;  11           -     25506
'         -  -     February 2003
R'EGIONAL INTEGRATION
AND DEVELOPMENT
., [               -   
lku
LMAURICE SCHIFF and L. ALAN WVINTERS
&ADDVEOMN
, U- ,.., .. .i..
..   .    .






Regional Integration
and Development
Maurice Schiff
and L. Alan Winters



© 2003 The International Bank for Reconstruction and Development / The World Bank
1818 H Street, NW
Washington, DC 20433
Telephone: 202-473-1000
Internet: www.worldbank.org
E-mail: feedback@worldbank.org
All rights reserved.
1 2 3 4 05 04 03
A co-publication of the World Bank and Oxford University Press.
The findings, interpretations, and conclusions expressed here are those of the author(s)
and do not necessarily reflect the views of the Board of Executive Directors of the World
Bank or the governments they represent.
The World Bank cannot guarantee the accuracy of the data included in this work. The
boundaries, colors, denominations, and other information shown on any map in this work
do not imply on the part of the World Bank any judgment of the legal status of any territory
or the endorsement or acceptance of such boundaries.
Rights and Permissions
The material in this work is copyrighted. No part of this work may be reproduced or trans-
mitted in any form or by any means, electronic or mechanical, including photocopying,
recording, or inclusion in any information storage and retrieval system, without the prior
written permission of the World Bank. The World Bank encourages dissemination of its
work and will normally grant permission promptly.
For permission to photocopy or reprint, please send a request with complete informa-
tion to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923,
USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com.
All other queries on rights and licenses, including subsidiary rights, should be addressed
to the Office of the Publisher, World Bank, 1818 H Street NW, Washington, DC 20433,
USA, fax 202-522-2422, e-mall pubrights@worldbank.org.
Library of Congress Cataloguing-in-Publication Data
Schiff, Maurice W.
Regional integration and development / Maurice Schiff and L. Alan Winters.
p. cm.
Includes bibliographical references and index.
ISBN 0-8213-5078-1
1. Trade blocs. 2. Regionalism. 3. Commercial policy. 4. Free trade. 5. International
economic relations. I. Winters, L. Alan. ll. Title.
F1418.7 .S35 2002
337.1-dc2l
2002074094



Contents
Preface   xi
Abbreviations and Acronyms  xv
Chapter 1. Regional Integration Agreements: An Overview     1
RIAs in History  4
Why Regionalism?   o
Why Another Book on Regionalism?   10
Summary    13
Regionalism as Trade Policy
Making the Most of Regionalism
Regionalism and Investment
Growth and Location
Integration of Domestic Policies
Regionalism as Politics
Regionalism and the Rest of the World
Rules of Thumb for Regionalism
Appendix. Selected Regional Integration Agreements with
Developing Country Members   26
Notes    29
Chapter 2. How Trade Blocs Increase Trade and Competition      31
Increased Trade between Members of Trade Blocs  32
Is More Trade Good, or Bad? Trade Creation and
Trade Diversion  33
Growth of Trade over Time  36
Not All Changes Come from Regionalism  40
A Change of Perspective: Imperfectly Competitive Markets  46
iii



iv    REGIONAL INTEGRATION AND DEVELOPMENT
Larger Markets, More Competition  50
Appendix. The Simple Analytics of Trade Creation and
Trade Diversion  54
Notes   59
Chapter 3. Making the Most of Regional Integration   63
Free Trade with Whom?  64
Why Not with Everyone?
Choosing Partners: The "Natural Trading Partners" Fallacy
Choosing Partners: Comparative Advantage
Neighborhood RIAs
North-South or South-South RIAs
How Many RIAs?    75
Free Trade Areas and Customs Unions  78
Trade Deflection and Rules of Origin: More Protection
Indirect Trade Deflection: Exporting Protection
Customs Unions Offer Lower Trading Costs and Greater
Integration
External Trade Policy  82
Setting External Tariffs in an FTA: A Race to the Bottom?
Trade Policy Institutions in Customs Unions Can Increase
Protection
Lobbies Bias RIAs toward Trade Diversion
RIAs Open a New Environment for Lobbying
RIAs and Protection: Summing Up
Integration and Taxes  94
Fiscal Compensation
Tax Competition
Notes   100
Chapter 4. Stimulating Investment   101
Investment Policies  103
Investment Planning: A Dead End
Bilateral Investment Treaties
Treatment of Investment in Current Regional Arrangements
Multilateral Investment Agreements
Integration as an Aid to Credibility-Not an Automatic
Effect  107
RIAs Allow Bad Policy to Be Punished
RIAs Can Affect the Incentives for Good Policies
RIAs Can Signal Government's Reform Intentions-
If Genuine
Regional versus Multilateral Routes to Credibility
RIAs as Investment Stimuli  113
Integration Affects Incentives to Invest
Investment Does Not Necessarily Mean Growth in RIAs



CONTENTS      V
Regional Integration and Foreign Direct Investment  117
Motives for FDI
Evidence of the Positive Effects of Integration
Notes   121
Chapter 5. Growth and Location     123
South-South and North-South Regionalism as Stimulants to
Growth    124
Knowledge and Institutions as Keys
Trade, Convergence, and Spillovers
Choosing the Wrong Partners Can Harm Growth
FDI and Knowledge Spillovers
Cross-Country Evidence on Openness and Growth
Agglomeration and Inclustrialization 137
The Balance between Centripetal and Centrifugal
Forces
Trade Liberalization Can Aid Industrialization
Intermember Distribution: Divergence Is Likely in
South-South RIAS
Notes   144
Chapter 6. Integrating Domestic Policies  147
Defining Policy Integration  149
The Baseline: National Treatment
Beyond National Treatment: Policy Integration
Modes of Policy Integration: Coordination,
Harmonization, and Recognition
The Economics of Policy Integration  154
Transactions Costs
Policy Integration to Increase Competition
Spillovers
Compensation and Enforcement
Local versus Regional versus Global Cooperation
Policy Integration to Date: More Promise Than Reality  174
Coverage
Depth of Integration: A Function of Political Objective
Prospects for Policy Integration in RIAs
Conclusion   183
Notes   183
Chapter 7. Regional Integration as Politics  187
Regional Integration as a Means of Reducing Frictions between
Antagonistic Neighbors  188
Trade as a Promoter of Peace
RIAs as Promoters of Peace
RIAs Are Not Always Effective Routes to Peace



Vi    REGIONAL INTEGRATION AND DEVELOPMENT
RIAs and Social and Political Pressures: Potentially Helpful, but
Not a Panacea   196
The Role of Regional Integration in Strengthening Democracy and
Political Institutions 198
Regionalism and the Nation-State 201
Regional Integration to Deal with Outside Threats and
Regional Hegemons
Regional Integration and Negotiations with the
Outside World
Notes   206
Chapter 8. Trade Blocs and the Rest of the World  209
Trade Discrimination: Still Significant  210
Trade Diversion and Excluded Countries  212
Evidence of Trade Diversion
Loss of Exports to Trade Diversion
The Effect of Large RIAs on Nonmembers' Terms
of Trade
The Road to Multilateralism: Are RIAs Stepping Stones, or
Millstones? 221
Multilateralism as a Process
Negotiating Power of RIAs
Regionalism and Tariff Levels
Has Regionalism Spurred Multilateral Negotiations?
Domino Regionalism
Regionalism as Insurance
RIAs as Negotiating Partners: Do They Promote
Free Trade?
Do RIAs Make It Easier to Tackle Tough Issues?
Open Regionalism: Little More Than a Slogan
Regionalism and the WTO   244
GATF, and All That
The Rules for RIAs: Useful, but Not Infallible
Proposals for Improving the Rules on RIAs: "Feasible"
and "Desirable" Do Not Overlap
Conclusion: Rules Are Not the Answer
Notes   256
Chapter 9. Rules of Thumb for Regionalism    261
Annex.     Selected WTO Provisions on Regional Integration
Arrangements    267
Bibliography   271
Index     299



CONTENTS      vii
LIST OF BOXES
Box 1.1   Defining Terms: What's in a Name?  2
Box 1.2   What the Politicians Say: Hopes and Rationales for RIAs  7
Box 1.3   What the Treaties Say: Aims and Objectives  8
Box 2.1   Computable General Equilibrium (CGE) Modeling: An Explanation
and a Health Warning  48
Box 3.1   Rules of Origin Are Protectionist  80
Box 3.2   Restaurant Bills, Universalism, and Protection  87
Box 3.3   Pressure Groups and M/lERCOSUR  91
Box 3.4   Taxes: A Race to the Bottom  99
Box 4.1   NAFTA and the Credibility of Mexico's Policy Reforms  109
Box 4.2   Credibility on Deep Integration: The European Single Market
Programme    110
Box 4.3   Will RIAs Enhance Credibility in Africa?  112
Box 5.1   Modeling the Agglomeration Effects of RIAs  139
Box 5.2   The Influence of Regionalism on Within-Country Location: North
America   144
Box 6.1   Defining Terms: "Deep" and "Shallow" Integration  151
Box 6.2   Harmonization and Mutual Recognition as Substitutes or
Complements: The EU Experience  154
Box 6.3   When Is Nondiscrimination Discriminatory? The Economic Effects
of Recycling Requirements  159
Box 6.4   Antitrust Law Competition by U.S. States in the 19th Century
165
Box 6.5   Labor Policies in U.S. States: Spillovers, but No "Race to the
Bottom"   166
Box 6.6   Multilateral Policy Integration and the Enforcement Role of RIAs
172
Box 7.1   Trade and Peace: The Political Lineup Follows the Economic
Lineup   189
Box 7.2   Modeling Trade and Security Externalities  191
Box 7.3   Democracy in MERCOSUR     199
Box 7.4   Trade Preferences Do Not Inevitably Lead to Political Integration
202
Box 7.5   CARICOM and International Negotiations  205
Box 8.1   Trade Diversion and Investment Switching in Europe  220
Box 8.2   Stepping Stones, or Millstones? A Summary  222
Box 8.3   Regionalism and Protection  227
Box 8.4   Is Open Access the Key to Benign Regionalism?  233
Box 8.5   Insurance Policies  237
Box 8.6   The Overburdened Negotiator  240



viii   REGIONAL INTEGRATION AND DEVELOPMENT
LIST OF FIGURES
Figure 1.1 Regional Integration Agreements, Notified and Active, 1948-April
2002   3
Figure 2.1  Share of Imports from Partners in Selected RIAs, One Year before
and Five Years after Implementation  32
Figure 2.2 Intrabloc Imports as a Share of GDP, Selected RIAs, One Year
before and Five Years after Implementation  37
Figure 2.3  Extrabloc Imports as a Share of GDP, Selected RIAs, One Year
before and Five Years after Implementation  38
Figure 2.4 Intrabloc Import Propensities, Selected RIAs, One Year before and
Five Years after Implementation  39
Figure 2.5  Extrabloc Import Propensities, Selected RIAs, One Year before and
Five Years after Implementation  39
Figure 2.6  EU Trade, 1980-96  43
Figure 2.7  Effect of RIAs on Trade within and among Blocs, Selected Periods,
1980-96   45
Figure 2A.1 The Market for a Single Homogeneous Good  55
Figure 2A.2 An RIA with a Small Partner  57
Figure 3.1  Overlapping Blocs in Africa  76
Figure 3.2  Potential Gains from Additive Regionalism for Chile from Joining
Additional RIAs  77
Figure 4.1  Net Inflows of Foreign Direct Investment, Mexico, 1985-96  120
Figure 5.1  Equalizing Exchange: Trade Liberalization and Income
Convergence, 1947-81   128
Figure 8.1  Average Prices of Brazil's Imports from Argentina Relative to Prices
of Imports from the United States, 323 Commodities, 1990-96
217
Figure 8.2  Average Prices of U.S. Exports to Brazil Relative to Prices of U.S.
Exports to the Rest of the World, 1,356 Commodities, 1991-96
218
Figure 8.3  Average Prices of Exports from the Republic of Korea to Brazil
Relative to Prices of Korea's Exports to the Rest of the World,
99 Commodities, 1990-96   219
Figure 8.4  Domino Regionalism: Changes in Welfare as RIAs Form  232
LIST OF TABLES
Table 2.1  Sources of Welfare Benefits in MERCOSUR's Imperfectly
Competitive Sectors  52
Table 2.2  Welfare Effect for Senegal of UEMOA and Unilateral Liberalization
53
Table 2A.1 Trade Effects of RIAs between Developing Countries, One Year
before and Five Years after Implementation of Internal
Preferences  59



CONTENTS      ix
Table 3.1  Trade Shares, Selected RIAs  67
Table 3.2  Revenue Implications of a Free Trade Area: Customs Duties before
and after Formation of the Southern African Development
Community (SADC)    95
Table 4.1  Net Inflows of Foreign Direct Investment, MERCOSUR, 1991-98
121
Table 6.1  Tariff Equivalents of Restrictions on Services Trade  160
Table 8.1  Tariff Averages and Peaks in Selected Developing Countries and
Customs Unions  :211
Table 8.2  Estimates of Potential NVelfare Effects of Selected Asia-Pacific RIAs
216
Table 8.3  Peru's Cattle Imports before and after Formation of the Andean
Pact   217
Table 8.4  Domino Regionalism: Strong, Intermediate, and No Expansion
234






Preface
T       he growth of regional trading blocs has been one of the major
developments in international relations in recent years; virtually
all countries are now members of at least one bloc. In addition to
the boom in numbers, the past 10 years have also witnessed qualitative
changes in regional integration arrangements. Three of these merit
special attention:
*    The move from a closed to a more open model of regionalism.
The new wave of regional integration agreements (which in-
cludes the revival of some old agreements) generally exhibits a
more outward-looking stance and greater commitment to boost-
ing, rather than controlling, international commerce.
*    The recognition that, in addition to reducing tariffs and quotas,
effective integration requires the removal of other barriers, or
what has come to be known as "deep integration."
*    The advent of trade blocs in which high-income countries and
developing countries are equal partners-so-called North-South
agreements.
This evolution pointed to the need for the World Bank to undertake a
new analysis of regional integration agreements with a focus on develop-
ing countries. There are two reasons for this emphasis. First, developing
countries are turning to regionalism as a tool for development, and the ef-
fectiveness of this tool needs to be assessed. Second, regionalism is part of
the global economic environment, and its effects on developing countries
need to be better understood. It also seemed desirable to ground the new
analysis more firmly in empirical results than had been done previously.
xi



xii   REGIONAL INTEGRATION AND DEVELOPMENT
The analysis presented in this book assesses the effects of regional in-
tegration arrangements (RIAs) on trade and competition (chapter 2); ex-
amines how to make the most of RIAs (chapter 3); considers the impact
of RIAs on investment and foreign direct investment (chapter 4); care-
fully examines the dynamic economic effects of integration (chapter 5);
assesses the needs and opportunities for deep integration (chapter 6); ex-
plores the intersection between the economic and political effects of inte-
gration (chapter 7); and examines the relationship between regionalism
and multilateralism (chapter 8). Although regionalism is too complex
and sui generis to generate universal and operational policy rules, the ev-
idence presented in the first eight chapters is reduced in chapter 9 to a set
of consistent policy messages that apply to most circumstances. These
rules of thumb are not inviolable, but they should not be broken lightly
(chapter 9).
The book is the outcome of a research project on regionalism and de-
velopment, led by the authors and carried out at the World Bank starting
in 1996. The main themes of that research were recently presented in a
companion volume, the World Bank Policy Research Report Trade Blocs
(2000). That report was based on the same ideas and research as are de-
scribed here, but organized differently and presented in a straightfor-
ward way for a broad audience. The present book underpins those
themes more firmly by going into more detail than could Trade Blocs
and by placing greater emphasis on the analytical and intellectual struc-
tures that can be used for exploring the effects of regionalism. The re-
search project has also generated more than 100 research papers, more
than half of which have been published. Most of these papers are avail-
able at <http://www.worldbank.org/research/trade/>.
The research project and this book benefited from contributions and
comments by Azita Amjadi, Soamiely Andriamananjara, Richard Bald-
win, Fred Bergsten, Jagdish Bhagwati, Magnus Blomstrom, Eric Bond,
Won Chang, Paul Collier, Jo-Ann Crawford, Jaime de Melo, Dean
DeRosa, Robert Devlin, Simon Evenett, Raquel Fernandez, Michael Fin-
ger, Faezeh Foroutan, Jeffrey Frankel, Anju Gupta, Dominique Hachette,
Bart Kaminski, Walter Kennes, Moonhui Kim, Ari Kokko, Pravin Krishna,
Praveen Kumar, Sam Laird, Philip Levy, Dorsati Madani, Will Martin,
Francis Ng, Marcelo Olarreaga, Pier Carlo Padoan, Arvind Panagariya,
Joost Polak, Jonathan Portes, Diego Puga, Sherman Robinson, Peter
Robson, Andre Sapir, Jeffrey Schott, Elena Seghezza, Isidro Soloaga,
T. N. Srinivasan, Sherry Stephenson, Costas Syropolous, David Tarr,
Peter Tulloch, Athanasios Vamvakidis, Tony Venables, Yanling Wang,
Shang-Jin Wei, John Whalley, and Alexander Yeats and by three anony-



PREFACE   xiii
mous referees. Special thanks go to Bernard Hoekman, who co-authored
chapter 6. Naturally, none of the individuals who contributed to or com-
mented on the book is responsible for its remaining shortcomings.
Melissa Edeburn, the production editor, and Nancy Levine, the copy
editor, did outstanding jobs. Excellent assistance was provided by Mary-
Ann Arouna, Rosie Bellinger, Janet Ellis, Maria Lourdes Kasilag, Audrey
Kitson-Walters, and Rebecca Martin.
Thanks to Lunia, Dominique, Oliver, and Zhen Kun for their love and
support throughout a very long project.






Abbreviations and Acronyms
ACP            African, Caribbean, and Pacific (Lome Convention)
countries
AERC           African Economic Research Consortium
AFTA           ASEAN Free Trade Area
APEC           Asia-Pacific Economic Cooperation
ASEAN          Association of Southeast Asian Nations
BIT            bilateral investrnent treaty
CACM           Central American Common Market
CAP            Common Agricultural Policy (EU)
CARICOM        Caribbean Community and Common Market
CBI            Cross-Border Initiative (Africa)
CEAO           Economic Community of West Africa (became UEMOA)
CEFTA          Central European Free Trade Area
CEMAC          Economic and Monetary Community of Central Africa/
Communaute economique et monetaire d'Afrique Cen-
trale
CEPGL          Economic Community of the Countries of the Great
Lakes/Communaute economique des pays des grands
lacs
CER            Closer Economic Relations agreement (Australia and
New Zealand)
CET            common external tariff
CGE            computable general equilibrium (model)
CMEA           Council for Mutual Economic Assistance
COMESA         Common Market for Eastern and Southern Africa
xv



xvi    REGIONAL INTEGRATION AND DEVELOPMENT
CRTA          Committee on Regional Trading Agreements (WTO)
CUSFTA        Canada-United States Free Trade Agreement
EAC           East African Cooperation (formerly East African Com-
munity)
EC            European Community
ECOWAS        Economic Community of West African States
ECSC          European Coal and Steel Community
EEA           European Economic Area
EEC           European Economic Community
EFTA          European Free Trade Association
EPA           economic partnership agreement
EU            European Union
FDI           foreign direct investment
FTA           free trade agreement; free trade area
FTAA          Free Trade Area of the Americas
G3            Group of Three (Colombia, Mexico, and Venezuela)
GATS          General Agreement on Trade in Services
GATT          General Agreement on Tariffs and Trade
GCC           Gulf Cooperation Council
GDP           gross domestic product
GNP           gross national product
GSP           generalized system of preferences
HS            Harmonized Commodity Description and Coding
System
ICSID         International Centre for Settlement of Investment
Disputes
IMF           International Monetary Fund
IOC           Indian Ocean Commission
IPRs          intellectual property rights
ITA           Information Technology Agreement
LAFTA         Latin American Free Trade Area
LAIA          Latin American Integration Association
LATN          Latin American Trade Network
MAI           Multilateral Agreement on Investment
MERCOSUR      Common Market of the South/Mercado Comuin del Sur
MFN           most-favored-nation
MITI          Ministry of International Trade and Industry (Japan)
MRA           mutual recognition agreement
NAFTA         North American Free Trade Agreement



ABBREVIATIONS AND ACRONYMS   xvii
OECD          Organisation for Economic Co-operation and
Development
PTA           preferential trade agreement
R&D           research and development
RIA           regional integration agreement
ROW           rest of the world
SAARC         South Asian Association for Regional Cooperation
SACU          Southern African Customs Union
SADC          Southern African Development Community
SADCC         Southern African Development Coordination Confer-
ence (renamed SADC)
SAPTA         SAARC Preferential Trading Arrangement
SMP           Single Market P'rogramme (EU)
TFP           total factor productivity
UDEAC         Union douaniere et economique de l'Afrique Centrale
(renamed CEMAC)
UEMOA         West African Economic and Monetary Union/Union
economique et monetaire ouest-africaine
UNCTAD        United Nations Conference on Trade and Development
VAT           value-added tax
WTO           World Trade Organization
Note: The use of the designations EEC, EC, and EU depends on the time
period and historical context.






CHAPTER I
Regional Integration Agreements: An Overview
The growth of regional trade blocs has been one of the major developments
in international relations in recent years. Virtually all countries are members
of a bloc, and many belong to more than one. Over a third of world trade
takes place within such agreements-nearly two-thirds, if Asia-Pacific Eco-
nomic Cooperation (APEC) is included. Regional agreements vary widely,
but all have the objective of reducing barriers to trade between member
countries-which implies discrimination against trade with other countries.
At their simplest, these agreements merely remove tariffs on intrabloc trade
in goods, but many go beyond that to cover nontariff barriers and to extend
liberalization to investment and other policies. At their deepest, they have
the goal of economic union and involve the construction of shared execu-
tive, judicial, and legislative institutions. For want of a better term, we lump
them all into the category of regional integration agreements (RIAs), as de-
fined in box 1.1. The appendix to this chapter lists the principal RlAs.
During the past decade, the move toward regionalism became a head-
long rush. Figure 1.1 highlights the dramatic increase in agreements noti-
fied to the General Agreement on Tariffs and Trade/World Trade Organi-
zation (GATT/WTO) in the 1990s.' Of the 194 agreements that had
been notified by the beginning of 1999, 87 dated from 1990 or after.
The past decade also witnessed qualitative changes in RIAs. There
have been three major developments:
1. The move from 'closed regionalism" to a more open model, in
line with prevailing views about national economic policy. Many
of the trade blocs that were formed between developing countries
in the 1960s and 1970s were based on a model of import-
1



2       REGIONAL INTEGRATION AND DEVELOPMENT
Box 1.1 DefiningTerms:What's in a Name?
Who but a staunch protectionist could have anything against a 'free trade agreement"? 'Preferential
trade agreements" sound less benign, while "discriminatory trade agreements," yet another name
for the same thing, sound nasty.
Martin Wolf, Financial Times, October 28,1996
This book Is primarily about preferential trading agreements between groups of countries. We refer
to them neutrally as regional integration agreements (RIAs) to avoid any unsubstantiated pejorative
Implications and to convey that arrangements can extend well beyond international trade into areas
such as investment, domestic regulation, domestic policies, standards, infrastructure, and politics.
Indeed, we shall argue that cooperation in these areas Is frequently more important than in intema-
tional trade and that in some cases governments should eschew using trade agreements as a vehi-
cle for cooperation on nontrade issues. We use the shorthand term "regionalism" to cover the
process of creating RIAs.
Although most RIAs are between neighboring countries, this is not universally true-as illustrated,
for example, by the free trade agreements between the United States and Israel, Chile and Canada,
and Mexico and the European Union (EU) and by Greece's accession to the EU. Nor is i required for
most of the analytical arguments that we make.
We do not attempt to cover all arrangements between countries. For example, we do not deal with
alliances for defense or political cooperation, not because they are not important but because we
have no particular expertise-nor the World Bank any particular mandate-in these areas.
substituting development, and regional agreements with high
external trade barriers were used as a way of implementing this
model.2 New-wave RIAs (some of which are old agreements res-
urrected) are generally more outward looking and more commit-
ted to boosting rather than controlling international commerce.
2.    The recognition that effective integration requires more than sim-
ply reducing tariffs and quotas. Many other types of barrier have
the effect of segmenting markets and impeding the free flow
of goods, services, investments, and ideas, and wideranging policy
measures-going well beyond traditional trade policies-are
needed to remove them. Such "deep integration" was first actively
pursued in the Single Market Programme of the European Union
(EU), but its elements are now finding their way into the debate
on other regional agreements.
3.    The advent of "North-South" trade blocs in which high-income
countries and developing countries are equal partners.3 Perhaps
the most important example is the North American Free Trade
Agreement (NAFTA), formed in 1994 when the Canada-United



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW        3
Figure 1.1 Regional Integration Agreements, Notified and Active, 1948 to April 2002
25
z 0
e   Ii" le le le le 4?  ,p 4? 4?4e le 4 4l 4? 4? 4? 4?db le le le l?e 4? 14 84 le 4P le
* Notdied and active n3 Currently inactive
Note. Information on mactive RlAs for 2000-02 is not available.
Source: World Trade Organization data.
States Free Trade Agreement (CUSFTA) was extended to Mexico.
The EU also has North-South arrangements, including the Europe
Agreements that link the EU with the transition economies of
Eastern Europe, a customs union with Turkey, and agreements
with many Mediterranean countries. In addition, the EU is com-
mitted to negotiating reciprocal trade agreements (economic part-
nership agreements, or EPAs) with the African, Caribbean, and
Pacific (ACP) countries.
These developments have occurred against the backdrop of globaliza-
tion: new technologies and more liberal trading regimes have led to
higher trade volumes, larger investment flows, and increasingly footloose
production.
All these considerations reinforced the need for the World Bank to
undertake a new analysis of regional integration agreements-one that



4    REGIONAL INTEGRATION AND DEVELOPMENT
would take political effects into account, carefully examine dynamic eco-
nomic effects, assess thc needs and opportunities for deep integration,
capture the new potentials created by North-South agreements, and be
firmly grounded in empirical analysis.
A new analysis would also have to focus on developing countries, for
two reasons. First, developing countries are turning to regionalism as a
tool for development, and the effectiveness of this strategy needs to be
assessed. Second, regionalism is part of the global economic environment
and affects developing countries, whether or not they participate in it.
Understanding its implications can help them better prepare for and cope
with regionalism.
Macroeconomic and exchange rate policies are important for the sus-
tainability of RIAs and affect their perceived or actual desirability. Un-
sustainable macroeconomic policies may lead to backsliding on coun-
tries' commitments to liberalize intrabloc trade or to converge to a
common external tariff in a customs union. And unexpected exchange
rate changes may affect the short-to-medium-term distribution of the ex-
pected gains from an RIA, possibly reducing its attractiveness. For rea-
sons of space and focus, however, we do not deal with these issues, limit-
ing our analysis instead to the real aspects of regional integration.
Another consideration is that our interest is in the long-term implications
of RIAs, and we view macroeconomic disequilibria as essentially tempo-
rary phenomena.
RIAs IN HISTORY
RIAs have been around for hundreds of years. For example, a customs
union of the provinces of France was proposed in 1664; Austria signed
free trade agreements with five of its neighbors during the 18th and 19th
centuries; and the colonial empires were based on preferential trade
arrangements. Customs unions were precursors to or were embodied in
the creation of new states in, for example, Germany (the Zollverein),
Italy, and the United States.
The 1930s saw a great fragmentation of the world trading system as
governments struggled with the slump in demand without the benefit of
global economic institutions to provide liberal focal points. One of the
"solutions" adopted was regional preferences. The exact causal relation-
ship between restricted trade and declining incomes during this period is
still debated, but fragmentation into closed blocs must have fostered in-
efficiency and frustrated recovery from the Great Depression.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW   S
Partly in response to the experience of the 1930s, and partly under the
influence of U.S. idealism and internationalism, the post-World War II
system established equal treatment of all partners (nondiscrimination) as
a fundamental principle of the trading system. Exceptions were permit-
ted, both on pragmatic grounds and for reasons of principle, and among
these exceptions was the ability to create trade blocs-free trade areas
(FTAs) and customs unions. Aside from reinforcing existing colonial
links, this concession was little used at first, but over its first decade it
contributed to the political reconstruction of Europe through the cre-
ation of the Benelux customs union in 1947, the European Coal and
Steel Community (ECSC) in 1951, and the more far-reaching European
Economic Community (EEC) in 1957.
The survival and apparent success of the EEC led to a spurt of region-
alism between developing countries in the 1960s. This spurt was mostly
driven by the import-substitution creed that protection was required for
industrialization-and hence for prosperity-and that this policy would
be less costly if extended over a larger economic area. The RIAs were
generally very protectionist and intcrventionist in the sense of trying to
determine administratively which industries to have and where they
should be located. They involved numerous controls and restrictions on
economic activity and, consequently, yielded rather modest economic re-
sults. In addition, the degree of implementation was often low, in part
because of disagreements on where industries should be located. By the
late 1970s, the ineffectiveness of these RIAs had become evident. None
seemed to have contributed strongly to development; some had col-
lapsed; and the strains of the debt crisis made those that survived largely
moribund.
In the 1980s a huge change in attitudes toward international trade
and competition took place. Led by the EU's Single Market Programme,
a new wave of apparently more liberal RIAs emerged. These were in-
spired by a set of hopes and aspirations, the analysis of which is the sub-
ject of this report.
The recent growth in regionalism was dominated by the EU's activi-
ties: the extension of the Single Market Programme to neighboring coun-
tries that were not yet members of the EU, through the European Eco-
nomic Area; the signing of Europe Agreements with the countries of
Eastern Europe; the accession of three new members in 1995; and the de-
velopment of a more active and formal Mediterranean policy that poten-
tially included RIAs with nearly every Mediterranean country. In fact, of
the 87 notifications of RIAs to the WTO since 1990, only 13 had no Eu-
ropean partner.



6    REGIONAL INTEGRATION AND DEVELOPMENT
But Europe was not the only continent involved. In the Americas the
Canada-United States Free Trade Agreement of 1988 was extended to
Mexico in 1994 through NAFTA; Common Market of the South (Merca-
do Comiun del Sur-MERCOSUR) was formed in 1991 and the Group of
Three (G3) in 1995; and the Andean Pact and the Central American Com-
mon Market (CACM) were resurrected in 1991 and 1993, respectively. In
1992 the countries of the Association of Southeast Asian Nations
(ASEAN), after 25 years of political cooperation with limited trade coop-
eration, formed a meaningful FTA, the ASEAN Free Trade Area (AFTA).
Since then, additional countries have joined AFTA, which has also started
talks with China. The Republic of Korea and Japan, among others, are ne-
gotiating an FTA. In West Africa the trade blocs re-formed in more liberal
and more tightly organized blocs. The Common Market for Eastern and
Southern Africa (COMESA) replaced a preferential trade agreement
(PTA), and many of its members also took part in the Cross-Border Initia-
tive (CBI). The Southern African Development Coordination Conference
(SADCC) transmogrified into the Southern African Development Commu-
nity (SADC), which is a trade and economic cooperation association
rather than a defense organization. East African Cooperation (EAC)
sprang up where the East African Community had failed. In North Africa
the Mahgreb and Mashraq groups have renewed their integration efforts.
WHY REGIONALISM?
Many factors lay behind the recent spurt in regionalism. Some were ex-
plicitly stated goals, as summarized in boxes 1.2 and 1.3. Other objec-
tives could not be so publicly admitted, and yet others represented more
fundamental causes. Among the objectives, stated and implicit, were:
*  Governments' wish to bind themselves to better policies-
including democracy-and to signal such bindings to domestic
and foreign investors
*  A desire to obtain more secure access to major markets
*  The pressures of globalization, forcing firms and countries to seek
efficiency through larger markets, increased competition, and ac-
cess to foreign technologies and investment
*   Governments' desire to maintain sovereignty by pooling it with
others in areas of economic management where most nation-
states are too small to act alone
*  A desire to jog the multilateral system into faster and deeper ac-
tion in selected areas by showing that the GATT was not the only



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW                       7
Box 1.2 What the Politicians Say: Hopes and Rationales for RIAs
Trade
"The implementation of the Arab Free Trade Agreement will be a major factor that will help the King-
dom expand its export markets."
Hani Mulki, Jordan's minister of industry, trade, and supply, Xinhua News Agency,
September 7, 1998
'The regional integration did not hamper world trade liberalization, but rather was an essential com-
ponent of it, enabling countries at differing levels of development to participate more effectively in
world trade."
Colombia's representative to the meeting of the WTO Working Party on MERCOSUR,
October 10-11, 1995
Investment
"To promote foreign direct investments in ASEAN. The synergies of both [ASEAN and the ASEAN Free
Trade Area, AFTA] will ensure that ASEAN remains as a highly attractive and globally competitive in-
vestment region."
Datuk Ajit Singh, secretary-general of ASEAN, Straits limes,
February 25, 1997
"Such a healthy atmosphere is a major attraction to foreign industrial investment.... We encourage
Arab investment through which we can find an entry to unify their economies through establishment
of free trade markets."
Mohamed Al Amadi, Syna's minister for economy and foreign trade, on the Arab Free
Trade Zone, Egyptian Gazette, November 6, 1998
"To clearly regulate the growing trade between our countries, encourage investment and create
jobs.'
Carlos Salinas de Gortarl, president of Mexico, on NAFTA, State of the Nahon Address,
November 4, 1994
Security and democracy
"Formation of Euro-Mediterranean forum will give impetus to the partnership at all political, eco-
nomic, security and development levels."
Ahmed Fathi Sorous, chairman of the Arab Parliamentary Union and speaker of the
Egyptian People's Assembly, interview with Sawt al Arab Radio, October 28, 1998
"NAFTA has contributed to the prosperity and stability of our closest neighbors and two of our most
important trading partners. NAFTA aided Mexico's rapid recovery from a severe economic recession,
even as that country carried forward a democratic transformaton of historic proportions."
William Clinton, president of the United States, letter to the U.S. Congress,
July 1997
Market access
"New Zealand's worst fear is for the world to divide into trading blocs ... in none of which New
Zealand has a home. So we are keen to keep working with ASEAN."
Helen Clark, prime minister of New Zealand, on the New Zealand-Singapore-FTA, Far
EastEconomic Review, August 17, 2000
(Box continues on next page.)



8      REGIONAL INTEGRATION AND DEVELOPMENT
Box 1.2 (continued)
Deep integration
'The Europe Agreement with the EC is to provide an appropriate framework for both a political dia-
logue and a comprehensive cooperaton in a broad range of areas and a step towards the Czech Re-
public's ultimate objective of full membership in the EC."
Representative of the Czech Republic to the WTO Committee on Regional Trade
Agreements, May 30,1997
Growth and development
"Will boost technology transfer from the north to the south Mediterranean countries."
Ahmed Fathi Sorous, chairman of the Arab Parliamentary Union and speaker of the
Egyptian People's Assembly, interview with Sawt al Arab Radio, October 28, 1998
"To achieve development and economic growth, alleviate poverty, enhance the standard and quality of IHfe
of the peoples of southem Africa, and support the socially disadvantaged through regional integration."
Chris Stals, governor of the South African Reserve Bank, on the SADC, May 1, 1997
Box 1.3 What theTreaties Say: Aims and Objectives
Trade and Income
"To enhance the competitiveness of their firms in global markets"
NAFTA, 1992
"To create an expanded and secure market for the goods and services produced in their territories.
To reduce distortions to trade"
G3 Treaty (Colombia, Mexico, and Venezuela), 1994
"To modernize their economies in order to expand the supply and improve the quality of available
goods and services, with a view to enhancing the living conditions of their populations"
MERCOSUR Agreement, 1991
Investment
"To ensure a predictable commercial framework for production activities and investment"
G3 Treaty, 1994
'Prerequisite for the stimulation of domestic, regional and foreign direct investment and the expan-
sion, growth and the development of the economies of each member state and the region as a
whole'
Kinshasa Resolution on the Establishment of COMESA, 1998
'A stimulus to the development of the national economies by expanding investment and production
opportunities, trade, and foreign exchange eamings"
Agreement on the Common Effective Preferential Tariff Scheme for the ASEAN Free
Trade Area, 1992
Development
"To ensure in particular that these arrangements encourage the development of the less advanced
members of the customs union and the diversification of their economies"
Southem African Customs Union (SACU) Agreement, 1969
(Box continues on next page.)



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW                             9
Box 1.3 (continued)
Democracy and human rights
"To involve the peoples of the Region centrally in the process of development and integration, particularly
through the guarantee of democratic rights, observance of human rights and the rule of law"
SADC Treaty, 1992
"To strengthen democracy and respect for human rights, sustainable and balanced economic and
social development, to combat poverty and promote greater understanding between cultures"
Barcelona Declaration adopted at the Euro-Mediterranean Conference, 1995
Regional cooperation and coordination
'To establish a firm foundation for common action to promote regional cooperation in South-East
Asia in the spirit of equality and partnership and thereby contribute towards peace, progress and
prosperity In the region"
ASEAN Declaration, 1967
"To foster coordinated action by the Parties in international economic fora, particularly in those re-
lated to the processes of Latin American integration"
G3 Treaty, 1994
Regional ind global integratlon
"To contribute to the harmonious development and expansion of world trade and provide a catalyst
to broader international cooperation"
NAFTA, 1992
"To facilitate the accession of Chile to the North American Free Trade Agreement ... To contribute to
hemispheric integration"
Canada-Chile Free Trade Agreement, 1996
game in town and by creating more powerful blocs that would
operate within the GATT system
*    A   desire   to  help   neighboring       countries     stabilize   and    prosper,
both for altruistic reasons and to avoid spillovers of unrest and
population
*    The fear of being left out while the rest of the world swept into re-
gionalism, either because this would be actually harmful to ex-
cluded countries or just because "if everyone else is doing it,
shouldn't we?"
Among the more fundamental causes, we would list the following:
*    The collapse of Soviet hegemony, which led the countries of East-
ern Europe and the Baltic to embrace democracy and capitalism
and those of Western Europe to seek ways of cementing and ac-
celerating their transition.
*    The change in understanding of the role of openness in develop-
ment, coupled with a natural political desire to limit the feared
adjustment costs of unilateral nondiscriminatory liberalization.



10    REGIONAL INTEGRATION AND DEVELOPMENT
*  The need to create a domestic dynamic for the reforms required to
achieve greater openness while at the same time minimizing the
political problems of disrupting existing sources of incomes and
rents.
*  The changed attitude of the United States toward trade blocs,
from active hostility to a broadly enthusiastic stance. This shift
both fostered RIAs and reduced the diplomatic pressure-overt
(through the GATT) and covert-for countries to desist from
forming them. It stemmed at least in part from an expressed frus-
tration with the slowness of the multilateral process. Also impor-
tant, if less public, were the increasing influence of business lob-
bies in U.S. policymaking, the decreasing competitiveness of U.S.
industry, and a lessened willingness to bear the costs of managing
the global system without receiving direct payoffs in the form of
markets. With the end of the Cold War, the overriding political
justification for bearing those costs disappeared, and the debate
focused more directly on mercantilist objectives, as it always had
in most smaller countries.
This book asks whether the hopes expressed for regionalism are justi-
fied, or have been justified by recent evidence, and how their fulfillment
might be made more likely. In doing so, it takes the economic arguments
summarized above and subjects them to more careful scrutiny than they
have usually received.
WHY ANOTHER BOOK ON REGIONALISM?
The literature on regionalism is massive, and a proposal to add to it re-
quires some justification. Our rationale is fourfold:
1. Policymakers need practical advice presented in the form of rea-
soned argument and evidence. The body of knowledge and analy-
sis that could provide concrete guidance on whether and when
to join RIAs, and on how trade groups should be structured to
maximize the welfare of the group and its individual members,
has not been pulled together before in a single volume. The
tradeoffs involved in forming or joining an RIA are complex
and difficult to isolate. Although this book provides few universal
or easy solutions to these problems, it does attempt to set the
issues out clearly and to suggest the analytical paths that policy-
makers should take in thinking about their own integration
arrangements.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW      11
2. Regionalism is still a very fertile area for research, with new re-
sults and interpretations emerging every day. Reading these, let
alone assessing them, is not the job of policymakers (they have
better things to do), and so a straightforward presentation of the
more important recent developments is desirable.
3. World Bank research has yielded new results on regionalism that
should be made widely available. Among these contributions are:
* The first coherent discussion of the credibility benefits of RIAs
* Tests of the role of political factors in setting an RIA's internal
and external trade policy
* Analysis of the effects of RIAs on industrialization in develop-
ing countries
* A comprehensive assessment of progress on "deep" (policy) in-
tegration within RIAs
* Findings on the effects of recent RIAs on members' trade pat-
terns and trade policies
* A formal analysis of the economic implications of establishing
RIAs to promote peace and security
* Evidence on the growth or nongrowth effects of RIAs
* A new approach to assessing the effects of RIAs on excluded
countries.
4. Although the existing literature is huge, there is no consistency in
the methodologies and intellectual bases of the research or in its
conclusions. Thus, in addition to summarizing what economists
think they know and assessing how much confidence to place in
that knowledge, we also seek to set forth the internal structure of
the various arguments. The intention is to allow policymakers to
judge for themselves which results apply to their particular cir-
cumstances and how well they fit together into a coherent view of
the RIA under consideration. In particular, we want to help poli-
cymakers avoid the regrettably common "pick-and-mix" ap-
proach of selecting a convenient set of results from the literature
without regard for their rnutual consistency.
One consequence of our ambitious agenda is that we are able to re-
solve fewer issues empirically than we would wish. There are not many
RIAs of sufficient longevity and consistency of application to provide
convincing historical evidence, and each case has so many different char-
acteristics and is so confounded by other factors (such as developments
in national politics, economic policy, and the world economy) that disen-
tangling the various effects becomes difficult. Accordingly, while we re-
port several innovative pieces of empirical analysis-for instance, on the



12    REGIONAL INTEGRATION AND DEVELOPMENT
effects of RIAs on excluded countries and on members' propensities to
protectionism-we also rely heavily on examples and on a priori reason-
ing. We attempt to gauge the plausibility of such pieces of evidence in
reaching overall conclusions, but we are conscious that they are more
ambiguous than is ideal.
As noted above, developments in the EU and its predecessors have un-
derlain both of the major spurts of regionalism in the postwar era, and
Europe has pursued regional integration farther and deeper than any
other (voluntary) arrangement outside the historical instances of nation
building. Europe has also generated most of the intellectual advances in
the field. We draw a great deal on the European experience, but we seek
throughout to apply its lessons to developing country issues.
Our approach is essentially economic, reflecting our comparative ad-
vantage. Our concern is with whether regionalism aids economic devel-
opment. We cast the net broadly, however-examining, for example, the
implications of the argument that RIAs can be aimed at political objec-
tives, how RIAs affect policy credibility, and the design of institutions for
setting trade policy in an RIA. Since these issues hinge on tradeoffs of the
sort that economics deals with-how much a reduction of tension is
worth in terms of production inefficiency-they fall well within our
ambit. We do not, however, give a full account of the political and social
aspects of regionalism.
The main themes of World Bank research and policy advice on re-
gionalism were presented in a companion volume, the Policy Research
Report Trade Blocs, published in July 2000. That report was based on
the same ideas and research as are described here, but it was organized
differently and was presented in a straightforward way for a broad audi-
ence. It had four themes:
*   Many RIAs are political in origin.
*   Nonetheless, they entail economic costs and benefits, either via
their effects on competition or by affecting patterns of trade or of
industrial location.
*  These economic aspects are important and certainly influence an
RIA's chances of survival. Careful design can maximize the net
economic benefits of an RIA.
*   Regionalism must be viewed in the context of a global trading
system that has delivered great benefits to the world over the past
50 years.
This book goes into more detail on the arguments underlying the four
themes than could the 2000 report, and it consequently focuses more
sharply on the analytical and intellectual structures that have been used



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW     13
for exploring the effects of regionalism. The book is not, however, an ac-
ademic tome. It is designed for policymakers, and it attempts to lay out
the various arguments clearly and accessibly. Its final chapter summa-
rizes as "rules of thumb" the practical lessons that emerge from the
analysis.
Regionalism is a complex business and depends heavily on particular
features of particular cases. There are no rules for policy toward RIAs
that are both universal and operational-the universal rules are so broad
as to be nonoperational, and the operational rules are too specific to be
universal. The rules of thumb presented in chapter 9 represent our judg-
ment of the general tendencies within good policy advice. Although they
are not carved in stone, policymakers should think long and hard before
breaking them.
SUMMARY
The remainder of this chapter offers a brief guide to the contents of the
book to help readers find their way around the material and to bring the
various parts into perspective. IPolicy conclusions will be evident from
the chapter discussions but are brought together in chapter 9.
Regionalism as Trade Policy
Chapter 2 examines the basic economics of tariff preferences-what hap-
pens when a country reduces its tariffs on imports from a subset of its
partners. Forming an RIA almost always increases trade between mem-
bers. But does it "create" trade by allowing cheaper products from bloc
partners to substitute for more expensive domestic production, or "di-
vert" trade by substituting intrabloc imports for imports from outside
the group? The latter can happen when partner goods, because they no
longer face tariffs, have a competitive edge over "outside" goods that
would otherwise be cheaper. The preference-granting country ends up
paying more for the imports, with the increased payments to producers
in partner countries being financed by monies that had initially accrued
to the government as tariff revenues. Part of this extra cost is a simple
transfer from taxpayers in the importing partner to producers in the ex-
porting partner, but because the real cost of imports has risen (the part-
ner is less efficient than outside producers), real resources are wasted by
the diversion. If trade diversion predominates across the board, an RIA
can reduce the welfare of both partners.



