, transfer pcrsonal income. as if there exists a higher ratio of costs to benefits of adjust- This would appear to be the first systematic confirma- ing those instruments. In more open economies, Romer's tion in the literature that open economies do less automatic results suggest, the relatively active use of monetary policy fiscal stabilization. It is consistent with the notion that runs the risk of destabilizing the terms of trade. New because more fiscal impulses leak abroad, they are less results similarly suggest that larger leakages lead more inclined to undertake the same costly countercyclical open economies to undertake less automatic fiscal stabi- adjustments in government expenditure and taxes as less lization. If these findings are correct, they suggest that open economies. integration has important implications for the conduct of The discussion to this point has focused on fiscal policy; stabilization policies. I now turn to monetary policy. Romer (1993) has examined the connections between monetary policy and openness in Applicability to Mercosur a framework emphasizing time-consistency problems. Tak- There is considerable evidence that limited macroeconomic ing the Barro-Gordon model as his point of departure, he policy harmonization, especially between Brazil and posits that the temptation for the central bank to engineer Argentina, has created strains within Mercosur. These a surprise rmonetary expansion in an effort to raise output problems have generally been associated with exchange- will be less in more open economies. If the first variable to rate-based stabilizations, which have created problems of respond to l he money surprise is the nominal exchange rate real overvaluation and intense competitive pressure on and prices are slower to adjust, as in Dornbusch (1976), import-competing producers, on the one hand, and inade- then the rnore open the economy, the greater the welfare quate monetary and fiscal control leading to rapid currency loss due to this short-run deterioration in the terms of depreciation and real undervaluation, with incidental ben- trade, and the greater the incentive for central bankers to efits for exporters, on the other. The consequent political restrain the temptation to unleash a monetary surprise. problems have been most severe when policy imbalances Romer's evidence, based on a multivariate regression analy- affect industries like motor vehicles, which are both polit- sis, is supportive of the hypothesis that policy is less infla- ically powerful and seen by the respective governments as tionary in more open economies.20 integral to their national industrial strategies. This result is consistent with the implications for mon- In a sense, these problems can be traced to the fact that etary policy of the literature on optimum currency areas effective inflation stabilization in Argentina, which dates (see, Munclell 1961). More open economies, according to to the Convertibility Law of 1991, coincided with the ini- this theory, have a greater incentive to harmonize their tialing of the Mercosur Agreement during that same year. monetary policies with those of their trading partners and Given the failure of previous stabilization efforts, even so to stabilize their exchange rates. The more open the econ- radical a change in the macroeconomic policy regime as omy, the less adequate are the means-of-payment and unit- Argentina instituted in 1991 took time to gain credibility. of-account services provided by the domestic currency, and Inflation persisted. Wholesale prices rose by around 6 per- the greater the incentive to peg the nominal exchange rate cent in 1992 and 3 percent in 1993. In Brazil, meanwhile, in order to import these services from abroad. We should inflation continued unabated and, typically for a hyperin- therefore expect to see more open economies make less use flating economy, the exchange rate rose even faster than of their monetary autonomy for stabilization purposes, prices (aside from periodic, unsuccessful attempts at stabi- other things being equal. lization). As early as 1992 this led Argentina to impose The empirical evidence is consistent with this view. For anti-dumping measures against Brazilian farm machinery example, Bayoumi and I (1997), analyzing a sample of and spark plugs. This was followed by similar measures industrial countries, find that countries which trade more against steel, refrigerators, paper, agricultural machinery heavily tend to intervene to stabilize their (nominal and and textiles. By early 1994, policy divergences had created 237 TRADE: TOWARDS OPEN REGIONALISM serious problems for a variety of Argentine industries. In to bring vehicles from one market to the other subject to the first week of January, Argentina introduced protection- zero tariffs, providing that they sent vehicles in the same ist measures against Brazilian chemicals. At a meeting of value in the other direction-hardly meaningful trade lib- the Brazilian and Argentine presidents later that month, eralization.24 the latter complained that Argentina had become a Clearly, this "auto row" was not entirely a product of "dumping ground" for cheap Brazilian industrial products macroeconomic policy imbalances; the priority the Brazil- like textiles, paper and steel.2' ian government attaches to developing a world-class auto- Further evidence can be found in the reaction to the sta- motive industry would have fanned conflict with bilization of Brazil's exchange rate in the second half of Argentina even in their absence. There is no question, 1994. Improved policy harmonization allowed Brazilian however, that those conflicts were inflamed by macroeco- and Argentine officials to negotiate a bilateral deal for the nomic problems. This point is reinforced to the extent that automotive industry-one of the sensitive sectors granted conflicts were not limited to the automotive sector. To cite special treatment under Mercosur-allowing Brazil easier one recent example, in March of this year Brazil, concerned access to the Argentine parts market and eliminating the about overvaluation and a growing trade deficit attribut- quotas both countries placed on imports of trucks and able to exchange-rate-based stabilization in conjunction vans. By the beginning of the next year Argentine officials with limited fiscal control, placed restrictions on the could reaffirm their commitment to Mercosur partly financing of imports. This elicited strong criticism from because "the partners are now following broadly similar Argentina, which saw these measures prompted by macro- economic policies. Argentina had been fearful that Brazil's economic policy imbalances as a threat to Mercosur.25 undervalued currency (and its overvalued one) would have Thus, the history of Mercosur confirms that poorly coor- hit its industry hard. Now Argentines say that the strength dinated macroeconomic policies can strain a free-trade of the Brazilian currency ... gives them a chance to increase agreement.26 Repeatedly, exchange-rate-based stabiliza- their exports to their neighbors. "22 tions and sharp exchange-rate movements associated with But as with most exchange-rate-based stabilizations, sudden loss of fiscal and monetary control have threatened the Brazilian program led to a widening of the trade political support for free trade within the region. The two deficit. The competitiveness and import-penetration prob- largest members of the FTA, Brazil and Argentina, have at lem shifted from Argentina to Brazil. In particular, stabi- various times been the source and the