14    REGIONAL INTEGRATION AND DEVELOPMENT
Distinguishing between trade creation and trade diversion is not
straightforward empirically, and we discuss some of the difficulties. The
deepest of these is deciding what trade would have looked like had no
RIA been formed-the counterfactual, or what Europeans refer to as the
anti-monde. The evidence on the balance between trade creation and di-
version in trade blocs is mixed. We offer new research showing that di-
version is significant but that for many of the more recent RIAs it has
been dominated by the effects of the partners' reductions of barriers to
imports from nonpartners that have accompanied the RIA.
The traditional analysis of trade creation and diversion is based on a
view of the world in which intercountry trade is driven entirely by differ-
ences in productivity and in factor endowments. But trade can also arise
from product differentiation and economies of scale, which reduce costs
as production grows. Then import barriers become even more costly be-
cause competition between firms is weakened and consumers lose from
the resulting decreases in output and increases in price. International
trade offers an important means of increasing competition by allowing
new suppliers to enter markets. RIAs, by fostering trade between mem-
bers, can generate such benefits because of the combination of larger
firm size (which increases economies of scale) and a larger number of
firms (which increases competition). When several national markets are
merged, the number of producers in each country may fall, while the
number of sellers with reasonable access to each market rises because
producers from partner countries now have access. For instance, if each
country in a two-country RIA starts off with three firms, even if each
loses one firm to intra-RIA competition, there will still be four (now larg-
er) firms producing and selling throughout the bloc.
These so-called procompetitive effects are believed to have operated
strongly during the course of European economic integration, but there
is not yet enough empirical evidence showing that developing countries
will be able to benefit greatly from them. The uncertainty partly reflects
developing countries' production structure, which is likely to include
fewer goods for which differentiation and economies of scale are im-
portant, and partly the fact that significant increases in competition
eventually depend on far more than merely removing tariffs and import
quotas. (See the discussion of policy integration in chapter 6.) And
of course, if the aim is to sell in large markets and buy from firms
that supply large markets, no market is larger than the world as a
whole; there will be greater procompetitive effects from nondiscrimina-
tory trade liberalization than from discriminatory or restricted liberal-
ization.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW     15
Making the Most of Regionalism
Why is regionalism so popular if it is just a pale imitation of nondiscrim-
inatory free trade? Chapter 3 describes the basic international trade as-
pects of the question and goes on to examine how the design features of
an RIA might affect its net benefits.
Governments may prefer RIAs to unilateral trade liberalization for
several reasons. They may be better able to exploit market power
against outsiders by coordinating trade policies. They may place a high
value on access to partners' markets and feel that it is better ensured by
making access to their own markets conditional on reciprocity. They
may prefer preferential to nonpreferential access to partners' markets.
And they may want to exploit the regional market as a base for pro-
tected industrialization. We do not find the last two arguments con-
vincing.
Clearly, the benefits of regionalism are likely to depend on finding the
best partners. The current notion of a "natural" trading partner is not
useful in this regard. We show that whether South-South or North-North
RIAs generate economic divergence or convergence between member
countries depends on the model used and the initial policy stance. We
also show that there are several reasons for the obvious tendency for
trade blocs to form between neighboring countries, including the desire
to reduce trade costs by relaxing or abolishing border formalities and to
facilitate the collection of tax revenues.
One of the main themes of this book is our preference for North-
South over South-South RIAs for developing countries. If a developing
country is going to pursue regionalism, it will almost always do better to
sign up with a large rich country than with a small poor one. In trade
terms, a large rich country is likely to be a more efficient supplier of most
goods and a source of greater competition for local producers. This rec-
ommendation is not absolute: several conditions need to be met if North-
South RIAs are to be strongly beneficial. Then, too, the North-South op-
tion is not open to all developing countries.
Many countries are members of several RIAs. If all these arrange-
ments are compatible, multiple membership may be beneficial-if the
RIAs are desirable per se. But there are also dangers. For example, the
RIAs may imply conflicting policies vis-a-vis third countries; they may
have different regulations governing imports of the same commodity
from different sources; or they may have different technical standards-
all of which increases the complexity, cost, and uncertainty of trade.
There is also a danger in hub-and-spoke arrangements in which one



16    REGIONAL INTEGRATION AN) DEVELOPMENT
country carries on free trade with many others that do not have free
trade between them. In this case, the "hub" country has considerable ad-
vantages as a location for economic activity and may suck firms and in-
vestment out of the spokes.
An important distinction is that between free trade areas and cus-
toms unions. The latter have, in addition to the internal free trade char-
acteristic of FTAs, a common external tariff or, strictly, a common trade
policy toward third countries. FTAs are easier to create and can be in-
stitutionally very light, whereas customs unions require the negotiation
of the common external tariff and coordination of all future trade pol-
icy changes. FTAs do face the danger of trade deflection, in which goods
enter the member state with the lowest tariff and then move tax free to
other members. Except for the extra transport cost, this is efficient eco-
nomically because it lowers the effective tariff, but it undermines mem-
bers' own tariff structures. Governments seek to prevent this from hap-
pening by imposing rules of origin to ensure that only locally produced
goods are exempted from tariffs. Such rules are often cumbersome and
protectionist and can greatly reduce the value of the FTA. Customs
unions avoid this problem and can induce a greater degree of integra-
tion. Finally, FTAs tend to be more liberal than customs unions in their
trade policy with respect to nonmembers; they have less market power,
and members may compete with one another to reduce tariffs, thus en-
larging their own share of the RIA's imports and increasing their tariff
revenue.
Lobbying influences tariffs, and it is plain that it will afflict FTAs and
customs unions too. For both, there is a presumption that the exceptions
to internal free trade negotiated as the RIA is formed will tend to reduce
the degree of trade creation and therefore the potential benefits. In the
case of customs unions the agreement will change the environment for
future lobbying, but whether toward or away from greater protectionism
is impossible to say a priori.
The chapter concludes with a discussion of RIAs and tariff revenues.
If countries have to use tariffs for revenue purposes, it may appear to pay
small developing countries to combine into a single market to reap
economies of scale and enhance competition while raising revenue
through tariffs on trade with the outside world. In fact, this is far from
being a watertight argument. Moreover, there are other potentially com-
plex revenue issues to consider, such as precisely how to replace the rev-
enue lost when tariffs on intrabloc trade are removed and how to avoid
reduction of taxes below optimal levels as a result of fiscal competition
between members.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW     17
Regionalism and Investment
As is clear from boxes 1.2 and 1.3, promoting investment is a prominent
objective of many RIAs. The logic is that larger markets, more competi-
tion, and improved policy credibility will increase the incentives for in-
vestment and so raise incomes. This argument is relevant to all invest-
ment but is most explicitly applied to foreign direct investment (FDI).
Chapter 4 briefly describes RIAs' policies toward investment and asks
whether the arguments advanced for their having positive effects are
justified.
Early RIAs were activist and interventionist, coopting regional inte-
gration into import substitution at a regional level. Such policies almost
completely failed and have been superseded by a much more market-
friendly approach that places more emphasis on policies guaranteeing
the fair treatment of investment. These guarantees are often embodied in
bilateral investment treaties (BITs) or in the investment chapters of RIA
agreements. BITs typically contribute to investment by prohibiting cer-
tain policies rather than by requiring policies that actively encourage in-
vestment, but they may play an important role in facilitating investment
flows.
Far more positive in intent is the argument that RIAs add credibility to
government policies in general and thus help increase investment and at-
tract FDI. We argue that South-South RIAs are unlikely to do this and
may in fact hinder investment if they are not accompanied by liberaliza-
tion of trade with the rest of the world. North-South RIAs, on the other
hand, can enhance a southern country's credibility, but typically only if
the RIA is likely to enhance economic performance in its own right and if
the northern partner is willing to enforce investment-encouraging "club
rules." The latter is more likely to occur if the policies on which a devel-
oping country wants to gain credibility are specified explicitly in the
agreement and if the northern partner has an identified interest in the
southern partner's success-for example, in allevia'tion of pressures to
migrate. In general, this interest will be greater the closer geographically
are the northern and southern partners.
Some recent analysis holds that the rate of return on capital (and on
investment) could well rise in all members of an RIA regardless of their
capital abundance. For example., if tradables are more capital-intensive
than nontradables, then opening up boosts the demand for capital; lower
tariffs and trading costs for capital equipment might reduce the price of
investment goods; and creation of a more efficient financial sector could
reduce borrowing costs. Unfortunately, there are few empirical studies of



18    REGIONAL INTEGRATION AND DEVELOPMENT
the impact of RIAs on investment; most trade blocs are so new that the
data are simply not there. The limited evidence tends to suggest that
RIAs have mildly positive effects on investment, but there is no evidence
that this translates into increased economic growth. A possible reason is
that additional investment may be attracted into inefficient sectors which
benefit from the RIA's high external trade barriers. Firmer evidence is
available that RIAs boost FDI, especially inflows of investment from
nonmember countries.
The real key to investment is the general policy stance in areas such as
sound macroeconomic policies, well-defined property rights, and effi-
cient financial and banking sectors. Regional integration may foster in-
vestment if it significantly increases policy credibility and market size,
but it needs to be accompanied by good policy overall.
Growth and Location
If RIAs can stimulate investment, might they also have a beneficial effect
on growth and industrialization? Chapter 5 moves beyond investment to
examine the economics of endogenous growth and of agglomeration (or
clustering).
Modern growth theory-the theory of endogenous growth-
emphasizes the role of knowledge in fostering productivity and growth.
Knowledge can be effectively transferred from one country to another
through international contacts and trade. Wealthy countries are knowl-
edge rich and so are likely to provide far more access to technology than
can poorer trading partners. RIAs that switch imports from richer to
poorer sources are therefore likely to have a perverse effect on countries'
growth rates. The boost that RIAs can give to members' growth rates by
supporting institutional reform also seems likely to be stronger when
developing countries join with richer partners than with poorer ones.
Developing countries may, of course, increase their access to knowledge
and technology through unilateral trade liberalization that does not dis-
criminate against rich countries, and many institutional reforms can also
be achieved through the multilateral system.
The direct evidence on RIAs and growth is subject to methodological
reservations but is actually pretty consistent. There is some evidence that
North-South RIAs can stimulate growth in the southern partner, little ev-
idence that RIAs between developed countries stimulate growth, and no
evidence that RIAs between developing countries do so. Casual examina-
tion of the recent performances of, say, Mexico, Poland, and Portugal
and a formal analysis of NAFTA suggest that serious North-South inte-



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW     19
gration may foster the southern partners' growth, reinforcing our views
about the relative merits of the two types of partners for developing
countries. The most important message, however, is that the general pol-
icy stance of the southern partner is crucial for capturing long-term
gains. Broadly speaking, Portugal reformed on joining the EU and bene-
fited; Greece did not reform and did not benefit.
The same preference for North-South over South-South RIAs arises
from the discussion of agglomeration. Although economists have long
been aware that industry tends to cluster in particular locations, they
have only recently learned to model the phenomenon formally and
so begin identifying precisely which combinations of conditions must
be satisfied for clustering to occur. The theory arose from attempts to
understand the possible effects of the enlargement and deepening of
the EU (Krugman and Venables 1990) and so is directly applicable to
the case of RIAs. The theory is still very young; the models do not
yet appear to be very realistic and have to date not been accompanied
by much empirical evidence. Thus, they are more parables than
forecasts. They do, however, shed light on qualitative factors, and the
issues they address are so important for policymakers and publics alike
that we think it useful to explore their implications for developing
countries.
Creating an RIA is likely to affect the incentives for industry to ag-
glomerate. Usually, it will encourage clustering by increasing market size
and allowing more effective exploitation of the links between firms. An
RIA may attract industry into member countries at the expense of non-
members (although if the RIA is small, such effects will also be very
small). RIAs also frequently cause industry to relocate from one member
to another. For RIAs between poor countries, this seems likely to in-
crease intermember inequalities because under an RIA firms find it easier
to agglomerate in the more prosperous countries while still selling in the
other member countries. For RIAs involving richer members, the results
are less clear-cut, and it is possible that poorer members will experience
industrialization as a result of joining an RIA. When agglomeration ef-
fects are considered, integration with richer neighbors (in a North-South
RIA) looks far better for developi[ng countries than does South-South in-
tegration.
Integration of Domestic Policies
As trade barriers have declined, policymakers have come to a better un-
derstanding of the importance of domestic regulation for economic inte-



20    REGIONAL INTEGRATION AND DEVELOPMEN-T
gration. The issues surrounding integration of domestic policies are the
subject of chapter 6.
Cooperation on domestic policies can substantially increase the gains
from forming a trade bloc. It can lift barriers that insulate national mar-
kets for similar goods and services and deliver economic benefits many
times those available from mere trade agreements. Policy integration-
intergovernmental cooperation in designing and applying domestic poli-
cies on taxes, health and safety regulations, the environment, standards,
and so on-can increase competition in domestic markets by reducing
transactions costs and allowing new suppliers to enter markets. Cooper-
ation on domestic policies can also assist in overcoming market failures
and can help ensure that trade restrictions are not reimposed through the
back door.
Except for the EU, however, RIAs generally have only "shallow" pol-
icy integration aims. Their objective is not economic union (which re-
quires institution building of a type we term "deep integration") but an
increase in competition through elirnination of policy interventions and
reduction of market segmentation. A number of recent and pending
RIAs-for example, APEC, MERCOSUR, and the proposed Free Trade
Area of the Americas (FTAA)-are discussing an intermediate level of in-
tegration that entails close governmental cooperation to harmonize do-
mestic regulations and policies, but no supranational authority. Without
specific timetables for action and further negotiation, however, neither
new nor existing RIAs are likely to make great progress. Experience sug-
gests that such negotiated policy integration is very demanding, both
politically and technically. Moreover, RIAs are not the only game in
town; indeed, decisions by individual countries to adopt policies em-
ployed elsewhere and multilateral efforts to set international technical
and regulatory standards have been more common than regional efforts.
Developing countries can do much to achieve the benefits of policy inte-
gration unilaterally by adopting international standards and by recogniz-
ing the regulatory norms of their major markets, such as the EU and the
United States.
One puzzle is why governments combine policy integration and trade
integration in the same institution. The trade component of an RIA can
provide specie to help overcome opposition to institutional and domestic
policy reforms, and policy integration can assist the implementation and
enforcement of RIA trade policies. Such linkage does not, of course, ex-
cuse the selection of suboptimal trade reforms and domestic policies, but
if the reforms and policies are desirable in their own right, combining
them might be politically efficient.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW   21
Wherever possible, policy integration that reduces regulatory costs
should extend beyond RIA partners to nonmembers so that the increase
in competition is maximized and policy integration benefits all trading
partners. Formal intergovernmental agreements such as mutual recogni-
tion arrangements for product standards and testing may be necessary
for policy integration, but special efforts should be made to ensure that
these do not perpetuate or increase discrimination. Explicitly discrimina-
tory policy integration should be resisted.
Competition between regulatory regimes, coupled with mutual recog-
nition, can be a useful route to policy integration. Some element of har-
monization may be called for to prevent adverse spillovers-such as a
threat that the competitive relaxation of merger regulation will lead to a
"race to the bottom"-or to safeguard public health and safety. These
measures are generally best limited to minimum standards based on
global norms. Regional standards that diverge from global norms will be
optimal only if region-specific characteristics exist.
The WTO and other multilateral institutions can play a large role in
policy integration. The WTO agenda is as broad as-and often broader
than-that of most RIAs, and more can be done in the WTO context
than is often recognized by trade bloc proponents. It could help further
by adopting rules extending the most-favored-nation (MFN) principle to
policy integration initiatives such as customs clearance documentation
and procedures that do not require formal intergovernmental equivalen-
cy or recognition agreements.
Policy integration proposals are very specific and should be evaluated
on their individual merits. Care should be taken that they are for
the general good and that, other than as part of a necessary coalition
for achieving reform, efforts to link them to trade liberalization are
resisted. While integration of domestic policies may sometimes necessi-
tate formal agreements, there is no fundamental reason why it
should require trade preferences. The EU and the United States, for in-
stance, have drawn up a series of mutual recognition agreements
for sectoral product standards completely outside an RIA context. To
date, developing countries have been entirely excluded from such initia-
tives.
The issue of cooperation in the area of regional public goods such as
natural resources, infrastructure, and energy differs from regional inte-
gration and is not addressed in this book. Whether and under what cir-
cumstances regional integration can help in attaining a cooperative solu-
tion regarding regional public goods is examined in Schiff and Winters
(2002a).



22    REGIONAL INTEGRATION AND DEVELOPMENT
Regionalism as Politics
Countries often form trade blocs for noneconomic reasons, such as na-
tional security, peace, and help in developing political and social institu-
tions. These are public goods and so are unlikely to be efficiently provid-
ed in the absence of some form of intervention, such as an RIA. These
political objectives can be important for RIAs-sometimes overwhelm-
ingly so-but it is still desirable that they be achieved efficiently and that
policymakers pay heed to their economic costs. Chapter 7 examines
some of the political objectives of regionalism, discusses their economic
implications, and assesses whether trade preferences are necessary for at-
taining them.
Political benefits such as peace and security can sometimes swamp the
simple material considerations that usually determine economic policy.
And because such benefits are typically shared by only a limited number
of countries-usually, neighbors-it makes sense to pursue them on a re-
gional basis rather than multilaterally. These political concerns are thus
relatively more important in analyzing RIAs than for some other interna-
tional issues. We show that under some circumstances the formation of
an RIA may be an effective way to deal with security tensions between
neighboring countries; the argument is essentially that mutual trade fos-
ters peace between countries and that regionalism fosters trade. We show
that for an RIA designed to enhance security, the optimum external tariff
(on imports from nonmembers) declines over time and as integration
deepens.
Joining an RIA with large democratic countries can help a develop-
ing country achieve or uphold democracy if the RIA imposes "club
rules" such as democracy and civil rights on its members. The size and
location of the partners matters because larger partners are generally
able to impose greater costs on (withdraw greater benefits from) recal-
citrants than are smaller ones, and because a democratic partner coun-
try is likely to be far more concerned about possible spillovers (such as
migration) from events in a sizable and nearby developing country than
from elsewhere.
A concern for many RIA members is whether increased regional inte-
gration will weaken the nation-state. We argue that for small and even
medium-size countries, pooling sovereignty and undertaking collective
action can enhance the effectiveness of the state by helping solve
economic problems, by strengthening countries against third-country se-
curity threats, and by increasing international influence by lowering ne-
gotiation costs or increasing bargaining power in dealings with the rest



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW      23
of the world. But, as we argued above, cooperation of this kind does not
usually require trade preferences.
Regionalism and the Rest of the World
RIAs are, by nature, exclusive clubs. After all, every country in the world
is excluded from nearly every RIA in the world, and every RIA excludes
nearly every country. The discrimination by RIAs against excluded coun-
tries is real and, according to the latest evidence, can cause significant
trade diversion. Trade diversion is mainly a cost to the partners who pay
more for their imports, but chapter 8 shows that in two sets of circum-
stances it can be costly to the excluded countries that lose exports.
First, if exports generate supernormal profits, losing them is costly be-
cause the income lost exceeds the value of the resources freed up by not
having to produce the exports any more. This can happen, for example,
if the exporter is able to charge monopoly prices or (a more important
case) where exports are taxed, because the tax-inclusive price received
for the exports exceeds the cost of the resources used up in production.
Export taxes are rare, but trade theory teaches us that import taxes are
export taxes; it is the act of trading (turning exports into imports) that is
taxed, and it does not matter which side of the transaction formally faces
the tax. Thus, a country with significant tariffs will lose welfare if its ex-
ports fall exogenously. Another way of seeing this is that when exports
fall, imports must fall as well to maintain the trade balance, and the re-
sult is a loss in tariff revenue.
The second case in which trade diversion hurts exporters is when the
decline in demand forces down export prices-that is, worsens the ex-
porter's terms of trade. We present new empirical evidence of the signifi-
cance of this effect from Brazil's experience following the creation of
MERCOSUR.
Possibly even more important than the static losses suffered by ex-
cluded countries is the issue of whether RIAs are stepping stones toward
the ultimate goal of globally freer trade, or millstones around the neck of
progress toward this goal. The world of multiple trade blocs is still too
new to permit a definitive empirical answer, and economic theory is not
completely clear on the matter, but we see significant dangers that re-
gionalism could start to undermine the multilateral trading system. Cer-
tainly, regionalism hits at one of the cornerstones of multilateralism as a
framework for international affairs-equal treatment for all.
Most analyses of the effects of bloc formation on the tariffs imposed
by noncooperating governments suggest that tariffs tend to increase with



24    REGIONAL INTEGRNriON AND DEVELOPMENT
the spread of regionalism. It is asserted that in some cases regionalism
has brought other countries to the negotiating table to agree on new
rounds of multilateral trade liberalization; for example, the formation of
the EEC is said to have led to the Kennedy Round. We argue that the de-
sired outcome is far from certain and that using such coercive tactics to
get others to reduce their tariffs is extremely dangerous. One such act
may call forth another, giving rise to "domino regionalism," which seems
to lie behind much of the spread of regionalism during the 1980s and
1990s. Under such conditions, one cannot infer that because regionalism
is spreading, it is benign. If everyone else is in a gang, you may want to
belong to one yourself, but that does not make gangs a good thing.
One of the problems with the theory of domino regionalism is that al-
though enlarging an RIA might increase the incentives for new members
to join, it does not necessarily increase the incentives for existing mem-
bers to let new ones in. Particularly if RIAs discriminate against exclud-
ed countries, insiders will want to stop expansion well short of the whole
world; there is no point in being inside if there is no one outside to ex-
ploit. It is sometimes argued that an alternative would be to insist on
open access to all RIAs-any country that could adhere to the rules of an
RIA could join it and reap its benefits. In practice, given that accession
has to be negotiated (because the rules of nearly all RIAs entail more
than just tariff reductions), there is no operational way to insist on such
access. We find the concept of "open regionalism" to be more of a slogan
than an analytical tool since, depending on the definition used, it either
reduces to something else (such as multilateralism) or does not separate
"good" from "bad" RIAs.
In time, RIAs will obviously impinge on the trade negotiation process.
Some argue that if talks involved a few RIAs rather than many separate
countries, they would be simpler and quicker. Possibly so, but this ig-
nores the difficulties caused by having to agree positions within each
bloc before and during the negotiations. It is also suggested that RIAs
could serve to develop blueprints for technically complex issues before
they come to the global level or could provide a means of tackling politi-
cally difficult issues that cannot yet be agreed globally. Again, we do not
rule out the possibility of benefits through these routes, but they are far
from guaranteed.
Chapter 8 concludes with a discussion of how the WTO handles
RIAs. While far from perfect, the current rules for developed countries
are probably about as good as we can get, but they are currently very
poorly enforced. Those for developing countries are looser and make it
even easier to create welfare-reducing RIAs. We advocate unifying the



REGIONAI INTEGRATION AGREEMENTS: AN OVERVIEW     25
rules on the developed country model and enforcing them more actively.
The rules, however, cannot guarantee benign regionalism, and responsi-
bility for ensuring that RIAs do not harm their members falls squarely on
the governments involved. In this effort, governments would be aided by
much clearer analyses of the economic effects of regional blocs than the
WTO's current legalistic approach requires.
Rules of Thumb for Regionalism
Chapter 9 summarizes our views on what has been learned from the
study. As noted above, regionalism is too complex and sui generis to gen-
erate universal operational rules. We believe, however, that there are con-
sistent lessons from the analysis that apply in most circumstances. These
are collected as rules of thumb that are not inviolable but should not be
violated lightly. The rules are grouped into eight main messages:
*   Use RIAs as a way of fostering competition. If an RIA is neces-
sary, it should be used as a procompetitive instrument, with a
focus on incorporating provisions that will foster greater competi-
tion in domestic markets.
*   North-South dominates South-South. Not all partners are equal.
RIAs with high-income countries are more likely to generate sig-
nificant economic gains than are those with poorer ones.
*   Credibility gains require explicitness. RIAs can enhance the credi-
bility of economic and political reform programs, but generally
only if they explicitly include provisions and mechanisms that di-
rectly affect the policies of interest.
*   Only efficient RIAs are likely to help politically. RIAs can help
solve political problems, but if they are economically wasteful or
divisive, they could have opposite effects.
*   Regional cooperation does not generally require trade prefer-
ences. The existence of widespread intercountry spillovers calls
for cooperation between developing countries in areas other than
trade policy, such as regulatory reform and provision of infra-
structure. Usually, however, these goals should be pursued inde-
pendent of trade discrimination.
*   Beware of transactions costs in operating RIAs. Governments
should consider carefully the transactions and implementation
costs associated with different types of RIAs.
*   RIAs may have positive or negative fiscal implications. The fiscal
dimensions of RIAs are important for countries in which trade
taxes generate a significant share of government revenue.



26    REGIONAL INTEGRATION AND DEVELOPMENT
Do not rely on the WTO to ensure that RIAs are beneficial.
Countries should not rely on the WTO to ensure that RIAs are
beneficial to members and to outsiders. The WTO forbids some
destructive forms of regionalism, but its main contribution to-
ward constraining the potential negative implications of regional-
ism for nonmembers is as an instrument for pursuing global liber-
alization on an MFN basis.
APPENDIX. SELECTED REGIONAL INTEGRATION
AGREEMENTS WITH DEVELOPING COUNTRY MEMBERS
In this appendix, the name of the agreement is followed by the type of
agreement (common market, customs union, FTA, and so on) and by the
governing GATT/WTO provision. Article XXIV of the GATT sets out
the conditions under which the formation of FTAs or customs unions is
permitted. The GATT Enabling Clause, approved in 1971, allows mem-
bers to "accord differential and more favorable treatment to developing
countries without according such treatment to other contracting par-
ties." The years are the dates of notification to the GATT/WTO or, if not
notified, of founding.
RIAs between Developed and Developing Economies
European Union (EU), common market, Article XXIV; formerly Euro-
pean Economic Community (EEC), European Community (EC). 1957,
Belgium, France, Germany, Italy, Luxembourg, Netherlands; 1973, Den-
mark, Ireland, United Kingdom; 1981, Greece; 1986, Portugal, Spain;
1995, Austria, Finland, Sweden.
European Economic Area (EEA), FTA, Article XXIV. 1994, EU, Iceland,
Liechtenstein, Norway.
Euro-Mediterranean Economic Area (Euro-Maghreb), FTAs, Article
XXIV. Bilateral agreements: 1995, EU and Tunisia; 1996, EU and Mo-
rocco.
EU bilateral agreements with Eastern Europe, FTAs, Article XXIV. 1994,
with Hungary, Poland; 1995, with Bulgaria, Czech Republic, Estonia,
Latvia, Lithuania, Romania, Slovak Republic, Slovenia.
North American Free Trade Agreement (NAFTA), FTA, Article XXIV;
extension of 1989 Canada-United States Free Trade Agreement (CUSF-
TA), Article XXIV. 1994, Canada, Mexico, United States.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW  27
Asia-Pacific Economic Cooperation (APEC), regional but not preferen-
tial agreement; not notified to the WTO. 1989, Australia, Brunei Darus-
salam, Canada, Indonesia, Japan, Republic of Korea, Malaysia, New
Zealand, Philippines, Singapore, Thailand, United States; 1991, China,
Hong Kong (China), Taiwan (China); 1993, Mexico, Papua New
Guinea; 1994, Chile; 1998, Peru, Russian Federation, Vietnam.
Latin America and the Caribbean
Andean Pact, customs union, Enabling Clause. 1969 (revived in 1991),
Bolivia, Colombia, Ecuador, Peru, Venezuela.
Central American Common Market (CACM), customs union, Article
XXIV. 1960 (revived in 1993), El Salvador, Guatemala, Honduras,
Nicaragua; 1962, Costa Rica.
Common Market of the South/Mercado Comun del Sur (MERCOSUR),
customs union, Enabling Clause. 1991, Argentina, Brazil, Paraguay,
Uruguay.
Group of Three (G3), FTA, Enabling Clause. 1995, Colombia, Mexico,
Venezuela.
Latin American Integration Association (LAIA), Enabling Clause; for-
merly Latin American Free Trade Area (LAFTA), Article XXIV, 1960.
Revived as LAIA, 1980, Argentina, Bolivia, Brazil, Chile, Colombia,
Ecuador, Mexico, Paraguay, Peru, Uruguay, Venezuela.
Caribbean Community and Common Market (CARICOM), customs
union, Article XXIV. 1973, Antigua and Barbuda, Barbados, Jamaica,
St. Kitts and Nevis, Trinidad and Tobago; 1974, Belize, Dominica,
Grenada, Montserrat, St. Lucia, St. Vincent and the Grenadines; 1983,
The Bahamas (part of the Caribbean Community but not of the Com-
mon Market).
Africa
Cross-Border Initiative (CBI), common policy framework, with the sup-
port of the International Monetary Fund (IMF), the World Bank, the EU,
and the African Development Bank; not notified to the WTO. 1992, Bu-
rundi, Comoros, Kenya, Madagascar, Malawi, Mauritius, Namibia,
Rwanda, Seychelles, Swaziland, Tanzania, Uganda, Zambia, Zimbabwe.
East African Cooperation (EAC), other, Enabling Clause; formerly East
African Community, 1967, broke up in 1977. Revived 1996, Kenya,
Tanzania, Uganda.



28    REGIONAL INTEGRATION AND DEVI'LOPMENT
Economic and Monetary Community of Central Africa/Communaute
economique et monetaire d'Afrique Centrale (CEMAC), customs union,
Enabling Clause, 1999. Formerly Union douaniere et economique de
l'Afrique Centrale (UDEAC);1966, Cameroon, Central African Repub-
lic, Chad, Republic of Congo, Gabon; 1989, Equatorial Guinea.
Economic Community of West African States (ECOWAS), FTA, En-
abling Clause. 1975, Benin, Burkina Faso, Cape Verde, C6te d'lvoire,
The Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Mauritania,
Niger, Nigeria, Senegal, Sierra Leone, Togo.
Common Market for Eastern and Southern Africa (COMESA), FTA, En-
abling Clause. 1993, Angola, Burundi, Comoros, Djiboutt, Arab Repub-
lic of Egypt, Ethiopia, Kenya, Lesotho, Malawi, Mauritius, Mozam-
bique, Rwanda, Somalia, Sudan, Swaziland, Tanzania, Uganda, Zambia,
Zimbabwe.
Indian Ocean Commission (IOC), integrated regional program for devel-
opment of trade; not notified to the WTO. 1982, Comoros, Madagascar,
Mauritius, Seychelles.
Southern African Development Community (SADC), FTA, Enabling
Clause; formerly known as the Southern African Development Coordi-
nation Conference (SADCC). 1980, Angola, Botswana, Lesotho,
Malawi, Mozambique, Swaziland, Tanzania, Zambia, Zimbabwe; 1990,
Namibia; 1994, South Africa; 1995, Mauritius; 1998, Democratic Re-
public of the Congo, Seychelles.
West African Economic and Monetary Union/Union economique et
monetaire ouest-africaine (UEMOA), customs union, Enabling Clause;
formerly Economic Community of West Africa (CEAO), 1973. 1994,
Benin, Burkina Faso, C6te d'lvoire, Mali, Niger, Senegal, Togo; 1997,
Guinea-Bissau.
Southern African Customs Union (SACU), customs union. 1910,
Botswana, Lesotho, Namibia, South Africa, Swaziland.
Economic Community of the Countries of the Great Lakes/Commu-
naut6 economique des pays des grands lacs (CEPGL), to promote region-
al economic cooperation and integration; not notified to the WTO.
1976, Burundi, Democratic Republic of the Congo, Rwanda.
Europe
Central European Free Trade Area (CEFTA). 1993. 1996, Czech Repub-
lic, Hungary, Poland, Slovak Republic, Slovenia; 1997, Romania; 1999,
Bulgaria.



REGIONAL INTEGRATION AGREEMENTS: AN OVERVIEW     29
Committee for Mutual Econornic Assistance (CMEA, also known as
COMECON). 1949, Bulgaria, Czechoslovakia, Hungary, Poland, Ro-
mania, Soviet Union; 1949, Albania; 1950, Democratic Republic of Ger-
many; 1962, Mongolia; 1972, Cuba; 1978, Vietnam. China attended
meetings as an observer between the late 1950s and 1961. The Federal
Republic of Yugoslavia negotiated a form of associate status in 1964.
Bilateral agreements between individual CEFTA members and individual
Baltic countries.
Middle East and Asia
Association of Southeast Asian Nations (ASEAN), 1967. 1977, ASEAN
Preferential Trading Arrangement, Enabling Clause; 1992, ASEAN Free
Trade Area (AFTA), Enabling Clause, Indonesia, Malaysia, Philippines,
Singapore, Thailand; 1994, Brunei Darussalam; 1995, Vietnam; 1997,
Lao People's Democratic Republic, Myanmar; 1999, Cambodia.
Arab Common Market, long-term aim is customs union, Article XXIV.
1964, Agreement for Economic Unity among Arab League States.
Gulf Cooperation Council (GCC), other, Enabling Clause. 1981,
Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, United Arab Emirates.
South Asian Association for Regional Cooperation (SAARC) Preferential
Trading Arrangement (SAPTA), preferential trading agreement, Enabling
Clause. 1995, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, Sri
Lanka.
NOTES
1. The General Agreement on Tariffs and Trade (GATT) was succeeded in
1994 by the World Trade Organization (WTO).
2. "Closed" regionalism was widespread, especially in Sub-Saharan Africa
and Latin America. Examples in Airica included the CEAO (created in 1973),
ECOWAS (1975), the CEPGL (1976), and the EAC (1967). In Latin America
they included the CACM (1960), LAFTA (1960), the Andean Pact (1969), and
CARICOM (1973). See the Abbreviations and Acronyms list for full names and
the appendix to this chapter for further information on these agreements.
3. These North-South agreements are to be distinguished from the preferential
trade agreements implied by colonial preferences.






CHAPTER 2
HowTrade Blocs Increase Trade and
Competition
A       well-crafted trade bloc can raise efficiency-and economic wel-
fare-in its member countries by facilitating consumer choice
j    Land increasing the competition that producers face. Dropping
tariff barriers enlarges markets and gives more efficient producers entry
into countries where their prices had been inflated by duties and other
trade barriers. But trade blocs can easily end up adding, rather than re-
moving, distortions to trade and efficiency. This chapter looks at the
basic structure and economics of free trade agreements (FTAs), whose
members do away with tariffs between themselves, and customs unions,
which, in addition, maintain a common external tariff against nonmem-
bers.
A trade bloc usually increases trade between its members. An impor-
tant issue, however, is whether it "creates" trade (by allowing cheaper
products from other bloc members to substitute for more expensive do-
mestic production) or "diverts" it (by substituting intrabloc imports for
imports from outside the group that were cheaper when both faced equal
tariffs).' Although the evidence on the balance between trade creation
and diversion in trade blocs is mixed, we offer some new research find-
ings showing that diversion must: be recognized as a serious possibility.
The balance between trade creation and diversion is an important de-
terminant of the overall benefits of an RIA. It is, however, based on a
view of the world in which intercountry trade is driven entirely by differ-
ences in productivity and factor endowments. But in fact, trade can also
arise from product differentiation and from economies of scale that re-
duce costs as production grows. In these circumstances competition be-
31



32      REGIONAL INTEGRATION AND DEVELOPMENT
tween firms is weakened, and consumers lose. International trade then
offers an important means of increasing competition by allowing new
suppliers to enter markets. RIAs can generate such benefits by fostering
trade between members, although there is not yet sufficient empirical ev-
idence to show that developing countries will be able to reap large gains
from this source.
INCREASED TRADE BETWEEN MEMBERS OF TRADE BLOCS
Reducing trade barriers between two countries is likely to increase their
trade with each other. Indeed, in many cases, as in CEFTA and the Arab
Free Trade Agreement, this is an explicit objective in the treaties estab-
lishing the RIAs. Figure 2.1 reports simple "before" (one year before im-
plementation) and "after" (five years after implementation) import sta-
tistics to see what happened as nine RIAs between developing countries
came into operation. In seven of the nine RIAs, intrabloc trade shares in-
creased, some very strongly. The exceptions, CARICOM and the GCC,
were among the slowest to implement internal tariff cuts. Overall, the in-
trabloc share increased by nearly half.
These results appear to be cause for celebration, since strong interna-
tional trade performance is now generally accepted as one of the main
Figure 2.1 Share of Imports from Partners in Selected RIAs, One Year before and Five
Years after Implementation
25-
20 -                                                      I Before
15   |                                                       After
10 
a.
5-
Not For    K         tR,t
Note For MERCOSUR, "before' is 1991 and "after' Is 1996; for Andean Pact 1, 1968 and 1974; for Andean
Pact 11, 1990 and 1996, for CACM II, 1990 and 1996; for CARICOM, 1972 and 1978; for the CEAO, 1965
and 1971; for UDEAC, 1965 and 1971; for AFTA, 1991 and 1996, and for the GCC, 1980 and 1986. Andean
Pact I refers to the original 1969 agreement and Andean Pact II to the revived agreement (1991); CACM II
refers to the Central American Common Market (revived). Other abbreviations used are MERCOSUR, Com-
mon Market of the South; CARICOM, Caribbean Community and Common Market; CEAO, Economic Com-
munity of West Africa; UDEAC, Union douaniere et economique de l'Afrique Centrale (now CEMAC); AFTA,
ASEAN Free Trade Area, GCC, Gulf Cooperation Council
Source: United Nations COMTRADE



How TRADE BLOCS INCREASE TRADE AND COMPETITION     33
determinants of economic prosperity. But before breaking out the cham-
pagne, we need to ask whether every increase in trade is desirable and
whether these increases are actually attributable to regionalism.2 To an-
swer these questions, we need both economic theory and a closer look at
the trade numbers.
IS MORE TRADE GOOD, OR BAD? TRADE CREATION AND
TRADE DIVERSION
The classical reason for gains from trade is that global free trade allows
consumers and firms to purchase from the cheapest source of supply, en-
suring that production is located according to comparative advantage. In
contrast, trade barriers discriminate against foreign producers in favor of
domestic suppliers. Domestic import-competing producers are induced
to expand even though their costs are higher than the cost of imports.
This misallocation starves domestic export sectors of resources, raises
their costs, and causes these sectors to be smaller than they otherwise
would be. Switching production from goods that a country can produce
efficiently to those it cannot reduces real incomes.
Since an RIA liberalizes trade, lowering at least some of the barriers,
does it not follow that it too will generate gains from trade? Unfortu-
nately, the answer is, not necessarily. The "gains from trade" argument
tells us what happens if all trade barriers are reduced, but it need not
apply to a partial and discriminatory reduction in barriers, as in an RIA.
The reason is that discrimination between sources of supply is merely
shifted, not eliminated. If partner country production displaces higher-
cost domestic production, there will be gains-trade creation. But it is
also possible that partner country production may displace lower-cost
imports from the rest of the world, resulting in trade diversion.
When partner supplies-now facing no tariffs and therefore
cheaper-displace sales of local output, the two classical sources of gains
from trade apply: real resources are saved by shifting production in the
direction of comparative advantage, and consumers benefit from facing
lower (undistorted) prices. Such trade creation unambiguously increases
economic welfare in the importing country. But RIAs can also divert
trade by allowing imports from partners to displace goods from outside
the bloc that were cheaper when both faced equal tariffs. Diverting trade
from cheaper to more expensive suppliers means that more resources are
used up to purchase the same output, which is clearly costly. Under some
circumstances, trade diversion also reduces consumer prices because con-
sumers no longer pay tariffs on their purchases; the benefits that this



34    REGIONAL INTEGRATION AND DEVELOPMENT
generates help to offset the additional cost of imports-conceivably, even
completely.
To see the effects of discrimination, it is helpful to think through a
simple example. Suppose that a country can import a good from a po-
tential partner country at $105 per unit and from the rest of the world at
$100 and that each pays $10 in duty, making the prices paid by con-
sumers $115 and $110, respectively. In this situation consumers obvious-
ly purchase from the rest of the world and pay $110. If the country joins
an RIA with the partner, the partner's imports come in duty free. The
price consumers pay for imports from the partner country then falls to
$105, while imports from the rest of the world still cost $110. The con-
sumers' response is obvious: they switch to the partner country, buying
the $105 good and saving $5. But because the government now loses $10
per unit (the revenue it was getting on each unit of imports from the rest
of the world), the net effect for the country is a loss of $5, and the RIA
has reduced real income. Another way of putting it is that the country
(not the consumers) used to pay $100 per imported unit and now pays
$105. This is the deleterious welfare effect of trade diversion.
This is just an example, and such circumstances clearly will not apply
in all sectors; there are sectors in which partner country costs are less
than those in the rest of the world and others in which the country under
study is an exporter. The next question is how to identify the circum-
stances in which trade diversion is more, or less, likely to be a problem.
First, notice that trade diversion can occur only if the country has a
tariff on imports from the rest of the world. The cost of trade diversion
cannot exceed the height of this external tariff. In our example, if the ex-
ternal tariff is initially low, there is not much tariff revenue to be lost,
and reducing the external tariff would not induce a switch in source of
supply. One clear policy implication is that the traditional advice to
countries to lower external tariffs applies as much to those inside RIAs as
to those outside.
Second, trade diversion arises only if partner country costs are out of
line with costs and prices in the rest of the world, and this will not be the
case if the partner itself has low trade barriers. For example, if the part-
ner had a duty of just $2 per unit, prices and costs in the partner country
could not exceed $102 (the price at which imports from the rest of the
world would be sold in the partner country). Preferential liberalization
would then cost the government $10 but would save consumers in the
importing country $8, creating a net loss of just $2 per unit and mitigat-
ing the cost of trade diversion. But, as discussed in the next chapter,
countries with high trade barriers usually impose rules of origin to pre-



How TRADE BLOCS INICREASE TRADE AND COMPETITION     35
vent trade deflection (reexport of imported goods from the low-tariff to
the high-tariff country), and prices could still differ by more than $2.
Third, our example assumes a rather artificial and "frictionless"
trade. In reality, products from different countries are not perfect substi-
tutes, and trade faces transport costs and other barriers apart from tar-
iffs. These factors will tend to rnake the change in the sourcing of im-
ports less sharp than in the example, again mitigating the costs of trade
diversion but also reducing the benefits of trade creation.
Finally, note that an RIA between two small developing countries is
likely to generate only trade diversion and no trade creation. This can be
seen most clearly in the case of homogeneous goods. Since small coun-
tries will typically not be able to supply all of their partners' needs for
imports, each member will continue to import some quantity of most
goods from the rest of the world. For these goods, domestic consumer
prices will continue to be fixed at the world price plus the import tariff.
Since neither of these changes with integration, consumption does not
change. Production, however, increases because each country can now
sell to the partner without paying the tariff. Thus, each member country
replaces cheaper imports from the rest of the world with more expensive
partner imports. The outcome is trade diversion and a loss for both
countries (Schiff 1997).
The welfare effect of an RIA on the bloc members as a group depends
on the balance between trade creation and diversion. Real resources are
saved if inefficient production is cut through trade creation but are lost if
imports are switched from low-cost to high-cost partner sources through
trade diversion. At the level of the individual country, how the RIA shifts
economic welfare between members is also important. It is therefore nec-
essary to consider effects felt via exports, as well as those stemming from
imports and tariff revenue.
The preceding discussion of trade diversion focused on imports, but of
course, at the same time as the home country is losing through trade diver-
sion, the partner country is increasing its exports. Is not the former's loss
just balanced by the latter's gain, so that, allowing for some trade creation,
the RIA as a whole is better off? The answer is that there may be an ex-
porter country gain, but it is less, per unit, than the importer country loss.
Recall that in our example consumers switch to imports from the
partner country. Partner country export sales expand, but how much of
a gain is this for the partner country? If exports are just selling at cost
($105 in the example), selling more units does not raise income in the
partner country (because some other sector must contract to release re-
sources for the expansion). If they are selling above cost, there will be a



36    REGIONAL INTEGRATION AND DEVELOPMENT
real income gain. But how much higher than cost can the price go? The
answer is that the price cannot go above $110 (if it did, consumers
would switch back to buying from the rest of the world), so the exporter
country gain per unit cannot exceed the gap between the price of imports
from the rest of the world and costs ($110 - $105 = a maximum of $5).
This line of reasoning suggests another way of thinking about trade di-
version. To return to our example (for the last time), the government has
given up $10 of tariff revenue per unit. We can see where this has gone; $5
per unit goes to the higher cost of producing partner country imports com-
pared with the cost of imports from the rest of the world, and the other $5
is divided between domestic consumers and partner country firms. The
shares depend on whether these firms are able to raise their prices in re-
sponse to having preferential access to the domestic market. It is often ar-
gued that an advantage of an RIA over unilateral liberalization is that
firms benefit from preferential access to partner markets. This is true, but
we now see that the gain comes only at the expense of consumers and gov-
ernment revenue. The RIA acts as an inefficient way of transferring some
of the country's tariff revenue either to domestic consumers or to partner
country producers. These transfers can be very large, depending as they do
on the total level of trade in the affected commodities, not just on the
amounts created or diverted by the RIA. Thus, they figure prominently in
the incentives that different groups have to lobby for RIAs (see chapter 3).
Transfers can also matter greatly in North-South RIAs because devel-
oping countries risk losing from an RIA with the North. First, develop-
ing countries typically have higher imports from their partners than ex-
ports to them and so the flows on which they lose revenue are larger than
those on which they gain. Second, they almost always have higher tariffs
than developed countries, and developed country firms are consequently
likely to gain more from increased access to the developing country's
market than vice versa. The developing countries can solve this problem
quite simply, however, by lowering their tariffs unilaterally.
GROWTH OF TRADE OVER TIME
The problem with looking just at the changes in trade shares in figure 2.1
is that they do not distinguish trade creation from trade diversion, both
of which increase the partner share. To separate these two phenomena,
we can exarnine shares in apparent consumption-the sum of all expen-
ditures in the domestic economy. At an aggregate level, shares in appar-
ent consumption are often approximated by corresponding changes in