subject of these prob- lization led to renewed problems for the Brazilian motor- lems. Thus, the corrosive effect of the lack of macroeco- vehicle industry. While Argentina was still only a minor nomic policy coordination in the Southern Cone can and source of Brazil's motor imports, the two countries were in has cut both ways. And yet even the very pronounced direct competition for foreign investment by U.S. and macroeconomic imbalances that have affected trade rela- European producers-foreign investment that is attracted tions between the Mercosur partners have not permanently only to platforms with access to Brazil's large domestic disrupted their efforts to deepen and broaden their free- market. Thus a sharp rise in Argentine exports to Brazil, trade agreement. even one starting from a relatively low base, could have major sectoral implications. In early 1995 these considera- The Case of the European Union tions led Brazil to double its tariffs on automobiles to 70 At this point, many readers will probably conclude that percent and simultaneously reimpose quotas, limiting none of the preceding arguments provides a compelling imports to 100,000 units for the second half of 1995.23 reason why the members of a free-trade area must coordi- Protests by Buenos Aires then led to some improvement in nate their macroeconomic policies. True, when countries Brazil's treatment of Argentina. When the World Trade follow unstable policies, risk-averse producers may fail to Organization condemned the quotas, Brazil removed them specialize along lines of comparative advantage, and their but retained the 70 percent import duties, which it com- countries may fail to fully reap potential gains from trade. mitted to phase out by 1999. In January 1996 the two Consumers and producers discomforted by terms-of-trade countries reached a bilateral agreement (an exception from fluctuations may lobby for policies that limit the extent to Mercosur) allowing producers operating in both countries which trade is actually free. But there is little that is con- 238 MACROECONOMIC POLICY COORDINATION tent-specific about these implications: Stable, coordinated tions (the theoretical possibility discussed in the first sec- policies are always better than unstable, uncoordinated tion above). Because intra-European trade is so extensive policies. There is nothing fundamentally different in the and interest groups are so well-organized, this political context of firee-trade areas. threat is especially potent in Europe. Only the history of the European Union (nee European Although examples of its operation can be found Community) is difficult to square with this conclusion. throughout the history of the European Community, this Since the inception of their regional arrangement, EU offi- dynamic is particularly clear in the response to the pound cials have regarded deepening their customs union and cul- sterling and Italian lira's forced departure from the Euro- tivating the close coordination of macroeconomic policies pean Monetary System in 1992. The depreciation of these as connected tasks. The Treaty of Rome, the document cre- currencies enhanced the international competitive position ating the European Economic Community, identified the of Italian and British exporters and intensified the profit exchange rates of the member states as matters of "common squeeze on their competitors elsewhere in the EU. In interest" (as mentioned in the introduction). The revalua- response, German machine-tool makers complained to tion of the Dutch guilder and German mark in 1961 high- their government that their Italian competitors were lighted the possibility that exchange-rate movements unfairly underselling them in international markets. The within the Community might create problems for its oper- Hoover Company shifted some of its operations from Dijon ation, prompting the creation of the Committee of Central to Scotland, eliciting complaints from French trade unions. Bank Governors to facilitate the closer coordination of French Prime Minister Alain Juppe warned that France monetary policies. In 1969, its customs union having been would "react against" EU member states that manipulated completed ahead of schedule, the Community reaffirmed their currencies. All this led Mario Monti, the EU's Com- its desire to move to monetary union in order to abolish missioner for the Internal Market, to warn that it was exchange-rate fluctuations and uncoordinated monetary "impossible to have a guaranteed single market in a situa- policies once and for all. Although this initiative led to the tion where currency fluctuations remained unchecked," formation of the Werner Committee, which reported favor- and that "the continuing devaluation of the lira would in ably on the prospects for monetary union and emphasized the long run lead to prolonged disruption" and pointed to the need for the close coordination of national fiscal poli- the need for "some sort of monetary arrangement to com- cies, it sooin became apparent that the reach of EC officials plement the single market.'27 A similar conclusion was exceeded their grasp. Still, in response to the collapse of the drawn by the editors of The Economist, who wrote, "as long Bretton Woods System they established the 'snake" to as Europe's currencies are free to move against one another, limit the flactuation of their bilateral exchange rates to 41/2 the single market will never be secure. The risk will percent. The Snake was succeeded in 1979 by the European remain that national governments will seek to protect their Monetary System, under which participating countries' countries' firms against rivals in countries that have just access to credit lines was expanded to facilitate their efforts devalued. The greater the volatility, the greater the pres- to hold their currencies within even narrower, 21/4 percent sure for national protection and the greater the danger to bands. Most recently, the members of the European Union, all the past achievements of the common market. "28 by adopting the Maastricht Treaty, committed to creating The susceptibility of Europe's customs union to a full-fledged monetary union, supplemented by various exchange-rate-induced protectionist pressures is only part measures for the supra-national surveillance of national fis- of the story, of course, for the EU is at the same time less cal policies. than and more than a customs union. It is less than a cus- toms union in the sense that the European Economic Com- Understanding the History munity's first concrete achievement was the Common Why this insistence in Europe on marrying a customs Agricultural Policy (CAP), itself a departure from free union with exchange-rate stability and macroeconomic trade. Deregulating the market in agricultural products policy coordination? Part of the answer lies in the per- was too radical a step for the member states, given the sym- ceived threat to free intra-European trade due to the pro- bolic value of farming and the political leverage of the tectionist response of sectoral interests to currency fluctua- farmers. The CAP reconciled the customs union with agri- 239 TRADE TOW'ARDS OPEN REGIONALISM cultural price supports by establishing minimum domes- capital flows and rendered it more difficult to hold Euro- tic-currency prices for a range of farm products. The prob- pean currencies within the narrow bands of the European lem was that a change in intra-European exchange rates Monetary System. Given the threat to the Single Market could unleash a flood of agricultural exports unless