How TRADE BLOCS INCREASE TRADE AND COMPETITION             37
shares in gross domestic product (GDP). If the share of imports in GDP
increases, the economy has become more open and there has been net
trade creation. If the share of nonpartner imports in GDP falls, trade has
been diverted. Figures 2.2 and 2.3 plot the changes in partner and non-
partner imports in GDP for our set of RIAs. Changes in overall openness
are the sums of these two figures and are given in Appendix table 2A.1.
The strongest increase in openness associated with an RIA is for
MERCOSUR, where imports more than doubled between 1990 and
1996, from 3.9 to 8.1 percent of GDP. CARICOM, the CEAO, and the
GCC also performed strongly. Over the periods examined, these blocs
display strong trade creation, albeit often from very low initial levels of
trade. Their shares of nonpartner imports in GDP also show           the
strongest increases, the very opposite of trade diversion. If this were all
attributable to regionalism-ancd if this outcome were all that there were
to regionalism-figures 2.2 and 2.3 would indeed be an endorsement of
this approach to trade policy.
This is, however, not the whole story: we need to know whether the
changes would have occurred anyway. That is, we need to identify what
would have happened in the absence of the RIA-what students of the
economics of integration call the anti-monde, or the counterfactual.
First, a simple numerical adjustment to the trade shares is necessary.
Countries that are growing faster will automatically tend to absorb more
Figure 2.2 Intrabloc Imports as a Share of GDP, Selected RIAs, OneYear before and Five
Years after Implementation
10
9                                            __ _ _ _
8-
7-
8 5-                                                      1 3Before
5-
(   4                                                    * After
a. 
2-
0  -  
~~~~
&i0   <      0q i           ol - ,C  0  00 
Note: For MERCOSUR, 'before' is 1991 and 'after" is 1996; for Andean Pact 1, 1968 and 1974; for Andean
Pact 11, 1990 and 1996; for CACM 11, 1990 and 1996; for CARICOM, 1972 and 1978, forthe CEAO, 1965
and 1971; for UDEAC, 1965 and 1971; for AFTA, 1991 and 1996; and for the GCC, 1980 and 1986. For the
full names of the RIAs, see the note to figure 2.1.
Source: World Bank data.



38     REGIONAL INTEGRATION AND DEVELOPMENT
Figure 2.3 Extrabloc Imports as a Share of GDP, Selected RlAs, OneYear before and Five
Years after Implementation
100-
80- 
o  60-                                       * Before
a.   40  _   l                                      After
C) ~~~  C," '~0 CCC
Note: For MERCOSUR, 'before' is 1991 and 'aftee is 1996, for Andean Pact I, 1968 and 1974, for Andean
Pact 11, 1990 and 1996; for CACM II, 1990 and 1996; for CARICOM, 1972 and 1978; for the CEAO, 1965
and 1971, for UDEAC, 1965 and 1971, for AFTA, 1991 and 1996, and for the GCC, 1980 and 1986. For the
full names of the RIAs, see the note to figure 2.1.
Source: World Bank data
of everyone's trade. Trade intensity indices allow for this by comparing
country A's share of B's total imports with A's share of all other countries'
imports.3 If A's share is higher in B than elsewhere, B's imports are biased
toward A. Such bias might exist for any of several reasons-for example,
commodity composition or geographic proximity. But because factors of
this sort tend to evolve relatively slowly, if sharp changes in trade inten-
sity are observed to accompany the creation of an RIA, we might infer a
causal relationship. As reported in appendix table 2A.1, intrabloc trade
intensities have increased in six of our nine RIAs, and extrabloc trade in-
tensities have declined in seven of the nine.
Next, we have to make allowance for domestic sales, using trade
propensity indices, which compare the share of imports from country A
in B's GDP with A's share of world trade. Trade propensity is the product
of trade intensity and openness (the ratio of B's imports to GDP) and so
reflects both the biases in trade patterns and the effects of RIAs on over-
all trade volumes.
Figure 2.4 shows a strong increase in the mean intrabloc import
propensity of our sample RIAs-120 percent over the six-year periods-
with notable increases in MERCOSUR, the GCC, and UDEAC. There
are declines for three blocs: CACM 11, CARICOM, and AFTA. For
CARICOM, which has liberalized its internal trade very cautiously, in-
trabloc trade did not increase strongly; for the other two, the strong in-
creases in their members' shares of world trade make the increases in in-
trabloc trade look less impressive.



How TRADE BLOCS INCREASE TRADE AND COMPETITION                       39
Figure 2.4 Intrabloc Import Propensities, Selected RIAs, OneYear before and FiveYears
after Implementation
100 -
0        0 
0       10__                      _
0,
&      0.1
0.
CL~~~~~~~~~~~~~~~~C
Note: For MERCOSUR, 'before" is 1991 and 'after' is 1996, for Andean Pact 1, 1968 and 1974; for Andean
Pact It, 1990 and 1996; for CACM II, 1990 and 1996; for CARICOM, 1972 and 1978; for the CEAO, 1965
and 1971; for UDEAC, 1965 and 1971; for AFTA, 1991 and 1996; and for the GCC, 1980 and 1986. For the
full names of the RlAs, see the note to figure 2.1.
Source: World Bank data.
The extrabloc import propensities (figure 2.5) all show modest in-
creases, with a mean increase of 30 percent. That is, in each of our sam-
ple RIAs the first five-year period is associated with an increase in the
share of GDP spent on imports from nonmembers that is larger than the
increase in nonmember shares of world trade. Again, this suggests a
good deal of progress toward opening members' economies and indicates
that regionalism has been associated with net improvements in economic
management.
Figure 2.5 Extrabloc Import Propensities, Selected RlAs, OneYear before and FiveYears
after Implementation
X 01g                                                           After
t      Ci       ;   Ci
Note: For MERCOSUR, "before" is 1991 and "after" is 1996; for Andean Pact 1, 1968 and 1974, for Andean
Pact II, 1990 and 1996; for CACM II, 1990 and 1996, for CARICOM, 1972 and 1978; for the CEAO, 1965
and 1971; for UDEAC, 1965 and 1971; for AFTA, 1991 and 1996; and for the GCC, 1980 and 1986. For the
full names of the RIAs, see the note to figure 2.1.
Source: World Bank data.



40    REGIONAL INTEGRATION AND DEVELOPMENT
NOT ALL CHANGES COME FROM REGIONALISM
Although the trade changes just discussed were undoubtedly of consider-
able benefit to the economies concerned, the no-change anti-monde
(counterfactual) calculations are far too simple to allow us to conclude
immediately that regionalism was working. Specifically, we must ask
whether the changes are due more to the nondiscriminatory trade liber-
alizations that many of the members undertook around the same time
than to regionalism per se.4 To answer this question, we need a more so-
phisticated way of representing the anti-monde. Members of RIAs may
improve their trade policies and performance over time (relative to their
starting points), but that can be an extremely undemanding benchmark.
Once we recognize the possibility of undertaking nondiscriminatory lib-
eralizations, we see that RIAs can still entail large distortions relative to
trade patterns elsewhere in the world or relative to the members' own
potential in a nondiscriminatory regime.
The most popular of these more sophisticated approaches, the gravity
model, explains trade between any two countries as a function of their
GDPs (richer economies both export and import more); their popula-
tions (larger economies depend proportionately less on trade than small-
er ones); the distance between them (as a proxy for transport costs, cul-
tural similarity, and business contacts); and physical factors such as
sharing a land border or being landlocked or an island. To these factors,
economists add variables to represent the additional trade that occurs if
both countries are members of a given RIA and, sometimes, the ways in
which RIA members' imports from and exports to the rest of the world
differ from the trade patterns of countries with no effective regional ties.
Gravity models are estimated on the basis of data from many countries.
They assume that in the absence of an RIA, member trade bears the same
relationship to GDP, population, distance, and so on as does the "arm's-
length" trade of the sample countries. This clearly depends on the set of
countries and years used to estimate the model. Estimates of trade creation
and diversion can be derived for a single year. As with trade propensities,
however, pairs of countries may trade heavily for extraneous reasons-cul-
tural similarities, compatible commodity compositions, and the like-and
so more convincing results are derived by looking not at snapshots but at
changes over the period during which the RIAs are being formed.
One of the first gravity models for RIAs (Aitken 1973) examined the
effect of the EEC and the European Free Trade Association (EFTA) on
European trade. Starting in 1951 (to obtain a picture of pre-EEC forces
at work), he estimated gravity models for a number of years and investi-



How TRADE BLOCS INCREASE TRADE AND COMPETITION    41
gated how the coefficients on intra-EEC and intra-EFTA trade evolved.
Both the EEC and EFTA substantially increased their intrabloc trade as
they reduced internal barriers, starting in 1961 and 1964, respectively.
By 1967, Aitken estimated, intra-EEC trade was 80 percent higher than
it would otherwise have been. and intra-EFTA trade was 50 percent
higher. Bayoumi and Eichengreen (1997) found similar results for
1952-92: the effects on intrabloc trade were concentrated in the early
years of EFTA and the EEC, and whereas EFTA was predominantly trade
creating, the EEC displayed both trade creation and diversion.
In the first gravity study of developing country RIAs, Aitken and
Obutelewicz (1976) found that the 1959-71 association agreements be-
tween 18 African countries ancd the EEC significantly increased mutual
trade, with the effect increasing progressively through the period of asso-
ciation.
Frankel, Stein, and Wei (1997) examined intra- and extrabloc trade
for eight RIAs-the EC, EFTA, NAFTA, MERCOSUR, the Andean Pact,
AFTA, the CER (the Closer Economic Relations agreement between Aus-
tralia and New Zealand), and the East Asian Economic Caucus-for the
period 1970-92. Their results for individual years show that increases in
intrabloc trade are generally accompanied by significant drops in trade
with the rest of the world (trade diversion).5
Finally, to explore the "new wave" of regionalism, a recent World
Bank study estimates trade effects for every year during the period
1980-96 (Soloaga and Winters 2001). It explicitly tests whether the
changes in the effects are statistically significant and extends the gravity
approach by looking for three separate effects on the trade of each RIA:
the effects on intrabloc trade, on extrabloc imports, and, uniquely, on ex-
trabloc exports.6 It considers nine major blocs, comparing members' per-
formance both between blocs and with the performance of 17 countries
that were not members of significant arrangements at the time.
They find that in Europe, where further integration was not accompa-
nied by the major liberalization of external trade, trade diversion is evi-
dent in the form of increased intrabloc trade and reduced extrabloc trade
(relative to expected values). In RIAs between developing countries, on
the other hand, the evolution of trade over the period appears to have
been dominated by external liberalizations. All trade increased over the
period (relative to expectations), but the increases in intratrade were sta-
tistically no larger than those in extratrade.
Figure 2.6 summarizes the results for the EU for the period. During those
years, the EU experienced significant deepening (the Single Market Pro-
gramme [SMP]) and enlargement (to Portugal and Spain), and so we might



42    REGIONAL INTEGRATION AND DEVELOPMENT
expect to find integration effects. In 1980 the EU shows unusually strong
trade with nonmember countries, which is not unusual for gravity models
based on large samples of countries. Its substantially lower-than-expected
trade between members is also not unusual, but the effect looks particularly
large because of the way Soloaga and Winters measure intrabloc effects.
They assume that in the absence of an RIA, trade between members would
be subject to the same forces as extrabloc imports and exports, and so to
capture the effects of the RIA they subtract from the "gross" intrabloc trade
coefficients the two extrabloc trade coefficients. Thus, in 1980 EU extrabloc
imports were higher than gravity model predictions by 1.86, and extrabloc
exports by 0.55.7 Intra-EU trade was higher than gravity model predictions
by 0.63, but in the absence of the RIA, trade between members would pre-
sumably have been subject to the same forces as other (extrabloc) trade; that
is, imports and exports would have been higher, by 1.86 and 0.55, respec-
tively. When we allow for these general effects, trade between EU members
is seen to be lower than it would have been expected to be given those coun-
tries' general trading behavior by 1.78 (= 0.63 - 1.86 - 0.55).
As noted above, the best estimate of the effects of integration is the
change in the intrabloc and extrabloc trade factors as integration occurs.
We therefore need to look at the evolution of the effects, as shown in fig-
ure 2.6. Clearly, as integration deepens and Portugal and Spain enter the
EU, the trade effects decline absolutely: that is, intrabloc trade grows,
and extrabloc trade falls relative to expected values. This is consistent
with the existence of trade diversion, as well as possible trade creation.
It is interesting to ask whether these changes are statistically signifi-
cant or could have arisen by chance. To answer this, Soloaga and Win-
ters conducted a separate exercise pooling observations for three periods,
1980-82, 1986-88, and 1995-96. These broadly capture the before and
after movements for the new and revived RIAs of the "new wave."
Figure 2.7 summarizes the results for all nine RIAs considered. The in-
dividual graphs are conceptually the same as in figure 2.6 except that
they contain estimates for only three periods rather than all 17 years. As
in previous studies, all the initial (1980-82) non-European intrabloc
trade effects are positive, suggesting that even before the blocs were
formed or revived, members traded disproportionately with each other.
Subsequently, although the trade effects evolved as RIAs were introduced
or revived, none of the changes was statistically significant. That is, after
allowing for changes in members' trade policy in general, intrabloc trade
has not changed significantly over the period of new-wave regionalism.
There was, however, some evidence of effects on overall trade. The EU's
and EFTA's propensities to import were significantly lower relative to ex-



How TRADE BLOCS INCREASE TRADE AND COMPETITION                43
Figure 2.6 EU Trade, 1980-96
2.5
2.0                          -           Single Market Act:
1.                                   Portugal and Spain join EC
W 10_*a.5
1.0
C 0.5
E
-0.5
2
r  -1  5  -
-2.0
-2.5 -
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
Note: The taxis shows values of coefficients on RIA dummy variables, in logarithmic equations. A coeffi-
cient of bimplies that the RIA raised trade by ([exp(b) - 1] ^ 100}% above the'expected'value.
Source: Soloaga and Winters (2001).
pected values in 1995-96 than in 1980-82, which is consistent with trade
diversion, while the four Latin American RIAs showed positive trends in
overall imports; these trends were statistically significantly for the CACM
and MERCOSUR. The EU and EFTA appeared to suffer significant declines
in their overall exports. NAFTA shows the same patterns as the European
blocs, but its changes in trade patterns are not statistically significant.
The conclusion from this work is that regionalism has had a smaller ef-
fect on developing countries' recent trade flows than the nondiscriminatory
liberalizations that these countries undertook. That is, as Latin American
countries, in particular, liberalized their trade in general over the 1990s, we
would have expected trade between them to have increased strongly any-
way, with stimuli coming as both the exporter and the importer liberalized.
Once we allow for this, it appears that members did not increase their trade
with each other significantly as a result of their RIAs. By contrast, during
the 1980s and 1990s, deepening and enlargement in EFTA and the EU were
not accompanied by major external liberalizations (except for the Tokyo
Round), and for these blocs, trade diversion is quite evident.
Looking at members' extrabloc trade, it is useful to distinguish the
blocs between developing countries that simultaneously liberalized their



44    REGIONAL INTEGRATION AND DEVELOPMENT
trade policies in general (ASEAN, LAFTA, the CACM, the Andean Pact,
and MERCOSUR in figure 2.7) from the others (the EC/EU, EFTA, the
GCC, and NAFTA). Since the mid-1980s, imports from nonmembers
have increased relative to expected values (obtained from the gravity
model) for the Andean Pact, the CACM, LAFTA, and MERCOSUR (sig-
nificantly so for the last two). This result reflects their unilateral trade
liberalization, which dominated any trade diversion from the RIA. The
same applies to ASEAN since 1980. Similarly, with one exception, these
liberalizing blocs' exports to the rest of the world have increased relative
to expected levels; this is to be expected from a trade liberalization and
confirms the dominance of the general liberalization over the discrimina-
tory RIA. The exception is MERCOSUR, where the export and import
effects move in opposite directions, suggesting that external trade devel-
opments over the 1990s were dominated by competitiveness factors. In
this case, given that the export data suggest an increasing focus on intra-
bloc trade, it would be premature to interpret the strong increases in im-
ports as indicating a sustainable liberalization that outweighs regional
biascs.
Turning to the cases in which the period since 1980 has not seen a
large liberalization, trade diversion is directly evident. Both the EU and
EFTA trade more with nonmembers than would be expected from the
gravity equations, but the excess has declined significantly since 1980.
The same is true for NAFTA since 1986-88, although in this case the es-
timated changes are not significantly different from zero.
The gravity model compares members' actual trade not with what it
was before the RIA-frequently, an undemanding standard-but with
what it would have been had members exhibited "average" behavior as
defined by the other countries in the sample. A related indicator is a com-
parison between trade within the RIA and members' export performance
in third-country markets, where they receive no preferences. Yeats
(1998) found that MERCOSUR members' exports to partners increased
most strongly in products where extrabloc exports were weak and pro-
tection was relatively high. Since member exports compete with the same
third-country suppliers inside the bloc as outside it, Yeats inferred that
the greater success of some of the exports within the bloc was attributa-
ble to trade preferences. Thus, he concluded, MERCOSUR tariff barriers
were diverting trade from cheaper goods made outside the bloc to more
expensive internal ones. This is not the same as comparing the RIAs with
their members' pre-RIA distorted positions, but it certainly suggests
some forgone opportunities from nondiscriminatory trade.
These more sophisticated analyses show a darker side of regionalism.
Although RIA partners trade more with each other than objective criteria



How TRADE BLOCS INCREASE TRADE AND COMPETITION                               45
Figure 2.7 Effect of RIAs on Trade within and among Blocs, Selected Periods, 1980-96
ASEAN                          MERCOSUR                          GCC
16s                               30                             3.0
S°E° 05°                                          a   1 525       aC EE o2.0             4
-1.0 22.0                                                              1.0
Ax                               15                         S     .-3n=.-~--
SE  0.5                          .5   1
E            dE                   E E 10             °              E O 0.
re >                                       A° 5 \ r°e E ~° 5 e E 3 0 -
aX-05                es                0.0                         0 \*-ff    e                  e -2 0
2~~~~~~~~~~~~~8
-1.5 1980-82  1 988-88  995-98    -2.0        988 88     98      -5.0
1980-82 1986-88 1995-96         1980-82 1986-88 1995-96         1980-82 1986-881995-96
NAFTA                              EFTA
1.5                     _        1.2
<   1.0                   --         08
0.5                          j 04
in 02
0.0                           -   Q
-1 .0                            -0.8
1980-82 1986-88 1995-98          1980-82 1986-88 1995-96
CACM                           Andean Pact
5.00                       -        3.0
4 0C                       -        25
VW 300      +                   _   O    20
1i 2.00                         _        1.5
56 E                                    10
mt  i.oc                      ---   a~
-a o.oc  ______________    5.0.5
_-0.00                         -°       -.
' ox  1980-82  198118-88 1995-98  -1e O O      *       5.0
-1 .5
1980-82 1988-88 1995-98
LAIA                           EC/EU
2.0                               1.5
a  1.0-                      _    a   1.0
0.5
oo~   ~~~S-
ie                                E   fi .0e ° 
-E                            0~~~~~~-
-3.0                              -2.0
1980-82  1986-88  1995-98         1980-821986-881995-96
---- Intra-EUTrde           -EU-imports            EU-Exports
Note: ASEAN, Association of Southeast Asian Nations; CACM, Central American Common Market; EC/EU,
European Community/European Union; EFTA, European Free Trade Association; GCC, Gulf Cooperation
Council; LAFTA, Latin American Free Trade Area; MERCOSUR, Common Market of the South, NAFTA,
North American Free Trade Agreement.
a The y-axes show values of coefficients on RIA dummy variables, in logarithmic equations. A coefficient of
b implies that the RIA raised trade by {[exp(b) - t] 100}% above the 'expected' value.
Source: Soloaga and Winters (2001).



46    REGIONAL INTEGRATION AND DEVELOPMENT
would predict, there is little evidence that the excess increases when an
RIA is signed or deepened, and there are signs that the degree of open-
ness to the rest of the world can decrease, albeit from levels that exceed
normality. One way of reconciling these results with those on trade
propensities discussed above is that, over time, the world has become
more open. Thus, while RIAs have usually (although not inevitably) been
associated with improvements relative to a member's pre-RIA positions,
they have not kept up with trends elsewhere, and so members have fallen
behind over time.
A CHANGE OF PERSPECTIVE: IMPERFECTLY
COMPETITIVE MARKETS
In the analysis above, trade is driven by comparative advantage based on
differences in productivity or factor endowments; producers make ho-
mogeneous products, and individual firms have no power to influence
the prices at which they sell. Economists refer to this situation as "per-
fect" (in the sense of complete) competition. In the real world, however,
competition is often imperfect. In the first place, many products are dif-
ferentiated, and consumers differ in their preferences for the varieties.
This attenuates the effects of price changes on patterns of demand (not
everyone wants to switch to a variety that has become cheaper). Produc-
ers then acquire a degree of market power; each effectively becomes a
monopolist for its own particular variety. Rather than being able to sell
only at the exogenously given price determined by an anonymous mar-
ket, firms can vary their prices and sales according to the demand for
their products. Furthermore, individual varieties of goods face not only
downward-sloping demand curves but also increasing returns to scale (at
least at low levels of output).
These changes in perspective alter economists' views of the returns to re-
gionalism and of how to quantify them and set the scene for a richer menu
of integration effects than was found above. Most obviously, they allow us
to explain intraindustry trade-the phenomenon in which a country imports
and exports different varieties of the same good. Intraindustry trade was
hardly acknowledged in economic literature until the 1960s, when empirical
studies of Western European RIAs (Verdoorn 1960; Balassa 1974) showed
that the bulk of trade expansion within Benelux and the newly established
EEC was intraindustry rather than interindustry or intersectoral. Intraindus-
try trade has subsequently become the leading component of manufactured
trade between developed countries, and some studies have suggested a posi-



How TRADE BLOCS INCREASE TRADE AND COMPETITION       47
tive link between RIAs and intraindustry trade. For developing countries,
however, intraindustry trade seems rather less important, and there is no
firm empirical evidence on the connection between it and regionalism.
Product differentiation also changes the analysis of the welfare effects
of an RIA. We argued earlier that if two small countries form an RIA,
trade diversion inevitably dominates the welfare calculation and that in
any market a country can gain only if it ceases to import the good from the
rest of the world. (See also the explanation in the appendix to this chapter.)
If such imports continue, the country's internal price remains anchored at
the world price plus its tariff, and there is no incentive for its domestic be-
havior to change at all. Displacing all imports from the rest of the world is
a very tight condition that makes gains from RIAs seem unlikely. And this
tight condition is necessary, but not sufficient, to obtain gains from RIAs.
The critical assumption in the analysis is that the good is homoge-
neous-every unit is identical regardless of its provenance-so that there
is only one price. If, however, domestic, partner, and third-country vari-
eties of the good differ slightly, each has its own price and faces its own
demand curve. As partners are exempted from the tariff and supply more
goods, the price of the variety they produce will have to fall for them to
sell the extra units. This allows at least a chance of domestic consumer
and producer effects before third-country trade stops completely. Thus,
differentiated goods make gains from RIAs more likely than if goods are
homogeneous, although because they weaken the link from reductions in
partner prices to reductions in domestic and third-country prices, they
tend to reduce the size of the gains when gains are recorded.8
Clearly, goods have different degrees of heterogeneity, but unfortunately
we have no empirical information about the importance of differentiation in
the medium to long run. The ability of firms in most sectors to copy each
other in the long run, and the large changes in market shares that are ob-
served over multiyear periods, suggest that the homogeneous goods view is
likely to offer reasonably robust predictions of long-run effects. Thus, its
basic insight-that if nonmembers face tariffs, member and domestic suppli-
ers of related goods will be able to charge significantly higher prices than if all
countries had equal access-will generally be a satisfactory guide to policy.
The most direct application of product differentiation in the analysis
of RIAs is in the use of computable general equilibrium (CGE) models.
These are fully articulated computer (mathematical) models of the
economies under study that include, for example, tariff rates, factor mar-
kets, product supplies and demands, and welfare indicators. Using CGE
models, researchers can simulate the effects of policy changes associated
with RIAs. Because such models typically contain a great deal of micro-



48       REGIONAL INTEGRATION AND DEVELOPMENT
economic detail, they can be used to predict changes in production in
each sector, as well as changes in factor prices and real incomes.
CGE models have become increasingly sophisticated as researchers have
refined the technique.9 First-generation models assume that all markets are
perfectly competitive but that products still differ by country of origin, so
that the costs and benefits of RIA membership arise only from trade diver-
sion and trade creation, as discussed above. Second-generation models in-
clude increasing returns and imperfect competition and so incorporate the
scale and competition effects that are discussed in the next section. Third-
generation models contain some dynamics, allowing for capital accumula-
tion and sometimes also technical progress; they are discussed in chaprer 4.
The conclusions from these models are, broadly, that there are gains
from regional integration but that these gains are small (Francois and
Shiells 1994; Harrison, Rutherford, and Tarr 1996). In the first-generation
models the interaction between trade diversion and trade creation has ef-
fects that are typically only a fraction of 1 percent of GDP. In second-gener-
ation models the effect generally increases to around 2-3 percent of GDP,
and in third-generation models the gain is approximately 5 percent of GDP.
The strength of these models is that they have sufficient microeco-
nomic structure to allow the effects of a policy change to be traced out in
detail and its real income effects to be calculated. Moreover, they are
about the only tool available for predicting the likely effects of an RIA
Box 2.1 Computable General Equilibrium (CGE) Modeling: An Explanation and a
Health Warning
According to Robinson and Thierfelder, two influential economic modelers, "empirical studies ...
overwhelmingly show that aggregate trade creation dominates trade diversion . . . [and that] in many
cases, there is no absolute trade diversion" (Robinson and Thierfelder 1999). This statement is not
proved: it relies not on data from actual instances of regionalism but on predict ons constructed from
CGE models, which tend to emphasize creation over diversion. The reason for this emphasis lies in
the models' assumption that products are differentiated, not homogeneous.
The assumption of heterogeneous goods is a major feature of almost all CGE models, which are the
workhorses for predicting the effects of RlAs. Partial equilibrium models, such as those in the ap-
pendix to this chapter, consider a single market in isolation. General equilibrium recognizes that
markets interact in complex ways so that everything depends on everything else. Demand for any
one good depends on the prices of all other goods and on income. Income, in turn, depends on
wages, profits, and rents, which depend on production, which depends on sales (that is, demand).
Prices depend on wages and profits, and vice versa; supply and demand must be equal in all mar-
kets, including factor markets; and imports must be paid for by exports plus foreign borrowings. Not
only do CGE models allow us to reflect product differentiation, but in the past 10 years they have
been extended to incorporate market power, accumulation, and even technical advances.
(Box continues on nextpage.)



How TRADE BLOCS INCREASE TRADE AND COMPETITION                             49
Box 2.1 (continued)
CGE models specify all their economic relationships in mathematical terms in a form that allows the
model to predict variables such as prices, output, and economic welfare, given information about
technology, preferences, and policies. In principle, all the relationships in a model could be estimat-
ed from detailed data on the economy over many years. In fact, however, the number and complex-
ity of the relationships far outweigh the data available, and so most CGE models are designed on the
basis of theory, intuition, and casual empiricism. Having specified most of the relationships, the
modeler manipulates a subset of parameters so that the model will replicate detailed data for one
base year. The modeler then re-solves the model, changing a few key features of the base year. For
instance, to explore an FTA, tarffs are set to zero on trade flows between the partners; the difference
between the base and the new predictions is the predicted effect of the FTA.
CGE modeling is a powerful tool that allows economists to explore numerically a huge range of issues
on which econometric estmaton would be impossible. One such application is to forecast the effects
of RIAs that are yet to be formed. We draw liberally on CGE models in this report-for example, on work
by Flares (1997) and by Harrison, Rutherford, and Tarr (1996, 1997, 2002). Although these models em-
body our best efforts at predicting the effects of future RIAs, it is well to recognize their limitatons:
* CGE simulations of an RIA are not unconditional predictions but, rather, "thought experiments" about
what the worid would be like if the RIA had been in existence and fully operative in the base year.
* Although CGE models are quantitative, they are not empirical. Indeed, they are almost wholly
theoretical, with no possibility of rigorous testing against experience. At best, a high proportion
of their individual relationships will be based on data and observation, but some components
(much less the overall model) can hardly ever be tested. At worst, CGE models represent theory
supplemented by incomplete base year data and 'guestimated" parameters. In some cases,
whole relationships are 'guestimated," as in the assumption by Hinojosa-Ojeda, Lewis, and
Robinson (1995) that productivity in NAFTA is related to the share of exports in total production.
* Conclusions about trade policy are very sensitive to the values chosen for tariffs and trade re-
strictions in the base case and simulation experiments. Unfortunately, Information on these
items-especially on nontariff barriers---is very hard to find and to aggregate satisfactorily to
the levels required by the models.
*  As noted above, the assumpton of heterogeneous goods tips the model toward finding benefits
from trade preferences.
before it is implemented. But they have a major weakness: they are not
fitted to data as carefully, or subject to the same statistical testing, as
econometric models. Critical relationships are often specified with no
empirical justification; many crucial variables cannot be measured satis-
factorily; the level of sectoral detail is often rather low, which implies a
good deal of averaging for the data on important variables; the results
are usually very sensitive to measurement error in these hard-to-measure
variables; and the specification of the behavioral relationships is usually
very simple. These factors probably tend to bias CGE models toward
finding benefits from RIAs. The cost of the CGE model's microtheoretic
consistency is a complexity that makes it impossible to establish rigor-
ously the relationship between the model and reality. This is not to con-
demn such models, but it does imply caveats, as noted in box 2.1.



50    REGIONAL INTEGRATION AND DEVELOPMENT
LARGER MARKETS, MORE COMPETITION
The second way in which the existence of imperfect competition changes
economists' perceptions of the benefits of RIAs is via economies of scale,
which are frequently the cause of imperfect competition. Many countries
are too small for activities that are subject to large economies of scale to
reach an efficient size. This might be because insufficient quantities of
specialized inputs are available or because markets are too small to gener-
ate the sales necessary to cover costs. Even if the economy is large enough
to support one optimally sized firm, such a firm would be a monopoly,
with all the associated drawbacks. Regional integration offers one route
for overcoming the disadvantages of smallness: by pooling resources or
combining markets, countries can benefit from a combination of scale ef-
fects and changes in the intensity of competition. But regionalism is not
the only or, indeed, necessarily the most effective way of overcoming the
handicap of smallness; unilateral trade liberalization, which shares the
markets and resources of the whole world, is likely to be more powerful.
There is plenty of evidence that the number of firms operating in most
developing countries is relatively small. Rodrik (1988) reports that meas-
ures of concentration (measures of firms' market power) in manufactur-
ing sectors in large developing countries are typically between 50 and
100 percent higher than in developed countries. This is potentially dan-
gerous, although it does not inevitably mean that competition is inade-
quate; entry costs in developing countries may also be relatively low,
which imposes competitive pressure on incumbents because excess prof-
its will be eroded by new entrants.'0
In principle, an RIA combines markets, making it possible to reduce
monopoly power as firms from different countries are brought into more
intense competition with each other. This can yield four types of gain.
The first is the textbook gain from increased competition: firms are in-
duced to cut prices and to expand sales, benefiting consumers as monop-
olistic distortion is reduced."I
The second source of gain is that market enlargement allows firms to
exploit economies of scale more fully. In a market of given size, there is a
tradeoff between scale economies and competition: if firms are larger,
there are fewer of them, and the market is less competitive. Enlarging the
market shifts this tradeoff, as it becomes possible to have both larger
firms and more competition. For example, there might be an initial situ-
ation in which two economies each have two firms in a particular indus-
try and these firms exploit their duopoly power, setting prices well above
marginal cost. After the formation of the RIA, there are four firms in one
combined RIA market. This increases the intensity of competition and



How TRADE BLOCS INCREASE TRADE AND COMPETITION      51
may induce merger (or bankruptcy) of some firms, perhaps leaving only
the three most efficient firms. The net effect is increased competition,
larger firm scale, and lower costs. "Triopoly" competition is likely to be
more intense than the original duopolies, and the surviving firms, being
larger and more efficient, can better exploit economies of scale.
The third source of gain arises if each firm produces a different variety
of the product. In the example just given, consumers throughout the RIA
now have a choice of three rather than two varieties.
The final source of gain is in the possible reductions in internal ineffi-
ciencies that firms are induced to make. If the RIA increases the intensity
of competition, it may induce firms to eliminate internal inefficiencies
(so- called X-inefficiency) and raise productivity levels (Horn, Lang, and
Lundgren 1995). Since competition raises the probability of bankruptcy
and hence of layoffs, it also generates stronger incentives for workers to
improve productivity and increases labor turnover across firms within
sectors (Dickens and Katz 1987).
There is a good deal of evidence that nonpreferential trade liberaliza-
tions achieve these gains. A number of studies have found that openness
to trade reduces price-cost margins, an indicator of competitive pressure
in the industry (Roberts and Tybout 1996).12 There is also evidence of an
association between trade liberalization and increases in efficiency, and
between trade liberalization and a reduction in the dispersion of efficien-
cy levels, as low-efficiency firms adapt or are eliminated.'3 Tybout (2000)
concludes that most of the efficiency gains from openness come from re-
ductions in inefficiencies rather than from scale effects.
As for regional integration, we have much less direct evidence. The
most extensively studied RIA is the EU. Here the static gains from com-
pleting the integration of the European market under the Single Market
Programme (SMP) were predicted to range up to about 5 percent of
GDP, split about equally between the traditional effects described earlier
in this chapter and the scale and competition effects just considered
(Catinat and Italianer 1988; Emerson 1988).14 These estimates were
based on extrapolations of calculations from a handful of industries and
assumed a significant increase in competitive pressure.'5 Indeed, they
were characterized as somewhat "heroic" (Winters 1992a).
The ex post evidence available to date is less flattering to the SMP, al-
though, in fairness, it is very provisional. The EU's Single Market Review
(CEC 1996) estimated its economic effects at only 1 to 1.25 percent of
GDP in 1994.16 Moreover, there is good evidence that general external
trade liberalization has been more important than regional integration in
achieving competitive gains. Jacquemin and Sapir (1991) show that the
procompetitive effects of trade occur not through high levels of intra-EU



52     REGIONAL INTEGRATION AND DEVELOPMENT
trade but only where there is a high degree of import competition from
firms outside the EU.
Turning to developing countries, several arguments suggest that the po-
tential competitive gains may be larger than for high-income economies.
The small size and relatively closed structure of many developing countries
mean that there is scope for more fully exploiting economies of scale and
for removing local monopoly power. It is often thought that the services
sector, with its high levels of restrictions and scope for economies of scale,
offers considerable potential gains from opening to competition. That,
however, would require the effective integration of services markets within
the RIA, which, as shown in chapter 6, is far from common.
A number of studies have calculated the potential (rather than actual)
gains that might be expected from the competition and scale effects.
Hunter, Markusen, and Rutherford (1992) construct a CGE model of the
U.S. and Mexican automobile industries and simulate the possible effects
of NAFTA; they predict large increases in output for Mexico, increases in
the scale of individual firms, and reductions in price-cost margins. A study
of MERCOSUR (Flores 1997) based on a similar methodology suggests
GDP gains of 1.8, 1.1, and 2.3 percent for Argentina, Brazil, and Uruguay,
respectively, the larger economies gaining less because they are already
closer to reaping economies of scale. Flores also uses an interesting decom-
position of the sources of welfare gain. As shown in table 2.1, for imper-
fectly competitive sectors, the most important effects are the direct re-
source savings from reduced trade and transport costs and the decline in
firms' markups. The benefits of increased variety, trade diversion, and
changes in export and import prices have only minor effects on welfare.
Quite different results are obtained by Dissou (2002), who uses a simi-
lar methodology to examine the effects on Senegal of the West African
Economic and Monetary Union (UEMOA), an RIA of small, least-devel-
Table 2.1 Sources of Welfare Benefits in MERCOSUR's Imperfectly Competitive Sectors
(percentage of total)
Lower
Trade   Lower                        Export  Import
Sector        Costs  Markups   Diversity  Diversion  Prices  Prices  Total
Steel         45.2     71 2     -9.2      -31      -5.0     0 9    100
Machinery     28 5     82 2     -1 2      -0.4     -9.4     0.3    100
Automobiles   48.5     69.5      0.0     -11.3     -7.2     0.5    100
Other vehicles  505    78.1     -2.3      -8.8    -18.0     05     100
Chemicals     37.5    73.2      -1 0      -4 4     -6 8     1.5    100
Source: Flores (1997).



How TRADE BLOCS INCREASE TRADE AND COMPETITrION                53
Table 2.2 Welfare Effect for Senegal of UEMOA and Unilateral Liberalization
(percentage deviation from base run)
Eliminatlon of    Improved
Full                 Tariffs on      Access to        CETon
Implementation       Regional        Regional       Nonregional    Full Unilateral
of Customs Union     Exports        Markets Only    Exports only   Liberalization
(1)                    (2)             (3)             (4)             (5)
0.57             -0.18            0.15            0.49            1.26
Note: CET, common external tariff; UEMOA, West African Economic and Monetary Union. The import, mar-
ket access, and CET effects do not add up to the customs union effect, 0.57, because of interaction effects.
Source: Dissou (2002).
oped countries (see table 2.2). UEMOA has a positive effect (0.57 percent,
measured as the deviation from the base run) on Senegal's welfare, but on
the import side, elimination of tariffs on intra-RIA imports has a negative
effect (20.18 percent) because of significant trade diversion. Improved re-
gional market access raises Senegal's welfare by 0.15 percent, which does
not offset the welfare loss of 0.1 8 percent from trade diversion. The small
size of the welfare gain, 0.15 percent, is likely to be attributable to the
small intrabloc trade flows. UEMOA's common external tariff (CET) is
generally lower than Senegal's pre-UEMOA tariffs, resulting in an addi-
tional welfare gain of 0.49 percent. Thus, the main benefit from UEMOA
for Senegal is the reduction in tariffs with respect to the rest of the world,
nor the removal of tariffs on intrabloc trade. And the gain from unilateral
trade liberalization (1.26 percent) is more than twice the 0.57 percent gain
from UEMOA, even though unilateral trade liberalization does not pro-
vide the benefit derived from improved market access.
Interesting as the estimates from these studies may be, it should be re-
called that since they are based on CGE models, they reflect not actual
outcomes but economists' predictions of the likely effects of integration.
There is virtually no direct ex post evidence on the competitive effects of
RIAs among developing countries. Indirect evidence, however, shows that
these gains depend largely on the existence of complementary production
between members and the interpenetration of each other's markets-that is,
on intraindustry trade. Although analysis of the importance of intraindus-
try trade in developing country RIAs is fragmentary and partial, simple sta-
tistics show strongly that developing countries, even in the middle-income
range, generate far lower levels of such trade than do developed ones.
A well-documented case study of duplication of plant in the tire industry
in Central America (Willmore 1974, 1976) illustrates the difficulties devel-
oping countries can encounter in reaping economies of scale in practice. A



54    REGIONAL INTEGRATION AND DEVELOPMENT
plant in Guatemala had the capacity to meet the entire CACM demand,
and there was another sizable plant in Costa Rica. Rationalization did not
occur, possibly because the external tariff remained high, enabling the firms
and governments (which were generally reluctant to see their firms go
under) to collude to impede effective competition. The same occurred with
cement plants in the Economic Community of West Africa (CEAO).
So, regional integration schemes may offer developing countries sub-
stantial potential gains through competition and scale effects. The gains,
however, are not automatic, and making sure that they are achieved calls
for careful policy design. In particular, it requires easing barriers to entry
(for example, for foreign direct investment) and allowing competition
free rein even when it hurts. There is much more to competition than the
mere removal of tariff barriers. In chapter 6 we return to competitive ef-
fects, when we examine policy integration. We show that, to date, few
RIAs have been able to advance very far in this direction. It should also
be kept in mind that gains from competition and economies of scale may
be attainable through unilateral and multilateral trade liberalization.
APPENDIX. THE SIMPLE ANALYTICS OF TRADE CREATION
AND TRADE DIVERSION
Appendix figure 2A.1 illustrates the simplest analysis of trade creation
and trade diversion, based on the conventional interpretation of Viner
(1950). For more detailed analyses, see De Rosa (1998); Robson (1998).
The figure represents the demand for and supply of a good on a market.
The good is assumed to be homogeneous across suppliers. We assume that
the industry is perfectly competitive and undistorted so that the supply
curve faithfully reflects the actual marginal costs of production. It is as-
sumed that the country is small and cannot affect the price of the good on
world markets. This implies that the import supply curve is horizontal: no
matter how much is demanded by consumers in the country, the world
price for the good remains unaffected. Equilibrium requires that demand
equal supply. Without trade barriers, consumers will import Q6 - Q, from
the lowest-cost supplier (country A), with amount Q, supplied by local pro-
ducers. If the government imposes a tariff, the internal cost of imports will
rise by the amount of the tariff. In the figure this is represented by an up-
ward shift of the import supply curve, as perceived by domestic purchasers,
from pA to pA + T. The domestic price of imports is now the world price
plus the tariff. The total quantity demanded falls from Q6 to Q4; sales of
domestic industry rise from Q, to Q3; and imports fall to Q, - Q3.



How TRADE BLOCS INCREASE TRADE AND COMPETITION     55
Figure 2A.1 The Market for a Single Homogeneous Good
Pntce
Domestic
supply
curve
/      \        ~~PB+T
pA + T
A          BI        C          D               p
E     F    G         H                          P
Demand
curve
Q    2     )3               Q4     Q5 Q6          Quantlty
The impact of the tariff on economic welfare is negative. Domestic
consumers pay more for the good; they lose the sum of areas A + B + C
+ D + E + F + G + H + I + J. From a national perspective, this is par-
tially offset by the gain that accrues to domestic producers from the high-
er price (areas A + E) and the tariff revenue that is obtained by the gov-
ernment (equal to area C + H-the tariff multiplied by the amount
imported). The net loss to society due to the tariff is therefore equal to
the sum of triangular areas B + F + G and D + I + J. These triangles are
measures of the distortion in resource allocation that the tariff creates:
domestic firms produce too much (relative to free trade), and consumers
consume too little.
If it is now assumed that our country concludes an RIA with country
A and so eliminates the tariff on imports from A, the same benefits
emerge as under general free trade. It is as though country B had never
existed-it figures in this commerce neither before nor after the RIA. If,
however, the RIA is with B, the higher-cost supplier, the price of imports
to domestic buyers falls from pA + T to pB. Consumption expands from
Q, to Qs; consumers gain A + B + C + D; producers lose A; and the



56    REGIONAL INTEGRATION AND DEVELOPMENT
government loses revenue C + H. The net welfare effect is then B + D
- H, which could be either positive or negative. Imports from country
A cease, replaced entirely by those from B. Of the increase in imports
from partner country B, Q3 - Q, displaces local sales and constitutes
trade creation in a pure Vinerian sense, generating welfare gains of area
B. Q, - Q4 represents the increase in consumption induced by the fall
in the internal price; this can also be thought of as trade creation and
generates welfare gains of D. As in the case of nondiscriminatory tariff
cuts, both parts of trade creation are beneficial nationally. The final ele-
ment of imports from B, however, Q4 - Q3, is trade diversion: imports
that initially came from A at price pA now come from B at price pB, cost-
ing the country welfare equivalent to area H. This is welfare reducing be-
cause the country is paying more for its imports: consumers pay less for
them, to be sure, but only because the government is no longer collect-
ing a tariff.
Appendix figure 2A.1 illustrates a number of points made in the chap-
ter 2 text. For example, the net effect of the RIA in this market depends
on the relative sizes of trade creation and trade diversion; the cost of di-
version is determined by how inefficient the partner is relative to world
best practice; and the existence of diversion depends on the tariff on A
being greater than that cost difference.
For future reference, we observe here that neither A nor B experiences
noticeable welfare effects from the RIA. The fact that each is prepared to
sell any amount to our country at a fixed price suggests that they would
be equally happy selling somewhere else or not producing at all (the price
they get from our country just matches the opportunity cost of supplying
the good) and that therefore they are indifferent about whether they sell
it or not. The most common cause of this outcome in reality is that A and
B are very large relative to our market: our demand is just not large
enough to matter.
We should also note that it is possible that the barrier to A's and B's
exports to us is not a tariff, or indeed any other barrier that produces a
revenue (rent) for someone, but a resource-using, or frictional, barrier.
For example, it may be that goods have to be extensively tested before
admission or that there are large amounts of paperwork to be done. In
these cases the positive analysis-what actually happens to quantities
and prices-is exactly the same as for the tariff, but the welfare analysis
is different. The triangles associated with trade creation remain as be-
fore, but with frictional barriers there are no trade diversion losses. Al-
though supply is switched to a higher-cost supplier, there were no rents
from trade in the first place, and so there was nothing to lose. The whole



How TRADE BLOCS INCREASE TRADE AND COMPETITION    57
of area C + H was just wasted in friction in the first equilibrium. Hence,
exempting supplies from B from such barriers represents a real resource
saving (at least, provided that doing so does not undermine any other
public policy objective achieved through testing or paperwork require-
ments). Of course, it would be even better to exempt A, as well as B,
from the barrier so that imports are available at real cost pA rather than
pB. This analysis shows that it is better to remove a frictional barrier of x
percent than a tariff of x percent: the effects on the private sector are the
same, but the first measure entails no loss of government revenue be-
cause the government was never getting any.
Appendix figure 2A.2 extends the argument to the case of an RIA be-
tween two small countries. This time, we draw the demand curve for im-
ports, representing domestic demand less domestic supply for the good
(that is, net demand)."7 We can buy from a small country, or from a large
one (the rest of the world). At first, when we impose equal tariffs on
each, we divide purchases between them. The small country (supplier B)
is able to supply some units verv cheaply, but as we demand more, our
demand becomes large relative to the small country's capacity to supply,
and we drive up its price. The rest of the world (supplier A) is large and
can supply effectively any amount we require at a fixed price. Thus, pro-
Figure 2A.2 An RIA with a Small Partner
Demand for
imports
/ S+T      S    \
A  B~~~~~~~~~~~~~~~~~
<  /   o/          \          p2~~~~~p+T
A~~~~~~~~~~~~~P
MIB            ,t@7M              Quantity of Imports



58    REGIONAL INTEGRATION AND DEVELOPMENT
ducers in B are happy to supply us when their costs are below pA, but
beyond that level we buy from A at price pA. Of course, since we have
to buy some units from A, B's firms charge us pA even though they
would supply for less (we have no alternative, so they can collect some
rent). Thus, the initial equilibrium involves imports M, of which MB
comes from B and M - MB from A. The domestic price is pA + T, but
of this, only pA flows to producers overseas; T remains at home with the
government.
Now imagine that we join an RIA with B. Now B's producers no
longer face the tariff, but if we have to continue to buy some supplies
from A at price pA, the internal price remains pA + T. (A is the marginal
supplier and so determines the price.) If the internal price does not
change, no domestic behavior changes, and so there is no chance of our
gaining anything through production or consumption effects. The sourc-
ing of imports, however, does change because producers in B, exempt
from the tariff, can now earn pA + T from exports rather than just pA.
Naturally they increase their sales, to MB'. Moreover, because we are no
longer charging tariffs to B, the areas A + B + C in the figure switch
from being tariff revenue to being a real cost of imports paid to produc-
ers abroad. This is a straight loss for us.
Our country's loss in this case depends on three things: the height of
the remaining tariff on A, the initial volume of imports from B, and the
extent of trade diversion (the increase in imports from B). Note that we
lose the tariff revenue on the initial level of imports from B even if there
is no trade diversion! But of course, the extra cost of imports to us repre-
sents a gain for B's producers-they get higher prices-so would not all
these effects cancel out for the RIA as a whole? The answer is, no, if there
is any trade diversion. B's producers get a straight transfer on the initial
level of exports, which does cancel out (area A), but as they expand their
sales, their real costs rise and so between MB and MB' the costs of pro-
duction exceed those that would apply if we bought from A. Of their
earnings on the increased level of exports, C is absorbed by extra costs
and B is surplus. Thus, overall, B's producers gain by A + B, while our
government loses A + B + C. Given that there is no possibility of trade
creation, we can be certain that in this simple case an RIA between small
countries will be welfare worsening.
A real RIA covers many goods, and so we would expect to gain export
revenue on some to compensate for the losses on others. The calculations
above show that the RIA will lose overall, but it is possible that one or
the other partner will be a net gainer. This is more likely if a country ini-
tially has lower tariffs or smaller imports from its partner.



How TRADE BLOCS INCREASE TRADE AND COMPETITION                      59
Table 2A.1 Trade Effects of RIAs between Developing Countries, OneYear
before and FiveYears after Implementation of Internal Preferences
(percent)
Import Shares  Trade Intensity   Import/GDP Ratio      Trade Propensity
Intra-  Extra-  Intra-  Extra-  Intra-  Extra-  Total  Intra-   Extra-
RIA            bloc   bloc    bloc    bloc    bloc   bloc   Imports   bloc     bloc
MERCOSUR
1990         14.5   85.5   10.52    0.87   0.57    3.34     3.91    0.41     0.03
1996         20.2   79.8   14.12    0.81   1.64    6.48     8.10    1.15     0.07
Andean Pact I
1968          4.3   95.7    2.90    0.97   0.54   12.15    12.69    0.37     0.12
1974          7.2   92.8    6.16    0.94   1.01   13.14    14.15    0.87     0.13
CACM 11
1990          9.1   90.9   81.19    0.91   2.63   26.22    28.85   23.42     0.26
1996         12.6   87.4   64.33    0.88   3.78   26.29    30.07   19.34     0.26
Andean Pact 11
1990          6.8   93.2    6.84    0.94   0.86   11.77    12.63    0.86     0.12
1996         13.6   86.4   15.55    0.87   2.01   12.80    14.81    2.30     0.13
CARICOM
1972          5.0   95.0   13.59    0.95   2.76   52.05    54.81    7.45    0.52
1978          3.8   96.2    7.66    0.97   3.10   78.01    81.11    6.21     0.78
CEAO
1965          3.3   96.7   11.68    0.97   0.48   14.39    14.87    1.74    0.14
1971          4.5   95.5   20.78    0.96   0.91   19.28    20.19    4 20     0.19
UDEAC
1965          1.4   98.6    8.92    0 99   0.24   17.59    1783     1.59     0.18
1971          4.0   96.0   30 09    0.96   0.79   18.82    19.61    5.90     0 19
AFTA
1991         15.9   84.1    3.80    ( 88   7.73   40.99    48.72     1.85    0.43
1996         17.7   82.3    2.82   ().88   9.34   43.40    52.74     1.49    0.46
GCC
1980          6.6   93.4    0.81    1.02   1.47   20.96    22.43    0.18     0.23
1986          5.2   94.8    2.15    0.97   1.50   27.71    29.21    0.63     0.28
Note: The dates given for each RIA are the respective "before" and "after" implementation dates. Andean
Pact I refers to the original 1969 agreement and Andean Pact II to the revived agreement (1991); CACM II
refers to the Central American Common Market. CARICOM, Caribbean Community and Common Market;
CEAO, Economic Community of West Africa; UDEAC, Union douaniere et economique de l'Afrique Centrale
(now CEMAC); AFTA, ASEAN Free Trade Area; GCC, Gulf Cooperation Council.
Source: World Bank data.
NOTES
1. The concepts of trade creation and trade diversion, first formulated by
Jacob Viner, continue to dominate the discussion of RIAs (see Viner 1950).
2. It should also be recognized that RIAs take different lengths of time to be-
come fully operational, so that the period chosen here (one year before the RIA
to five years after) is only a crude mndicator of their effects.



60     REGIONAL INTEGRATION AND DEVELOPMENT
3. Trade intensity and propensity indices are discussed in Anderson and
Norheim (1993).
4. Debate rages as to whether nondiscriminatory liberalizations should be at-
tributed to the process of regionalism. This makes a huge difference to how one
appraises certain RIAs. The issue is taken up in chapter 3, where we consider the
effect of RIAs on trade policy making.
5. The model that Frankel, Stein, and Wei (1997) prefer (they do not say why)
estimates a single effect for each RIA, applying equally to 1970, 1980, 1990, and
1992. This suggests that RIAs had positive impacts on all trade flows except for
the EC's intrabloc trade and EFTA's and NAFTA's trade with nonmembers. The
authors conclude that there is little trade diversion. But estimiating a constant ef-
fect over 1970-92 seems inappropriate, given the changes that were occurring in
RIAs at that time.
Frankel and Rose (2000) use a gravity model to examine the effect of curren-
cy unions and FTAs on trade and find strong intraunion and intra-FTA effects,
with the latter equal to between 70 and 100 percent of the former.
6. Statistically, this is equivalent to estimating effects on extrabloc imports,
extrabloc exports, and intrabloc trade. What Soloaga and Winters refer to as an
overall import effect is numerically the same as an extrabloc import effect, and
similarly for exports. The size of the intrabloc trade effect differs between the
two parameterizations, however. The effect in the (alternative) version of this
endnote would equal that of the text model plus the two extrabloc effects (see the
example in the text).
Statistically Soloaga and Winters actually proceed by estimating effects on
total imports, total exports, and intrabloc trade, but they show that their esti-
mates may be simply translated into statements about extrabloc imports, extra-
bloc exports, and intrabloc trade of the sort made in the text.
7. In the gravity model these coefficients are added to the logarithm of trade; the
effect is obtained by multiplying trade by exp(l.86) and exp(0.55), respectively.
8. As shown in appendix figures 2A.1 and 2A.2, although triangle gains are
obtained from the reduction in the price of imports from partner countries, the
RIA generates rectangles of losses because of a reduction in imports of substitute
products from excluded countries. In other words, trade diversion results in a
loss in tariff revenues and welfare, and the net impact of the RIA onl welfare is
ambiguous.
9. The classification of CGE models is taken from Baldwin and Venables
(1995), which surveys some of these studies in greater detail.
10. Tybout (2000) finds no evidence that price-cost margins are systematical-
ly higher in developing than in developed countries, although he reports a num-
ber of instances where this is so.
11. Dlankov and Hoekman (1997) report the positive effects of trade reform
on competition in Bulgaria.
12. See also Levinsohn (1993); Harrison (1994); Foroutan (1996); Krishna
and Mitra (1998).



How TRADE BLOCS INCREASE TRADE AND COMPETITION        61
13. See Nishimizu and Page (1982); Tybout, de Melo, and Corbo (1991);
Haddad and Harrison (1993); Tybout and Westbrook (199S); Harrison (1996).
14. Later research added, even more speculatively, dynamic growth effects on
top of these estimates (Baldwin 1989; McKibbin 1994). These effects are consid-
ered in chapter 4.
15. The original industry studies in this area were undertaken by Smith and
Venables (1988).
16. This is not an easy calculation because many of the measures of the SMP
were not adopted until late in the process, EU data were available only until
1994, and several other shocks such as recession and German reunification
clouded the picture.
17. An alternative assumption is that there is no domestic production, so that
all consumption has to be met from imports.






CHAPTER 3
Making the Most of Regional Integration
O   ne    important message that emerges from the analysis in chap-
ter 2 is that, in purely trading terms, an RIA cannot provide
any benefits that member countries cannot attain through
nondiscriminatory tariff reductions. When we look at the effects of
RIAs on reallocation of factors of production (assuming that their
quantity and productivity are not affected by regional integration) or on
patterns of consumption, there appears to be no economic rationale for
forming trade blocs. Nondiscriminatory tariff reductions would bring a
country all the gains of trade creation without the costs of trade diver-
sion and would thus be a superior option. When transfers between gov-
ernment revenue and producers are taken into account, one member
might do better in an RIA than under nondiscrimination, but only by
driving the other partner even farther below that level. Thus, the argu-
ments in chapter 2 cannot explain why both members would agree to
an RIA.
Then why are RIAs so popular? Several explanations are offered in
later chapters-for example, that regionalism helps reduce the chances of
conflict with neighbors or facilitates the negotiation of agreements to
share regional resources-but here we consider whether extending the
analysis of purely trade or border factors could explain this popularity.
We find that trade-based factors offer little justification for preferring
RIAs to free trade, although we do find reasons why RIAs formed with
other motivations tend to be between neighbors and, more controver-
sially, why developing countries will generally gain more from RIAs with
developed countries than with other developing countries.
63



64    REGIONAL INTEGRATION AND DEVELOPMENT
The judgments behind these conclusions are necessarily empirical and
pragmatic. Although economic theory suggests that RIAs can always be
designed to be welfare improving, in practice the necessary information,
the incentives, and the legal rights to manipulate the instruments of trade
policy to guarantee such outcomes are missing. Ultimately, it is not theo-
retical possibilities or empirical generalities that define the appropriate-
ness of an RIA but practical issues such as which partners are actually
chosen and the level of tariffs against the rest of the world. These are the
subject of this chapter.
FREE TRADE WITH WHOM?
One of the rare questions on which trade bloc economists agree is the
commonsense proposition that an RIA between any set of countries
could be designed to leave the rest of the world indifferent, at least one of
the members better off, and the rest no worse off (Kemp and Wan 1976;
Panagariya and Krishna 2002). Unfortunately, this result requires that
external tariffs be set so that members' total trade with the rest of the
world (commodity by commodity) is maintained at preunion levels and
that lump-sum taxes and transfers take place between partners. The re-
quired external tariffs are almost impossible to calculate-and might
conflict with WTO obligations-while the transfers are never feasible in
practice.'
Given these practical constraints on ensuring that any given RIA is
welfare improving, and the reality that forming an RIA with one set of
partners precludes or at least complicates creating one with others, the
question of why an RIA might be attractive cannot be divorced from the
question of who belongs to it. We therefore consider in this section both
rather generic arguments in favor of discrimination and rather specific
ones about the particular attractions of RIAs with certain types of
partner.
Why Not with Everyone?
One argument for preferring an RIA-specifically, a customs union-to
free trade is that if a trade bloc's joint economy is large, coordinating its
international trade policies will allow it to strike better bargains with
trading partners.2 That is, it can reap terms-of-trade benefits not avail-
able to its members acting independently. This rationale is examined in
chapter 8, where we consider the effects of RIAs on the rest of the world,



MAKING THE MOST OF REGIONAL INTEGRATION     65
for such gains are necessarily at the latter's expense. There is very little
research on such benefits, but Chang and Winters (2002) suggest that
they may have been important for MERCOSUR. In general, however,
improving the terms of trade is not commonly believed to be a big moti-
vator of trade policy stances, and besides, most RIAs, especially among
developing countries, are pretty small.
A second argument concerns market access. An RIA not only affects
imports but also promises the benefits of duty-free access for exports to
partner markets. This is an important consideration for a country that is
considering joining an RIA, but the benefits the country derives from
such market access can only corne at the expense of its partner-which,
for its part, would be better off if it were to eliminate its tariff. The ben-
efits are the exact counterpart of the tariff revenue losses discussed in
chapter 2. Both parties cannot gain simultaneously from such transfers.
Another aspect of the same argument is that countries may feel better
able to liberalize their own trade if others are doing so at the same time.
If improved access to another market depends on opening one's own, a
new constituency for trade reform may be created among exporters. If
countries simultaneously reduce tariffs on each other's exports, adverse
terms-of-trade changes are less likely than if only one party liberalizes
(Bagwell and Staiger 1999), and if several liberalize together, the need for
any one of them to devalue its currency to maintain its foreign balance is
reduced. Also, adjustment might be easier if export sectors are expanding
as import-competing sectors contract. These are all reasons for coordi-
nating trade reforms across countries, and RIAs provide one way of
doing this. But concerted unilateralism, under which several countries in-
troduce MFN liberalization together, would also fit the bill, as several
APEC member countries argued in the early days of that agreement.
A third reason for favoring RIAs over MFN free trade arises if mem-
ber countries wish to industrialize behind protective barriers. Forming a
group that provides scope for intrabloc industrial specialization reduces
the cost of protection and can generate welfare gains that would not be
open to members through unilateral tariff liberalization (Cooper and
Massell 1965). The first wave of RIAs among developing countries in the
1960s and 1970s set import substitution as a major policy objective, and
a number of RIAs, including LAFTA and ASEAN, appear to have been
explicitly motivated by gains from pursuing industrialization on a re-
gional basis. But these policies must be based on noneconomic govern-
mental objectives or public-goods arguments, given that there are few
grounds for concluding that economic gains can be derived from protect-
ed industrial development.



66    REGIONAL INTEGRATION AND DEVELOPMENT
If industries can be rendered competitive at higher levels of output
when preferential markets are available, it might appear that an RIA
could be justified. But it has then to be explained why the industry would
need the support of a protected regional market if mere expansion of
output (perhaps through scale economies) suffices to make its costs com-
petitive. The issues surrounding the dynamics of industrialization and
long-term comparative advantage are addressed in more detail in chap-
ters 4 and 5.
Choosing Partners: The "Natural Trading Partners" Fallacy
Ever since customs theory began, with Viner (1950), economists have
puzzled about who the perfect partners would be. For many years, this
proved to be a sterile debate, with every proposal turning out to be, at
best, a restricted and special case. (Panagariya 1997 lists the most impor-
tant contributions to the discussion.) Recently, World Bank research has
yielded new findings which, if they do not close the book on the contro-
versy, at least shine a brighter light on the question.
Early research focused on whether potential RIA members were com-
plementary in their production and consumption patterns and whether
they were already major trading partners. The first criterion proved very
difficult to pin down. In the early 1990s the second appeared to gain
considerable currency, along with the notion that countries close enough
to each other to reap savings on transport costs were natural partners
(Wonnacott and Lutz 1989; Summers 1991).
Before assessing the validity of the argument that countries which al-
ready trade disproportionately with each other are ideal partners for an
RIA, we examine whether this has been the case for most RIAs. The left-
hand side of table 3.1 shows, for a sample of RIAs, the share of intrabloc
trade in partner countries' total trade at the time of bloc formation. The
right-hand side shows the share of an individual country's total trade
that is carried out with a bloc it is joining. It is clear that the intrabloc
share of trade is much larger in North-North RIAs (EEC, 38.6 percent;
CUSFTA, 30.2 percent) and in North-South RIA's (NAFTA, 42.1 per-
cent) than in South-South RIAs. The largest intrabloc share in South-
South RIAs is in ASEAN (16.7 percent), followed by MERCOSUR (12.9
percent), the EAC (12.8 percent), and CEFTA (10.3 percent). The other
RIAs have much smaller intrabloc trade shares: below 1 percent for three
Sub-Saharan African RIAs (UDEAC, the SADC, and the CEAO), 1.0
percent for the CACM, 2.3 percent for the Andean Pact, 3.2 percent for
SAARC, 3.3 percent for the G3, and 4.2 percent for the GCC. This



MAKING THE MOST OF REGIONAL INTEGRATION                         67
Table 3.1 Trade Shares, Selected RIAs
(percent)
Share of                                  Country's Trade with EU (or
RM and                Intrabloc Trade   Country, RMA,             CACM) as Share of Country's
Year Formed            in Total Trade   and Year Joined                   Total Trade
Andean Pact (1969)          2.3         Costa Rica-CACM (1962)                2.5
ASEAN (1967)               16.7         Czech Rep.-EU (1995)                 46.7
CACM (1962)                 1.0         Estonia-EU (1995)                    16.7
CEAO (1973)                 0.9         Hungary-EU (1994)                    41.5
CEFTA (1993)               10.3         Iceland-EU (1994)                    60.0
CUSFTA (1989)              30.2         Latvia-EU (1995)                     21.7
EAC (1996)                 12.8         Lithuania-EU (1995)                  23.0
EEC (1 962)a               38.6         Morocco-EU (1996)                    74.4
G3 (1995)                   3.3         Norway-EU (1994)                     61.3
GCC (1981)                  4.2         Poland-EU (1994)                     49.4
MERCOSUR (1991)            12.9         Romania-EU (1995)                    41 8
NAFTA (1994)               42.1         Slovak Rep.-EU (1995)                26.4
SAARC (1985)                3.2         Slovenia-EU (1995)                   53.7
SADC (1980)                 0.1         Tunisia-EU (1995)                    56.8
UDEAC (1966)                0.9
Note: ASEAN, Association of Southeast Asian Nations; CACM, Central American Common Market; CEAO,
Economic Community of WestAfrica (became the Western African Economic and Monetary Union, UEMOA);
CEFTA, Central European Free Trade Area; CUSFTA, Canada-United States Free Trade Agreement; EAC, East
African Cooperation; EEC, European Economic Community; EU, European Union; G3, Group of Three
(Colombia, Mexico, and Venezuela); GCC, Gulf Cooperation Council; MERCOSUR, Common Market of the
South; NAFTA, North American Free Trade Agreement; SAARC, South Asian Association for Regional Coop-
eration, SADC, Southern African Development Community; UDEAC, Union douanimre et economique de
l'Afrique Centrale (became the Economic and Monetary Community of Central Africa, CEMAC).
a. The EEC was formed in 1957, but the earliest bilateral trade data obtained are for 1962.
Source: World Bank data.
evidence leads us to conclude that RIAs are typically formed between
neighboring countries but not necessarily between countries that are al-
ready major trading partners.
As the table shows, except for Costa Rica and the CACM, the coun-
tries that signed FTAs with the EU were major trading partners with it.
As more countries joined the EU, staying on the outside became ever
more costly, and the incentive to join increased. This phenomenon of
"domino regionalism" is examined in more detail in chapter 8.
Attractive though it would be to have a shortcut to assessing RIAs, the
concept of "natural trading partners" has little or no operational con-
tent. To be useful, it would have to be easily applied and to identify trade
that we wish to stimulate artificially by means of preferences. That a
trade flow is already large says nothing about the need to stimulate it: the
advantages of proximity and conformable commodity composition are



68    REGIONAL INTEGRATION AND DEVELOPMENT
already reflected in its size. In fact, some flows are large because of exist-
ing distortions and need to be curtailed rather than boosted.
The argument on transport costs is more sophisticated but is ulti-
mately of little more practical use. RIAs can be advantageous if they
allow two countries to trade a good with each other that, previously, one
exported to the rest of the world and the other imported at a higher
price. This situation could arise either because the rest of the world levies
tariffs or export taxes on its trade or because of transport costs (Wonna-
cott and Wonnacott 1981). By trading together and so reducing the mar-
gin between the buying and selling prices, the RIA improves one or both
countries' terms of trade, but only if at least one of them ceases its trade
with the rest of the world. Amjadi and Winters (1999) show that al-
though for MERCOSUR high transport costs on trade with the rest of
the world offer scope for large terms of trade gains, very few goods seem
likely to experience the required change in trade patterns. Similarly, with
the advent of NAFTA, Mexico reduced rather than expanded the set of
goods (defined at the six-digit level of the Harmonized Commodity De-
scription and Coding System trade classification) for which the United
States or Canada, or both, were its only suppliers.
Schiff (2001), revisiting the natural trading partners debate, argues
that both advocates and detractors of the hypothesis have largely missed
the point because, although they have analyzed the target country's rela-
tionships with its potential partners and with the rest of the world, they
have failed to consider relationships between the last two. This is impor-
tant because it constrains the patterns of trade and prices that could ac-
tually rule in the world prior to the formation of the RIA. Once this is
taken into account, the size of the preexisting trade flow is no guide to
"naturalness"-the relationship is neither necessarily positive (as advo-
cates claim) nor necessarily negative (as detractors claim). Schiff does,
however, argue that for a given target country, larger partners are better
than smaller ones because they are more likely to satisfy the target coun-
try's import demand without increasing prices, and they are less likely to
cease importing the target's export goods from the rest of the world and
hence to reduce their internal price of these goods below the world price
plus their tariff on third countries. He also argues that the RIA is likely to
be better off if its members are complementary in the sense that under
nondiscriminatory trade policies, each would import what the other ex-
ports rather than what the other imports.
In a recent paper, Krishna (forthcoming a) estimates the welfare effect
of a U.S. reduction in its preferential tariffs with respect to the imports of
various trading partners and examines the relationship between these



MAKING THE MOST OF REGIONAL INTEGRATION    69
welfare effects and the distance, as well as the volume of trade, between
the United States and its trading partners. He finds no correlation be-
tween geographic proximity and the welfare effect of these RIAs or be-
tween trade volume and welfare effect. Thus, he obtains no support for
the natural trading partners hypothesis.
Choosing Partners: Comparative Advantage
The second line of inquiry in the question of choosing partners is pursued
by Venables (2000, 2002), who considers the comparative advantage of
RIA members relative to each other and to the rest of the world. For sev-
eral special models, comparative advantage is important in defining a
country's ideal partners and helps to determine whether RIA members' in-
comes converge or diverge-a topic we discuss more fully in chapter 5.
Suppose that two developing economies both have a comparative dis-
advantage in manufactures relative to the rest of the world but that for
one the disadvantage is less than for the other. Kenya and Uganda can
serve as examples. Their comparative disadvantage in manufactures
could stem from many alternative sources-technological, geographic,
or institutional differences. Let us assume that it is because of different
endowments of capital: Kenya has little capital per worker relative to the
world average, but Uganda has even less. The initial position is one in
which each country has some manufacturing that serves local consumers
and survives because of relatively high tariff protection.
Now suppose that, starting from a position in which they levy identi-
cal tariffs on all imports, these two countries form a customs union.
Since Kenya has a comparative advantage in manufacturing relative to
Uganda (although not relative to the rest of the world), it will expand
manufacturing output to supply the whole customs union, or at least
move in that direction, while Uganda's manufacturing sector will con-
tract. These developments move Kenya's production structure farther
away from its true comparative advantage while moving Uganda's closer.
Surprisingly, however, Kenya gains from the reallocation, while Uganda
may lose and will certainly do less well than Kenya. The reason is that
Uganda suffers trade diversion: some mnanufactures that were previously
imported from the rest of the world are now imported at greater cost
from Kenya. For Kenya, by contrast, there are gains from being able to
supply manufactures in the Ugandan market, protected from competi-
tion with the rest of the world.
The implication is that countries with comparative advantage closer
to the world average do better in an RIA than do countries with more



70    REGIONAL INTEGRATION AND DEVELOPMENT
extreme comparative advantage. Interposing the "intermediate" country
between the "extreme" one and the rest of the world distorts the extreme
country's trade, causing a switch in import supplier. But the intermediate
country does not experience this switch in supply; its trade with the ex-
treme country and its trade with the rest of the world are less close sub-
stitutes and therefore less vulnerable to trade diversion. This, in turn, im-
plies that a customs union between two poor countries will tend to cause
divergence (the richer country-Kenya in this case-gains, while the
poorer country, Uganda, loses) but that a customs union between two
rich countries will tend to cause convergence (the poorer country is clos-
er to the world average and gains more than its partner). The same basic
forces therefore mean that regional integration between rich countries
causes their incomes to converge, whereas integration between poor ones
causes divergence.
What if the countries lie on opposite sides of the world average and
form North-South RIAs? The model is, unfortunately, ambiguous (Ven-
ables 2002: figures 4-6). If the developing (southern) partner is not too
far from the world average, it will almost certainly gain, but if it is very
poor (that is, very poorly endowed with capital or skills), it is likely to
lose unless its partner is correspondingly very rich (very well endowed).
Venables shows that for these results to hold, trade patterns must be
very particular. If goods are homogeneous (as in the Ricardian and
Heckscher-Ohlin models), the partners must not trade prior to the forma-
tion of the customs union and must trade afterward ("corner solutions").
Alternatively, products may be differentiated by origin (the Armington as-
sumption) such that every country is the sole supplier of its own particu-
lar varieties of each good, which is another form of corner solution.3
But the results do not hold in general. In particular, if both partners
are small and trade with the rest of the world both before and after for-
mation of the customs union, we are back in the world described in the
chapter 2 discussion of trade creation and diversion, in which integration
can have no beneficial effects. To see this, keep the assumption of equal
tariffs and assume a Heckscher-Ohlin model with homogeneous goods,
where both Kenya and Uganda initially import manufactures and export
agriculture. Before the customs union is formed, producers of manufac-
tures obtain the world price plus the tariff by selling at home but receive
only the world price by selling to their prospective partner. Thus, they
sell only at home, and they import the excess demand from the rest of the
world (ROW). With equal domestic prices in the two countries, margin-
al costs are the same as well. Of course, given that Kenya has relatively
more capital, it produces more manufactures than Uganda, but that is



MAKING THE MOST OF REGIONAL INTEGRATION    71
immaterial as long as it imports from the ROW After the customs union
is formed, both countries can get the same price at home and in the part-
ner's market. But since marginal costs were already equal to the domestic
price before the RIA was formed, the RIA has no effect on prices or trade
flows; neither Kenya nor Uganda gains or loses.
In fact, if Kenya had the higher tariff on manufactures initially and the
customs union adopted that tariff unionwide, we would see convergence,
not divergence. Provided that both countries imported manufactures
from the ROW initially, the marginal cost of producing manufactures
would be lower in Uganda because its tariff is lower. Thus, once the cus-
toms union was formed, Uganda would export manufactures to Kenya
(until marginal costs were equalized), while Kenya would reduce its im-
ports from the ROW; Uganda would gain, Kenya would lose, and South-
South integration would result in convergence. This should not come as
a surprise: since countries would already have exploited or arbitraged
away their comparative advantage positions by trading on the world
market, the only thing left to exploit once the RIA was formed would be
trade policy differences.
Another restriction in this model is the number of potential partners
and goods (Krishna forthcoming b). Once we assume several potential
partners in the North that are producing differentiated products, North-
South RIAs may result in welfare losses because the high substitution be-
tween northern goods may result in large amounts of trade diversion. As
discussed in chapter 2, the southern partner can minimize the costs of
trade diversion by reducing tariffs unilaterally.
Neighborhood RIAs
One of the most striking features of international economic integration is
that the great majority of RIAs involves neighboring countries (hence the
term "regional"!). The main exceptions are APEC (if it can be considered
a RIA), the Group of Three (Mexico does not border either Colombia or
Venezuela), various plurilateral FTAs (the EU with South Africa, with
Mexico, and with Chile), and bilateral FTAs such as those between
Canada and Chile, Chile and Mexico, and Israel and the United States.
The possible benefit of regionalism in reducing the chances of conflict
with neighbors might explain some of this, as might its role in helping
them negotiate agreements to share regional resources. Other possible
reasons include a history of cooperation with, and better knowledge of,
neighboring countries and a desire to replace past tensions with an insti-
tutional framework that promotes cooperation.



72    REGIONAL INTEGRATION AND DEVELOPMENT
One pure trade possibility is that neighbors together constitute a re-
gional market for certain goods that, for reasons of taste or excessive
transport costs, are not tradable with the rest of the world. If an RIA
includes all potential suppliers of such a good, it is equivalent to multi-
lateral free trade and hence, for small countries, is bound to be welfare
improving. This is a more demanding criterion than simply observing
that members do not import this good from the rest of the world, which
might just reflect existing tariffs. The situation is also relatively rare; in
1993, for example, out of a total of 4,858 types of goods, the four
MERCOSUR countries imported only 124 types exclusively from each
other (Amjadi and Winters 1999). These "local" markets were concen-
trated in agriculture and accounted for just half of 1 percent of total im-
ports. Enlarging the criterion to goods for which over 95 percent of im-
ports comes from other members adds only 155 types, or 2.2 percent of
total imports. Poorer countries may well have higher proportions of
their trade in such local markets, but even so, this does not seem likely
to be a significant potential benefit of regionalism. Moreover, even if
trade were restricted to close neighbors, that is not an argument for
forming an RIA: nondiscriminatory liberalization would have the same
beneficial effect.
Another important reason for forming a customs union between
neighbors is to reduce the transactions costs involved in border formali-
ties. These are often more important hindrances to trade than customs
duties and are far more likely to be avoidable between contiguous than
between distant countries. Even for neighbors, however, eliminating
these costs can be a complex process; the EU took from 1957 until the
mid-1990s to get close to having "invisible borders" between even a sub-
set of its members.
A third incentive for neighborhood customs unions is to ensure that
countries get their proper shares of tariff and other fiscal revenues when
goods are in transit or are transferred. (That is, the customs union con-
tains agreements to transfer tariff revenues from the country of collection
to the country of final destination. Such agreements can be contentious if
the partners do not trust each other fully.) These administrative consider-
ations, which are particularly important for landlocked countries, played
a major role in the origins of four long-standing African RIAs: UDEAC,
the CEAO, the EAC, and SACU. A customs union can also help avoid
the costs of smuggling or tax competition. Senegal has long sought to
form a customs union with The Gambia to remove the incentive for its
citizens to smuggle in imported manufactures from that country. Robson
(1998) estimates that for many years such smuggling generated a signifi-



MAKING THE MOST OF REGIONAL INTEGRATION      73
cant proportion (an estimated 25 percent in 1980) of The Gambia's rev-
enues from import duties-at Senegal's expense.
North-South or South-South RIAs
North-South RIAs mostly involve arrangements between developing
countries and the EU or, more recently, North America. Those two trad-
ing areas include a high proportion of the world's most efficient produc-
ers of many products, operate behind relatively low tariffs for manufac-
tures, and are capable of supplying the bulk of the needs of the southern
economies. To see the advantages this has for RIA formation, consider
the polar case of a good for which domestic supplies and imports from
partners and nonpartners are perfect substitutes.
The southern partner in a North-South RIA reduces its tariffs on im-
ports from a supplier large enough to satisfy many of its needs at little
more than the prevailing international cost. If it imports only from the
northern partner after the bloc is formed, its domestic prices fall to
northern levels and it benefits from increased consumption and reduced
production of high-cost domestic substitutes. In effect, in the commodi-
ties for which this is true, the South can enjoy gains from a North-South
RIA much like those from unilateral liberalization on an MFN basis. In
the language of appendix figure 2A.1 in the previous chapter, it is as if pA
and pB were very close together, so that losses from diversion (area H)
are small.
There are two critical issues here. First, how frequently and by how
much do the northern partner's costs exceed international minima? Al-
though the EU and the United States are generally efficient producers
and are subject to international competitive pressures, there are still
plenty of products for which they are not least-cost producers. Even if
their cost disadvantage were only, say, 5 percent, it would apply to a
high proportion of imports and so would represent a significant loss of
income relative to multilateral free trade. As early as 1977, Roemer
identified the tendency for metropolitan countries to sell far broader
ranges of goods in their colonies and former colonies than they were
able to do on world markets. This is the direct analogue of the 5 per-
cent margins.
Second, one needs to be confident that prices would actually fall to the
internal levels of the northern partner. The story above implicitly as-
sumes that the partner becomes the only supplier, but as we saw in chap-
ter 2, if imports do continue from third countries which continue to face
the tariff, domestic prices will riot fall, and there will be no gains. The



74    REGIONAL INTEGRATION AND DEVELOPMENT
southern country will just lose tariff revenue. But even if third-county
imports do cease, what makes northern firms cut their prices to their do-
mestic levels? The answer is, competition, and if anything occurs to cur-
tail competition between northern firms in the southern market, the price
cuts will be curtailed or wholly absent. Among the things that could cur-
tail competition are collusion (tacit or otherwise) not to compete in these
"captive" markets; economies of scale in selling (for example, consign-
ment size) so that it is not worthwhile for more than one or two firms to
sell in the southern market; and product differentiation. The last increas-
es the chances that the posttariff prices of northern goods will fall some-
what, but it curtails the effects of such price decreases on competing do-
mestic suppliers.
The southern government would, of course, lose tariff revenues-to
domestic consumers to the extent that consumer prices fell, and partly to
northern producers if their prices exceeded those at which imports oc-
curred before the RIA. As we saw in chapter 2, the former is just a trans-
fer within the southern country, but the latter represents a transfer of in-
come to the northern partner.
The most striking example of trade dependence between developing
and developed partners is probably NAFTA. In 1991, before NAFTA
was a real possibility, 70 percent of Mexico's imports came from the
United States, and for 614 out of the 4,854 trade headings in which im-
ports occurred, the United States or Canada was the only source. By
1996 the share of imports from the United States had risen to 78 percent,
but, in line with the general liberalization of the economy accompanying
NAFTA, the number of imports bought only from NAFTA partners had
decreased to 296. Without information on prices, it is impossible to
quantify the effects, but Mexico's continuing extensive imports from
nonpreferred suppliers makes it likely that its internal prices are signifi-
cantly influenced by the remaining tariffs against them. Moreover, be-
cause Mexico's tariffs are typically well above U.S. tariffs, Mexico's loss-
es of tariff revenue on imports are likely to far exceed its gains on its
exports to the United States. Thus, Mexico's gains from NAFTA in the
simple trade dimension are likely to be considerably less than those pos-
sible if tariffs were removed from all import sources.
North-South RIAs such as NAFTA and the Euro-Mediterranean
agreements are typically FTAs rather than customs unions. When these
RIAs do involve customs unions, as in the Turkey-EU agreement, the
southern partner is effectively obliged to adopt the lower northern tariff,
implying a substantial measure of trade liberalization. This is a further
source of gain.



MAKING THE MOST OF REGIONAL INTEGRATION    75
HOW MANY RIAs?
For an individual country, an attractive strategy may be to maintain
many RIAs simultaneously, especially with its main trading partners.
This would combine duty-free access to multiple markets with zero tar-
iffs on imports from multiple sources. Such a strategy, pursued by sever-
al countries in a given region, can result in overlapping RIAs and in what
is known as "spaghetti-bowl regionalism." We return to this at the end
of the section. Putting aside the complications associated with overlap-
ping RIAs and the very real danger that different FTAs have different ad-
ministrative requirements such as rules of origin (as discussed in the next
section), combining a number of RIAs could effectively substitute for free
trade. Several countries now pursue essentially this strategy. For exam-
ple, Chile is party to 12 trade agreements (APEC and LAIA, plus bilater-
al links with Argentina, Bolivia, Canada, Colombia, Ecuador, the EU,
MERCOSUR, Mexico, Peru, and Venezuela). Panama belongs to nine;
Mexico to eight (with eight more under negotiation); Bolivia, Costa
Rica, and Nicaragua to five each; and El Salvador, Guatemala, and Hon-
duras to four each. The picture is equally complex for Eastern Europe:
the Slovak Republic belongs to nine RIAs, the Czech Republic and Slove-
nia to eight, Estonia to six, and Hungary, Latvia, Poland, and Romania
to five each. RIA memberships in Africa are shown in figure 3.1, which
reveals a pattern of overlapping blocs.
The EU has a wide range of RIAs with European and Mediterranean
partners and, counting its nonreciprocal trade agreements, actually of-
fers tariff preferences to all but 10 of its trading partners (Winters 2000).
The results for Chile from a multiregional CGE model suggest strong
benefits from such "additive regionalism" (figure 3.2). Chile's current
agreement with MERCOSUR appears to have a potential welfare-
reducing effect, at least on the sirnple trade criteria applied here. An FTA
with NAFTA would be more attractive because of the better market ac-
cess. Indeed, all of the positive benefits from NAFTA are attributable to
the gaining of tariff-free access to the U.S. market for nongrain crops;
without this, Chile would lose. Joining both MERCOSUR and NAFTA
offers overall gains of 1.48 percent of GDP. This exceeds the sum from
the two separate FTAs by 0.87 percent, which is broadly a measure of
the extent to which liberalizing two sources of imports simultaneously
reduces trade diversion.4 Adding the EU and the rest of Latin America to
the cocktail adds value strongly, both through further reductions in trade
diversion and by increasing Chile's access to export markets. These esti-
mates should not be taken too seriously as absolute values, but the basic



Figure 3.1 Overlapping Blocs in Africa
CEMA C        Nile Basin        a
CC S oTm      rn~e      COMIESA        IGAD
ECOWA
CILSS Pen mae n InesaeCmne       nDogtCnrlI h         ae;CMS,Cmo             aktfrEsenadSuhr
0
Afnca: AC. Eat Africa CoprbnCCA    cnmcCmut fCnrl fia              tts   CWScnmcCmuiyo
z
0 olos0
Note. AMU, Arab-Maghreb Union; CBI, Cross-Border Initiative; CEMAC, Economic and Monetary Community of Southern Africa,
CILSS, Permanent Interstate Committee on Drought Control in the Sahel; COMESA, Common Market tor Eastern and Southern
Africa: EAC, East African Cooperation; ECCA, Economic Community ot Central Atrican States; ECOWAS, Economic Community of
West African States; IGAD, Intergovernmental Authority on Development, IOC, Indian Ocean Commission; SACU, Southern Atrican
Customs Union; SADC, Southern Atrican Development Community, UEMOA, West African Economic and Monetary Union.



MAKING THE MOST OF REGIONAL INTEGRATION           77
insight that-provided it is adnministratively straightforward-additive
regionalism could generate benefits is worth noting.
One problem with the incremental approach is that Chile's early
partners thought they were negotiating access to a market that was
effectively protected by 11 percent tariffs against everyone else, and they
were prepared to "pay" for that privilege with their own concessions.
Then they discovered that Chile was selling essentially the same deal to
other partners. In two cases this led to protests, which Chile addressed
by accelerating tariff reductions on the complainers' exports. The solu-
tion to partner complaints about: unforeseen liberalization was yet more
liberalization. This was a good outcome for Chile, but it is evident that
the threat of this kind of problem could undermine the whole approach.
Managing multiple RIAs is clearly not easy. However, if credibility and
administrative skills are high, if one's country is an attractive partner, if
the FTAs cover the bulk of trade, and if the cost of trade diversion in-
curred before a large number of FTAs is in place is not high, it could be a
useful strategy.
Figure 3.2 Potential Gains from Additive Regionalism for Chile from Joining Additional
RIAs
(percent)
9 00
8 00
7 00
6,00
5.00
X .0  BenefRs        
o 4 00
3 00                                                        _
2.00__
100
0 00
M           N          MtN       MiN+E       M+NdEtl.
fTA
Note: E = EU, European Union; L = Latin America; M = MERCOSUR, Common Market of the South;
N = NAFTA, North American Free Trade Agreement.
Source: Harrison, Rutherford, and Tarr (2002).



78    REGIONAL INTEGRATION AND DEVELOPMENT
Moreover when, in chapter 4, we look at the dynamic effects of RIAs,
we shall see that such effects might also arise from multiple RIAs. The best
example is "hub-and-spoke" regionalism. If one country (or group of
countries) has RIAs with a number of countries that maintain barriers be-
tween each other, then this hub country becomes the preferred location for
investment-firms can reach more markets tariff free than they can from
any of the other locations-and this will tend to bid up factor prices and
raise real income in the hub. The world's largest hub is the EU, which has
separate RIAs with nearly all other European and many Mediterranean
countries, most of which do not grant each other free trade.
A quite different situation arises if a country's multiple RIAs conflict.
All members of a customs union must set the same external tariffs. Yet
Bolivia, Colombia, Ecuador, Peru, and Venezuela are in the Andean Pact
(a customs union), while Colombia and Venezuela are also in the Group
of Three (an FTA with Mexico), and Bolivia has formed an FTA with
MERCOSUR. Similarly, as was shown in figure 3.1, Namibia and Swazi-
land belong to COMESA while also belonging to SACU, a customs
union with South Africa and Botswana, and Tanzania belongs both to
the SADC and to the EAC, a customs union with Kenya and Uganda.
Since these obligations are formally contradictory, it may be unclear
which will prevail in practice, and special conditions and exclusions may
have to be formulated. In addition, the administrative aspects of the var-
ious agreements tend to differ, notably as concerns rules of origin. The
outcome is a mass of complex rules that traders have to cope with (and
frequently seek to avoid).
Sub-Saharan Africa seems to combine the worst of the various ele-
ments examined. As shown in figure 3.1, most of its RIAs overlap with
others, and a number of countries belong to RIAs with contradictory ob-
ligations. Expending the region's scarce administrative resources on ne-
gotiating and managing multiple and complex RIAs is unlikely to be effi-
cient. Finally, the host of regional RIAs is unlikely to provide a
significant increase in market access or in trade liberalization for the
countries in the region, most of whose trade is with Organisation for
Economic Co-operation and Development (OECD) countries.
FREE TRADE AREAS AND CUSTOMS UNIONS
A central issue for countries planning to integrate their trade is whether
to choose a free trade agreement or a customs union. Of the 162 RIAs
notified to the GATT/WTO as of August 1998, 143 were FTAs, which



MAKING THE MOST OF REGIONAL INTEGRATION     79
have zero internal tariffs but no harmonization of external tariffs, and 19
were customs unions, which have a common external tariff as well as in-
ternal free trade.5 Under most circumstances, customs unions are more
efficient than FTAs and allow greater market integration, but they also
require more coordination and place tighter constraints on individual
member policies and sovereignty.
Trade Deflection and Rules of Origin: More Protection
A major worry for FTA members is trade deflection, the redirection of
imports from third countries through the FTA member with the lowest
external tariff. If unconstrained, this device reduces the effective tariff of
every member to that of the lowest plus the transport cost involved in in-
direct importing, which is wasted. The usual solution is rules of origin,
the apparently reasonable requirement that goods qualifying for tariff-
free trade should be produced in a member country rather than just pass
through member countries. In practice, rules of origin often become in-
struments of protection.
Rules of origin can lead to trade diversion when exportables between
FTA members are not wholly produced within a partner country but rely
partly on inputs imported from nonmember countries. Under rules of
origin, exports have to derive a certain proportion of their value from
local content or undergo certain production processes within the FTA to
obtain duty-free treatment. Trade diversion will result if the rules of ori-
gin create an incentive for producers in one partner to purchase higher-
cost inputs from another even though cheaper inputs can be had from
the rest of the world. Rules of origin can also artificially increase domes-
tic sales if the favored input is domestically produced.
Although trade between Canada, Mexico, and the United States had a
high regional content even before the three countries formed NAFTA
(Cordoba 1996), NAFTA's rules of origin have serious protective effects
in certain sectors, shifting their trade and investment patterns from
lower-cost to higher-cost sources. Most clothing produced in Mexico
gains tariff-free access to the North American market only if its inputs
are virtually 100 percent sourcecl in North America (WTO 1995). In the
automobile industry the origin requirement of 62.5 percent local content
has induced Japanese automobile manufacturers with plants in Canada
to invest to produce components in the United States rather than import
cheaper components from Japan. The rules of origin also require the
tubes in color televisions to be of North American origin if the televisions
are to receive duty-free treatment. Since the inception of NAFTA in



80      REGIONAL INTEGRATION AND DEVELOPMENT
1994, five television tube factories have been planned or established in
North America by Japanese or Korean firms, probably at the expense of
expansion in Southeast Asia (Stephenson 1996). Even tomato catsup is
affected (box 3.1).
Rules of origin effectively allow protection to be "exported" from
one member to another even if each country maintains its preintegra-
tion structure of external tariffs. The more restrictive the rules of origin,
the greater the scope for trade diversion on intermediate products. At
some point, however, the rules become so restrictive that producers opt
to source inputs from outside the free trade area and forgo duty-free
access.
Rules of origin also pose governance problems for developing coun-
tries. They take considerable effort to negotiate and are opaque and
complex to operate.' The EU's agreement with Poland has 81 pages of
small print in its rules-of-origin section, and NAFTA has about 200
(Krueger 1997). FTAs require controls on products crossing internal
frontiers to ensure compliance with rules of origin and the payment of
customs duties for noncomplying imports, and this can pose significant
administrative costs. Herin (1986) estimated these costs at 3 to 5 per-
cent of f.o.b. (free on board) prices for EFTA-EC trade. Documentation
and verification-and their costs-must be a continuing part of FTA
arrangements in order for importing countries to avoid tariff revenue
losses and keep the protection system effective. Finally, rules of origin
allow customs authorities-and individual customs officers-a good
deal of discretion. The administrative cost of ensuring that this discre-
tion is not abused is considerable, and the cost of failing to do so is
even higher.
Box 3.1 Rules of Origin Are Protectionist
To Chile's dismay, the rules for tomato catsup changed when the Canada-United States Free Trade
Agreement (CUSFTA) evolved into NAFTA. Under CUSFTA, catsup processed from imported tomato
paste qualified for duty-free treatment in intemal trade, but under NAFTA rules the tomato paste it-
self must be produced within a NAFTA member in order for the catsup to qualify for free entry. In
1992 Chile was the leading foreign supplier of tomato paste to the United States, and the catsup pro-
duced from the tomato paste enjoyed free entry under CUSFTA. Mexico and Chile together account-
ed for over 80 percent of U.S. tomato paste imports, in roughly equal quantities. Under NAFTA, cat-
sup made out of Chilean paste can no longer circulate duty-free, and Chile's share dropped to 5
percent, while Mexico's share rose to 75 percent.
Source: Palmeter (1993).



MAKING THE MOST OF REGIONAL INTEGRATION      81
Indirect Trade Deflection: Exporting Protection
Rules of origin may prevent the FTA member with the lowest tariff rate
from importing goods and sending them on duty free to another. But
they do not prevent a low-tariff partner from meeting its own require-
ments for a product from the rest of the world and then transferring a
corresponding amount (or all) of its own production to its partners. This
is termed indirect trade deflection, and its consequences for efficiency de-
pend on the capacity of the low-tariff country to supply all its partner's
needs. If it can meet the partner's entire import requirements at its own
tariff-inclusive price out of its own production, it will render the part-
ner's high tariffs ineffective and at the same time earn additional rents. In
the limiting case that every good fell into this class, indirect trade deflec-
tion would render a free trade area equivalent to a customs union that
takes the lowest preunion tariff as the basis for the CET (Robson 1998).
If, however, the low-tariff country cannot meet the full import needs
of its partner, the latter continues to import its marginal requirements
from the rest of the world, and the domestic price remains anchored to
the world market price plus its unchanged tariff. Then, although the ex-
porter enjoys additional rents-the high-tariff country's forgone duties-
there would be no beneficial effects on resource allocation in the high-
tariff partner, and real resources would be wasted as trade was diverted
from third-country to partner suppliers.
Unfortunately, there has been no empirical investigation of indirect
trade deflection within RIAs, but it has been observed working across
trade blocs. Trade deflection through the EU and thence into the United
States has modified, or even nullified, the effect of increased U.S. nontar-
iff barriers for textiles and clothing (Hamilton 1988).
Customs Unions Offer Lower Trading Costs and Greater Integration
Customs unions have common external tariffs and so do not formally
need rules of origin. Provided that they also define and enforce any non-
tariff protection measures at the bloc level, they can, at one stroke, avoid
all the administrative costs and distortions associated with rules of ori-
gin. Harmonizing nontariff barriers, however, is a demanding require-
ment; for example, throughout its first 30 years the EU allowed members
to maintain their own quotas on certain third-country imports such as
clothing, footwear, and steel and to use border measures to prevent those
goods from crossing internal borders (Winters 1993). The EU-Turkey
customs union allows the parties to impose antidumping duties on each



82    REGIONAL INTEGRATION AND DEVELOPMENT
other, which automatically requires the presence of border formalities
and rules of origin to define partner goods. Similarly, if, say, industrial or
safety standards differ between members, border controls or their equiv-
alent will be necessary for enforcing them.
In effect, a customs union needs to have not only a common external
tariff but also a trade policy that is common in all respects. Customs
unions are thus a good deal more complex to create than FTAs. Although
they offer greater market integration and lower costs, they also require
more ongoing coordination. Adopting a CET means reconciling the in-
terests of member states and then establishing continuing political
arrangements to deal with subsequent adjustments-for example, modi-
fications stemming from global trade talks, or the imposition of tempo-
rary safeguards, antidumping duties, or antisubsidy duties. Those
arrangements may imply substantial loss of sovereignty over trade policy
instruments and revenue sources. FTAs also need an initial reconciliation
of interests, effected through the character and restrictiveness of the rules
of origin that are adopted, but once these are settled, only relatively light
institutional arrangements are needed.
EXTERNAL TRADE POLICY
The net benefits of RIA membership depend directly on the external
trade policy stance. There are strong arguments for pursuing a policy of
external openness in conjunction with regional integration.
To begin with, trade diversion is more likely, and more costly should it
occur, the higher are external trade barriers. It is more likely because the
relative price differences created by preferential liberalization will be
greater with a higher external tariff, inducing trade diversion in more
sectors. And it is more costly, since a higher external tariff will provide
greater incentives for inefficient sectors to expand. Producers are able to
charge high prices (because the tariff protects them from world competi-
tion) and capture what was previously tariff revenue on trade between
members. Fundamentally, and just as for individual countries, the gains
from competition with low-cost suppliers-gains to consumers, gains
from developing an efficient industrial structure, and competition-
induced efficiency gains at the firm level-may be forgone if tariffs
inhibit such competition. These are arguments both for low tariffs, on
average, and for tariff schedules that are relatively uniform, avoiding
peaks. Very high rates in particular sectors are almost certain to produce
diversion, as in EU agriculture.



MAKING THE MOST OF REGIONAL INTEGRATION    83
A common complaint about such advice is that forming or deepening
an RIA causes adjustment costs and that simultaneous external liberal-
ization magnifies these costs to an unacceptable level. The problem with
this argument is that adjustment costs are only worth paying at all if the
adjustment is in the right direction. High tariffs or tariff peaks might in-
duce costly economic changes that move the country away from eco-
nomic efficiency. Besides, simulation studies of regional integration in-
volving developing countries and large high-income nations or blocs
(such as the EU) suggest that the adjustment costs associated with RIA
implementation are as high as those that would arise if trade liberaliza-
tion were implemented on a nondiscriminatory basis (Rutherford, Rut-
strom, and Tarr 2000).
These are compelling arguments for a genuinely liberal external policy
to accompany RIAs-but "should" does not always determine policy. An
important question is whether RIAs change the balance of incentives and
forces for external liberalism. If so, one could start to answer the ques-
tion of whether regionalism is a stepping stone to multilateralism or a
millstone around its neck. In part, the answer is determined by the inter-
actions between the RIA and outside countries (as discussed in chapter
8), but it also depends on the RIA's internal incentives and constraints on
trade policy.
As so often, there is no simple empirical regularity that would allow
us to resolve these questions. In 1994-15 years later than originally in-
tended-the Andean Pact finally set a common external tariff (which nei-
ther Bolivia nor Peru accepted). An average CET of 12.8 percent re-
placed average tariffs of 11.1 percent in Colombia, 10.2 percent in
Ecuador, and 11.8 percent in Venezuela (Echavarria 1998). Under
NAFTA, Canada reduced tariffs on around 1,500 items imported from
nonmember countries, while Mexico raised tariffs on 503. Since the
founding of the EEC in 1957, the average common external tariff on
manufactures has fallen from about 13 percent to about 3 percent in
2001 (following the Uruguay Round). Clearly, it is necessary to consider
the parts of the argument separately and in more detail.
Setting External Tariffs in an FTA: A Race to the Bottom?
FTA members decide their own tariff policies while keeping intrabloc
tariffs at zero. The interactions between members may reduce external
tariffs for three reasons. First, if tariffs on fellow members are con-
strained to zero, the optimal level on closely competitive goods from
third countries will be relatively low, to reduce trade diversion. Second, if



84    REGIONAL INTEGRATION AND DEVELOPMENT
there is trade deflection, high-tariff countries lose tariff revenue. If they
reduce their tariffs to just below the level of their partners, they can re-
capture the revenue without affecting internal prices or resource alloca-
tion. A series of such moves in a competition between members will tend
to lower external tariffs (Richardson 1995). Third, if duties on inputs
used to produce exports to other members cannot be rebated, high-
import tariffs render exporters of final goods uncompetitive. This con-
cern apparently lay behind Canada's decision to reduce 1,500 tariffs on
inputs in 1995, shortly after the establishment of NAFTA.
There are, however, three counteracting forces that can increase pro-
tection. First, rules of origin essentially export protection, allowing pro-
ducers in one country to benefit from protection in another. Second, the
result of fierce FTA competition in the intra-FTA segment of their busi-
ness might lead firms to seek protection against third-country imports in
another segment. This requires the firm to provide different goods or dif-
ferent varieties in the two segments, but that is not unusual. And it might
be reinforced by the third force for protection: as tariffs on partners are
fixed rigidly at zero, resources for lobbying against third-country im-
ports become more plentiful. The balance between these pro- and anti-
protection forces will vary from case to case, but there is at least some-
thing in the view that an FTA could encourage liberalism.
Trade Policy Institutions in Customs Unions Can Increase Protection
The situation in customs unions is quite different. Creating a customs
union provides an (unavoidable) opportunity to review tariff structures
and create new institutions for determining trade policy. National
tariffs must be harmonized at some agreed level, and in the process
international obligations-notably those toward the WTO-must be
respected. As explained in chapter 8, WTO rules are not a particularly
good guide to economic policy, and customs unions are best served
by using the domestic advantages of a liberal unilateral trade policy
as their guidebook. Tariffs should be low, and the number of rates very
few. Nonuniformity across products tends to increase administrative
costs and encourage discretion and corruption, so if the CET is more
uniform than the rates it replaces, there will be an additional benefit to
RIA members.
Unfortunately, the forces created by a customs union do not lead in
particularly liberal directions, although care in establishing the union
and its institutions can help. By coordinating their trade policies, mem-
bers of a customs union may be able to increase their negotiating power



MAKING THE MOST OF REGIONAL INTEGRATION     85
against the rest of the world. If they are able to negotiate effectively as a
bloc (which does not always happen), this will change the nature of
world trade talks, probably in the direction of greater protection (see
chapter 8).
In the most hegemonic of customs unions, SACU, South Africa simply
decided trade policy and compensated the smaller countries for the costs
it imposed on them. Similarly, Brazil dominated the determination of
MERCOSUR's external tariff (Olarreaga and Soloaga 1998). Large
countries are generally less dependent on international trade and more
prone to protection than smaller ones.
Even in more evenly balanced customs unions such as the EU, it may
pay to allow one member disproportionate power over certain negotia-
tions. If tariffs in the customs union and the rest of the world are strate-
gic substitutes (that is, if as one bloc increases its tariff, the optimal re-
sponse of the other is to lower its own), letting the more aggressive
member "lead" negotiations with the rest of the world on an issue will
result in higher customs union tariffs and lower tariffs abroad than
if the union negotiated collectively. In this outcome, the lower tariff
in the rest of the world increases welfare in the customs union, possibly
by enough to offset the costs of the union's own tariff. If, however,
tariffs are strategic complements, the customs union would gain by let-
ting its less aggressive member lead, resulting in lower tariffs and higher
welfare all round (Gatsios and Karp 1991, 1995). Unfortunately,
there are no general rules for determining which tariffs are complements
or substitutes (even in simple models), so it is not clear whether delega-
tion within the customs union would lead to higher or lower external
tariffs.
Moreover, once one recognizes that tariff setting is a continuing
process (a repeated game, in economic jargon), the situation becomes
even more complex. A more aggressive leader will be able to extract a
more favorable deal because its threats to retaliate (with the whole of
the customs union's resources) will be more credible. Whether this leads
to lower protection overall depends on whether a more aggressive cus-
toms union can achieve a more liberal outcome by virtue of its readi-
ness to retaliate or whether it actually has to use its retaliatory muscle.
The smaller the union, the less likely it is to achieve a liberal outcome
overall.
By convention, the EU allows countries disproportionate influence, up
to and including veto power, over policy in areas in which they claim
vital interests. Given that a country's interest in a sector is commonly
related to that sector's share in its GDP, it is easy to imagine this feature



86    REGIONAL INTEGRATION AND DEVELOPMENT
enhancing the interests of producers. The effect on a customs union's
trade policy depends in part on whether a sector's having a high share of
a member's GDP reflects comparative advantage or past policy distor-
tions. If the former, one might expect relatively liberal stances; if the lat-
ter, protection will be strongly defended.7 One encouraging aspect of this
situation is that because any trade creation will tend to relocate produc-
tion in a sector toward relatively more efficient members, over time this
could reduce protectionist pressure.
Even with genuinely intergovernmental decisionmaking in a customs
union, several features can lead to protectionist biases in the process of
aggregating individual members' preferences into a common policy. Pol-
icy will generally be made by bureaucrats and ministers representing
their own governments, and this can create a protectionist bias. The in-
centives for bureaucrats, who can reap no direct reward from the profits
they create, tend toward the protectionist (Messerlin 1983), and adding
layers of intergovernmental decisionmaking tends to swing influence
away from voters and toward official preferences for administrative con-
venience and a quiet life (Scharpf 1988). If there is no single political lo-
cation at which the costs and benefits of protection come together-such
as the presidency in the United States-the costs, which are usually