support from currency fluctuations, this pointed to the need for a prices were adjusted whenever exchange rates changed, single currency to complement the Single Market. This something that farmers who valued price stability and pre- vision found expression in the Delors Report of 1989 and dictability were unwilling to accept. Until the CAP was the Maastricht Treaty adopted by the European Council in reformed in the 1990s in the direction of lump-sum pay- December 1991. ments, this departure from free trade did as much to sup- The other effect of the Single European Act, besides port the case for intra-European exchange-rate stability as increasing the difficulty of holding national currencies free-trade initiatives themselves. within narrow bands, was to magnify the political backlash If the EU is less than a customs union, it is also more against misalignments. The Single Act mandated changes than one in the sense that trade integration has always been in non-tariff barriers, public procurement policies, foreign part of a broader political agenda. The European Commu- investment regulation and policies toward distressed nities were created as a postwar strategy to lock Germany industries, all of which worked to intensify international into Europe. Not only was trade seen as a means of culti- competition. By eliminating barriers to intra-EU competi- vating understanding and interdependence between Ger- tion, these initiatives magnified the shifts in trade, com- many and its neighbors, but commercial integration was petitiveness and profitability caused by currency swings. supposed to be a first step toward political integration.29 This in turn increased the threat of a protectionist back- This commitment to pursue political as well as economic lash. In a sense, then, the Single Act led inevitably to mon- integration predisposes European policymakers toward ini- etary unification by increasing the difficulty of holding tiatives designed to broaden cooperation from trade to steady the exchange rates between Europe's separate exchange rates and macroeconomic policies. It is impossi- national currencies and by magnifying the threat to ble to understand why some EU member states, Germany Europe's customs union from the consequent currency in particular, are prepared to contemplate economic and swings. monetary union except as part of a broader bargain in For readers concerned with Mercosur, it is worth con- which they sacrifice their monetary autonomy in return for trasting Europe's Single Market with the North American an expanded foreign-policy role in the context of a more Free-Trade Area. There, too, the liberalization of interna- politically integrated European Union. The Maastricht tional capital flows has increased the difficulty of holding Treaty is not just a monetary agreement, after all; it has exchange rate within narrow bands-witness Canada's three "pillars" concerned with (1) monetary union, (2) a long-standing policy of floating and Mexico's abandon- common foreign policy, and (3) justice and home affairs. ment, under duress, of its crawling band. There, too, free trade has magnified the political response to currency Understanding the Maastricht Treaty swings, a point evident in the protests heard in Texas and The debate over the Maastricht Treaty underscores this New Mexico when the peso depreciated in late 1994 and essential point: In Europe, the Single Market and monetary early 1995. But integration between the three NAFTA union are linked precisely because the EU is more than a members is not yet so deep that these currency swings pro- customs union. Indeed, the Single Market itself is more voke protectionist pressures sufficiently intense to disrupt than a customs union. The Single European Act of 1986 the operation of the free-trade area.3' While NAFTA con- sought to bring down unemployment and banish tains a variety of side agreements, it does not approach the "Eurosclerois" by simplifying regulatory structures, inten- deep integration of Europe's Single Market. sifying competition, and helping European producers to exploit economies of scale and scope. It mandated the elim- Implications for Mercosur ination of all barriers to the free mobility of labor and cap- Four implications for Mercosur follow from our discussion ital. This entailed removing the remaining capital con- of the connections between the political economy of FTAs trols. But the elimination of controls liberated speculative and the stance of macroeconomic policy. 240 MACROECONOMIC POLICY COORDINATION 1) In the short run, the need for macroeconomic another, and the potential for a protectionist backlash due policy coordination is limited. So long as Mercosur to currency fluctuations will mount. This wilL encourage remains just an FTA and is not accompanied by measures efforts to coordinate exchange rate and macroeconomic poli- abolishing restrictions on factor flows, subsidies for domes- cies among the member states of the FTA. tic industry and preferential public procurement like those 4) The options for more closely harmonizing the featured in the Single European Act, the impact of cur- macroeconomic policies of the Mercosur countries are rency swings on the profitability of competing national relatively limited. Feasible policy coordination is likely to industries will not be as pronounced as in Europe's Single focus on harmonizing policy rules and guidelines. In prin- Market. Hence, exchange-rate-induced lobbying for subsi- ciple, feasible strategies (moving from the most ambitious dies and protection will not be as intense.31 There will be to the least) include a common currency, a common cur- less need for the close coordination of monetary and fiscal rency board peg, explicit coordination exercises and tacit policies than in the European Union. NAFTA may be a coordination achieved by harmonizing the rules that guide better analogy for the problem that will face Mercosur. the formulation of policy. In practice, only the last strategy While the post-1994 depreciation of the Mexican peso pro- is likely to be feasible. I consider the alternatives in turn. voked complaints of unfair currency practices from U.S. a) Monetary Unification. A single currency for the Merco- producers in border states who found their profits squeezed sur countries might seem to be the ultimate implication of by the decline in Mexico's dollar-denominated labor costs, this argument. This is unrealistic, however, over any hori- that pressure was not so intense as to actually lead to the zon relevant for policy planning. A national currency is an adoption of protectionist measures. essential concomitant of sovereignty, and nation states are 2) In the short run, measures to regulate interna- loath to give it up. Besides its symbolic value, the currency tional capital flows in and out of the Mercosur region is the revenue source of last resort in time of war. Histori- will enhance governments' ability to stabilize ex- cally, therefore, political unification has always preceded change rates. So long as Mercosur countries can continue monetary unification. If Europe succeeds in reversing this to take measures to regulate international capital flows order, its achievement will be unprecedented. And that (taxing capital inflows, taxing funds that exit soon after achievement will have been possible only because the their arrival, raising reserve requirements on banks that members of what is now the European Union have capital- borrow offshore), they will retain some capacity to limit ized on the powerful impetus for political