spread thinly and widely, will tend to be dominated by the more concen-
trated and obvious benefits, and a bias toward producers and protection
will be introduced. If policies are enshrined in the customs union's con-
stitution, as in the EU's Treaty of Rome, which defines the need for and
objectives of the Common Agricultural Policy, or if they acquire their
own bureaucracies-again, as in the case of EU agricultural policy-
interventions are legitimized and defended from within, and reform can
become very difficult (Winters 1997b).
Problems can also arise if the costs of policies are allocated across
countries in proportions different from their benefits. If a government
wishes to minimize the danger of its nationals having to bear the costs of
protection (consumer costs and higher factor prices) with no payoff for
its own producers, and if it believes that it may not be able to defeat the
calls for protection, it will tend to join protectionist coalitions and seek
protection for its own industries. The solution is to ensure that some in-
stitution at the level of the customs union (a hegemony or a central body)
takes responsibility for all the costs and benefits of the whole package
(see box 3.2).
Institutional responsibility for policies must be clear. In the EU the
struggle between the European Commission, representing the center, and
the national capitals for control of nontariff barriers to imports arguably
resulted in more active use of those instruments than the parties might



MAKING THE MOST OF REGIONAL INTEGRATION                    87
Box 3.2 Restaurant Bills, Universalism, and Protection
Institutions for making decisions about protection can impart a protectionist bias to the outcome.
Two simple but instructive models of institutional failure are the restaurant bill problem and univer-
salism.
The restaurant bill problem
Four acquaintances go to a restaurant and decide, before ordering, to split the bill four ways. Is it
surprising that the bill is higher than when each pays for his own meal?
Similarly, suppose a country's benefits from a policy regarding a product are proportional to its
share of the customs union's output of that product, while its costs are proportional to its share of
GDP. A production subsidy financed from general taxation will have this characteristic, as will,
approximately, a tariff, which benefits producers at the expense of consumers. If each country has
a veto, or if consensus is valued highly, countries sitting down to settle on a package of price-
increasing policies on several products will press for inclusion of any good for which their share of
production exceeds their share of GDP. As each country is likely to have several products for which
this Is true, the easiest package to build-and to sell at home to local pressure groups-will in-
crease the prices of nearly all products even if, overall, each country would have preferred "no
change" to the outcome actually agreed on.
Universalism
Imagine that protection for steel is being discussed and that each of three member states in a cus-
toms union produces one type of steel. If an, one type is protected, the government of the country in
which it is produced receives all the "benefits" (profits, employment, political convenience, and so
on), but each member bears some of the costs-which exceed total benefits, since protection is in-
efficient. Net costs are zero if the measure is rejected (Shepsle and Weingast 1981).
The difficulty facing a government in deciding how to vote is expressed in the position, "We oppose
this measure in principle, but if it passes, we want our share of the spoils. " The worst outcome is that
protection passes but the country's producers are on the outside. Thus, If governments fear that they
might not be able to stop a protectionist measure, they will vote for it, but on the condition that it
cover their own producers.
Simple arithmetic (Winters 1994) shows that this situation is quite likely to arise in small groups of
decisionmakers, and experience confirms this insight. Universalism drove the Introduction of the
U.S. Smoot-Hawley tariff of 1929, with individual congressmen acquiescing to protection for their
fellows' 'pet" industries in return for recipiocal treatment for their own (Schattschneider 1935). In
the EU, the United Kingdom opposed the inadequate McSharry agricultural reforms in 1992 but, in-
stead of being adamant, used most of its influence to ensure that large farms (of which it has rela-
tively many) were not excluded from Income support mechanisms (as McSharry had originally pro-
posed). In 1995 the EU Council of Ministers passed 92 of its 94 common trade policy decisions
unanimously. Given that trade policy typically.redistributes real income from one member to another,
this suggests that something like universalism was going on (Bilal 1998).
otherwise have wished. The commission had to prove that it could use
nontariff barriers to meet national objectives in order to win control of
them; although its measures were possibly less restrictive than national
ones would have been, they applied to all members rather than just one
(Winters 1993).



88    REGIONAL INTEGRATION AND DEVELOPMENT
Lobbies Bias RIAs toward Trade Diversion
Lobbying is a fact of life for trade policymakers, and creating an RIA
gives interest groups new threats to manage and new opportunities to in-
fluence policy. When industrial associations were asked for their views
on the FTAA, the Florida Citrus Mutual said that the proposed RIA
would "mean the end of the U.S. industry" and that "citrus products
must be exempted from further tariff cuts." The Rubber and Plastic
Footwear Manufacturers Association asserted that eliminating duties
would "cause havoc" and spell the death of that industry. Exempting
footwear "would have virtually no impact on any country's balance of
trade figures and would in no sense violate the [WTO requirement that
the FTAA cover] substantially all trade."8 When NAFTA was signed,
some U.S. industries were granted 15 years to adjust to free trade; they
included industries such as citrus that already had the highest protection.
In addition, new and more protective rules of origin were formulated for
textiles and clothing and for autos.
In 1998 UEMOA debated the details of its CET within a previously
agreed tariff structure of 5 percent on intermediate goods, 10 percent on
capital goods, and 20 percent on consumer goods. Aside from the addi-
tion of some temporary surcharges, one of the notable features of this de-
bate was the conclusion that cement was a consumer good worthy of 20
percent protection. In 1997 private sector protests led Madagascar to
postpone tariff cuts promised under the IOC (Lecomte 1998: 5). The
Andean Pact allows members to charge tariffs below the agreed CET
level-provided that there is no local producer of the good concerned
(Echavarria 1998).
One does not need much experience of the world to recognize the fin-
gerprints of lobbyists on these stories, and in fact, lobbying is a particu-
lar problem when forming an RIA. Not only does it lead to resistance to
internal free trade and liberal external policies, but it does so in ways
that make trade diversion more likely.
Governments respond to pressures from domestic interest groups to
increase popular support and their chances of reelection. Lobbying for a
sector or an interest provides a "public good" in the economist's sense
that the policies sought apply to all group members regardless of how
much they contribute to the lobbying effort. The larger the group of ben-
eficiaries, the more difficult it is to prevent people from free-riding (ben-
efiting without contributing), and the smaller the benefit that any indi-
vidual receives, the less likely he is to take the trouble to enter the debate.
For these reasons, consumers find it very hard to organize a lobby, and
lobbying is dominated by producers, who typically organize along



MAKING THE MOST OF REGIONAL INTEGRATION    89
sectoral lines. This effectively gives profits a double weight in the gov-
ernment's calculations: once as a source of income in the traditional as-
sessment of national economic welfare and again as a source of lobbying
support, which governments value in its own right. The heavy private
sector involvement in current regional plans-APEC and the FTAA-
shows how seriously firms take the opportunity to influence this element
of the business environment. RIAs are undoubtedly better for being busi-
ness friendly, but lobbying can bias an RIA too far toward producer
objectives.
Trade creation can be a mixed blessing for a negotiating government:
it generates surpluses for consumers at home and for exporters in the
partner country but reduces them for one of the main lobbying groups-
domestic import-competing producers. Trade diversion generates no
such reduction in profits, and although it yields fewer consumer gains,
that may matter less to governments concerned about reelection or about
keeping the support of business elites. If two governments can exchange
trade-diverting concessions, trade diversion becomes attractive politics
even if it is bad economics.
This bias toward trade diversion has two major implications. First, re-
gionalism may be attractive where multilateral liberalization is not.
Farmers and policymakers in the EU accept internal free trade as man-
aged through the Common Agricultural Policy but are not ardent sup-
porters of global rrade talks on agriculture. Brazil is keen to see MER-
COSUR have internal free trade in information technology products, but
it was the main holdout on the (nearly) global Information Technology
Agreement of 1996. Second, exceptions or delays in achieving free trade
tend to be concentrated in sectors with strong trade creation potential-
the very sectors that promise economic gains (Viner 1950; Grossman and
Helpman 1995). If the creation of RIAs is driven primarily by the indus-
trial attractions of trade diversion, regional liberalization cannot con-
tinue all the way to global free trade because the last step in that direc-
tion would necessarily generate only trade creation.
These dangers are elegantly demonstrated by Krishna (1998) for a
simple three-country world in which policy is determined solely by its ef-
fects on profits. He shows that, taking just two of the countries, the more
trade diverting is an FTA between them, the stronger its backing and
hence the more likely it is to come about. He then shows that following
the formation of such a two-country RIA, the backing for further (multi-
lateral) liberalization with the third country is reduced. An implication of
this outcome is that a multilateral liberalization which was feasible be-
fore the FTA might cease to be feasible afterward. Very simply, an ex-
porter might find any liberalization of foreign markets worth lobbying



90    REGIONAL INTEGRATION AND DEVELOPMENT
for, but having once achieved a regional liberalization, he might find that
the (possibly negative) incremental returns to lobbying for global liberal-
ization did not warrant the effort. Thus, if the world should attempt to
achieve multilateral free trade via regionalism, progress would stop at
the intermediate stage: regional stepping stones would leave us stranded
in midstream.9
Another factor favoring lobbying for an RIA rather than for MFN
free trade is uncertainty. A country's exporters can be confident that they
will benefit from an RIA partner's tariff concessions if the latter are re-
stricted to RIA members, even if they are inefficient. If the market is
opened to all comers, these exporters can be much less confident that
they will be the beneficiaries. Hence, although global negotiations will
open more markets than regional negotiations, for any given market
exporters are likely to lobby harder for regional than for MFN liberal-
ization.
It is sometimes argued that multinational companies are a bulwark
against protectionism because they operate on both sides of tariff bor-
ders. Multinationals do frequently seek lower barriers to their imported
inputs, and this can effectively discipline local pressures for protection.
But multinationals also frequently show strong protectionist instincts. In
the early 1990s, when the Europe Agreements were under negotiation,
West European motor vehicle producers penetrated Eastern European
markets in a fairly noncompetitive way, with one firm setting up in each
market. Their entry encouraged (and was allegedly conditional on) fierce
import restrictions. Similarly, as Brazil raised barriers to nonregional
production, world auto producers expanded their capacity there. Box
3.3 provides further evidence of the role of pressure groups in defining
MERCOSUR's CET and in obtaining exceptions both to it and to inter-
nal free trade.
RIAs Open a New Environment for Lobbying
Besides providing an opportunity to influence the new tariff structure it-
self, the creation of an RIA also alters the environment in which pressure
groups seek changes in trade policy. Unfortunately, we do not have much
of an idea about the direction in which this factor will push.
Some scholars argue that lobbying pressure is diluted by customs
unions. It costs more to lobby for a 1 percent increase in one's own tariff
in a customs union than in a single country because there is more oppo-
sition to overcome (de Melo, Panagariya, and Rodrik 1993; Panagariya
and Findlay 1996) or a larger number of representatives to influence



MAKING THE MOST OF REGIONAL INTEGRATION                      91
Box 3.3 Pressure Groups and MERCOSUR
A highly nonuniform common external tariff (CET), with its numerous exceptions, makes MERCOSUR
an ideal case study of the role of pressure groups in RIAs (Olarreaga and Soloaga 1998). Negotiation
of the CET culminated in the Ouro Preto Protocol of December 1994. Each member was allowed an
exceptions list. As shown in the table, the Initial CET applied only to about 75 percent of the universe
of 9,119 tariff lines. Left out were capital goods, computer and telecommunication equipment, auto-
mobiles, and sugar. Convergence to the CET is to be achieved for most goods by 2006, but there is
no agreement on convergence dates for sugar and automobiles. The Ouro Preto Protocol also estab-
lished a list of deviations from intra-MERCOSUR free trade. These were scheduled to disappear by
2000.
Olarreaga and Soloaga set about to explain variations and deviations from the CET across 27 Indus-
tries, using measures of political and interest group activity such as wages, industrial concentration
indices, labor-capital ratios, import penetration, intraindustry trade, and trade creation. Among their
findings were the following:
* Sectors where significant trade creation is likely tend to be exempted from internal free trade.
Internal free trade is also resisted more successfully by sectors with high employment shares.
This reflects government desire to avoid large-scale labor adjustments, as well as the voting
strength and trade union presence of large sectors. High third-country import penetration, which
puts competitive pressure on domestic firms, also leads to high protection.
* Political considerations explain 58 percent of the variation in the CET across industries. The ne-
gotiated CET in any sector is correlated with the share of capital remuneration in value added in
that sector and with the share of sector-specific capital in total inputs. If labor is fairly mobile and
new firm entry is difficult, existing capital receives most of the benefits of protection. The re-
searchers also found a positive correlaton with industry concentration: more concentrated in-
dustries find it easier to organize lobbying efforts and are thus more effective lobbyists.
* The CET in a given sector mainly reflects the preferences of the member country that has the
greatest producton in that sector. Thus, Brazil's preferences are the main determinants of the
structure of the CET, as that country represents at least 70 percent of MERCOSUR production in
each of the 27 sectors considered.
Deviation from the CET and from Intemal Free Trade among MERCOSUR Members
Deviation from the CET         Deviation from Intemal free trade
Share of total                       Share of total
Number of      tariff lines  Number of tariff      tariff lines
tariff lines   (percent)        lines               (percent)
Argentina     1,540           17              231                  2.5
Brazil        1,605           18               17                  0.2
Paraguay      2,101           23              293                  3.2
Uruguay       1,961           22              407                  4.4
Note: Several other deviations from the CET are not shown here: they include preexisting special promotion
regimes and the tax-free areas of Manaus, in Brazil, and Tierra del Fuego.
(Richardson 1994). Given the lower individual returns and the greater
difficulty of controlling free-riding in a larger, more diverse group, the
sum of member lobbying activity falls as a result of integration. The
lobby from, say, Senegal does not wish to devote resources to lobbying
for protection for producers in Togo. Such spillovers will probably



92    REGIONAL INTEGRATION AND DEVELOPMENT
reduce average protection, even if some members succeed in gaining
higher tariffs on some goods than they would individually.
All these analyses presuppose, however, that fragmented lobbies face
a unified customs union government, whereas the reality might be ex-
actly the opposite. The RIA's governments might be fragmented, while
some lobbies' power may be enhanced by integration (Winters 1993;
Bandyopadhyay and Wall 1999). For example, each member might start
with a lobbying game in which industry and agriculture more or less
cancel each other out. But if integration enables agriculture lobbies
to cooperate (because they produce the same things) while industry lob-
bies compete (because they produce different things), the customs union
may end up with agricultural protection. Overall, therefore, although
dilution effects will undoubtedly be present, they will not always
predominate.
An RIA that generates trade creation can create a liberalizing momen-
tum, and this will reinforce dilution effects. Industries that shrink be-
cause of lost protection eventually lose lobbying power. By squeezing rel-
atively inefficient sectors and promoting more efficient ones, regionalism
might ease future liberalization. Hathaway (1998) attributes the decline
in U.S. footwear protection since the 1970s to the sector's declining abil-
ity to lobby effectively as it contracted from 216,000 employees in 1960
to 58,000 in 1990. The industry association found it ever harder to ob-
tain import restrictions and instead started to focus on helping firms
adopt new technologies. In 1990 its president decided to "stop spending
one more penny or one more minute . .. on import restrictions," and the
association began admitting importers into its ranks. In 1991 it decided
not to oppose tariff cuts on footwear under the U.S. Caribbean Basin Ini-
tiative. Its reaction to the FTAA, however (cited above), suggests that
seeking special protection is a hard habit to kick.
Furthermore, an RIA may well concentrate lobbying activity against
outside countries. Once an RIA has been firmly agreed, there is no point
in lobbying for protection against partner suppliers. This may well turn
attention to third-country suppliers, and it also means that lobbying
resources will be released to make lobbying against those suppliers
cheaper than before. The result is likely to be greater pressure on exter-
nal tariffs (Panagariya and Findlay 1996).
It is certainly the case that lobbying activity has increased hugely in
Europe since the advent of the EU and that it has become much more in-
tensely focused on Brussels. The number of lobbying organizations in
Brussels grew from 300 in 1970 to about 3,000 in 1990 (Anderson and
Eliassen 1993). Expenditure on lobbying was approximately $150



MAKING THE MOST OF REGIONAL INTEGRATION     93
million in 1990 and was rising rapidly. By 1998, there were 13,000 pro-
fessional lobbyists in Brussels, approaching one for every European
Commission staff member (The Economist, August 14, 1998). Magee
and Lee (1997) offer one of the few formal attempts to quantify their ef-
fects, by analyzing external protection in France and Italy as the EC
widened and deepened between 1968 and 1983. The main effect was the
supposedly exogenous liberalization trend throughout the OECD coun-
tries, which broadly halved average tariffs, from 15 to 7.5 percent. In ad-
dition, however, Magee and Lee identified a 1.7 percentage point in-
crease in the average due to greater pressure from industries suffering
from trade creation and where mergers had increased concentration.
This rise was partly offset by a 1.1 percentage point decrease attributable
to the dilution of lobbying efforts as the political arena enlarged from the
national to the EU scale and as some industries grew rapidly as a result of
increased market size and efficiency gains.
It is not possible to say a priori whether, once it is formed, a customs
union will be more or less vulnerable to lobbying. Much will depend on
the institutions it adopts.
RIAs and Protection: Summing Up
Unfortunately there is too little empirical evidence to balance the argu-
ments on RIAs and protection accurately; whether RIAs will increase
or decrease protection remains an empirical matter. Foroutan (1998)
suggests that RIAs have not increased their protection levels recently,
but the evidence is not particularly strong. (These results are discussed
in box 8.3 in chapter 8, when we examine external pressures on RIA
trade restrictions.) Whatever the net effect, there are likely to be bene-
fits from establishing good trade policy institutions at the customs
union level.
Members should avoid excessively bureaucratic decisionmaking
methods, ensure that a single institution takes responsibility for the
whole package of measures, and refrain from writing protectionist poli-
cies into an RIA's basic documents. Even where members opt for an ex-
plicitly intergovernmental approach and eschew a central authority for
the customs union, they should establish a central body that is responsi-
ble for analyzing the collective interest in trade policy and widely dis-
seminating the results. This body should be so composed and constituted
that it will avoid equating benefits solely with production or identifying
high intrabloc trade shares with the collective interest.



94    REGIONAL INTEGRATION AND DEVELOPMENT
INTEGRATION AND TAXES
Many developing countries are heavily dependent on trade taxes as a
source of revenue; some African countries raise as much as half of gov-
ernment revenues this way.'0 Membership in an RIA erodes these rev-
enues-directly, as tariffs on intra-RIA trade are reduced, and indirectly,
when trade diversion occurs, as importers switch away from external im-
ports subject to tariffs. Revenue loss can also arise if a customs union sets
tariffs below a country's pre-RIA levels. As we saw above, the loss of
government revenue lies at the heart of the trade diversion argument: a
consequence of trade diversion is that revenue is transferred to partner
producers, causing a direct loss to the given country. If the government
has difficulty in mobilizing alternative revenue sources and is already
spending its revenue efficiently, then even losses that are transferred to
domestic consumers (as when tariffs on nonmembers fall), could be
costly. Formation of an RIA may therefore need to be accompanied by
steps to improve the take from domestic excise, sales, or value-added
taxes (VATs)."
How much revenue has typically been lost as a result of RIA forma-
tion? In practice, countries that are less dependent on trade taxes often
end up losing larger amounts of revenue. This paradox arises because
intra-RIA trade volumes are typically high in RIAs in which dependency
on trade taxes has been quite low, such as the EU, while countries with
higher trade tax dependency have tended to form RIAs with countries
with which they have relatively little trade.
There are, however, exceptions to this. Cambodia derived 56 percent
of its total tax revenues from customs duties prior to its entry into AFTA
(the ASEAN Free Trade Area), with two-thirds of these duties levied on
imports from ASEAN countries (Fukase and Martin 2000). Entry into
ASEAN provided a powerful stimulus for the introduction of a VAT in
early 1999. Turning to Africa, it is estimated that when UEMOA is fully
operational, it will roughly halve Senegal's tariff take, from CFA 100 bil-
lion to between CFA 39 billion and CFA 63 billion. Of this, the bulk is
attributable to reductions in external protection, with intrabloc trade
preferences accounting for no more than CFA 5 billion."2 In the SADC,
some of whose countries are heavily dependent on trade with South
Africa, substantial amounts of revenue are involved. Table 3.2 provides
estimates of the revenue cost of moving to free internal trade, which
would approximately halve customs revenue in Zambia and Zimbabwe,
costing the governments 5.6 and 9.8 percent in government revenue, re-
spectively. These are substantial revenue losses and point to the need to



MAKING THE MOST OF REGIONAL INTEGRATION               95
Table 3.2 Revenue Implications of a FreeTrade Area: Customs Duties before and after
Formation of the Southern African Development Community (SADC)
(percent)
Estimated change
Customs Duty as Share of
Member Country        Total Tax Revenue     In Customs Duty   In Total Tax Revenue
Malawi                     14.3                 -36.7               -5.3
Mauritius                  29.8 *               -18.2               -5.4
South Africa                3.6                   4.9                0.2
Tanzania                   24.0                 -8.3                -2.0
Zambia                     12.3                 -45.3               -5.6
Zimbabwe                   18.4                 -53.3               -9.8
Note: There are discrepancies between the customs duty revenues reported by customs departments and
those reported in budget numbers. For example, Malawi reported duty revenues for fiscal 1996 of 1,505.2
million and 2,028.7 million kwacha, but its customs department reported 615 million kwacha. For consis-
tency, we have used the numbers reported by customs. The projections assume that each country's aver-
age tariff rates against SADC members are zero.
Source: IMF staff calculations.
ensure that alternative tax systems are in place before eliminating sources
of trade tax revenue.
It is important to inquire what form revenue replacement takes, not
only because inefficient taxation reduces economic welfare but also be-
cause the net benefits of signing an RIA can depend critically on how dis-
tortionary taxes are. Under Chile's FTA with MERCOSUR, it is predict-
ed that VAT rates will need to be increased by half-from an average of
7.9 percent to around 12 percent-to maintain revenue neutrality (Har-
rison, Rutherford, and Tarr 1997). But in fact, Chile's VAT is fairly dis-
tortionary. If Chile were to improve collection and eliminate ad hoc ex-
emptions to create a fully uniform VAT, it could halve the nominal
average rate and increase real incomes by 0.3 percent while at the same
time making tariff revenue replacement cheaper and trade reform more
attractive. Rather different results apply to the Egypt-EU FTA (Konan
and Maskus 2000). If lost revenue is replaced by current discriminatory
taxation on capital, Egypt gains about 0.3 percent of GDP under the
FTA. Cleaning up the tax system, however, would yield total gains of 1.4
percent, and the marginal effect of the FTA is then to reduce welfare by
0.2 percent. So, an FTA may be better than current policy, but once the
gains from tax reform are reaped, it is actually harmful.
For poor economies, import tariffs represent one of the cheapest and
easiest ways of collecting government revenue, especially if the country
has only a few well-defined outlets for international trade. The observa-



96    REGIONAL INTEGRATION AND DEVELOPMENT
tion above that in most cases RIAs between such countries pose few
threats to revenue raises another intriguing possibility: perhaps small
countries can obtain a given tariff revenue at lower economic cost by
combining in an RIA than by acting independently. The tradeoff is
straightforward to describe (if not to quantify). An RIA between small
countries may offer benefits in terms of larger markets and more compe-
tition than nondistortionary tariff policy but at the same time will re-
quire higher taxes on imports from third countries.'3 If the former gains
outweigh the latter losses (plus, of course, the trade diversion to which
small-country RIAs are condemned), the FTA will help revenue-
constrained countries.
There are several caveats to this argument, however. First, if an RIA is
better than nondistortionary tariff policy and yet still entails an increase
in tariffs against the rest of the world, there are clearly significant gains
to be had by developing other sources of revenue, and this course should
receive high priority. Second, revenue tariffs, whether discriminatory or
not, should have different patterns from protective tariffs. If varied
across commodities, they will be high on those with inelastic demand for
imports and will have no correlation with whether there is local produc-
tion. Probably better from a practical perspective would be a single tariff
rate that is uniform across commodities. Uniform tariffs eliminate distor-
tions between different imports, significantly reduce administrative costs
(which were, after all, the reason for using tariffs to raise revenue in the
first place), and substantially reduce the scope for lobbying or corrup-
tion. Third, when there is no domestic production, a tariff is equivalent
to an excise tax (which is levied on all sources of a good, including do-
mestic sources). Then it is better to call the measure an excise tax to
make clear that should domestic production grow up behind the tariff
wall, it will be taxed at the same rate as imports. This will prevent the
emergence of inefficient domestic activity in response to the side-effect
protection associated with a revenue tariff. Fourth, as we have argued
extensively, even without revenue considerations there is no presumption
that an RIA between small countries will improve economic welfare; in-
deed, the balance of the argument is to the contrary.
Fiscal Compensation
A problem for many trade blocs among developing countries is that
some members account for disproportionate shares of the bloc's import-
competing production and hence of the tariff-free trade within the bloc.
This problem was evident in the CACM, where El Salvador and



MAKING THE MOST OF REGIONAL INTEGRATION    97
Guatemala accounted for the bulk of industrial production; the CEAO
(where Cote d'Ivoire is in this position); the East African Community
(Kenya); and SACU (South Africa). The other members, whose losses of
tariff revenues were translated into higher prices for industrial imports
from the more industrialized countries (that is, who suffered trade diver-
sion), felt that they needed compensation.
These concerns are magnified if trade policies, not just trade, are un-
balanced. An extreme example is the Lebanon-EU FTA, currently under
negotiation. In 1995 Lebanon's imports from the EU were 3,547 billion
Lebanese pounds, with an average tariff rate of about lS percent, and its
exports were 143 billion pounds, with an average tariff of perhaps 4 per-
cent.14 Potential revenue losses from the FTA are about 532 billion
pounds, or approximately one-sixth of government revenue, but the
gains on exports come to only about 5 billion pounds. Arguments for
similar discrepancies have been made for Mexico in NAFTA (Bhagwati
and Panagariya 1996) and for the South Africa-EU agreement (Teljeur
1998).
In some RIAs, such as NAFTA, compensation is explicitly ruled out in
the negotiation process. Where it is not, the only practical way to esti-
mate the warranted compensation in an existing RIA is to compare actu-
al revenue with what would be collected if members imposed the exter-
nal tariff on all their imports, including those from other members. This
is essentially how the SACU agreement handles compensation for
Botswana, Lesotho, Namibia, and Swaziland (the BLNS countries). But
such estimates are essentially accounting exercises that make no al-
lowance for the fact that the quantities and pretax prices of tradable
goods change as a result of an RIA-as does welfare. Compensation
based on forgone revenues represents a political settlement to allow for
the most obvious sources of disparity in the costs and benefits of RIAs,
using a transparent and analytically plausible formula. But no simple for-
mula can provide more than a crude approximation for offsetting the
disparities in costs and benefits in a broader appraisal of an RIA. Nor, in
the face of economic change, can any such formulas maintain the balance
of advantage initially agreed by political negotiation.
One broader argument that governments sometimes make in assessing
an RIA is that protection helps maintain value added and employment in
manufacturing and thus that irnports from partners are economically
costly even when they displace higher-cost domestic output. This was
among the arguments that led to SACU's inclusion of an enhancement
factor in its revenue distribution formula, which initially increased com-
pensation about 40 percent above estimated revenue losses.1" The plausi-



98    REGIONAL INTEGRATION AND DEVELOPMENT
bility of such development benefits and the role of RIAs in achieving or
frustrating them are examined in the next chapter.
Two major problems beset such compensation arrangements and
make it difficult to devise mutually acceptable compensation formulas.
First, although the transfer benefits accrue to the private sector (which
may be partly foreign owned), compensation has to be provided out of
public revenues by authorities who often face practical constraints on
revenue raising. Second, the costs that are compensated are direct, imme-
diate, and obvious, while any wider benefits from integration tend to be
diffuse and uncertain and accrue only in the longer term. The failure of
compensation has often been accompanied by failure of the RIA itself, as
happened to the East African Community and the CEAO, or by the with-
drawal of members-Chad from UDEAC and Honduras from the
CACM. In the light of this experience, unless prospective partners are
structurally balanced, both actually and potentially, the formation of an
RIA is highly risky without the simultaneous adoption of low external
tariffs to minimize the transfers implied in duty-free intra-RIA trade.
Low external tariffs, of course, also increase economic efficiency.
Tax Competition
Except for the very largest countries, RIAs constrain members' fiscal dis-
cretion and sovereignty. This is most obvious for customs unions, where
the CET transparently removes tariffs from the policy domain of any sin-
gle country. But there will also be constraints on a wider set of instru-
ments in both customs unions and FTAs. Measures to minimize transac-
tions costs, maximize the tax base, and avoid smuggling will all create
pressures for a degree of indirect tax harmonization.
We have already seen how these factors have stimulated the formation
of some RIAs as countries have sought to manage their neighbors' tax
policies and halt smuggling by eliminating fiscal borders. Here we con-
sider the opposite side of the coin, in which constraints are unwelcome
side effects of RIAs signed for other reasons.
If goods flow unimpeded across an RIA's internal borders, govern-
ments may be tempted to lower indirect tax rates to capture revenue
from cross-border shopping. This tax competition could lead to a wide-
spread reduction in indirect tax rates. Such an outcome could be wel-
come in that it brings about a downward convergence of effective tax
rates and leads to efficiency gains by reducing consumption distortions.
An alternative view, however, is that disregarding such fiscal spillovers
leads to a welfare loss for the RIA as a whole because tax receipts and



MAKING THE MOST OF REGIONAL INTEGRATION     99
total expenditure fall below optimal levels. This view is frequently ar-
gued by the larger continental members of the EU; box 3.4 gives an illus-
tration of the same problem in classical times.
Tax harmonization has been a major issue within the EU over many
years, generating fierce controversy and even at times threatening to de-
rail the Single Market Programme. Recognizing that some harmoniza-
tion is indispensable to the operation of a single market, members have
agreed to adopt a minimum standard VAT rate of 15 percent, with no
more than two reduced rates. No agreement, however, was reached on
precisely which goods and services fall into which tax bracket or on the
use of higher than standard rates. Efforts to harmonize excise taxes on,
for example, alcohol or gasoline have formally been even less successful,
although tax competition driven by demands from users or producers
has induced a measure of de facto harmonization.
Fiscal discretion has long been an issue in SACU, whose smaller
members have virtually no role in determining tariffs or other indirect
taxes in the customs union area. But even where they formally retain
the power to set rates in a closely integrated customs area, the pressures
to harmonize VAT rates at or close to South African levels are so strong
that the BLNS countries still have little control over their indirect tax
regimes.
Outside the EU and SACU--for instance, in UEMOA and UDEAC
(now CEMAC)-tax harmonization has not been successful. For devel-
oping countries, the moral is to not underestimate the political difficul-
ties of tax harmonization-and to recognize that trying to eliminate mild
differences across borders may be more trouble than it is worth. In the
United States and Canada, for example, differences in state and provin-
cial sales taxes of the order of 5 percentage points are an irritant rather
than a major problem.
p    s in e a! )niE7htgportgwsest, I  ., lie  nali yiefeidyaway Th-odesiiU
Ii  U.  S     S "        S~~---u U gg   -_
mc tliTi ofl t 8hrbor cevenuesiia sStgifcan aniartlypa gbedcastI h iipp
~8h   *hd ueafjsc xpjroc~e!dsIt  __eaerectLAdS



100     REGIONAL INTEGRATION AND DEVELOPMENT
NOTES
1. Srinivasan (1997) attempts to calculate such external tariffs, but under very
restrictive circumstances.
2. Recall that free trade agreements have internal free trade, while customs
unions have both internal free trade and a common external tariff (trade policy)
against nonmembers. The section "Free Trade Areas and Customs Unions" in
this chapter contains a fuller discussion.
3. Note that these corner solutions are a necessary but not sufficient condition
for the results to hold.
4. The benefits of access to MERCOSUR and to NAFTA markets are basical-
ly independent of each other, hence the change in the overall benefits must reflect
the combination of the import effects.
5. Of course, many RIAs contain other elements as well; this classification
refers only to their policies on trade in goods.
6. Agreement on the rules for clothing and autos was a major issue in NAFTA
(Krueger 1997).
7. A sector with comparative advantage may prefer high protection so that it
can reap high rents on sales within the customs union, but at least it could sur-
vive with lower protection. It may also seek liberalization in other import sectors
to reciprocate for any liberalization that it seeks in its export markets.
8. International Trade Reporter 15 (31, August 5), 1998.
9. Models in which citizens vote directly on their own interests can generate a
similar outcome (Levy 1997), especially if the partners are similar and relatively
capital abundant compared with the rest of the world (for example, the EU).
10. Trade taxes include export and import taxes, trade monopoly profits re-
mitted to governments, and taxes and official profits on foreign exchange trans-
actions.
11. Not every MFN tariff cut reduces tariff collections. The initial duty may
be above the revenue-maximizing rate, especially when exemptions and evasion
are taken into account. Thus, revenue replacement is not always as daunting as
may at first appear.
12. These estimates, from a CGE model, are made as though UEMOA had
been fully operational in 1996. They vary according to whether quantities of
trade are allowed to change in response to the RIA and, if so, according to what
assumptions (Ng and Winters 1998).
13. Relative to the pre-RIA position, the partners import less from third coun-
tries (trade diversion) and need to make up the revenue that is forgone by raising
tariffs on third countries.
14. Data are from Martin (1996); Fuleihan (1997); and Moukarbel (1997).
15. A further argument was that SACU lacks any institutional arrangements
for controlling administrative interventions by South Africa that hinder the in-
dustrial development of the smaller member countries, or for adjudicating dis-
putes relating to the operation of the customs union.



CHAPTER 4
Stimulating Investment
I nvestment is a key component in economic development and has be-
come one of the main objectives of countries pursuing regional eco-
nomic integration. The logic is that larger markets, more competition,
and improved policy credibility will increase the incentives for investment
and by that means raise incomes both directly, by increasing the capital in-
tensity of production, and indirectly, by encouraging technical progress.
These arguments are relevant to investment from all sources, but they are
applied particularly often to regionalism as a means of attracting foreign
direct investment (FDI). In NAFTA, for example, stimulating FDI flows is
an explicit objective. This chapter briefly describes the policies of trade
blocs toward investment and asks whether the arguments advanced for
such policies having positive effects are justified.
The first section explores explicit investment policies. In early RIAs,
these policies were almost always activist and interventionist, coopting
regional integration into regionwide import substitution. Such policies
almost completely failed and have been superseded by a much more
market-friendly approach that places more emphasis on policies guaran-
teeing the fair treatment of investment. These guarantees are often em-
bodied in bilateral investment treaties or, where there are trade and other
links, investment chapters in RIA agreements. They typically foster nega-
tive integration-that is, they preclude certain policies rather than re-
quire policies that actively encourage investment-but they may play an
important role in facilitating investment flows.
We then examine the common argument that RIAs add credibility to
government policies other than investment rules and thus help increase
101



102    REGIONAL INTEGRATION AND DEVELOPMENT
investment and attract FDI. We argue that South-South RIAs are unlike-
ly to provide much added credibility and may in fact have the opposite
effect, especially if they are not accompanied by liberalization of trade
with the rest of the world. North-South RIAs, by contrast, can enhance a
southern country's credibility, but typically only if the RIA is likely to im-
prove economic performance in its own right and if the large northern
partner is willing to enforce investment-encouraging "club rules." The
latter is more likely to be true if the policies on which a developing coun-
try wants to gain credibility are specified explicitly in the agreement.
Of course, investment does not depend only on explicit policies. An RIA
will stimulate investment if it raises expected returns or lowers costs. Clas-
sic economic analysis predicts that a South-South RIA should have no im-
pact on returns to capital but that a North-South RIA will increase the rel-
ative price of exportables in both countries, thus raising wages and
decreasing the return to capital in the South and causing investment there
to fall. More recent analysis holds that the rate of return on capital (and
investment) could well rise in both countries regardless of capital abun-
dance. RIAs typically reduce the transactions costs of tradables more than
those of nontradables, so if tradables are more capital-intensive than non-
tradables, RIAs will increase the demand for capital and raise its rate of
return. Lower tariffs and trading costs for capital equipment can also re-
duce the price of investment goods, increasing the rates of return and ac-
cumulation. Furthermore, integration may result in a more efficient fi-
nancial sector, especially for a southern member of a North-South RIA,
reducing lending margins and the cost of funds and increasing investment.
Unfortunately, few empirical studies of the impact of RIAs on invest-
ment are available; most trade blocs are so new that the data are simply
not there. Ex ante simulations show, however, that increased investment
is most likely in North-North RIAs, somewhat less likely in North-South
RIAs, and least likely in South-South RIAs. Where we do have ex post
evidence on investment, it tends to suggest mildly positive effects, but
there seems to be no evidence that this translates into higher economic
growth. Firmer evidence is available for FDI, which seems frequently to
boom after an RIA is signed. We explore this and explain why it is likely
that RIAs stimulate inflows of investment from nonmember countries
but have ambiguous effects on intrabloc flows.
Throughout the chapter, we argue that general policy reforms such as
sound macroeconomic policies, well-defined property rights, and
efficient financial and banking sectors are likely to be far more important
in influencing investment and FDI than merely joining an RIA. Reforms
such as macroeconomic stabilization, market liberalization, and privati-



STIMULATING INVESTMENT    103
zation should raise the returns to all factors and foster investment. Re-
gional integration will promote investment if it significantly enhances
policy credibility and increases market size, but only if it is accompanied
by good policy overall.
The discussion of investment leads naturally to questions about the lo-
cation of industry and economic growth. These topics are dealt with in
chapter 5.
I 4VESTMENT POLICIES
This section traces the evolution of investment policies in RIAs since the
early days and discusses current policies embodied in bilateral and multi-
lateral agreements.
Investment Planning: A Dead End
Until about 1980, theories of economic development placed great store
on the stimulation of investment: and on the role of the public sector in
directing it (and often financing it). The prevailing theories of import
substitution called for governments to plan industrial structure carefully
and to ensure that the untoward effects of competition did not disrupt
the search for economies of scale. Naturally, regional economic integra-
tion was seen as a potential tool for this endeavor. Recognizing that indi-
vidual national markets were too small to support large-scale industries,
policymakers planned to meet regional demand through one or a few
large regional projects instead of numerous smaller national ones. Thus,
for example, the 1982 Preferential Trade Agreement, an RIA of eastern
and southern African countries (and the predecessor of COMESA, which
was created in 1993), called for member states to cooperate to promote
collective self-reliance and complementary industrial development and to
expand trade in industrial products.
The idea was that if countries desired a given level of industrial activity
but were indifferent about its precise composition, an RIA would reduce
the cost of achieving it. The RIA would allow each member to specialize
in particular industries and reap the economies of scale associated with
supplying the whole RIA rather than have to maintain a wider set of
smaller industries to serve just their national markets (Cooper and Massell
1965; Johnson 1965). As we saw above, there is nothing wrong with the
logic that larger markets allow greater efficiency, but application of the
theory proved top-heavy. Planners looked to intraregional specialization



104    REGIONAL INTEGRATION AND DEVELOPMENT
to support their comprehensive regional development plans, which typi-
cally put the pursuit of scale economies above all other objectives. Na-
tional industrial development plans restricted entry and were coordinated
across countries to avoid duplication and the underutilization of industrial
capacity. This regional "rationalization" of industry was, effectively, just
infant industry protection and import substitution on a regional level.
These policies were almost universally failures. RIA members were
often unable to agree on industry location.' In one of the most notorious
examples, C6te d'lvoire, Ghana, and Togo agreed in 1976 that they
could reap economies of scale if Togo built a single cement plant to serve
the region. The countries disagreed, however, on how to manage the in-
dustry, and each country ended up with its own cement plant supplying
its own domestic needs (Robson 1998). Agreeing to forgo industrial sec-
tors so that partner countries can establish them imposes tangible and
immediate costs on governments in return for uncertain benefits in the
future, when some other project is assigned to them in turn. Understand-
ably, without a history of trust and experience in working together, and
without any great confidence in the market mechanism, governments
were reluctant to do this.
The sorry experience of managed industrialization, the evolution of
economic orthodoxy toward reliance on markets, and the increasing mo-
bility of international capital have combined to move most recent RIAs
(and reincarnations of old ones) away from outright industrial planning.
Instead, countries have looked at integration as a means of generating in-
vestment by making the market more attractive and by improving policy
credibility. In addition, many have sought to alter the environment for
FDI directly by changing the regulations governing it, either on an MFN
basis or just within the bloc. In fact, many countries have signed bilater-
al investment treaties independent of trade agreements.
Bilateral Investment Treaties
Bilateral investment treaties (BITs) have become an important part of the
regulation of worldwide investment flows.2 There are now well over 2
thousand BITs linking countries on all continents and at all levels of de-
velopment. They are important in their own right and as models for and
precursors to regional investment treaties, most of which are part of
broader RIAs. BITs are generally short agreements that provide "shallow"
or "negative" integration for investment-that is, they seek to remove bar-
riers and alleviate uncertainty affecting investment rather than stipulate
positive actions to encourage it.3 They are almost always reciprocal and



STIMULATING INVESTMENT    105
typically contain sections dealing with their scope of application (defini-
tions of investment, nationality, and so on), the admission of investment,
general standards for the treatment of investment once it has arrived, and
dispute settlement. Most also contain a number of specific provisions.
Most BITs explicitly defer to domestic law on the admission of invest-
ment; governments retain almost complete discretion to manage the sec-
tors in which FDI occurs and the shares of ownership that foreigners
may hold. The United States has generally been more progressive, insist-
ing, in most of its BITs, on national treatment on the admission of
investment (no less favorable treatment for U.S. residents than for do-
mestic residents), subject only to explicit specific exceptions. During the
1990s, when FDI assumed a much more prominent role in development
thinking, this approach began to find favor elsewhere, as well.
On standards of treatment once investment has arrived within the
host country, virtually all BITs call for fair and equitable treatment, secu-
rity of ownership, and freedom from unreasonable or discriminatory re-
strictions on the operation of investments. Nearly all guarantee MFN
treatment (the principle that no foreigner should be treated more favor-
ably than the partner country's residents), although, as with trade, this is
subject to exceptions permitting even more favorable treatment of part-
ners in an RIA. Most BITs also specify national treatment, and, in fact,
the treatment of foreigners is often more favorable than for domestic res-
idents-for example, where domestic residents face foreign exchange re-
strictions but foreign firms effectively do not. Nearly all BITs contain
provisions on the transfer of funds associated with FDI and on expropri-
ation. Here, too, the United States has led the way in introducing provi-
sions that restrict the imposition of performance requirements and
permit the international mobility of key personnel.
Dispute settlement provisions vary but usually define arbitration pro-
cedures that in recent BITs are often based on international standards
such those of the World Bank Group's International Centre for Settle-
ment of Investment Disputes (ICSID). An interesting aspect of the provi-
sions is that they sometimes provide for private entities in one partner
country to take action against the government of the other.
Treatment of Investment in Current Regional Arrangements
At the regional level, some treaties are just multicountry extensions of
BITs-that is, freestanding investment agreements. These include, for
example, the 1970 Agreement on Investment and Free Movement of Arab
Capital among Arab Countries, and the 1993 Colonia Protocol on the



106    REGIONAL INTEGRATION AND DEVELOPMENT
Promotion and Reciprocal Protection of Investment within MERCOSUR.
Others, of more interest to us, form part of broader RIAs that include trade
and sometimes other provisions. For example, the EU, the Andean Pact,
LAIA, NAFTA, and COMESA all have investment provisions.
The most far-reaching investment provisions are those of the EU,
where the objective has been to create a common market with a single
market for capital and investment. Restrictions on investment and the
movement of people are long gone, and the remaining frictions are in
areas such as differences in taxation and company law. The European
Single Market Programme sought to address some of these issues, but the
process is not yet complete. An important part of the integration of cap-
ital markets in the EU is its common competition policy, which restricts
the worst excesses of investment incentives (known as state aids).
Just as it has been very progressive in its BITs, the United States has been
instrumental in advancing the treatment of investment in RIAs. NAFTA
contains a deep and innovative investment chapter, especially considering
that it is only a free trade agreement rather than a customs union. It has
become a model for other associations such as the Group of Three and for
APEC, with respect to its nonbinding investment principles.
NAFTA provides for national treatment in establishment; MFN treat-
ment in establishment and operation; a ban on new performance re-
quirements and a phaseout of old ones; guarantees of convertibility at
market-exchange rates of funds for repatriating profits, disinvestments,
and so on; and a ban on expropriations except for public policy reasons
on a nondiscriminatory basis and with full compensation. This is a far-
reaching menu, even though the signatories made a number of excep-
tions in the annex to the investment chapter, including general national
security exceptions and sector- and country-specific ones. NAFTA also
has extensive dispute settlement provisions that permit private action
against governments.
In 1994 APEC members agreed on an even more extensive set of prin-
ciples governing investment but were able to do so only on a nonbinding
basis. These principles basically represent a goal toward which national
regimes might seek to evolve in the fullness of time. They offer no con-
crete protections at present and so probably have only a very slight (psy-
chological) effect on intra-APEC investment flows.
Multilateral Investment Agreements
Many bilateral and multinational agreements other than RIAs embody
national treatment. Prominent among them are the OECD's Codes on



STIMULATING INVESTMENT   107
Liberalization of Capital Movements and Current Invisible Operations,
and its National Treatment Instrument, which requires that nonresident
enterprises which have been permitted to establish in host countries be
treated in the same manner as domestic firms. A number of developing
countries have agreed to sign these codes. Recently, the OECD tried to
liberalize investment further and establish binding dispute settlement
procedures in the Multilateral Agreement on Investment (MAI). This ini-
tiative failed both because the OECD countries turned out not to be
ready for such a step and because, for a variety of reasons, the MAI
aroused substantial opposition among nongovernmental organizations
(Henderson 1999). Investment policy has also been proposed as a subject
for future WTO negotiations under the Doha Development Agenda, but
only if all members first explicitly agree on the form and content of those
talks at the WTO ministerial meeting in 2003.
INTEGRATION AS AN AID TO CREDIBILITY-NOT AN
AUTOMATIC EFFECT
A country intent on restoring or accelerating growth must improve its
economic management, but this may not be enough if it suffers from low
credibility as a result of, say, a history of bad policies. Potential investors,
domestic or foreign, are likely to be suspicious of the government's stat-
ed intentions if they have been burned in the past by sudden tariff
changes, tax increases, or nationalizations. Thus, the benefits of econom-
ic reform will be slow to arrive if credibility has to be built up over time.
In fact, reform may well fail without credibility: investors may not re-
spond, may respond perversely because they anticipate reversal, or may
launch speculative attacks against the reforms, and interest groups that
lose under the reform may attempt to reverse it.
In developing countries, poor credibility can stem from a number of
factors. Governments driven by interest groups or reelection pressures
may be tempted to reverse reforms, or elections may bring a party
opposed to the reforms into power. Kleptocratic regimes can make op-
portunistic raids on investors, and ethnically riven societies may pursue
destructive redistributions. Unless governments can make a convincing
case that they will refrain from such actions, investment is likely to re-
main low.
But it is difficult-and takes a long time-for a country to raise its
standing by itself. When reform is hampered by a lack of faith, govern-
ments must try to reduce uncertainty, and one way of doing this is to



108    REGIONAL INTEGRATION AND DEVELOPMENT
anchor reforms through a credible binding commitment. Estonia, for
example, with no track record to appeal to, established a currency board
to take monetary policy out of the government's hands.
RIAs Allow Bad Policy to Be Punished
Another route might be to sign an RIA that locks in reforms by changing
the incentives for bad policy. Such an outcome may come about either
because the RIA increases the rewards for good policy or the costs of a
bad policy directly or because it permits "punishment" by other RIA
members if the country breaks "club rules" (Fernandez and Portes
1998). In the latter case the partners must have the power and commit-
ment to enforce the necessary reforms. The partners need to be large and
stable enough, and to have a sufficiently strong interest in the RIA, to
make it worth their while to discipline the target country. This is a diffi-
cult combination to achieve, for the larger and more stable a country is,
the less likely it is to depend on any particular RIA for its prosperity.
A partner country is likely to be more concerned about a neighbor
than about a distant partner, and more willing to serve as a policy anchor
for it. Mexico's economic performance is more important to the United
States than Argentina's because Mexico is a more important trading
partner and Mexican social and political instability can affect perceived
U.S. welfare directly, especially via migration (see Francois 1997 and box
4.1). Similarly, the EU is likely to be more concerned about Poland or
Hungary than about Pakistan or Zimbabwe.
A large country can also support integration among its former
colonies. French sponsorship gave the two African monetary unions in
the CFA zone, which have survived for more than four decades, enough
credibility to resist a needed devaluation for many years. Similarly, the
British Colonial Office gave SACU, which goes back over 80 years, cred-
ibility for its first half-century; later SACU became a kind of North-
South RIA dominated by a hegemonic power, South Africa.
If reforms are clearly stipulated in an RIA treaty and a clear and
credible punishment incentive exists, a North-South RIA can provide a
strong commitment mechanism and raise the credibility of the reforms.
By writing the reforms into an RIA agreement, subsequent punishment
is given a formal legal basis, and if the reforms also affect the partner
countries' welfare, their imposition of punishments is legitimized politi-
cally.4 The obvious cases in which treaties define policies include trade
policies between the partner countries in an FTA, external trade policies
in a customs union, and perhaps aspects of investment policy such as



STnMULATING INVESTMENT                109
Box 4.1 NAFTA and the Credibility of Mexico's Policy Reforms
Did NAFTA increase the credibility of Mexico's policy reforms? One way to check is to look at Mexi-
co's behavior following its 1982 and 1994 debt crises. The 1982 debt crisis led to an enormous
increase in economic intervention by the government, which nationalized the banking system and
put in place foreign exchange controls, a universal regime Of import-licensing requirements, and
controls on foreign investment.
Mexico started to liberalize its economy around 1985, dismantling its import-licensing regime
and reforming its foreign exchange restrictions and its laws on foreign investment and intellectual
property. It acceded to the GATT in 1986, and it established formal consultation mechanisms on
trade and investment with the United States well before the establishment of NAFTA.
In 1985 Mexico's average tariff was 18.5 percent (100 percent for some products), and its aver-
age tariff on consumer goods was 45 percent. After it joined the GATT, its bound tariff rates fell
significantly, and its applied average tariffs by sector fell to between 4 and 13.1 percent. Tariffs were