integration lent the impact on their exchange rates of international capital by three wars in three generations. They have already spent mobility. They will have more capacity to hold their cur- nearly half a century building transnational institutions of rencies stab-le than European governments (which are pro- collective governance. Those institutions possess the legit- hibited from resorting to such devices by the Single Euro- imacy necessary to administer a common monetary policy pean Act). Less monetary and fiscal coordination will be and the capacity to extend the side payments needed to needed to achieve a given degree of exchange-rate stability make it politically acceptable. than if capital markets were totally integrated and open.32 Nothing similar is likely to occur for the foreseeable 3) In the longer run, the pressure for policy coordi- future in South America (or, for that matter, North Amer- nation will grow. There are good reasons to anticipate that ica or East Asia). Brazil in particular continues to insist the integration of Southern Cone financial markets with that Mercosur should remain a union of nation states and those of the rest of the world will only intensify over time, involve a minimum of transnational institution building. with, among other things, the penetration of international Its permanent secretariat is tiny, and decisionmaking banks into the region. If the experience of the European power is vested in an inter-governmental Common Market Union is any guide, intra-Mercosur trade will grow dispro- Council comprising the foreign and finance ministers of portionately. But as Latin American financial markets the member countries. The willingness to hitch unification become still more open and competition between the man- to economic integration evident in Europe requires a polit- ufacturing and service sectors of the Mercosur economies ical commitment that is not present in the Southern Cone. grows more intense, it will become more difficult to stabi- b) A Common Currency Board Peg. Convertibility laws lize the currencies of the participating countries against one like Argentina's, if adopted simultaneously in all the Mer- 241 TRADE TOWARDS OPEN REGIONALISM cosur countries, are a way of reconciling macroeconomic d) Common targets for policy. The least ambitious and policy with the need to maintain political support for the most realistic option for harmonizing policies is to harmo- FTA. Such laws create political barriers to changing the nize the rules guiding their formulation. This could exchange rate, and if adopted simultaneously in all the involve the adoption of common inflation targets for mon- major Mercosur countries would prevent currency fluctua- etary policy. This approach is akin to that recommended by tions from disrupting political support for the FTA. If each the Centre for Economic Policy Research (1996) for the of the Mercosur countries instituted an Argentine-style monetary relations between EU countries inside and out- convertibility law, tying its currency to the dollar, for side the monetary union. If the European Monetary Union example, there would be no scope for their currencies to starts without the participation of all 15 EU member fluctuate against one another and no threat to the FTA states, there will be the problem of preventing currcncy from currency volatility. instability between the "insiders" and "outsiders" from Currency boards and convertibility laws have costs as eroding support for the Single Market. Coordinated infla- well as benefits, of course. They limit policy flexibility and tion-targeting by the respective central banks has been the capacity of the authorities to act as lenders of last proposed as a way of preventing the kind of dramatic mis- resort. These drawbacks may be too high a price to pay for alignments most likely to arouse political opposition to securing an FTA. Even countries like Argentina will ulti- free trade. Experience suggests that considerable short-run mately move away from their currency boards as they seek exchange-rate variations are still possible between the cur- to regain some policy autonomy. Thus, currency boards are rencies of central banks following broadly similar policies unlikely to be a permanent solution.'3 toward inflation. But the hope is that the currencies of c) Ongoing coordinzation exercises. A less ambitious countries adopting common inflation targets would not option would be ongoing coordination exercises akin to the become seriously and persistently misaligned. The knowl- annual Group of Seven summits of finance ministers and edge that their price levels will move together over the central bankers and exceptional agreements like the Plaza long run implies that the exchange rate between their cur- and Louvre Accords. The Plaza, it is worth recalling, was a rencies should display long-term stability. And if the pol- response to the soaring external value of the dollar in the icy rule is transparent, stabilizing expectations should pre- first half of the 1980s and to the perception that the diffi- vent serious short-run deviations from that long-run culties it created for the U.S. Rust Belt would provoke a equilibrium. For inflation targeting to have this beneficial protectionist reaction in the Congress. It entailed sterilized effect, however, it must be credible. This requires, among intervention to bring the currency down. The Louvre other things, that other policies, especially fiscal policies, Agreement two years later involved more far-reaching do not seriously distort relative prices and create doubts adjustments in monetary and fiscal policies, again to pre- about the stability of the monetary regime. It requires vent exchange-rate fluctuations from eroding support for coordination among domestic policymakers, in other the global trading system. Subsequent G-7 summits have words. But it does not also require continuous cooperation been designed to construct the policy consensus necessary with their international counterparts, the obstacles to for more regular coordination of national macroeconomic which are likely to be even more formidable. policies, again with the stability of exchange rates and the trading system in mind. Conclusions: Free Trade and Macroeconomic The problem is that these exercises have hardly been a Policy rousing success. The significance of both the Plaza and What does this analysis suggest about whether countries Louvre Accords has been questioned (the Plaza because the should contemplate an FTA with partners who have not yet dollar had begun declining before the agreement, the Lou- adopted sound and stable macroeconomic policies? One vre because few significant adjustments in macroeconomic answer is that they should pay macroeconomic policies policies actually followed). G-7 summits have produced abroad no mind and negotiate an FTA irrespective. This few concrete results, especially in recent years (Bergsten follows from the proposition that divergences between and Henning 1996). It is hard to imagine that a South undistorted domestic prices on the one hand and import American counterpart to the G-7 could do better.34 prices on the other offer the home country gains from trade 242 MACROECONOMIC POLICY COORDINATION whatever the source of the divergence.35 If import prices Bayoumi, Tamim and Paul R. Masson (1995). "Fiscal Flows in the decline because policy in the partner country causes its United States and Canada: Lessons for Monetary Union in exchange rate to depreciate, that is a source of gains from Europe." Eeropean