further capped by Mexico's entry into NAFTA in 1994. Thus, through NAFTA and the GATT, Mexico's
govemment severely limited its ability to raise tariffs and to dismantle its trade and investment re-
forms.
Whereas the debt crisis of 1982 led to nationalizations, exchange controls, and a dramatic
increase in protection to deal with the balance of payments crisis, the same did not occur in 1994.
With the support of multilateral agencies and the United States-Mexico's largest trading partner by
far-intervention was aimed at stabilizing the economy without closing it. And although protection
did increase on some non-NAFTA imports, the general thrust toward openness was not heavily
compromised.
What explains the different reactions to the two crises? Some have argued that the shift in the
prevailing intellectual climate toward liberalization was sufficient (Bhagwati and Panagariya 1996).
This was clearly Important, and possibly necessary, but equally plausible is that Mexico had locked
in its policies through accession to the GATT and to NAFTA in a way that made reversal very costly.
national treatment. Agreements might also include various deep inte-
gration measures and so generate credibility in these areas, as well (see
box 4.2).
What RIA treaties typically do not include is macroeconomic and
general domestic policy constraints. Whalley (1998b) argues that a de-
sire to increase the credibility of domestic reforms was central to Mex-
ico's negotiations on NAFTA. "Mexican negotiators were less con-
cerned to secure an exchange of concessions. . . . The idea was clearly
to help lock in domestic policy reform through this process" (71-72).
But, in fact, NAFTA        covered neither macroeconomic policies (which got
completely out of hand in the run-up to the implementation of the
agreement in 1994) nor privatization and deregulation, which actually
have moved rather slowly in Mexico. Indeed, one might argue that the
intense focus on NAFTA in the early 1990s encouraged an overopti-
mistic view of Mexico's economic prospects, diverted attention from
macroeconomic management, and discouraged macrostabilization for



110      REGIONAL INTEGRATION AND DEVELOPMENT
Box 4.2 Crediblity on Deep Integratlon:The European Single Market Programme
Holmes and Smith (1998) argue that one effect of the removal of intraregional trade barriers under
the European Single Market Programme (SMP) was a reduction of the risk premium on investment.
In their view, trade policy commitments by EC (now EU) governments lessened future uncertainty
and induced additional investment and growth. Had these commitments merely covered reductions
in tariffs, it is not clear that they would have had to be Implemented regionally rather than multilat-
eraily. But they went much deeper, to issues such as standards, customs formalities, and domestic
regulation in which EU institutions are critical and over which the new power of enforcement granted
to central bodies-the European Commission and the European Court of Justice-added consider-
able leverage. Baldwin, Francois, and Pcrtes (1 997) forecast a similar reduction in risk premiums for
the Eastern European countries as they accede to the EU over the next decade.
fear of puncturing the political enthusiasm for NAFTA on either side of
the border. By contrast, the Europe Agreements between the EU and the
transition economies of Central and Eastern Europe do cover certain
domestic policy reforms (for example, on competition policy), as do
the RIAs between Australia and New Zealand and between Iceland and
the EU.
RIAs Can Affect the Incentives for Good Policies
An RIA may increase the returns to investment-friendly policies by en-
larging the potential market. Similarly it may increase the costs of "anti-
investment" policies or fiscal laxity because the domestic market is more
exposed to competition. It will be clear from this that genuinely liberal
RIAs can increase the credibility of promises of good policy, whereas
RIAs that do little to increase competition and exposure will not do
much for credibility, either. That is, direct credibility effects will tend to
magnify the static effects discussed in chapter 2. They are the icing on the
procompetitive cake of an RIA, but no cake means no icing.
An RIA can also help lock in the larger partner's trade policies. Most
RIAs have dispute settlement mechanisms, which can improve the secu-
rity of market access by providing a forum for dealing with disputes.
Such mechanisms, however, are frequently very political and smack of
managed trade. A small country may have little power to resist effec-
tively if its larger partner reneges on an agreement by, for example,
applying antidumping measures incorrectly to help some domestic in-
terest group. Moreover, even where the small partners can insist on ad-
dressing the problem, solutions frequently take the form of guaranteed
access for politically determined quantities rather than restoration of



STIMULATING INVESTMENT   111
free competition. Security of access can be particularly attractive if trade
wars are feared. In the early stages of negotiating NAFTA, the president
of Mexico at the time, Carlos Salinas de Gortari, declared, "what we
want is closer commercial ties with Canada and the United States, es-
pecially in a world of big regional markets being created. We don't want
to be left out of any of those regional markets" (cited in Perroni and
Whalley 1994).
RIAs Can Signal Government's Reform Intentions-If Genuine
Closely related to the incentives to pursue good policy is the use of an
RIA to signal that a government has changed its spots. If entering an RIA
entails (political) sunk costs in terms of challenging interest groups that
will lose from increased competition, and if it requires liberal or sound
policies to make sense, entry provides the government with a signaling
device, for only a government with genuinely liberal intentions would
sign. So, in the presence of asymmetric information about the type of
government, an RIA could improve credibility.
This argument is a persuasive explanation for some developing coun-
try governments' recent interest in RIAs. There has been a shift in per-
ceptions of the policy requirements for economic growth, and after
several decades of pursuing inward-looking policies, governments clearly
require a means of signaling genuine changes in their attitudes. What is
less evident is whether RIAs are the best means of signaling and whether
signaling is sufficient to justify RIAs. First, there is the "icing on the
cake" argument mentioned above. If an RIA is genuinely liberal-that is,
if it increases domestic competition and creates trade-it is beneficial
(relative to the status quo) in those terms alone, and credibility is an
extra, albeit one that may be more valuable than the fundamental static
gains. Second, there are other means of signaling and winning credibility:
trade policy binding at the WTO; acceptance of the IMF's Article VIII;
domestic rhetoric; and constitutional limitations, for example.5 These
measures do not carry the trade-diverting risks of RIAs.
Moreover, a country's policy credibility will not benefit from entering
an RIA if its government is not already heavily committed to reform.
Compare Greece, which did not undertake much-needed macroeconom-
ic reforms after joining the EU, wvith Portugal and Spain, which did. Por-
tugal and Spain gained credibility and benefited from increased FDI;
Greece's credibility-and FDI performance-remained low. This exam-
ple confirms that even strong, deep North-South RIAs are not sufficient
to ensure good policy or credibility. Indeed, Alogoskoufis (1995)



112       REGIONAL INTEGRATION AND DEVELOPNIENT
suggests that accession to the EC actually exacerbated Greece's policy
problems by providing transfers which enabled the country to postpone
needed reforms. A further disappointment of this kind is likely to be the
economic partnership agreements (EPAs) between the EU and the ACP
countries (see box 4.3).
Box 4.3 Will RlAs Enhance Credibility in Africa?
African countries desperately need policy credibility. An influx of investment and technology would
assist the growth strategies that are absolutely fundamental for alleviating poverty successfully.
Unfortunately, RIAs are unlikely to help in this regard, and if they divert attention from more impor-
tant domestic issues that might enhance credibility, they will be positively harmful.
We have already argued that even if combining small poor economies into a trade bloc has
attractions, enhancing credibility is not one of them. Thus, only North-South RIAs are of direct inter-
est here, and, of these, only those with the EU are likely to be significant. (U.S. trade with Africa is
small and is dominated by oil; the EU is far more important for Africa in terms of traditional links,
location, and the depth of trading and investment flows.) In the Cotonou Agreement of 2000, the EU
and the African, Caribbean, and Pacific (ACP) states agreed to turn their nonreciprocal trading agree-
ment, the Lome Convention, into fully reciprocal North-South RIAs, in the form of economic partner-
ship agreements (EPAs). Prominent among the reasons for this step was the ACP states' desire for
credibility, using the EU as an external agent of restraint (Collier 1996). Our analysis suggests that
they are unlikely to succeed in this. (Several other reservations are stated in Winters 2000 and Schiff
and Winters 2002b.)
Even on the narrow issue of the instruments negotiated under the RIA, such as tariffs on intra-
bloc trade, the incentives for the EU to discipline a small African country have to be considered. The
EU clearly has the necessary market power, but does it have the will? Its traditions on trade policy,
even between its own members, are fairly pragmatic, and since the EPAs use high-level political
bodies to manage intrabloc relations, it is easy to imagine managed trade solutions and temporary
derogations emerging, rather than punishment. The internal politics and public relabons associated
with the EU's excluding a traditional supplier from its market because, as the press might put it, "it
just happened to need to raise tariffs to close a fiscal deficit," would be very difficult-the more so
because most of the ACP countries are former colonies of EU members. Finally, ACP countries are so
small and so distant that EU governments have no significant material interest in their success or
their markets.
Even if the EPAs enhance the credibility of the ACP countries' low barriers to EU imports, they will
have litle effect on their trade policy toward third parties, which is at least as important because of
the potential losses from trade diversion. FTA members set their own trade policies, so there is no
formal discipline and not much of a political, moral, or legal case for the EU to press ACP partners to
be liberal in this respect. There are incentives such that when tariffs on imports from the EU are fixed
at zero, those on the other suppliers will be lowered, but these are not guaranteed to prevail (see the
discussion in chapter 3). Similarly, although the EU is trying to induce some of its other association
partners to liberalize their mutual trade (for example, within the Central European Free Trade Area
and in the Mediterranean area), it has not pressed for liberalization of these partners' third-country
trade, which would entail a loss for the EU.
The EPAs are not strong devices for signaling ACP countries' liberal intentions, simply because
these countries generally appear to have few such designs. In the extended negotiations, ACP
countries resisted having to open their markets to EU imports, demanded very long adjustment
periods, and initially sought to obtain a waiver in the WTO to permit the current (WvTO-inconsistent)
Lome arrangements to continue indefinitely. (The actual waiver obtained in 2001 covered the period
to 2008.)



STIMULATING INVESTMENT    113
Regional versus Multilateral Routes to Credibility
There are several reasons why RIAs might provide better or more credi-
bility than can multilateral agreements, specifically the WTO. First,
many countries have bound their tariffs at the WTO at levels significant-
ly higher than their applied tariffs and can raise them without violating
their WTO commitments. Given the weakness of the lock-in mechanism,
domestic and foreign investors will remain hesitant to commit resources.
Second, RIAs allow countries to make commitments on matters that
are difficult to negotiate multilaterally, including aspects of deep integra-
tion, such as harmonization of investment codes. Thus, not only are
more policies explicitly constrained, but there are more areas in which
punishment can be exercised.
Third, for commitments to be credible, there must be a high degree of
certainty that retaliation will follow violations. More countries are af-
fected by a country's actions at the WTO than in an RIA, which by defi-
nition has few members. RIA members therefore internalize a larger share
of the total loss from any violation-100 percent in a two-country RIA-
and have a stronger incentive to retaliate than countries at the WTO. In
addition, RIA members have less scope for free-riding on the punishment
(that is, for letting others take action) and more direct returns to making
the agreement credible overall (Fernandez and Portes 1998).
The WTO, however, does have proven enforcement procedures; it
brings major trading powers into the frame; and it is less subject to rene-
gotiations and internal political pressures. Also, as noted above, only
"good" RIAs will have any credibility effects. Thus, even if a developing
country has entered a North-South RIA, it is also likely to gain by acced-
ing to the WTO. Mexico used both the GATT and an RIA to lock in its
reforms (see box 4.1), as have the Eastern European countries, which
have sought WTO membership as well as Europe Agreements over the
past decade.
RIAs AS INVEISTMENT STIMULI
More than trade policy is needed to attract investment; the climate for
investment also matters. This section looks at how policy integration
within RIAs can influence investors' decisions and at whether integration
has a positive effect on growth.
Integration Affects Incentives to Invest
For policy changes to have a positive impact on investment, they
must raise expected returns or lower costs. Recent analyses (for example,



114    REGIONAL INTEGRATION AND DEVELOIPMENT
Baldwin and Forslid 1996; Baldwin and Seghezza 1996; Baldwin,
Forslid, and Haaland 1996) suggest that the rate of return on capital
(and on investment) can rise in all integrating countries regardless of cap-
ital abundance.' The researchers note that regional integration typically
reduces the transactions costs of tradables more than those of nontrad-
ables, shifting both demand and supply toward tradables. If, as is com-
monly believed, tradables are more capital-intensive than nontradables,
trade liberalization will raise the relative demand for capital and hence
its rate of return. The higher rate of return then increases the amount of
capital that people wish to use and so increases investment. (Extra in-
vestment is required to boost the capital stock, but even after the desired
levels are reached, investment will not fall back to its original level be-
cause the greater amount of stock requires more servicing and replace-
ment.) Moreover, increased competition in tradable goods sectors may
induce improvements in efficiency, lower markups, and a larger demand
for inputs in those sectors, further increasing the relative demand for
capital. Integration may also affect the prices of capital goods. Lower
tariffs and trading costs on imports of capital equipment may reduce the
price of investment goods, raising the rates of return and accumulation.
Increased competition from capital goods imports could also stimulate
the domestic capital goods industry to greater efficiency.
These arguments are less forceful in low-income than in middle-
income countries because low-income countries' capital goods already
generally enter with low or zero tariffs and there is no indigenous capital
goods industry before integration. Developing countries have often given
their industries high rates of effective protection by imposing zero or low
tariffs on capital goods and high tariffs on final goods. This policy gives
capital-intensive industries the highest effective protection, so that if
those industries lose most with integration (as they well might in a
North-South RIA), the result will be a reduction in the demand for capi-
tal and in investment. Additionally, of course, for a small country that
continues to import capital goods from the rest of the world, prices will
not change, as was explained in chapter 2.
Regional integration that goes beyond tariff reduction may raise the
efficiency of the financial sector, reducing lending margins and the cost
of funds and leading to higher investment. This is likely to provide some
benefits for developing countries. Two small developing countries may
gain by integrating their financial sectors because of the increased com-
petitiveness and because integration may allow them to diversify their
portfolios better and reduce their risk premia. These benefits will, how-
ever, be small if the countries have similar endowments and production



STIMULATING INVESTMENT    115
structures and thus highly correlated shocks. The benefits of increased
competitiveness are likely to be more important in an RIA with a large
partner that has a well-developed financial sector than in an RIA be-
tween developing countries-and the gains from nondiscriminatory lib-
eralization will be even larger. Financial integration is, however, difficult;
even the EU needed several decades to achieve it. And, as discussed in
chapter 6, the deeper integration needed to capture these benefits has
been mostly absent in South-South RIAs.
Overall, the theoretical arguments that an RIA will raise returns and
investment in developing countries are more persuasive for North-South
RIAs than for South-South ones. Furthermore, increased investment
must not be equated with increased economic welfare: investment that is
not well conceived can diminish welfare. The conditional nature of the
arguments made must also be kept in mind. RIAs may raise investment,
but to be certain, one must be sure that the conditions actually apply.
Whether an RIA will raise the return to capital depends on the capital in-
tensities in the various sectors and on the initial tariff structure. Opening
up the financial sector under unstable macroeconomic (fiscal and ex-
change rate) policies is likely to hurt more than help.
Finally, RIAs are not necessary for inducing investment. General re-
forms such as stabilization, market liberalization, and privatization
should raise the returns to all factors and are likely to be more than
enough to increase private investment. In the mid-1970s Chile began a
profound process of unilateral economic reform that led to a significant
increase in its savings and investment rates, approaching those of the
East Asian "tigers." (In fact, Chile left the Andean Pact to gain more
freedom for undertaking its reforms.) Neither is an RIA sufficient to in-
duce investment, as we saw with Greece. What seems to matter most is
the quality of domestic policies. A country suffering from low credibility
(say, as a consequence of a long history of illiberal policies and failed re-
forms) may be able to use a North-South RIA to raise its investment level
if it first undertakes serious reforms and then uses the RIA to give the
reforms credibility.
Investment Does Not Necessarily Mean Growth in RIAs
Baldwin (1989, 1992), in a major ex ante analytical study of investment
effects in an RIA, postulates a positive investment effect from the Euro-
pean Single Market Programme (SMP). Early estimates placed the static
gains from the SMP at around S percent of GDP (CEC 1988). Treating
these as pure gains in productivity that raise rates of return, Baldwin



116    REGIONAL INTEGRATION AND DEVELOPMENT
argues that investment will also increase as a result. Considering the in-
vestment increases in the context of a Solow (1956) growth model, in
which the steady-state (long-term) growth rate of the economy is given ex-
ogenously and is independent of government policy, Baldwin argues that
as capital accumulates, the economy can support a higher level of income,
and thus growth will increase temporarily as it approaches this level. Bald-
win terms this additional medium-term income effect, which adds perhaps
another 5 percent to GDP, the "medium-term growth bonus." But whereas
the static income effect of the SMP is essentially free (setting aside tempo-
rary adjustment costs) and so constitutes a pure welfare gain, the medium-
term income effect must be paid for by increasing investment-sacrificing
present consumption for higher future consumption. Thus the welfare
gain associated with the medium-term income effect is only about 0.5 per-
cent of GDP-only one-tenth of the income effect itself (Baldwin 1993).
The static effect would be spread over, say, 5 to 7 years following comple-
tion of the internal EC market, while it would take about 10 years for half
of the medium-term effect to be realized.
Integration in the EU seems likely to have a greater impact on invest-
ment and growth than will RIAs involving only developing countries.
The SMP not only deals with border measures such as tariffs and non-
tariff barriers but also entails "deep integration," which aims to elimi-
nate all barriers to the movement of goods, services, people, and capital
and pertains to a much larger market. Thus, it should have a greater im-
pact on contestability of markets and competitiveness of firms than
would the simple removal of trade barriers within a smaller bloc, such as
is found in most South-South RIAs.
The evidence on the impact on investment of RIAs involving develop-
ing countries is scarce and ambiguous. Brada and Mendez (1988) esti-
mate the effects of integration on capital formation in RIAs involving de-
veloped and developing countries (EFTA, the EEC, CMEA, LAFTA, and
the EAC) and find that except in the CMEA, investment increased but
that the effect on growth was very small. They conclude that the "dy-
namic effects of integration can neither explain the rapid growth of West
European countries in the 1960s nor serve as a strong justification for
encouraging nonmember countries to join existing schemes or to organ-
ize new ones" (167).
De Melo, Montenegro, and Panagariya (1993), by contrast, report
that African RIAs such as the CEAO and UDEAC experienced increased
investment rates following integration. In the CEAO, created in 1974,
the investment rate in member countries increased from 6.8 percent
of GDP in 1960-72 to 8.6 percent in 1973-85. The corresponding



STIMULATING INVESTMENT    117
increases for UDEAC, created in 1973, are from 15.9 to 18.3 percent,
and for the Andean Pact (1969), from 18.3 to 19 percent.
Such stimuli, however, are not unambiguously desirable when exter-
nal trade barriers are high; with high barriers, investment is likely to go
to the highly protected capital-intensive sector, where the private rate of
return on capital is high but the social rate of return is low or negative.
Hence, the higher rate of investnment need not raise economic growth or
improve welfare. De Melo, Montenegro, and Panagariya (1993) find no
significant differences in growth rates for members of RIAs compared
with other countries, and, in a multivariate regression analysis, they find
no effect of membership in an RIA on growth in developing countries.
Despite rising investment rates in the CEAO, UDEAC, and the Andean
Pact, income growth fell in all three. The declines in growth rates are not
necessarily attributable to the formation of the RIAs, since the growth
rates both of developing countries as a whole (78 countries) and of
OECD countries also fell in this period, probably because of the 1973 oil
shock. Still, the declines certainly do not support the view that regional-
ism boosts growth-a subject to which we return in chapter 5.
An important lesson from these results is that the current sacrifice
which additional investment entails need not result in additional future
benefits if investment occurs in a distorted environment. By contrast, in a
liberal policy environment where domestic relative prices reflect world
prices and private rates of return approximate social ones, additional in-
vestment is likely to be productive. This is another reason for lowering
external trade barriers as countries form RIAs.
REGIONAL INTEGRATION AND FOREIGN DIRECT
INVESTMENT
The formation of an RIA affects investors' decisions about where and
how to invest. The motives behind these decisions, and the evidence that
an RIA attracts investment, are examined in this section.
Motives for FDI
Although FDI is subject to the same forces that influence total invest-
ment, it also responds to various specific determinants. One common
motive for FDI is to boost local sales and market access. Access to large
or rich markets can be an important factor in attracting FDI into a coun-
try. In a survey conducted by Japan's Ministry of International Trade and



118    REGIONAL INTEGRATION AND DEVEi.OPMENT
Industry (MITI) on the motives behind Japanese manufacturing FDI in
Asia, North America, and Europe, 70 percent of respondents cited local
sales as a major factor (Kawai and Urata 1996). Similarly, the spurt of
FDI in ASEAN countries in the late 1980s is usually attributed to those
countries' rapid growth (Alburo, Bautista, and Gochoco 1992).7 A nec-
essary condition for this sort of investment is that the market not be as
readily accessible from outside the bloc. That is, FDI motivated by mar-
ket access is of the "tariff-jumping" variety. FDI for automobiles and au-
tomobile parts in ASEAN economies increased partly because these were
highly protected sectors and FDI was the only way to gain a foothold
(Bhalla and Bhalla 1997). The same happened in MERCOSUR. It is im-
portant to note that tariff-jumping FDI can be immiserizing-that is, it
can reduce domestic economic welfare. The tariff directs capital into sec-
tors in which the RIA does not have comparative advantage, and by
causing other factors of production to flow into those sectors, it reduces
output in those in which it does.8
These factors are more likely to stimulate FDI flows from outside the
RIA than between members. Firms originally located in a member coun-
try receive access to the whole market without relocation and so have
less incentive to invest in other members. Firms located in third coun-
tries, on the other hand, will have more incentive to locate new produc-
tion facilities in a member country and to service the other members of
the bloc through intra-RIA (preferential) exports. Such "platform invest-
ment" is particularly likely if there are increasing returns to scale in
production, making for lumpy investments that are viable only above a
certain size. The integrated market may become large enough to bear the
fixed costs of the establishment of new foreign affiliates. Thus, RIAs
might attract more foreign investment (for example, in the production of
consumer durables) into developing regions as a whole than can frag-
mented national markets. Whether these effects are more likely for RIAs
between larger countries or between smaller ones depends on the relative
sizes of the blocs' demands for a product and the minimum efficient scale
of production. Combining several small economies may make new FDI
attractive in small-scale industries, but these same industries may already
be present in larger countries even without an RIA. And it is not likely
that any developing country RIA could stimulate FDI in, say, the produc-
tion of large airliners.
A second important set of motives behind FDI is to take advantage of
local factors of production (such as local labor) and to set up export plat-
forms. In the MITI survey mentioned above, 44 percent of the Japanese
firms surveyed indicated that use of local labor was one of their motives.



STIMULATING INVESTMENT   119
These motives can be enhanced if a developing country forms an RIA
with a developed country, as a firm located in the developing country can
then take advantage of cheap inputs while obtaining free access to the
large developed economy. Although Mexico had long been used as an ex-
port platform to the United States, after 1994 NAFTA had a profound
impact on FDI into Mexico from countries outside the bloc, as this in-
vestment became a way to guarantee market access to Mexico's northern
partners (Blomstrom and Kokko 1997; Fernandez and Portes 1998).
About 80 percent of the vehicles produced by U.S. auto manufacturers in
Mexico in 1997 were for export, compared with 48 percent in 1994
(USITC 1997). Following the creation of NAFTA, Japan redirected part
of its FDI from the United States and Canada toward Mexico. Whereas
Japanese investments in the United States are geared primarily for the
local market (and are mainly directed toward manufacturing, commerce,
and banking and finance), those in Mexico, particularly in the automo-
bile industry, are intended more for the NAFTA continental market.
A third factor may be that the removal of internal barriers in the RIA
allows firms to allocate operations across member countries more effi-
ciently. Thus, if RIA members differ in their endowments, the RIA may
stimulate vertical FDI. This potentially important aspect of North-South
arrangements lies at the heart of Ethier's (1998) theoretical exploration
of the benefits of regionalism. With guaranteed preferential access to the
northern market, the southern country becomes an attractive location
for labor-intensive activities. Knowing this, it is more confident of receiv-
ing inflows of investment and is hence more comfortable with liberaliz-
ing than it would be under multilateral liberalization, where it would
face fiercer competition for inflows. This argument is equally applicable
to intrabloc FDI and to that from outside.
Evidence of the Positive Effects of Integration
There is strong circumstantial evidence that creating or deepening an
RIA stimulates FDI. For example, between 1946 and 1970, the number
of manufacturing subsidiaries of EC-member multinationals located in
other EC countries increased more than sixfold, from 68 to 434, while
those located in non-EC European countries increased only from 95 to
311 (Yannopoulos 1990). Similarly, the European Commission (1998)
found that intra-EU FDI rose more rapidly than investment outside the
EU following the introduction of the SMP. Germany and the United
Kingdom, in particular, switched their investment from the United States
to the EU beginning in the late 1.980s.



120    REGIONAL INTEGRATION AND DEVELOPMENT
It has also been widely documented that European integration has
made member countries more attractive to U.S., Japanese, and other
third-country firms. The creation of the single market in 1992 had a sig-
nificant impact on decisions by Japanese, Korean, and Taiwanese (China)
companies to establish operations in the EU (Motta and Norman 1996).1
Total inflows of FDI into member countries expanded from ECU 10 bil-
lion in 1984 to ECU 63 billion in 1989. During this period, inflows into
the United States also increased, but at a significantly lower rate (WTO
1995). The European Commission (1998) found that the EU's share of
worldwide inward FDI flows increased from 28 to 33 percent during
1982-93.
Turning to North-South RIAs, the data show that FDI to Mexico rose
from $4.3 billion in 1993 to $11 billion in 1994, the year NAFTA came
into force (figure 4.1). The experience of RIAs between middle-income
countries is similar: for example, following the signature in 1991 of the
Treaty of Asunci6n that established MERCOSUR, FDI in member coun-
tries increased from $3.5 billion in 1991 to $18 billion in 1996 and to
$38 billion in 1998. With nearly $11 billion in FDI in 1996 (up dramat-
ically from $1.1 billion in 1991), Brazil surpassed Mexico as the largest
FDI recipient in Latin America (table 4.1). A qualification to these fig-
ures is that one cannot know for sure that it was the RIA per se rather
than better policy in general that boosted FDI; in the case of MERCO-
SUR, the two occurred more or less simultaneously. And the welfare im-
plications are ambiguous because some of the boost in FDI, as in the au-
tomobile sector, was caused by trade and industrial policy distortions
such as very high tariffs.
Figure 4.1 Net Inflows of Foreign Direct Investment, Mexico, 1985-96
12 -
10-
08-
~  6
2
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
Source' World Bank, World Development Indicators, various years.



STIMULATING INVESTMENT        121
Table 4.1 Net Inflows of Foreign Direct Investment, MERCOSUR, 1991-98
(millions of current U.S. dollars)
Country    1991   1992    1993.   1994    1995    1996     1997     1998
Argentina  2,439  4,384   2,763   3,432   5,279   6,513    8,094    6,150
Brazil    1,103   2,061   1,292   3,072   4,859   11,200  19,652   31,913
Paraguay     84     137    111     180     184      246      270     256
Uruguay       0      1     102     155     157      137      126     164
Source: World Bank, World Development Indicators, various years
Although we have some evidence of the impact on investment of
North-North and North-South RIAs and of South-South RIAs among
large middle-income countries, rio such evidence is available for South-
South RIAs among small low-income countries. Furthermore, all the
cases cited in the literature indicate anticipation of the RIA followed by
something of a decline. This strongly suggests that RIAs change firms'
views of the optimal stock of investment in member countries. Thus, we
observe a big step-up during the adjustment period but only a small per-
manent increase in the flow commensurate with maintaining the higher
stock. There is no evidence to date that steady-state flows of FDI in-
crease.
One area not examined here is the potential of FDI to generate tech-
nological knowledge spillovers. This question is examined in chapter 5.
NOTES
1. Such pacts can also fail because the country assigned to produce certain
products lacks the capacity to do so. The Andean Pact assigned exclusive pro-
duction rights to different countries by sector in 1970. The program failed when
Bolivia produced only 2 of the 10 products assigned to it and Venezuela violated
the agreement and started making the products (Echavarria 1997). In other cases
the more industrialized member countries (for example, Kenya in the EAC and El
Salvador in the CACM) refused to accept constraints on the expansion of their
manufacturing sectors.
2. This section draws on the discussion in WTO (1996).
3. The terms "shallow" and "narrow" integration are discussed in chapter 6.
4. In the 21st century, a northern government that imposes punishments
which are neither legally based nor politically imperative is likely to stir up a
storm of protest about neocolonialism and interference.
5. Article VIII of the IMF's Articles of Agreement deals with the general obli-
gations of members, including avoidance of restrictions on current payments,
avoidance of discriminatory currency practices, convertibility of foreign-held



122     REGIONAL. INTEGRATION AND DEVELOPMENT
balances, information provision, consultation regarding existing international
agreements, and collaboration regarding policies on reserve assets.
6. These studies were developed mainly with the European Single Market Pro-
gramme in mind but may also apply to RIAs that involve developing countries.
7. The share of ASEAN in world inward FDI increased from 2.5 percent in
1985 to 4.3 percent in 1990.
8. Bhagwati (1987) has proposed another trade-oriented explanation for FDI,
which is grounded in the political economy of protection. In this interpretation,
"quid pro quo" FDI is made not to circumvent tariffs but to defuse the threat of
protection. Unlike the tariff-jumping kind, this type of investment occurs in an-
ticipation of the imposition of protection, in which case FDI that transfers tech-
nology and promotes employment is in the nature of a bribe paid to the host
country in return for allowing free access to the investor country's exports. This
may be significant in the EU.
9. Motta and Norman (1996) argue that this may have been a consequence of
the stricter application of antidumping legislation in the EU beginning about
1985.



CHIAPTER 5
Growth and Location
Economic growth lies at the heart of economic development, and for
many countries growth must be driven by industrialization. This chapter
considers two of the main concerns of developing country policymakers,
and ultimate justifications of regional integration agreements: economic
growth, and location of industries.
We ask, first, how RIAs affect long-run growth, and we find that the
preference for North-South over South-South RIAs based on the analysis
of static effects holds for dynamic effects as well. Modern growth theory-
the theory of endogenous growth-emphasizes the role of knowledge in
fostering productivity and growth. It stresses that knowledge can be effec-
tively transferred from one country to another through international con-
tacts and trade. Wealthy countries are knowledge rich and so are likely to
provide far more access to technology than are poorer trading partners.
RIAs that switch imports from richer to poorer sources are thus likely to
have a perverse effect on members' growth rates. RIAs might also help
countries boost their growth rates by supporting institutional reform, and
this effect too seems likely to be stronger when developing countries join
with richer partners than when they form an RIA with poorer ones.
The direct evidence on RIAs and growth is subject to some methodolog-
ical reservations but is actually pretty consistent. There is little evidence that
RIAs between developed countries stimulate growth, some recent evidence
that North-South RIAs may affect growth, and none that RIAs between de-
veloping countries do so. Casual examination of the recent performances
of, say, Mexico, Poland, and Portugal and a formal analysis of the effects of
North-South versus South-South RIAs suggest that serious North-South in-
123



124    REGIONAL INTEGRATION AND DEVELOPMENT
tegration may foster growth, reinforcing the earlier conclusions about the
relative merits of the two types of partner for developing countries.
Second, how do RIAs affect the development and location of industry
in developing countries, and can they materially affect countries' growth
prospects? The discussion of investment in chapter 4 suggested that RIAs
can stimulate investment and thus might have a beneficial effect on either
industrialization or growth. In this chapter we move beyond investment
to consider the economics of agglomeration (or clustering).
Although economists have long been aware that industry tends to
cluster at particular locations, they have only recently learned to model
this agglomeration formally, allowing them to begin identifying precisely
the combinations of conditions that must be satisfied for it to occur. The
theory arose from attempts to understand the possible effects of the en-
largement and deepening of the European Union (Krugman and Ven-
ables 1990) and so is directly applicable to the case of RIAs. The theory
is, however, very young; the models do not yet appear to be very realistic
and have to date not been accompanied by much empirical evidence.
They are more parables than forecasts, but they do offer considerable in-
sight into qualitative factors, and they address such a major concern of
policymakers and publics alike that we need to explore their implications
for developing countries.
Creating an RIA is likely to affect the incentives for industry to
agglomerate. Usually, it will encourage clustering because the RIA in-
creases market size and allows more effective exploitation of the links be-
tween firms. An RIA may attract industry into member countries at the
expense of nonmembers, although if it is small, this effect will also be very
small. RIAs also frequently lead industries to relocate from one member
to another. For RIAs between poor countries, this seems likely to increase
intermember inequalities because under an RIA firms find it easier to ag-
glomerate in the more prosperous countries while still selling in the other
member countries. For RIAs involving richer members, the results are less
clear-cut, and it is possible that poorer members will experience strong in-
dustrialization after joining an RIA. For developing countries, integration
with richer neighbors (North-South RlAs) looks better from an agglomer-
ation point of view than does South-South integration.
SOUTH-SOUTH AND NORTH-SOUTH REGIONALISM AS
STIMULANTS TO GROWTH
In the final analysis, the principal objective of trade policy is economic
growth. Any change that generated even a small increase in the long-term



GROWTH AND LOCATION      125
rate of economic growth would result in cumulative gains that would
easily swamp any static and mecium-term benefits of the sort discussed
so far. If openness stimulates growth, as is often claimed, is it not reason-
able to expect that partial openness (regionalism) would at least partial-
ly stimulate growth-especially since it frequently appears to affect
investment rates? The remainder of this chapter examines these claims
and finds little evidence for them for South-South RIAs but some evi-
dence for North-South RIAs. It should be kept in mind, though, that
even when regionalism does have beneficial effects, nondiscriminatory
openness typically offers them as well.
A well-structured RIA might increase a member country's underlying
growth rate, raising its development trajectory by, for example, increas-
ing credibility (as we saw in chapter 4) or reducing tensions between
countries (see chapter 7). These, in turn, could raise capital and labor
productivity, lead to additional flows of investment and knowledge, and
push an economy several notches up on the development path.
The scope for policy-driven growth is a contentious issue among
economists. The traditional (neoclassical) view of economic growth ad-
mits no means of influencing the long-term growth rate, which is set ex-
ogenously according to rates of population growth and technical
progress (Solow 1956). Even if policy can influence the rate of capital ac-
cumulation, it can affect the growth rate only temporarily. As capital
accumulates faster, it runs into diminishing returns, so that, in time, the
additional investment is completely absorbed by the additional deprecia-
tion associated with having extra capital stock, and the capital-output
ratio eventually stops increasing. Hence, although the level of income
increases, and so does the medium-term growth rate as the economy
approaches its new growth path, the long-term growth rate does not
increase.
More recent endogenous (self-generated) growth theory, however,
holds that the returns to capital-especially human capital (Lucas 1988)
and knowledge capital (Romer 1986, 1990)-do not diminish at the ag-
gregate level because of positive spillover effects. Thus, policies that
affect the accumulation of these factors can permanently raise the long-
term growth rates of output and income.
The theoretical literature on openness and growth has not generated
robust findings on the link between RIAs and growth. Although open-
ness is typically positively associated with growth, especially where trade
or FDI is a medium for transferring knowledge, this is not necessarily
true for regional integration. Recent empirical work, however, indicates
that North-South RIAs are likely to generate productivity gains to the
developing partner (see "Trade, Convergence, and Spillovers," below).



126    REGIONAL INTEGRATION AND DEVELOPMENT
Knowledge and Institutions as Keys
Accumulation of physical capital can have little effect on long-term
growth because physical capital eventually encounters diminishing re-
turns. Human capital is different. Although individual human capital is
also expected to run into diminishing returns, one person's return can be
positively affected by the average level of available human capital
through positive spillover effects. For instance, an engineer is likely to be
more productive if she can interact with and learn from other qualified
engineers than if she cannot. Thus, the average level of human capital
may rise without running into diminishing returns and can have an ii-
pact on long-term growth (Lucas 1988).
Knowledge is given the primary role in endogenous growth theory-
increases in knowledge capital are expected to have a permanent positive
impact on the rate of growth. Knowledge, once produced, has the public-
good characteristic that its use by one user does not prevent Its use by an-
other. Creation of knowledge has large spillover effects to others and is
therefore likely to display increasing rather than diminishing returns.
Another important factor is the legal, institutional, and regulatory
framework, which includes the quality and stability of the political
process, de facto property rights, and other institutional aspects. Olson
(1996: 7) reviews the factors determining per capita income and growth
and concludes that "the most important explanation of the differences in
incomes across countries is the difference in their economic policies and
institutions." Hall and Jones (1999) find that most of the variation in
output per worker across countries cannot be explained by endowments.
They conclude that a country's long-run economic performance is deter-
mined primarily by the institutions and policies that make up the eco-
nomic environment within which individuals and firms make invest-
ments, create and transfer ideas, and produce goods and services. Recent
studies by Acemoglu, Johnson, and Robinson (2001) and by Engerman
and Sokoloff (1997) find that institutions have had a fundamental role in
explaining growth in the last few centuries.
Given this general framework, how can policy affect long-term eco-
nomic growth? First, a government can promote investment in education
and other forms of human capital. Second, it can improve its political
and legal institutions so as to enhance incentives to accumulate and in-
novate. Regional integration may be able to help, in terms of both policy
integration (chapter 6) and international politics (chapter 7). If, by join-
ing an RIA and taking policy integration measures, a country improves
its legal and regulatory framework, it may obtain a growth benefit.
(There are, however, few examples outside the EU of countries actually



GROWTH AND LOCATION     127
engaging in policy integration.) Joining an RIA may also help a country
improve its political system if this is a condition for membership
(chapter 7).
Trade policy can play a major role in knowledge accumulation.
Knowledge has international public-good characteristics, with cross-
border spillovers through trade, FDI, scientific exchanges, and the like.
Since most developing countries are not major producers of scientific or
technical knowledge, it is important that they pursue foreign trade poli-
cies which enhance the acquisition of knowledge from abroad. If open-
ness helps, what sort of openness? Is general opening up of the economy
the best way to absorb foreign knowledge? Or can developing countries
do better through preferential trade liberalization?
Trade, Convergence, and Spillovers
Ben-David (1993) offers convincing evidence that increasing mutual
trade among affluent countries leads to upward convergence in per capi-
ta incomes. He shows that in Europe, the strong increases in trade asso-
ciated with increased integration coincided with a dramatic narrowing of
per capita incomes across countries. As shown in figure 5.1, there was an
almost continuous convergence in per capita incomes in Europe as inte-
gration proceeded, from 1947, when the Benelux customs union was cre-
ated, through 1981. Stages along the way included the creation of the
ECSC in 1951 and of the EEC in 1957, elimination of quotas in 1962,
and removal of internal tariffs in 1968. Income differences narrowed by
about two-thirds over the period, and the convergence was upward, with
the poorer countries experiencing faster growth than previously.
Do these results imply that a developing country has only to increase
its trade with affluent countries to raise its growth? Not according to this
approach (but see below). Ben-David (1994) finds benign convergence
among developed countries, with the poorer catching up with the richer;
no convergence between middle-income countries or between them and
other countries (including rich ones); and a malign convergence among
poorer countries. He also finds that convergence is common between
countries that are major trading partners (known as convergence clubs)
but not among random groups of countries (Ben-David 1998). This rein-
forces the view that trade is the mechanism through which convergence
occurs, although the underlying cause might be some other aspects of
openness, such as FDI, that are strongly correlated with trade. Ben-David
(1996) finds that the existence of convergence clubs is attributable more
to convergence in rates of total factor productivity (TFP) growth than in
investment rates; Henrekson, Torstensson, and Torstensson (1997) de-



128    REGIONAL INTEGRATION AND DEVELOPMENT
Figure 5.1 Equalizing Exchange:Trade Liberalization and Income Convergence, 1947-81
0.34 -     x1947 Benelux customs union
1951 creabon of ECSC
*0
Xc               \ 1957. creation of EEC
0 .2 8  - - - - - - -  - - - - - - - - - - - - - - - - -
QE                      959: tanff and quota reforms initiated
cE
.0
e  0.22 -                      -
2   ~ ~ ~ ~ ~ ~ ~  ~   \ ~~~~~~1968: final elimination of tariffs
_~~~~~~~~~~~ 738 EEC expands to 10 members
0.2
0;
1962, final elimination of quotas
0.1  i1981, EEC expands to 10 membersl
1945          1955         1965          1975
Note ECSC, European Coal and Steel Community; EEC, European Economic Community.
Source: Ben-David (1993).
rived similar findings for the EC and EFTA. These results suggest that (as
predicted by endogenous growth theory) convergence arises from the
contacts and information-the knowledge-generated by trade and FDI,
rather than from incentives to accumulate physical capital.
Karras (1997) examines whether integration facilitates convergence in
the per capita incomes of member countries by investigating convergence
during the period 1960-90 in three RIAs: ASEAN (5 countries), the EU
(15 countries), and LAFTA (7 countries). Three empirical tests all reveal
strong convergence in the EU (the initially poorer countries grow more
quickly), somewhat weaker convergence in LAFTA, and absence of con-
vergence, or even divergence, in ASEAN. Karras concludes that regional
integration does not guarantee convergence in the standards of living of
member countries but that convergence may be associated with the de-
gree of economic integration (as manifested tn reduction of protection,
enhanced internal trade, and increased policy coordination), which is
highest in the EU and weakest in ASEAN. Another possibility is that con-
vergence is more likely in North-North RIAs (the EU) than in South-
South ones (LAFTA and ASEAN).
A more formal stream of work on trade and producttvity is that by
Coe and Helpman (1995) and Coe, Helpman, and Hoffmaister (1997),
who seek to explain TFP levels in OECD countries and developing coun-



GROWTH AND LOCATION     129
tries, respectively. The authors construct an index of total knowledge
capital, measured by accumulated investment in research and develop-
ment (R&D), for each developed country. They assume that trading
partners obtain access to a country's stock of knowledge in proportion to
their imports from that country--total imports, in the 1995 paper, and
imports of machinery and transport equipment in the 1997 study.
In both exercises, access to foreign knowledge has a statistically sig-
nificant effect on productivity. For developing countries, Coe, Helpman,
and Hoffmaister find that TFP is related both to the openness of the
economy (the ratio of imports to GDP) and to the interaction of the lat-
ter with the access to foreign knowledge that foreign trade provides. An
economy benefits from foreign knowledge, first, according to how open
it is in general and, second, according to whether it imports mainly from
those countries that have the largest knowledge stocks. These results are
intuitively very attractive and suggest, again, that trade is a major con-
duit for spillovers between countries.
Lumenga-Neso, Olarreaga, and Schiff (2001) argue that the new ap-
proach developed by Coe and Helpman (1995) has not been carried
through to its logical conclusion because countries also obtain "indirect"
knowledge spillovers through trade. That is, France (say) obtains knowl-
edge spillovers from trading with other OECD countries, which means
that it can draw on more knowledge than it alone produces. Therefore,
any country (say, Belgium) that trades with France will obtain not just di-
rect knowledge spillovers from the knowledge France produces but also
indirect spillovers from the knowledge France acquires through its trade
with other countries. The authors find that these indirect spillovers are
larger than the direct ones and that both have a significant impact on
TFP. They also perform various tests and obtain results that are statisti-
cally superior to those obtained by Coe and Helpman.1
Coe and Helpman's approach has been extended in recent work fo-
cusing on developing countries. In an industry-level analysis, Schiff,
Wang, and Olarreaga (2002) show that southern countries' TFP re-
sponds more strongly to North-South trade than to South-South trade.
Another finding is that R&D-intensive industries in the South learn
mainly from trade with the North and that industries with low R&D in-
tensities learn mainly from trade within the South. Thus, North-South
RIAs tend to favor the development of R&D-intensive industries, while
South-South RIAs tend to favor the development of low-R&D-intensity
industries. The authors conclude that forming a South-South RIA may
delay the transformation of member countries to a high-R&D economy
by reducing technology spillovers from the North.



130    REGIONAL INTEGRATION AND DEVELOPMENT
Work on Latin America (Schiff and Wang 2002a) shows that the inter-
action between education and OECD knowledge spillovers has a positive
effect on TFP in R&D-intensive industries. The implication is that there
are virtuous growth cycles, with increases in education resulting in in-
creases in the TFP of R&D-intensive industries, greater demand for
skilled labor (which is typically complementary with technology), in-
creased demand for education, and so on. Moreover, given that OECD
R&D stocks grow continuously over time, the interaction effect between
education and OECD knowledge spillovers implies that education has
permanent effects on productivity growth in R&D-intensive industries.
Since these industries benefit mainly from spillovers from the North, the
interaction effects provide another argument for North-South rather than
South-South RIAs in Latin America. The FTAA and the Chile-EU FTA, as
well as prospective North-South RIAs such as the EU-MERCOSUR FTA,
should help in this regard.
Keller (2002) shows that the impact of technological knowledge on
TFP diminishes as the distance between the knowledge-exporting and
knowledge-importing countries increases. Although he does not specify
the channel of knowledge diffusion, this effect appears likely to apply to
trade. This would seem to suggest that in choosing a partner in the
North, a developing country is better off, other things being equal,
choosing a close rather than a distant one. According to this view, Mexi-
co is better off selecting Canada and the United States as trading partners
than Japan or the EU. This is supported by the findings of Schiff and
Wang (2002b) showing that the impact of foreign knowledge from the
United States and Canada on TFP in Mexico is several times larger than
that from the rest of the OECD.
Choosing the Wrong Partners Can Harm Growth
An important implication of Coe, Helpman, and Hoffmaister's (1997)
work is that any trade policy-including formation of an RIA-that
switches a developing country's imports of machinery and equipment
from countries with high stocks of knowledge to countries with lower
stocks may harm growth. Conversely, an increase in openness is likely to
result in faster TFP growth. Thus, countries seeking to accelerate TFP
growth should pursue trade policies that increase openness and avoid
switching trade from countries with high knowledge stocks to countries
with lower stocks.
The corollary is that a developing country contemplating forming an
RIA is usually better off choosing a partner with a high-and quickly



GROWTH AND LOCATION     131
growing-stock of knowledge. If an RIA among developing countries re-
sults in trade being diverted away from more to less knowledge-intensive
suppliers-say, if Argentina switches from U.S. and Japanese to Brazilian
capital goods-it may harm TFP growth. The static analysis of RIAs has
long observed that trade diversion is potentially harmful. Dynamic
analysis suggests that trade diversion may be harmful if the diversion is
from a source with a large stock of knowledge to one with a smaller
stock but beneficial if it is from a small knowledge source to one with a
larger stock of knowledge.
To give one illustration, Winters (1997a) applies the Coe, Helpman,
and Hoffmaister approach to simulate the impact on TFP growth of an
FTA between Lebanon and the EU. He conservatively assumes that sign-
ing a Euro-Mediterranean Agreement shifts 4 percent of Lebanese im-
ports from the United States and Japan to the four large EU economies,
which have smaller R&D stocks. This lowers the R&D capital stock to
which Lebanon has access by 12.5 percent, causing a I percent reduction
in Lebanon's TFP. But the agreement will also probably result in an in-
crease in the openness of the Lebanese economy, and this will offset the
dynamic diversion effect. Provided that the increase in openness is more
than 2.5 percent (from 8.5 to 8.7 percent), Lebanon would show a net
increase in TFP. Even so, the analysis raises the question as to whether
the EU is the best partner for Mediterranean or Eastern European coun-
tries from the viewpoint of knowledge spillovers, as other areas, such as
the United States and Japan, have typically generated more knowledge.2
Spillovers, however, are also affected by closeness in geography, lan-
guage, and history (Keller 2002), which may lower the cost of absorbing
and adapting knowledge from European sources. For example, im-
porters in Morocco may find it casier to learn how a machine operates if
it is imported from France rather than from Japan because of the lan-
guage bond and because Moroccans often have long-standing relation-
ships with French traders.
Schiff and Wang (2002b) show that in the case of NAFTA, trade di-
version in Mexico is dynamically beneficial because the diversion takes
place from a source (the OECD countries outside North America) with a
small effect on TFP to one with a larger impact-the United States and
Canada. The authors estimate elasticities of TFP with respect to foreign
knowledge of 0.36 for the United States and Canada and 0.04 for the
rest of the OECD. They then simulate the effect of NAFTA and find that
it raises TFP in Mexico by 5.5 to 7.5 percent, 0.5 percent of which is at-
tributable to trade diversion ancd the rest to trade creation.3
Although this analysis indicates that an RIA with a country or region



132    REGIONAL INTEGRArION AND DEVELOPMENT
that has a large or fast-growing stock of knowledge may lead to dynam-
ic gains, such gains might not last forever. The regions that generate the
most knowledge today may not do so in the future. For instance, the
United Kingdom was the leader during the Industrial Revolution, and
Japan led in various industries in recent decades. A country that had
RIAs with Japan and East Asia might have obtained significant dynamic
benefits before 1990 but, given the present crisis in that region, might
now be losing relative to one that had a nonpreferential trade policy.
From a long-term viewpoint, nondiscriminatory trade liberalization is
probably the optimal policy in a world where it is hard to predict which
region will generate the most knowledge in the future, and where it is dif-
ficult to join and leave RIAs. That is, unilateral free trade may be optimal
from a long-run dynamic perspective, as well as a static viewpoint.
Another reason why unilateral MFN liberalization may be optimal is
that knowledge travels not only from North to South (through trade and
FDI) but also between northern countries through licensing mechanisms,
trade, and FDI. Hence, the level of knowledge of a northern country de-
pends not only on its own prodtuction of knowledge but also on how
much knowledge it acquires from other northern countries. An OECD
country may produce relatively little knowledge itself but may have ac-
cess to additional knowledge through trade. This is likely to reduce the
cross-country variation in knowledge stocks in the North. Lumenga-
Neso, Olarreaga, and Schiff (2001) show that although the cross-country
coefficient of variation of produced technological knowledge stocks is
0.70 for OECD countries, it is only 0.29 with respect to available knowl-
edge stocks, where available stocks include spillovers from trade with
other OECD countries. Thus, a developing country may benefit from
trading with a northern country that generates little knowledge indige-
nously because the latter has acquired knowledge from other northern
countries. Given the present level of understanding of how knowledge
transfers are generated, nonpreferential trade liberalization seems likely
to be the more robust policy recommendation. For a few countries, how-
ever, North-South integration may make sense; Mexico is an example. If
the preferential route is chosen, a robust result seems to be that a poor
country will gain little or nothing from an RIA with another poor coun-
try and that a North-South agreement is preferable.
The theory of endogenous (self-sustaining) growth is predicated on
the assumption that knowledge spills over from one enterprise to anoth-
er without the receiving enterprise having to purchase it; it is an exter-
nality. This is what reconciles decreasing returns at the firm level with
constant or even increasing returns at the industry or economy level.



GROWTH AND LOCATION      133
Such externalities provide a justification for policy intervention, for, as is