Economic Review 39, pp. 253-74. trade like aniy other. For the same reason that economists Bergsten, C. Fred and C. Randall Henning (1996). Global Economic Leadership and the Groep of Seven, Washinigtoni, D.C.: Institute for generally oppose the use of anti-dumping duties, arguing International Economics. that, if a foreign country wants to sell its exports at low Bini-Smaghi, Lorenzo (1991). "Exchange Rate Variability and Trade: prices, that is a source of gains from trade for the importer, Why Is It So Difficult to Find Any Empirical Relationship?" they would reject arguments that "exchange dumping" (a Applied Economics 23, pp. 927-36. decline in irnport prices due to a policy-induced swing in Centre for Economic Policy Research (1996). Flexible Integration, the exchange rate) is anvthing other than a welfare benefit. London: CEPR. Chowdhury, Abdur R. (1993). "Does Exchange Rate Volatility Another, very different, conclusion is that governments Depress Trade Flows: Evidence from Error Correction Models." should hesitate to negotiate an FTA until the potential par- Review of Economic and Statistics LXXV, pp. 700-06. ticipants have all installed sound and stable macroeco- Clark, Peter B. (1973). "Uncertainty, Exchange Risk, and the Level nomic policies. Otherwise policy-induced real-exchange- of International Trade." Western Economic Jornal 11, pp. 302-13. rate-instabiLity will fan protectionist pressures and Cushman, David 0. (1983). "The Effects of Real Exchange Rate Risk undermine political support for the agreement. Thus, on International Trade."Jozernal of International Economics 15, pp. while a strictly delimited economic analysis points to one 45-63. oDeGrauwe, P, M. Janssens and M. Lelianert (1985). "Real Exchange answer, a political economy analysis points to another. Rate Variability from 1920 to 1926 and 1973 to 1982," Prince- Which conclusion one then prefers will depend on the ton Studies in International Finance No. 56. International Finance force one attaches to the endogenous protection channel. Section, Department of Economics, Princeton University. A final argument derives from the observation that open- DeGrauwe, P. and M. Rosiers (1984). "Real Exchange Rare Variabil- ness heightens the premium on good macroeconomic policy iry and Monetary Disturbances," unpublished manuscript. and reinforces the penalty associated with mistakes (Interna- Catholic University of Leuven. Dornbusch, Rudiger (1976). "Expectations and Exchange Rate tional Monetary Fund 1997). An FTA, like other forms of Dyam ."ona of Poiia Ecnm 4p.16-6 ' ~~~~~~~Dynamics."Journal of Political Economy 84, pp. 1161-76. opening, should therefore encourage stabilization and policy Dyer, Geoff (1997). "Fears Over Brazil's Deficit Played Down." reform. In this view, freeing trade should not wait for macro- Financial Times (April 4), p. 7. economic stabilization: instead it should encourage it. Edwards, Sebastian (1987). "Real Exchange Rate Variability: An Empirical Analysis of the Developing Country Case." Interna- References tional EconomicJoarnal 1, pp. 91-106. Arize, A. C. (1996). "Real Exchange-Rate Variability and Trade Eichengreen, Barry and Charles Wyplosz (1996). "Taxing Capital Flows: The Experience of Eight European Economies." Interna- Flows to Improve the Operation of the International Monetary tional Review of Economics and Finance 5, pp. 187-205. System," in Mabhu al Haq, Inge Kaul and Isabel Grunberg (eds.), Barham, John and Angus Foster (1994). "Wrapped up in a Mutual The Tobin Tax. New York: Oxford University Press. Embrace: A Look at the Progress towards Economic Integration Ethier, Wilfred (1973). "International Trade and the Forward in Latin America." Financial Times (17 January). Exchange Market." American Economic Review 63, pp. 494-503. Batra, R. N. (1975). "Production Uncertainty and the Heckscher- Frankel, Jeffrey A. and Shangjin Wei (1993). "Trade Blocs and Cur- Ohlin Theorem." Review of Economic Stoedies 42, pp. 259-68. rency Blocs," paper presented to the CEPR Conference on the Bayoumi, Tamim and Barry Eichengreen (1995). "Restraining Your- Monetary Future of Europe, La Coruna, Spain (December 11-12). self: The Impact of Fiscal Rules on Economic Stabilization." Staff Gagnon, Joseph E. (1989). "Exchange Rate Variability and the Level Papers 42, pp. 32-48. of International Trade," International Finance Discussion Paper _ (1996). "Optimum Currency Areas and Exchange Rate Volail- No. 369 (December), International Finance Division, Board of ity: Theory and Evidence Compared," unpublished manuscript. Governors, Federal Reserve System, Washington, D.C. International Monetary Fund and University of California, Goldstein, Morris and Geoffrey Woglom (1991). "Market-Based Fis- Berkeley. cal Discipline in Monetary Unions: Evidence from the U.S. _ (1997). "Exchange Rate Volatility and Intervention: Lessons Municipal Bond Market," IMF Working Paper WP/91/89. from the T'heory of Optimal Currency Areas," unpublished man- Washington, D.C.: International Monetary Fund. uscript. International Monetary Fund and University of Califor- Gotur, Padma (1985). "Effects of Exchange Rate Volatility on Trade: nia, Berkeley. Some Further Evidence." Staff Papers 32, pp. 475-512. 243 TRADE: TOWARDS OPEN REGIONALISM HelpmaDi, Elhanan and Assaf Razin (1978). A Theory of International Turnovsky, S. J. (1974). "Technological and Price Uncertainty in a Trade Under Uncertainty. New York: Academic Press. Ricardian Model of International Trade." Review of Economic Sted- Holly, Sean (1995). "Exchange Rate Uncertainty and Export Perfor- ies 41, pp. 201-17. mance: Supply and Demand Effects." Scottish Jouernal of Political World Bank (1997). Private Capital Flows to Developing Cocintries. Ecunomiiy 42, pp. 381-90. Washington, D.C.: World Bank. Hooper, Peter and Steven Kohlhagen (1978). "The Effect of Exchange Rate Uncertainty on the Price and Volume of Interna- tional Trade."Joarnal of International Economics 8, pp. 483-511. Notes International Monetary Fund (1984). "Exchange Rate Volatility and 1. While evidence from NAFTA would be revealing, this World Trade," Occasional Paper No. 28. Washington, D.C.: Inter- arrangement is too recent to provide much useful information yet; I national Monetary Fund. therefore concentrate on the EU case. However, the political and eco- _ (1997). World Econiomic Outlook (May). Washington, D.C.: IME nomic repercussions of the late-1994 devaluation of the peso are sug- Kemp, M. C. and N. Liviatan (1973). "Production and Trade Pat- gestive, and I briefly discuss them below. See also Ruffin (1974) and terns under Uncertainty." Economic Record 49, pp. 215-27. Turnovsky (1974). Kroner, Kenneth F and William D. Lastrapes (1993). "The Impact 2. See also Ruffin (1974) and Turnovsky (1974). of Exchange Rate Volatility on International Trade: Reduced 3. This is true because the relative price of the two goods will not Form Estimates using the GARCH-in-Mean Model." Jornal of in geteral equal the slope of the producrion possibilities frontier. The International Mloney and Finance 12, pp. 298-318. essential conclusions of this analysis carry over to the Heckscher- Kumar, Vikram and Joseph A. Whitt, Jr. (1992). "Exchange Rate Ohlin model, albeit with additional complications. See Batra (1975). Variability and International Trade." Federal Reserve Bank of 4. This assumption is not unproblematic, as Helpman and Razin Atlantic Economic Reviem (May/June), pp. 17-32. (1978) show, but it has the virtue of simplicity. I return to it below. Latin America Regional Reports (1995). "Getting There," MAexico 5. While I relax this assumption momentarily, it is a more plau- anid NAFTA Report. London: Latin American Newsletters, Ltd. sible starting point than the opposite assumption of risk-averse pro- (January 19). ducers and risk-neutral consumers. Firms are better placed to hedge Machlup, Fritz (1943). International Trade and the National Income price risk through recourse to future markets; they have the special- AMlelteplier. Philadelphia: Blakiston. ized knowledge needed to transact in those markets and the scale to Meade, James E. (1951). The Theory of International Economic Policy, Vol- defray transaction-related fixed costs. They can merge with firms in noine ): The Balance of Payments. Oxford: Oxford University Press. the sector whose relative price fluctuates inversely with their own or Mundell, Robert (1961). "A Theory of Optimum Currenicy Areas." establish a subsidiary there. Households are typically less well-placed American Economic Revieuw 51, pp. 657-64. to engage in these types of hedging behavior. For all these reasons it ___(1963). 'Capital Mobility and Stabilization Policy Under Fixed makes more sense to treat producers as risk-neutral. and Flexible Exchange Rates." Canadian joeernal of Economics and mksmr es otetpouesa iknurl Poial ScexiblenExc ange 29, es." pp.adian 475-85. of Eonoi-sand 6. Clark (1973) focuses on the relation between exchange-rate Obstfeld, Maurice and Alan M. Taylor (1996). "Nonlinear Aspects of uncertainty and the volume of trade in a simple static model. deriv- Obsteld,Maurce ad Aan M Tayor Q996) "Nnlinar Apect of ing a negative effective of the first variable on the secondi Goods-Market Arbitrage and Adjustment: Heckscher's Com- in ahege effctivof t ar iable on the sond. modity Points Revisited," unpublished manuscript. University of 7. These conclusions are strong because the model is simple. If CalifyPornia, Berkel tey,"anduNorthwesn Uaniersipt. hniversity of firms in our Ricardian world can reallocate resources after relative California, Berkeley, and Northwestern University. Rodrik, Dani (1996). "Why Do More Open Economies Have Bigger prices are revealed, they will be less inclined to insure against price Governments?" NBER Working Paper No. 5537 (April). uncertainty by producing a diversified portfolio of products. If they Rogers, John H. and Ping Wang (1995). "Real Exchange Rate iMove- can insure against risk to their profits through recourse to financial meors in High Inflation Countries," International Finance Dis- markets, their production decisions will be no different from the cer- cussion Paper no. 501 (February). International Finance Division, tainty case. If households can similarly insure against risk to their Board of Governors, Federal Reserve System, Washington, D.C. purchasing power (by, for example, purchasing financial assets whose Romer, David (1993). "Openness and Inflation: Theory and Evi- return is high when terms of trade are poor), their behavior will not dence." Qearterly Journal of Econoinics CVIII, pp. 869-904. differ from the certainty case. Helpman and Razin (1978) provide an Ruffiii, Roy J. (1974). "Coimiparative Advasitage under Uiicertainty." extensive analysis showing that uncertainty makes no difference for Jouirnal of Inteinational Economics 4, pp. 261-74. international trade if producers and consumers can insure against it Sapir, Andre, Khalid Sekkat and Axel A. Weber (1994). "The Impact on financial markets. of Exchange Rate Fluctuations on European Union Trade," CEPR 8. According to their estimates, eliminating exchange-rate Discussion Paper No. 1041 (November). volatility would have increased intra-EC trade, for example, by 0.77 Trefler, Daniel (1993). "Trade Liberalization and the Theory of percent in 1980. Doubling that actual level of volatility observed in Endogenous Protection: An Econometric Study of U.S. Import 1992 inside the ERM would have reduced the trade of the member Policy,"Journal of Political Economy (U.S.), 101: 138-60, February. countries by 0.25 percent. 244 MACROECONOMIC POLICY COORDINATION 9. This is also the conclusion of other recent authors; see for ity (not analyzed here). The negative coefficient results from a few example Kumar and Whitt (1992). Gagnon (1989) provides theo- outliers-countries with highly variable rates of money grDwth but retical support for this conclusion; he simulates a theoretical model only moderately variable real exchange rates. Omitting the three in which exchange-rate variability has a negative effect on the level countries with the most variable money-growth rate causes this vari- of trade, calibrating it to observed trade flows and real exchange able to switch signs, from negative to positive, though it continues rates, which clemonstrates that the effect of increasing exchange-rate to differ insignificantly from zero at standard confidence levels. variability on trade flows is very small. 16. These results were sensitive to specification, so too much 10. This finding, which comes through robustly for both the should probably not be made of them. Dropping out other regressors 1970s and 1980s, does not carry over to real exchange rates, however. sometimes reduced the coefficients on these two interaction terms Sapir, Sekkat and Weber (1994) consider also the impact of the level below the critical confidence levels. The only other variable that of the exchange rate, as opposed to its variability, interpreting long- sometimes entered with a statistically significant interaction term in standing movements in the level as misalignments. They find that this sensitivity analysis was the average inflation rate. Its coefficient changes in the level significantly affect the volume of exports from was negative, suggesting that higher average inflation had a smaller the large EU countries to the United States and from the United impact on real exchange-rate variability in more open economies. Kingdom to its EU partners; in contrast, they find no impact on the 17. His analysis uses data for 1983. He concentrates on non-tar- volume of trade between ERM member countries. This suggests that iff barriers on the grounds that successive GATT rounds had already the relationship between the exchange rate and trade is non-linear: reduced the level of tariff barriers to close to zero. Large exchange-rate changes, like those between the United States, 18. Goldstein and Woglom (1991) show that this behavior is the United Kingdom and the ERM countries, have a significant plausibly associated with the tendency for the costs of financing impact on trade, while relatively small changes, like those that are deficits to rise with their level. typically characteristic of the rates linking the currencies participat- 19. Unfortunately, it does not appear to be possible to estimate ing in the ERM, have no discernible effect. The authors interpret this the analogous model for emerging markets, given the difficulty of as evidence that the ERM has helped Europe avoid large, persistent assembling standardized data on the composition of government rev- exchange-rate movements capable of significantly dislocating intra- enues and expenditure. EU trade and provoking serious protectionist pressures. 20. This hypothesis is true holding constant other determinants I1. More recently, Rogers and Wang (1995) have reported results of the inflation rate. consistent with Edwards' findings; even for a sample of high-infla- 21. Barham and Foster (1994). tion countries, they find that real exchange-rate movements depend 22. Latin America Regional Reports (1995). on both monetary and, somewhat more surprisingly, real (fiscal, as 23. For a sense of the magnitudes, note that Brazil imported some well as other) shocks. well . other) shock200,000 automobiles in the first quarter of 1995. 12. These data were available for 30 countries: Bolivia, Burundi, 24. There are other significant departures from free trade as well, Cameroon, Central African Republic, Chile, Colombia, Costa Rica, a COr dvoie,Cyrus Dmiica Rpubic Euadr,Fii,. rece although bilateral trade will be duty-free from the beginning of Cote d'Ivoire, Cyprus, Dominican Republic, Ecuador, Fiji, Greece, 00 o eal e a ar' hpe nti oue Guyana, Mahlwi, Malaysia, Morocco, Nicaragua, Nigeria, Paraguay, 20. F ee S der ithsvlm Philippines, Sierra Leone, South Africa, Togo, Trinidad and Tobago, 25. See Dyer (1997). Uganda, Uruguay Venezuela, Zare and Zambia.26. Another fact consistent with this conclusion is that, as part of Uganda, Uruguay, Venezuela, Zaire and Zambia. 13. Following Edwards, wholesale price indexes were used for the the five-year program agreed to in December 1995 to perfect the partner countries, while the consumer price index was used to proxy free-trade area and customs union, governments committed to mov- for the domestic price level. ing toward harmonizing their economic policies. 14. It turns out that the results are insensitive to the exclusion of 27. The first passage is a direct quote from Monri, while the sec- one of the two nominal exchange-rate variability measures. When ond and third are the Financial Times' effort to paraphrase him. these are included as alternatives rather than simultaneously, only Financial Times, February 28, 1995, p. 6. the coefficien: on the nominal effective rate differs significantly from 28. The Ecosomist, February 11, 1995, p. 14. zero at standard confidence levels. None of the other coefficients is 29. lt is revealing in this context that the European Communities noticeably af.'ected except for that openness, which is significantly were designed to encompass not just the economic community but greater than zero at the 90 percent level in the equation using the also the Defense Community and the Atomic Energy Community. effective rate, suggesting that more open economies have more vari- 30. Recent months have seen a number of calls for Mexico to able real exchange rates. establish a currency board system under which the peso would be 15. Readers may be surprised by the negative, insignificant coef- irrevocably tied to the dollar-but not from fear that continued cur- ficient on the money growth variable. However, this is simply con- rency swings would undermine political support for the free-trade firmation of Edwards' results; he too finds that the money-growth agreement, but rather in response to doubts that Mexico can effi- variable is insignificant in the long-run real exchange-rate equations, ciently run its own financial affairs on a discretionary basis. although it tends to matter for short-run real exchange-rate variabil- 31. All things being equal. 245 TRADE: TOWARDS OPEN REGIONALISM 32. This is not to argue that such controls are perfectly effective devaluing, Argentine producers whose profits are squeezed and or even desirable from an efficiency-maximizing standpoint, only Argentioe workers who suffer unemployment will have an incen- that they enhance governments' ability to peg their exchange rates tive to lobby for hidden and overt import restraints, undermining and loosen the link between the exchange rate and monetary policy. the operation of the FTA. Rigidly pegging the exchange rate under See Eichengreen and Wyplosz (1996). On the use of capital controls a currency-board regime may, therefore, prevent the market from in developing countries, see World Bank (1997). venting competitive pressures and only aggravate the protectionist 33. In addition, there may be circumstances when an exchange- threat. rate adjustment is needed to eliminate a competitive imbalance 34. In the absence of an effective enforcement technology, coop- arising for other reasons-when it is the solution, not the problem. eration in the international economic policy domain rests on reputa- Consider a hypothetical case where Argentina's labor productivity tion-on the repeated interaction of the relevant institutional repre- declines exogenously but domestic prices and costs are slow to sentatives. The G-7 clearly has a leg up on this process and yet has adjust. If Argentina devalues in response, Brazilian producers will still found it hard to deliver results. have little reason to complain, since the depreciation of the peso 35. The critical qualifier is of course that domestic prices are would simply offset the deterioration in Argentina labor produc- undistorted; otherwise, we are in a second-best world where trade tivity. But if the convertibility law prevents Argentina from need not be welfare-enhancing for the home country. 246 Comment C LAUD IO LO SE R WOULD LIKE TO THANK PROFESSOR EICHENGREEN FOR HIS PAPER ON "FREE TRADE AND MACRO- economic Policy," which I have read with great interest. The world has become increasingly inter- |dependent with a rapid expansion of trade that has occurred hand in hand with that of capital flows. "Globalization" has become a common word describing economic enterprise. Or to put it differ- ently, not only are the issues addressed in this paper important at the moment, I dare say that they will stay with us for some time to come. In structuring my comments, I will first touch upon a that puts it on a collision course with sustainable growth. number of specific points of the paper, to help set the tone Indeed, in the daily practice of the Fund and in our discus- for our discussion, and then I will comment briefly on the sions with authorities, one cannot really think of sound role of the International Monetary Fund in the paper's con- trade policies and the opening up of economies to world text of macroeconomic policy and policy coordination, and trade and specialization in the absence of proper macroeco- its relations,hip to trade. nomic conditions. I believe that the assertion by the author The specific comments: refers more to the necessary assumption that has to be made Professor Eichengreen opens the paper with the assertion sometimes in the theoretical literature to keep models that most economists might be tempted to answer "none" tractable, rather than to the experience of any minister of if asked what are the connections between trade theory and finance who needs to balance trade and financial policies on policy on the one hand, and open economy macroeconomics a daily basis, and where a rapid weakening of the trade on the other. This happens notwithstanding the presence of account requires a swift macroeconomic response to keep the outspoken student in the back of the class sometimes the external sector in a sustainable position. reminding the instructor that there should be such a con- Having said this, in turn the balance between domestic nection. In lact, and without trying to be condescending, I savings and investment that is consistent with a sustain- wondered whether that outspoken student may have able path of growth, and macro policies could be compati- worked in the Fund for a while. One of the workhorses in ble with different trade regimes. However, in today's glob- the Fund's toolbox is the Stand-By Program, and another is alized world, it will be almost a necessary condition that its close (but longer term) relative, the Extended Arrange- the trade regime be compatible with a further integration ment. These programs are typically called upon initially in the world economy to help generate sustained growth when macroeconomic policies have become unbalanced, for the medium-term. leading to a severe external disequilibrium-i.e., a large A second technical detail in the paper that I found shift in the trade and current account position of a country interesting was the prominence afforded by the model to Claudio Loser is Director of the Western Hemisphere Department of the International Monetary Fund. 247 TRADE TOWARDS OPEN REGIONALISM the idea that consumers, more than producers, might form become ineffective and that they are absolutely ineffective pressure groups to ask for some domestic production of for capital outflows. In fact, these have harmful effects on importables, because of consumer price uncertainty over confidence and may precipitate the crisis that the authori- time. However, in today's world typically one thinks of ties attempt to avoid. consumers as comparatively small participants in the polit- Now let me switch to the more general policy issues ical economy of protectionism (I would say that they are that Professor Eichengreen addresses in the second part of the great absentees in the debates), and they might find it his paper, namely the applicability of macro coordination difficult to get organized to exercise such pressure. I won- to Mercosur. He uses as a parallel example the European der what the empirical evidence can tell us about the con- Community, exploring the influence that macroeconomic sumer-based protection model. Also, I wonder about the policies may have on exchange and trade relations and the assumption in the consumer-based protection model implications from that for the degree of policy coordina- regarding the ability of a country to affect the relative price tion that is necessary in regional trading blocks. of tradables. If countries are by and large price takers, how The issues discussed here are the regional version of the does the presence or absence of domestic production problems that the Bretton Woods negotiators struggled increase or reduce consumer price uncertainty over time? with when the IMF and World Bank were created. There I share strongly the idea that access to well-functioning was broad recognition that a minimum degree of policy financial (e.g., futures) markets helps reduce price uncer- coordination and dialogue was indispensable for bringing tainty over time and thus favors an increase in trade. As about acceptable trade relations, and to move away from you know, the Fund advocates the removal of capital con- isolationist competitive devaluations and protectionist straints, and we are now involved in a major exercise to trade and capital mobility policies that had proved so dam- bring capital convertibility into the jurisdiction of Fund aging to world economic welfare. Not surprisingly then, surveillance. We do this because we believe that, together the first article of the IMF's Articles of Agreement states with stable and transparent macroeconomic policies, such that one of the purposes of the Fund is "to facilitate the liberalization increases access to fmnds for investment in a expansion and balanced growth of international trade." In country's area of comparative advantage. Liberal capital practice we try to meet this challenge through our annual flows are also likely to contribute to the better functioning consultations with member countries, which are discussed of domestic capital markets by providing access to a wider in the executive board of the Fund in the process of multi- array of capital market instruments. The benefits of open lateral surveillance. Through this surveillance exercise, the capital markets thus not only favor trade deepening Fund offers a global forum for policy coordination and dia- directly through added investment in output capacities logue. This of course, is reinforced by the work of the third where a country might have a comparative advantage, but large international agency overlooking international coop- also indirectly through improved access to instruments eration, the World Trade Organization (WTO), which that assist in reducing uncertainty over time and helps in deals with the agreements to attain a more integrated the process of efficient external trade. In this context, I world of trade of goods and services. must say that I felt uneasy by the assertion made near the Now, groups of countries may have the opportunity to end of the paper that "remaining limits to capital mobility go beyond this stage of policy coordination by forming an will enhance the government's ability to stabilize exchange explicit regional trading area. The Fund has been cau- rates." Indeed, more often than not, our experience is that tiously supportive of such efforts provided that: (1) the the use of capital controls seeks to help stabilize the regional trading area does not become a regional fortress in exchange rate in the response to unstable and unbalanced which the common external tariff is in excess of the tariffs domestic macroeconomic policies. That is, an endogenous that were established by the individual member countries cause of exchange-rate instability, and not some exogenous prior to the formation of the trading bloc (according to the factor or set of factors that would warrant, at best tem- literature on trade diversion and trade creation); and (2) the porarily, the use of capital constraints. While some coun- policy of opening up trade and payments systems continue tries have made use of these measures, it is my strong belief not only between member states of the trading area but that beyond the short term, capital controls for inflows also for the trading area vis-a-vis the rest of the world. To 248 MACROECONOMIC POLICY COORDINATION do this with any measure of success, I cannot agree more trade and payments at very low rates of domestic inflation, with Professor Eichengreen's analysis that over time, the stabilizing expectations that result from this policy stance need for policy coordination will likely grow. That does not should prevent serious short-run deviations in the necessarily mnean that there is a need to form a single cur- exchange-rate equilibrium from that of the long run. This rency block, While such a notion has clear advantages, it is even more so the case in the demanding environment of requires a degree of coordination that in the short run may today's globalized world. In other words, Professor Eichen- entail high costs. The experience of NAFTA shows that green indeed hits the nail right on the head when he con- integration can take place if there is a good set of pre- cludes the paper by saying that there is no substitute for dictable ancl appropriate macroeconomic policies, even if stable and appropriately cautious macroeconomic policies the exchange rate and the pace of inflation differ among whatever the exchange rate and monetary regime. In this countries. The success with flexible exchange-rate policies regard, as has always been the case with the outspoken stu- of the Unired States with its major trading partners, dent at the back-or front-of the class, his impact Canada, Mexico and Japan, testifies to that effect. remains significant and helps us see the truth of the state- In this regard-and again, to paraphrase Professor ment that good economic policies reinforce each other, but Eichengreen-if the policy rules adopted by individual that one set of bad policies precludes other policies from countries are transparent and lead to increased openness of succeeding. 249 0 ¢ "-4 ( *t CL, C 9~~ ~