well known, in the presence of externalities, market outcomes are not
optimal. The work of Coe, Helpman, and Hoffmaister is perfectly con-
sistent with this view, which suggests benefits to an activist policy that
shifts imports toward high-knowledge suppliers.
In fact, however, the model also works in the absence of such exter-
nalities, that is, with developing country importers paying for and inter-
nalizing all the knowledge they receive through trade. They could then
be expected to choose their tracde patterns correctly, and even if they
could raise growth rates by choosing a different pattern of supply, they
are rational in not doing so because of the cost involved. In these cir-
cumstances there is no case for intervention to stimulate growth, and an
RIA that artificially disturbed the pattern of trade would be distor-
tionary. We do not claim that there are no spillovers in knowledge diffu-
sion through international trade, but we do want to caution against the
view that there definitely are. And even in the presence of knowledge
spillovers, unilateral MFN liberalization may be optimal because knowl-
edge advantages may not last.
FDI and Knowledge Spillovers
Blomstrom and Kokko (1998) describe three forms of knowledge
spillovers from the presence of FMI: (a) local firms improve their produc-
tivity by copying some technology used by affiliates of foreign firms
operating in the local market; (bh the entry of an affiliate leads to more
severe competition in the host economy, so that local firms are forced to
use existing technology and resources more efficiently; and (c) competi-
tion forces local firms to search for new, more efficient technologies.
Although these potential spillover effects seem analytically plausible, re-
sults from the empirical literature are mixed, at best. One reason is that
foreign firms may take action to limit spillovers to the host country's do-
mestic firms.
A fourth form of knowledge spillover is the effect on domestic up-
stream firms that sell to affiliates of foreign firms or on domestic down-
stream firms that buy from these affiliates-"backward" and "forward"
linkages. The effect of spillovers on the productivity of upstream and
downstream firms is empirically more promising because foreign firms
typically procure domestic inputs of better quality than do domestic
firms, produce better intermediate products to be used by downstream
firms, or require downstream firrns to improve the marketing and distri-
bution of their final products.



134    REGIONAI INTEGRATION AND DEVELOPMENT
Studies on FDI and productivity of domestic firms typically find no
positive intraindustry spillover effects and may even find negative ones.
Aitken and Harrison (1999) use data from Venezuela to study whether
domestic firms benefit from FDI and find (a) a positive relationship be-
tween increased foreign equity participation and plant performance, sug-
gesting that individual plants do benefit from foreign investment, and
(b) a decline in the productivity of domestically owned plants when for-
eign investment increases, suggesting a negative spillover from foreign to
domestic enterprises. A possible explanation is that FDI reduces the
share of the market left for domestic firms and that under economies of
scale, a smaller output implies a higher average cost. Another is that for-
eign firms typically select the best workers, leaving less productive work-
ers for the domestic firms. One issue left unresolved, which may affect
Aitken and Harrison's results, is causality: that is, does plant perform-
ance improve because of FDI, or does FDI go to better-performing
plants? Kokko (1996) uses Mexican data to examine whether there are
significant spillovers associated with competition effects from foreign
presence and finds no significant competition effect.
Haddad and Harrison (1993) find, for Morocco, no significant rela-
tionship between higher productivity growth in domestic firms and
greater foreign presence in the sector. Djankov and Hoekman (2000)
study the effects of FDI on Czech firms and find that foreign investment
has the predicted positive impact on local TFP growth of recipient firms
but that joint ventures and FDI appear to have a negative spillover effect
on firms that do not have foreign partnerships.
Empirical studies have found, however, that FDI generates interindus-
try spillovers through backward and forward linkages. A foreign firm
buying local inputs is likely to help the local upstream firm ensure that
these inputs possess some required characteristics or are of a minimum
quality. Similarly, the foreign firm may sell inputs that are tailored to the
downstream firms and raise their productivity. Kugler (2001) has found
this to be the case for Colombian firms, and similar evidence has been
found for other developing countries, including Indonesia.
Smarzynska (2000) shows, for transition economies in the early
1990s, that within high- and medium-technology sectors, foreign in-
vestors who are leaders in technology and marketing techniques tend to
engage in wholly owned projects rather than share ownership, in order
to minimize leakage. This result, however, does not hold in low-R&D in-
dustries. Smarzynska (2002) also finds for Eastern Europe and the for-
mer Soviet Union that weak protection of intellectual property rights
(IPRs) deters investors in high-R&D sectors that rely heavily on IPRs



GROWTH AND LOCATION      135
and encourages projects that focus on distribution rather than produc-
tion. A conclusion that can be drawn from this work is that IPR
protection is an important component of measures for attracting tech-
nology-intensive FDI. We saw in chapter 4 that domestic policy reforms
are essential to attract FDI; it would seem that to attract R&D-intensive
FDI, RIAs may require good IPR. laws.
FDI in the export sector can be expected to be more attractive than in
the import-competing sector. First, it is less likely to result in static wel-
fare losses (see chapter 4). Second, it is likely to be internationally com-
petitive, since it exports to the world market, or to benefit from prefer-
ential access to partners' markets. The formation of NAFTA has had the
effect of attracting FDI in export-oriented activities and is likely to result
in a "quality-ladder" phenomenon over time and in increased productiv-
ity. MERCOSUR, by contrast, has led to some tariff-jumping FDI, as in
the automobile sector, with production of goods that are mainly export-
ed to Brazil and are typically not competitive in OECD countries. These
conclusions are supported by the work of Balasubramanyam, Salisu, and
Sapsford (1996), who find that the growth effect of FDI is stronger in
countries which pursue export-promotion rather than import-substitu-
tion policies. The likelihood of FDI going to the export sector can be in-
creased by reducing protection to import-competing sectors.
Cross-Country Evidence on Openness and Growth
Despite the well-known shortcomings of cross-country regression as a re-
search technique-for example, its almost complete inability to establish
causation-huge use has been made of it as a tool for exploring the de-
terminants of economic growth. We review here some of the main stud-
ies on trade and growth based on cross-country regressions. Many of
them have included openness in one form or another, and nearly all have
identified a positive relationship between openness and growth. Several
of the earlier works in this genre (Dollar 1992; Sachs and Warner 1995a;
Edwards 1998) have received rather rough treatment recently at the
hands of Rodriguez and Rodrik (2001). The latter argue that these stud-
ies' measures of openness are so flawed that the relationship remains
unproved, especially as it pertains to the trade dimension of openness as
opposed to the macroeconomic (exchange rate) dimension.4 It is impor-
tant to note that Rodriguez and Rodrik do not argue that openness hin-
ders growth-there is no general evidence for this proposition-but
merely that the positive relationship is not yet fully established. Ro-
driguez and Rodrik's views have been challenged by Bhagwati and Srini-



136    REGIONAL INTEGRATION AND DEVELOPMENT
vasan (1999), who argue that since the case for openness was based on
far more than cross-country regressions, losing that strand of evidence
should not have a great effect on the overall conclusion. Commenting on
Rodriguez and Rodrik, Jones (2001) estimates some 100 specifications;
he concludes that trade restrictions are harmful to long-run income and
that the effects are potentially large, although there is a great deal of tin-
certainty regarding the magnitude of the effects. He also points out that
cross-country growth regressions are a coarse tool for addressing this
question.
In an influential study, Frankel and Romer (1999) go to great lengths
to sort out the causation between trade and growth. They show that the
part of trade attributable to purely exogenous factors such as population
size, land area, and distance seems to generate improved growth rates-
that large countries close to large markets grow faster. From this, they
tentatively infer that other aspects of trade-those attributable to poli-
cy-will also boost growth. The latter conclusion seems quite reason-
able, since many of the trade barriers that countries impose appear to be
analytically equivalent to transport costs, which Frankel and Romer
have shown to be inimical to growth.
All the studies mentioned above refer to nondiscriminatory liberaliza-
tion. The direct evidence that RIAs stimulate growth is weak. Henrekson,
Torstensson, and Torstensson (1997) use a cross-sectional regression to
suggest that European integration enhanced members' growth rates dur-
ing the period 1976-85, possibly by as much as 0.6-0.8 percent per year,
and that the influence operated through technology transfer rather than
through investment. Other commentators have failed to replicate these re-
sults, and Baldwin and Venables (1995) state that no study has identified
positive growth effects for non-European RIAs. Vamvakidis (1998) shows
that the EU's marginally significant positive impact disappears once one
has taken the openness of individual member countries into account.
Vamvakidis' results are probably more reliable than those of Henrekson,
Torstensson, and Torstensson because of his longer time period and be-
cause the latter sought to capture general openness via a price distortion
index based on Dollar (1992), which Rodriguez and Rodrik (2001) argue
is flawed. Vamvakidis does find beneficial effects on a country's growth
rate from having large, rich, open neighbors, but this is quite independent
of participation in RIAs. Cross-country growth regressions by de Melo,
Montenegro, and Panagariya (1993) and by Brada and Mendez (1988)
similarly find no growth effects from RIAs over the 1960s to 1980s.
The strongest and most direct approach to identifying the growth ef-
fects of non-European RIAs is that of Vamvakidis (1999), who uses



GROWrH AND LOCATION     137
panel data to explore whether countries' growth rates changed when
they liberalized their trade and compares the 10 years before liberaliza-
tion with the 10 years after. Vamvakidis finds strong evidence that
nondiscriminatory liberalizations boosted growth and that discriminato-
ry ones (that is, RIAs) did not. He did not explore the effects of the RIAs
created or revived in the 1990s because data for 10 years after liberaliza-
tion were not yet available. Leaving aside the possibility that new data
would show an improvement, Vamvakidis has provided strong evidence
that RIAs are neither good nor bad for growth. His results differ from
those based on the technology diffusion model, which indicate that the
southern member can gain from a North-South RIA.
AGGLOMERATION AND INDUSTRIALIZATION
Comparative advantage is not the only force that influences the location
of activity in an RIA. As economic centers start to develop, cumulative
causation mechanisms come into effect, leading to clustering (or agglom-
eration) of economic activity and extending the advantage of locations
that have a head start.5
Spatial clustering of economic activities is pervasive. Cities exist be-
cause businesses, workers, and consumers benefit from being close
together. Particular types of activity are frequently clustered, the most
spectacular examples being the concentrations of electronics industries in
Silicon Valley, cinema in Hollywood, and banking activities in the
world's financial districts. Clustering also occurs in many manufacturing
industries-for example, U.S. automobile manufacturing in the Detroit
area, and the medical equipment, printing machinery, and other indus-
tries studied by Porter (1998).
The Balance between Centripetal and Centrifugal Forces
Clustering or agglomeration typically arises from the interaction be-
tween centripetal forces, which encourage firms to locate close to each
other, and centrifugal forces, which encourage them to spread out. The
centripetal forces are usually classified in three groups (Marshall 1920):
*  Knowledge spillovers or other beneficial technological externali-
ties that make it attractive for firms to locate close to each other-
in Marshall's phrase, "the mysteries of the trade become no
mysteries, but are, as it were, in the air."
*  Various labor market-pooling effects, which encourage firms to
locate where they can benefit from readily available labor skills.



138     REGIONAL INTEGRATION AND DEVEI OPMENT
Linkages between buyers and sellers. Firms will, other things
being equal, want to locate where their customers are, and cus-
tomers will want to locate close to their suppliers. These linkages
are simply Hirs,chman's (1958) backward (demand) and forward
(supply) linkages. They create a positive interdependence between
the locational decisions of different firms that can give rise to a
process of cumulative causation, creating agglomerations of
activity.6
These centripetal, or agglomeration, forces can operate at an aggre-
gate level or can be much more narrowly focused. For example, aggre-
gate demand creates a backward linkage, drawing firms from all sectors
into locations with large markets. Other forces affect broad classes of
business activity: these include availability of basic industrial labor skills
and access to business services such as finance and telecommunications.
By contrast, knowledge spillovers affecting particular technologies, or
the availability of highly specialized inputs, might operate at an industry
level. In this case the forces work for clustering of the narrowly defined
sector rather than for clustering of manufacturing as a whole.
Pulling in the opposite direction are centrifugal forces, which include
congestion, pollution, and other negative externalities that might be as-
sociated with concentrations of economic activity. Perhaps the most
obvious indicator is the very high rents in major cities. Competition for
immobile factors will deter agglomeration, as the price of land (and per-
haps also of labor) is bid up in centers of activity. Similarly, the presence
of many firms in core areas increases competition and drives down mar-
gins. Finally, there are the demands of consumers who are not located in
the centers of activity; dispersed consumers will encourage dispersion of
producers, particularly if trade barriers or transport costs are high.
Trade Liberalization Can Aid Industrialization
Trade liberalization affects the balance of the centripetal and centrifugal
forces through at least three mechanisms (Puga and Venables 1997,
1998; see also box 5.1): (a) reductions in a country's import barriers im-
prove the market access of firms located in its partner countries;
(b) opening a country's markets to increased product market competition
from foreign firms reduces the profitability of local firms; (c) lower im-
port barriers mean cheaper imported intermediate goods and hence high-
er local profitability. The mix of these effects depends on the type and
extent of liberalization-unilateral, multilateral, or regional-and on the
relative strengths of the centrifugal and centripetal forces.



GROWTH AND LOCATION               139
Box 5.1 Modeling the Agglomeration Effects of RIAs
Puga and Venables (1998) outline a new approach for studying the effects of different types of trade
liberalization (multilateral, unilateral, regional, and so forth) on the Industrialization process in devel-
oping countries. Although their model is highly stylized, it provides the first formalization of phenom-
ena that have long been discussed at an intuitive level.
The authors consider a-three-country model containing a large northern country and two small
southern ones. Each country has two sectors: a perfectly competitive homogeneous commodity sec-
tor ("agriculture') and a monopolistically competitive industry ("manufactures") in which firms pro-
duce several varieties of differentiated products. In addition to real trade costs, ad valorem tariffs are
imposed on all trade flows in manufactures; trade in agricultural produce is free. For simplicity, Puga
and Venables choose their initial parameter values such that at the status quo ante equilibrium, the
two southem countries produce only agricullural products and no manufactures. From that starting
point, they solve the model for progressively lower tariff barriers. Trade liberalization always attracts
industry to the South, but the timing and magnitude of industrialization differ depending on the type
of liberalization and on the centripetal and centrifugal forces.
Under multilateral trade liberalization, industry initially starts up in only one of the two southern
economies; agglomeration economies make it unprofitable to maintain two new locations before one
has achieved a reasonable size. As trade barriers continue to decline, it becomes profitable for man-
ufacturing to set up in the other southern country as well, partly at the expense of the first, which
suffers a small fall In its share of world industry. At tariffs below that point, the southern economies
are identical, and further reductions in tariffs bring a steady relocation of industry from the North to
both economies. Unilateral liberalization by the developing countries also promotes their industrial-
ization because it makes for cheaper intermediate inputs from the North, but its effects are weaker
than those of multilateral liberalization because the market access dimension is lacking.
Membership in an RIA may or may not be better than unilateral reform. South-South RIAs will be
sensitive to the market size of member states, but above a critical level, industry establishes in both
partners because of the effective enlargement induced by the reciprocal reduction in intra-South
barriers. The spread of industry is uneven, however, with industry initially developing in one of the
countries and spreading to the second only when trade barriers are lower. The southern economies
attract less industry than they would under multilateral liberalization because they do not benefit
from greater access to the northem market or to intermediate goods produced in the North.
Compared with a South-South arrangement, a North-South RIA offers better prospects for the par-
ticipating southern country. Industry spreads more quickly and to a larger extent because the south-
ern economy benefits from both improved access to the northem market and the low cost of inter-
mediates. In this model, with only one northern market or supplier, the participating southern country
is better off in a North-South RIA than under multilateral liberalization. Both deliver the same market
access and cheap intermediates, but under the RIA, competition is weaker because the excluded
southern country is presumed to receive neither advantage. And even if the participating country is
better off than under multilateral liberalization, the South as a whole is worse off because the ex-
cluded South does not receive any industry.
Multilateral reduction of trade barriers can increase the profitability
of industrial location in developing countries by decreasing the price of
imported intermediate goods and offering firms better access to larger
developed country markets. It can, however, also reduce profitability by



140    REGIONAL INTEGRATION AND DEVELOPMENT
making markets more competitive. The first two effects, combined with
large wage differences between northern and southern countries, can
dominate and cause industry to move to the South. As the world trading
system has become more open, industrial production has gradually
moved from the United States, Western Europe, and Japan to developing
countries in Latin America, Southeast Asia, and Eastern Europe. The
spread has been quite uneven across developing countries, since the
stronger centripetal forces in middle-income countries make them more
attractive as industrial locations. For example, in 1995 Asia, with 33
percent of the developing world's total income, received 65 percent of all
FDI flows to developing countries, while Latin America (with 31 percent
of total income) got 27 percent and Africa (8 percent of total income)
only 5 percent.
Unilateral trade liberalization by one developing country can attract
industry and bring real income gains, despite more intense import com-
petition, if the availability of cheaper imported intermediate goods be-
comes a dominant force. Industry will develop sooner, and on a larger
scale, the greater is the share of intermediates in production and the larg-
er is the market in the liberalizing economy. The spread of industry to the
South, however, is likely to be slower than under multilateral liberaliza-
tion, as unilateral liberalization does not improve developing countries'
access to large northern markets.7
Preferential trade liberalization affects both the locational attractive-
ness of the bloc relative to the rest of the world and the relative attrac-
tiveness of individual bloc members. An RIA increases the locational
advantage of its members at the expense of the excluded countries; it al-
lows member country firms to sell their output and buy their inputs on
larger markets, setting off beneficial centripetal forces and pulling pro-
duction and industry into the bloc and away from countries outside the
RIA (Puga and Venables 1997, 1998).A The strength of these effects will
depend on the size and nature of the integrated market (that is, its com-
bined income and output and the depth of its integration), and many
developing country RIAs will be too small to have significant locational
effects. For example, none of the active African RIAs has a GDP larger
than Belgium's; UEMOA's GDP is less than 10 percent of Belgium's GDP,
CEMAC's is less than 4 percent, and even the SADC's is less than 65 per-
cent. Recalling the economies of scale arguments that are commonly
made for European integration, it is plain that for many industries, such
groups are well below the necessary size for attracting industry.
Moreover, even for industries where scale requirements are low, many
South-South RIAs have failed to attract activity because, at least until re-



GROwTII AND LOCATION     141
cently, they have lacked the institutional preconditions. In most African
RIAs, not only are markets too small but, in addition, infrastructure is
poor, a sound legal framework is lacking, progress in economic reform
has been slow, there is heavy government intervention, the private sector
is small and fragile, and economic and political instability are common.
We argued in the preceding chapter that addressing these matters is the
key to investment, and the same is true for industrial agglomeration.
Unilateral reforms to achieve a sound business environment are the sine
qua non of industrial development and are far more important than re-
gional integration issues of any hue.
Intermember Distribution: Divergence Is Likely in South-South RIAs
Turning to the internal distribution of industry, we need to ask how re-
gional integration affects the balance between centripetal and centrifugal
forces. Can membership bring about or amplify the clustering of eco-
nomic activity, and if so, might it widen income differentials between
partner countries?
By reducing trade barriers, membership in an RIA makes it easier to
supply customers from a few locations. This suggests that the balance of
forces may be tipped in favor of agglomeration, although the ensuing re-
location of industry could develop in several different ways.
One possibility is that particular sectors become more spatially con-
centrated. This is likely if the centripetal forces act at a narrow sectoral
level. For example, industries in the United States are much more spatial-
ly concentrated than in Europe, even controlling for the distribution of
population and of manufacturing as a whole. This difference suggests
that further regional integration in Europe could cause agglomeration at
the sectoral level-for example, Germany gets engineering, financial
services cluster in the United Kingdom, and so on. The possibility that
this might happen is generating some concern in Europe, although the
evidence for it is so far rather weak (see Midelfart Knarvik and others
1999). If it does happen, it will create considerable adjustment costs as
the industrial structure of different locations changes, but it will also
yield aggregate benefits because of the real efficiency gains from spatial
concentration. This sectoral agglomeration need not be associated with
increases in intra-RIA inequalities; each country or region may attract
activity in some sectors.
An alternative possibility is that instead of there being relatively small
sectors, each clustering in different locations, manufacturing as a whole
comes to cluster in one or a few locations, leading to deindustrialization



142    REGIONAL INTEGRATION AND DEVELOPMENT
of the less favored regions. This outcome is more likely, the smaller is
manufacturing as a whole-either because it is a small share of the
economies or because the economies are small overall. Small sectors need
to concentrate to create critical mass, and if, in addition, the share of
manufacturing is small, fitting the whole of manufacturing into one or a
few locations is less likely to encounter factor supply constraints and to
lead to rising prices of immobile factors such as land. Concentration will
also be more likely if linkages are broad, across many industrial sectors,
rather than narrowly sector specific. This situation is more apt to occur
in early stages of development, when a country's basic industrial infra-
structure-transport, telecommunications, and access to financial mar-
kets and other business services-is thinly developed and unevenly
spread.
This suggests the real possibility that RIA membership could lead to
agglomeration and divergence between member countries, reinforcing the
effects of comparative advantage discussed in chapter 3. It seems likely
that both comparative advantage and agglomeration are at work in some
South-South RIAs, with both forces leading industry to agglomerate in
the relatively richer and initially more industrialized members. As Nairo-
bi, Abidjan, and Dakar have attracted manufacturing, so they have start-
ed to develop business networks and the linkages that tend to lock manu-
facturing into the location. The process might be further accelerated by
the propensity of FDI to cluster in relatively few locations. Agglomeration
then accentuates the forces for divergence that we outlined above.
Industry agglomeration in a subset of member countries can create ten-
sions within an RIA. Several South-South RIAs have failed because of dis-
putes over the location of industry and the design of compensation schemes
for perceived losers. In the 1950s and 1960s severe frictions between
Kenya, Tanzania, and Uganda arose over the benefits from economic inte-
gration within the East African Community. Uganda and Tanzania con-
tended that all the gains were going to Kenya, which was steadily enhanc-
ing its position as the industrial center of the common market, producing
70 percent of the manufactures and exporting a growing percentage of
them to its two relatively less developed partners. By 1958, 404 of the 474
companies registered in East Africa were located in Kenya. By 1960,
Kenya's manufacturing sector accounted for 10 percent of its gross nation-
al product (GNP); in the other two states the share of manufacturing was 4
percent (Hansen 1969). The community collapsed in 1977, having failed to
satisfy the poorer members that they were getting a fair share of the gains.
In other circumstances agglomeration forces can help offset the diver-
gences associated with comparative advantage. For example, firms



GROWTH AND LocArION     143
choosing a location in NAFTA may want both the agglomeration bene-
fits of locating in the United States and the factor price advantages of
Mexico, but the latter may predominate. Thus, in North-South arrange-
ments the South could become a preferred location for assembly indus-
tries (for example, automobiles) or subcontracting industries such as
textiles, possibly in export-processing zones. This happened with the
growth of automotive assembly plants (maquiladoras) south of the U.S.-
Mexico border: major investments (exceeding US$500 million) in new
capacity and plant expansions in Mexico by Japanese, German, and Ital-
ian auto parts producers benefited from NAFTA-related changes in Mex-
ico's Auto Decree that liberalized foreign investment restrictions in the
sector and provided access to Mexican and Latin American markets
(USITC 1997).
These examples suggest that from the point of view of the participat-
ing developing economy, North-South arrangements are better than
South-South ones because they give a southern economy the benefits of
both improved access to a large northern market and low-cost northern
intermediates. For the same reasons, southern economies attract more in-
dustry under multilateral trade liberalization than under South-South
schemes. In fact, it is possible that a North-South RIA will generate more
industry for the southern economy than would a multilateral liberaliza-
tion, depending, among other things, on how important the northern
partner is in the South's total exports. Where this happens, however, it is
at the expense of excluded southern countries, which industrialize more
slowly because they neither gain market access nor liberalize their own
regimes (see box 5.1).
There are also likely to be locational effects within member countries.
The strongest example is probably Mexico, where liberalization and ac-
cession to NAFTA have shifted the center of economic gravity from
Mexico City to the northern border. Krugman and Elizondo (1996) have
argued that closed economies such as pre-1986 Mexico are more prone
to extreme agglomeration because the only linkages are domestic ones.
The Mexican trade liberalization vis-a-vis the rest of the world started to
undermine the dominance of Mexico City and to disperse industry to
other locations in the country. This was supplemented by the increasing
importance of the United States as a trading and production-chain part-
ner as it introduced border preferences and subsequently signed NAFTA
(see box 5.2).
An important final point is that agglomeration forces will be strongest
at intermediate levels of trade barriers (or transport costs). When barri-
ers are very high, each country will have its own industry to supply local



144       REGIONAL INTEGRATION AND DEVELOPME.NT
Box 5.2 The Influence of Regionalism on Within-Country Location: North America
If the main nonpolicy barriers to trade are transport costs, economic integration should increase eco-
nomic activity in border cities (Hanson 1996). The lure of frontier regions is stronger if production of
final goods uses foreign intermediate inputs. After the U.S. Automotive Products Trade Act of 1965
eliminated barriers on motor vehicle trade with Canada, motor vehicle production expanded along
the Michigan-Ontario corridor.
Integration with the United States has had strong effects on industrial location in Mexico, where in-
dustry has shifted toward states with good access to the U.S. market. This integration has been ef-
fectively under way since the mid-1 980s, when Mexico removed most barriers to foreign trade and
lifted restrictions on foreign ownership. Mexico has developed a large export-manufacturing indus-
try that specializes In the assembly and processing of foreign-made components. A large part of this
industry imports most of its inputs from, and ships most of its output back to, U.S. firms. The plants
are overwhelmingly concentrated in Mexican cities on or near the U S.-Mexico border. Employment
growth has been higher in regions that have larger agglomerations of industries with buyer-supplier
relationships with U.S. firms (Hanson 1998).
The growth of export production in Mexican border cities also makes U.S. border cities natural sites
for complementary manufacturing activities, and production activities in the United States are relo-
cating accordingly. U.S. border cities specialize in the industries that produce parts and components
for Mexican assembly plants. Hanson (1996) argues that NAFTA, by further lowerng trade barriers,
is likely to contribute to the further expansion of the binational production centers that are forming
along the border. He finds that employment growth in U.S. border cities is positively correlated with
employment growth in Mexican export assembly plants.
consumers. When they are very low, firms go where labor costs are
cheapest because they can bring in their inputs and ship their output at
very low cost, as with the production networks described in box 5.1. But
where barriers are intermediate, firms are reluctant to move away from
suppliers and other agglomeration benefits, yet they are able to supply
foreign markets through exports.
NOTES
1. Coe and Helpman (1995) and Coe, Helpman, and Hoffmaister (1997) as-
sume, rather than ascertain by testing, that imports from          industrial countries
provide the correct weights with which to combine those countries' stocks of for-
eign knowledge to reflect importers' access to it. Keller (1998) finds that Coe and
Helpman's results are little better than can be obtained through a random
weighting. Lumenga-Neso, Olarreaga, and Schiff (2001), who extend the Coe
and Helpman approach by taking indirect knowledge spillovers into account, ob-
tain results that strongly dominate those of Coe and Helpman and weakly domi-
nate Keller's. Coe and Hoffmaister (1999) show that the weights used by Keller
were not truly random      and find that Coe and Helpman's results are superior to



GROWrH AND LOCATION        145
those obtained with random weights, in which case R&D spillover effects are
nonexistent. Both sets of results suggest that trade is an important channel for
knowledge spillovers, although the testing of alternative hypotheses remains to
be conducted.
2. Schiff and Wang (2002c) show that Poland obtains larger TFP gains by
trading with the United States and Canada than by trading with the EU.
3. Schiff and Wang (2002b) teste(i whether the difference in effects was attrib-
utable to NAFTA itself by allowing the various parameters to differ in the post-
NAFTA period from the pre-NAFTA one. None of the parameters was found to
be significantly different in the later period.
4. Commenting on Rodriguez and Rodrik, Hsieh (2001) provides suggestive
evidence that restrictions on capital goods imports have important adverse ef-
fects on growth, particularly in developing countries. This is consistent with the
findings of Coe, Helpman, and Hoffmaister (1997).
5. This section is based on Fujita, Krugman, and Venables (1999).
6. This argument only works if there are increasing returns to scale in produc-
tion; if there are not, firms can put small plants in many different locations.
7. For many goods, such access is already pretty free, so unilateralism is not
completely stymied.
8. Of course, these effects are in addition to the (positive or negative) resource
allocation effects analyzed in chapters 2 and 3.






CHAPTER 6
Integrating Domestic Policies*
Cooperation on domestic policies can substantially increase the gains from
forming a trade bloc. It can lift barriers that insulate national markets and
deliver economic benefits many times those available from preferential
trade agreements. Intergovernmental cooperation in designing and apply-
ing domestic policies such as taxes, health and safety regulations, environ-
mental standards, and so on-what we call policy integration-can in-
crease competition in domestic markets by reducing transactions costs and
allowing new suppliers to enter rnarkets. Cooperation on domestic poli-
cies can also help in overcoming market failures and ensuring that trade
restrictions are not reimposed through the back door.
Most existing RIAs aim only to reduce market segmentation by con-
straining the use of national policies rather than by actively integrating
them. The main exceptions are the EU and the Closer Economic Rela-
tions (CER) agreement between Australia and New Zealand. A number
of other RIAs, including some still under construction, are beginning to
discuss policy integration. But without specific timetables for action and
further negotiation, neither new nor existing RIAs are likely to make
great progress. Experience suggests that negotiated policy integration is
very demanding, both politically and technically.
An RIA can assist policy integration by providing an institutional
framework and by making available sources of gain that may help over-
come opposition to policy reforms. Similarly, policy integration can assist
the implementation and enforcement of the trade component of an RIA,
* This chapter was co-authored by Bernard Hoekman.
147



148    REGIONAL INTEGRATION AND DEVELOPMENT
which (provided that the trade provisions are desirable) could be a major
gain. Regulatory policies are even more vulnerable to capture than are
trade policies, however. Care is required to prevent such policy-bundling
from merely creating rents for the few at the expense of the many.
Competition between regulatory regimes, coupled with partners' mu-
tual recognition of each other's standards or regulations, can be a useful
approach to policy integration, but it is not without its challenges. Har-
monization may still be called for in some dimensions to avoid adverse
spillovers, such as the threat of a "race to the bottom," or threats to pub-
lic health or safety. Such harmonization is generally best limited to mini-
mum standards based on global norms; idiosyncratic regional standards
will be optimal only if region-specific characteristics exist. Mutual recog-
nition agreements, whereby countries agree to treat goods meeting part-
ners' standards as if they met their own, can easily become barriers to
trade with nonpartners. Wherever possible, policy integration to reduce
regulatory costs should extend beyond RIA partners to nonmembers so
as to maximize the increase in competition. Certainly, formal intergov-
ernmental agreements are necessary for some elements of policy
integration, but special efforts should be made to avoid perpetuating or
increasing de facto discrimination.
Moreover, RIAs are not the only route to policy integration; much can
be done to reduce transactions costs and facilitate market access by adopt-
ing international standards and best practices unilaterally and by partici-
pating in multilateral fora with global membership. Indeed, unilateralism
and multilateralism have been more common than regional approaches to-
ward coordinating technical and regulatory standards. Developing coun-
tries can achieve many of the benefits of policy integration unilaterally by
adopting international standards and recognizing the regulatory norms of
their major markets or suppliers such as the EU and the United States.
For this reason, the WTO and other multilateral institutions play a large
role in policy integration. The WTO agenda is as broad as, or broader
than, that of most RIAs, and more can be done in the WTO context than
is sometimes recognized. The WTO could help further by requiring mem-
bers to apply the MFN principle to policy integration initiatives that do
not require formal intergovernmental equivalency or recognition agree-
ments, such as customs clearance documentation and procedures.
Policy integration proposals are very specific and should be evaluated
on their individual merits. Care should be taken to ensure that they are
for the general good. Except as part of a necessary coalition for achieving
reform, efforts to link them to regional trade liberalization should be re-
sisted. While integrating domestic policies may sometimes require formal



INTEGRATING DOMESTIC POLICIES  149
agreements, there is no fundamental reason why it should require trade
preferences. The United States and the EU, for instance, have drawn up a
series of mutual recognition agreements for sectoral product standards
completely outside an RIA context. To date, however, developing coun-
tries have been entirely excluded from such initiatives.
This chapter explores the role that regulatory policy interaction be-
tween RIA members could play in developing country RIAs. It draws
mainly on EU experience, since that is by far the most advanced example
to hand. Its main conclusion is that policy integration offers scope for
considerable gains, but only at the expense of very hard political and
technical work and at the risk of exacerbating rather than erasing
distortions. Because of the difficulties of achieving policy integration,
governments should not casually assume that deep integration will auto-
matically make their RIAs welfare-enhancing. They will need to be both
determined and well informed if integration is to be turned to good ef-
fect. Governments should also be alert to the opportunities for reducing
transactions costs and market segmentation held out by multilateral or-
ganizations and unilateral action.
DEFINING POLICY INTEGRATION
In this section we define policy integration and discuss the modalities
available for pursuing it. We distinguish between agreements that treat
foreign goods or firms exactly the same as domestic ones (which is not
policy integration) from those that involve negotiating how both sets of
goods or firms are treated (which is). We then distinguish between inter-
governmental coordination of policies, harmonization of standards
across countries, and recognition by governments of each other's stan-
dards as approaches to policy integration. We largely ignore the first ap-
proach, which is essentially ad hoc and nonbinding, and concentrate on
the remaining two. In addition, because our focus is on the international
trade dimensions of RIAs, we also ignore government cooperation on
nonregulatory matters such as infrastructure and water resource man-
agement and on issues of monetary integration. Both are hugely impor-
tant but lie outside the boundaries of this book.'
The Baseline: National Treatment
National treatment requires that foreign products or producers, once
they have entered a territory, receive the same treatment as domestic
counterparts with respect to taxes, health and safety standards, competi-



150    REGIONAL INTEGRATION AND DEVELOPMENT
tion rules, and so on. National treatment has always been a basic build-
ing block of international trade treaties. It ensures that liberalization
commitments cannot be circumvented by the discriminatory application
of domestic policies, such as an excise tax that is higher for foreign than
for domestic products.
National treatment does not constrain a government's policy sover-
eignty per se; it merely precludes discrimination in favor of domestic sup-
pliers and allows foreign goods, services, and factors to compete with
domestic ones on an equal basis. It can have particularly far-reaching ef-
fects if it is applied to services and factor markets. Although much rarer
than preferential agreements in goods, stand-alone agreements have been
used to create common markets for factors of production. For example,
Denmark, Finland, Iceland, Norway, and Sweden operate a common
labor market that has eliminated immigration controls for their resi-
dents. Similarly, the 1973 Trans-Tasman Travel Arrangement gives citi-
zens of Australia and New Zealand the right to work and reside in either
country. In both instances, the common labor market is underpinned by
agreements on health services and social security, illustrating that agree-
ments in the labor area are multidimensional packages.
As we saw in chapter 4, national treatment has also been important
for investment flows, assuring investors that governments will not dis-
criminate against them after they have invested. RIAs vary in the extent
to which they offer national treatment to investors, with the EU,
NAFTA, and the EU's Europe Agreements going farthest toward abolish-
ing performance criteria and related policies such as local content and
trade-balancing requirements. Only the EU has completely liberalized
market access and entry by foreign firms across the board, although U.S.
investment treaties normally seek such pre-establishment rights at a sec-
toral level and the Doha Development Agenda, agreed on at the WTO
ministerial conference in 2001, calls for WTO members to consider
extending these rights. As has been noted, however, RIAs are far from
necessary for establishing investment agreements. There exist many
stand-alone investment agreements and a few multilateral ones.
Beyond National Treatment: Policy Integration
Policy integration involves deliberate actions by governments to go be-
yond national treatment and seek to reduce the market-segmenting effect
of regulatory regimes through coordination, harmonization, or mutual
recognition of national policies and enforcement mechanisms (box 6.1).



INTEGRATING DOMESTIC POLICIES                  151
Box 6.1 DefiningTerms: "Deep" and "Shallow" Integration
The term "shallow" integration is sometimes used to describe integration based on national treat-
ment. "Deep" integration (the term favored by Lawrence and Litan 1990) then refers to efforts that
go beyond that. Tinbergen (1954) coined the terms "negative" and "positive" integration, which cor-
respond, basically, to shallow and deep integration. Negative integration entails removal and limita-
tion of policies, as distinguished from positive efforts to coordinate them.
The terms "deep" or "positive" are misleading if used to suggest that this type of cooperation is
more profound, far-reaching, binding, or beneficial than integration based on nabonal treatment. This
is not necessarily so. "Deep" cooperation may be limited to only a few areas of little economic con-
sequence. It may even be welfare reducing for one or more members.of an RIA if it involves the
adoption of rules that are suboptimal from a national perspective. In this volume we reserve the term
"deep integration" for agreements such as the EU that aim at achieving some measure of economic
union and that create supranational institutions for pursuing this goal. The more general term
"policy integration" is used for less far reaching cooperation on domestic policies and regulations.
It can be pursued unilaterally or cooperatively. International cooperation
is generally driven by market access concerns and by recognition that in-
creased competition will bring economic benefits. Experience suggests
that a mix of harmonization and mutual recognition is required to
achieve policy integration where the underlying regulatory policies have
national health, safety, or prudential objectives.
Modes of Policy Integration: Coordination, Harmonization,
and Recognition
The principal instruments of policy integration are coordination, har-
monization, and          acceptance      ("recognition") of foreign            regulatory
regimes. Countries can cooperate on domestic policies through com-
mercial agreements to facilitate trade and investment flows or by pur-
suing more far-reaching economic unions that require them to cede
sovereignty to supranational institutions which develop and enforce
rules. With the notable exception of the EU, existing RIAs are primarily
commercial instruments that do not aim at economic union. Policy in-
tegration, if pursued, is intergovernmental and issue specific, driven by
a realization that there are potential gains from cooperation in a par-
ticular area.
Coordination of Domestic Policies. Coordination is the least far-
reaching type of policy integration. It is limited to efforts by governments



152    REGIONAL INTEGRATION AND DEVELOPMENT
or regulatory bodies to cooperate in developing or implementing a norm
or rule, and it involves "voluntary and largely unenforceable alignments
of national policies and measures in particular fields" (Robson 1998).
Coordination may be based on formal agreements on, for example, the
use of the principle of positive comity in the application of competition
law, or may be ad hoc, as in cooperation on infrastructure projects.2 We
do not discuss coordination further in this book, as it implies no binding
commitment on the part of governments.
Harmonization of National Standards and Regulations. Harmoniza-
tion may involve unilateral adoption by one country (or RIA) of anoth-
er's set of rules, or negotiation of a common set of disciplines. Examples
abound of unilateral harmonization to the standard of another country
or RIA. These cases are often driven by market-size disparities. For ex-
ample, in 1992 Canada adopted U.S. auto emissions standards to ensure
that its automakers could realize economies of scale by avoiding separate
production lines for its home and U.S. markets. Switzerland, similarly,
adopted the EU regime on technical regulations and industrial standards
so that Swiss goods could enter and circulate in the EU on the same basis
as EU-produced goods (Messerlin 1998).
Many developing countries use legal regimes developed in Europe or
the United States, usually maintaining systems inherited from a colonial
past or military occupation. Others have deliberately adopted foreign
norms. Korea imported many German and U.S. product standards in the
1950s as part of a strategy for upgrading the quality of industrial pro-
duction and fostering exports.
Harmonization between more equal partners requires a deeper institu-
tional setting. Efforts to set common norms may be limited to intergov-
ernmental cooperation or may involve a decision to cede sovereignty to
common or supranational institutions for rule-setting. In the EU the
power to propose directives and regulations (that is, to propose rules)
has been delegated to the European Commission, and the European
Court of Justice has strong supranational powers of enforcement. In
NAFTA harmonization is limited to member states' acceptance of bind-
ing independent third-party adjudication or arbitration on disputes
about rules, which is much less intrusive than the EU arrangements.
Recognition of Foreign Regulatory Regimes and Conformity Assess-
ment Procedures. For a country that sets certain norms for goods and
services within its borders, the natural first thought is to test imports
on entry. It is, however, potentially very costly to devise and enforce



INTEGRATING DOMESTIC POLICIES    153
standards, and so there are incentives to cooperate with trading partners.
Unilateral recognition of foreign regulatory regimes is the simplest route:
a country just adopts international norms or the standards of a trading
partner. Thus, a government may decide that the professional qualifica-
tions of doctors trained and certified in certain countries are sufficient to
allow them to practice in the country (although other constraints may
still restrict entry by foreign services providers). Similarly, foreign certifi-
cation for certain imports may be accepted as proof of safety; the Under-
writers Laboratories (UL) mark, for example, is accepted in many
countries. The key to unilateral recognition is familiarity with and trust
in partner standards and certification systems.
Where familiarity or trust are weaker, or where, for nationalistic or
mercantilist reasons, partners are not willing simply to adopt each
other's standards, mutual recognition agreements (MRAs) are a possibil-
ity. These entail each member's recognizing its partners' standards as
acceptable inside its own boundaries, but without applying them to its
domestic suppliers. MRAs can cover either the standards themselves or
the conformity assessment systems used to establish compliance with
standards.
Harmonization as a tool of policy integration and the idea of relying
on the principle of mutual recognition to facilitate trade have been pur-
sued most intensively by the EU. MRAs may be required even if a unilat-
eral harmonization strategy is followed, as the trading partner whose
standards are emulated may not accept foreign test results or conformity
assessment systems as equivalent to its own even if the formal standards
are identical. Conversely, European experience suggests that mutual
recognition may require some degree of harmonization of either stan-
dards or testing procedures to ensure that the underlying norms satisfy
basic minimum standards, especially in areas where mandatory stan-
dards or regulations apply (see box 6.2).
Mutual recognition has proved a useful tool for increasing competi-
tion in European markets, but even there it has turned out to be hard
work (see Messerlin 2001), and it may not be viable at all for developing
country RIAs. The process relies heavily on mutual trust in the compe-
tence and ability of the institutions responsible for enforcing mandatory
standards and on a willingness to accept some compromise in setting
minimum standards. Even if developing countries adopt EU, U.S., or in-
ternational standards, significant institutional strengthening is likely to
be required before developed country governments will accept "home
country supervision." Alternatively, greater reliance could be placed on
third-party conformity assessment of goods and services.



154       REGIONAL INTEGRATION ANI) DEVELOPMENT
Box 6.2 Harmonization and Mutual Recognition as Substitutes or Complements:
The EU Experience
Article 36 of the Treaty of Rome permits EU members to maintain domestic policies that restrict trade
if necessary for protecting national health, security, or morals or the environment. Virtually all RIAs
have similar provisions. Initially, the EU sought to limit the market-segmenting effects of national
regulations under Article 36 through a process of harmonization. Many of the early efforts toward
harmonization centered on food standards-the first harmonization directive, issued in 1962, dealt
with food coloring. Progress toward harmonization was very slow, in part because adoption of an EU-
wide norm required unanimity. It took over a decade to reach agreement on the composition of fruit
jams and mineral water, and only nine directives on foodstuffs were adopted between 1962 and
1979. Differences in national norms, reflectng national tastes, histories, and legal regimes-and the
efforts of producer lobbies seeking to restrict competition from imports-made it difficult to achieve
the required consensus. For example, the Germans set great store by their Reinheitsgebot, a stan-
dard established in 1516 specifying that beer may have only four ingredients: malted barley, hops,
yeast, and water. Other countries permit preservatives or additives.
The result of such differences was that little progress was made toward reducing the market-
segmenting effects of national standards. Indeed, member states continued to adopt numerous in-
consistent and idiosyncratic product regulations. In a landmark 1979 case the European Court of
Justice found that a German ban on the sale of a French Cassis de Dijon used to prepare kir, an aper-
itif, could not be justified on the basis of public safety or health. This established the principle, later
incorporated into the 1987 Single European Act, that goods legally introduced into circulation in one
member state could not be barred from entering and being sold in another.
The EU's 'new approach" to standards, which dates from 1985, differentiates standards that
have health and safety (public interest) dimensions from those that do not. For the latter, govern-
ments must recognize other members' regulations as equivalent to their own. For the former, rather
than full harmonizaton, members agreed on a process of determining common minimum standards
("essential requirements"). Moreover, as part of the Single Market Programme, standards became
subject to qualified majority voting rather than unanimity. Between 1987 and 1995, 28 new-
approach directives were adopted specifying essential requirements. Meanwhile, under the old
approach, which also became subject to majority voting, adoptions increased from an average of
12.5 a year over 1958-85 to 20.8 a year over 1986-97 (Pelkmans 1990; Neven 1996). Even at these
rates, standards are likely to remain a significant barrier for some years.
Clearly, mutual recognition cannot wholly replace harmonization as an approach to standards. In
many cases harmonization is required to define "essential requirements," and governments will not
agree to mutual recognition if they feel that vital interests are at stake. Moreover, it is not correct to
argue, as Torrent (2002) does, that mutual recognition is merely another route to harmonization.
Where the differences between standards are minor or concern unimportant dimensions, govern-
ments may be prepared to accept other members' standards as long as their own are accepted else-
where in return, even if in a full harmonization negotiation they would feel obliged to hold out for
their own positions.
THE ECONOMICS OF POLICY INTEGRATION
Policy integration can benefit countries by reducing transactions costs.
This saves resources from being wasted on unproductive activities, low-
ers market access barriers, and integrates segmented national markets
for similar goods and services, all of which increases the degree of



INTEGRATING DOMESTIC POLICIES    155
competition. Well-executed policy integration can benefit both RIA part-
ners and, on occasion, producers in the rest of the world. In this section
we discuss five elements of the economics of policy integration: transac-
tions costs, increased competition, spillovers (externalities), compensa-
tion of losers, and identification of the appropriate geographic level for
cooperation. Most candidate policies for integration contain elements of
several of these issues.
Transactions Costs
Domestic regulations can segment markets by impeding foreign firms
from competing with national ones. This may be deliberate or may sim-
ply be a side effect of a policy objective. For example, health and safety
standards may entail duplicative testing and conformity assessment in
both exporting and importing countries. Customs procedures may also
be duplicative or redundant: tax authorities in an exporting country may
require data very similar to that demanded by the importer's customs of-
ficials, but in a different format. All such administrative requirements
impose transactions costs on enterprises that engage in international ex-
change and raise consumer prices.
Border Formalities. It has been estimated that in the early 1990s the
costs of border formalities affecting intra-EU trade (which was already
mostly duty free except for some goods originating in Portugal and Spain)
were equivalent to more than 1.2 percent of the gross value of internally
traded goods (European Commission 1997). And intra-EU trade proce-
dures were efficient compared with those in many other countries. In 1988
the EU had already adopted a Single Administrative Document (the EUR-
1 form), and many members had simplified their procedures to reduce cus-
toms burdens for large traders. Similarly, in intra-EFTA trade, Herin
(1986) estimated that complying with rules of origin cost an average of 3
percent of the value of a transaction, even though for such administrative
tasks, EFTA was probably one of the cheapest places in the world.
Standards Certification. Standards represent another potential source
of large gains from the reduction of transactions costs. Over 60 percent
of U.S. exports are subject to health, safety, and related standards in their
destination markets. Government-issued certificates are required for
45 percent of exports to the EU; private third-party certification is ac-
cepted for 15 percent; and manufacturers' self-certification suffices for
the rest (Wilson 1996: 7). Within the EU, about 75 percent of the value
of intra-EU trade in goods is subject to mandatory technical regulations



156    REGIONAI INTEGRATION AND DEVEIOPMENT
(European Commission 1996). Certification in regulated sectors can in-
volve frequent and redundant sampling of products-sometimes up to
100 percent of foreign goods-and this can effectively block imports.
Duplicative testing and certification requirements have rapidly be-
come more important as barriers to international trade. Unter (1998) es-
timates that redundant testing and conformity assessment procedures
faced by Hewlett-Packard increased sixfold over 1990-97. The EU now
requires third-party testing, certification, or quality system registration
for certain regulated sectors. It will accept certification only by organiza-
tions that have been approved by the member states as being technically
competent and that have been notified as such to the European Commis-
sion. This considerably raises the costs for non-EU manufacturers in
many sectors. Avoidance of these costs was a prime motivation for the
EU-U.S. MRA negotiations.
EU-U.S. trade talks on mutual recognition of conformity assessment
began in 1992, with the aim of ensuring that assessments performed by in-
dependent entities would be accepted in both markets. The EU sought as-
surance that U.S. certifiers were competent to test for EU "essential
requirements" and that European firms could test and certify to the
corresponding U.S. regulatory requirements. On its side, the United States
sought to eliminate the costs of exporters' having to obtain certification
by EU-approved bodies in addition to meeting U.S. requirements. The
EU's increasingly communitywide approach to standardization did, how-
ever, mean that U.S. firms now had to seek only one certificate for all EU
markets, which greatly increased the returns to negotiating an agreement.
Significant differences in European and U.S. testing and certification
systems made agreement difficult. The European system relies less on
firms' self-declarations of conformity, and more on mandatory third-
party testing and certification, than does the U.S. system. As of 1998,
member states had certified only 600 testing bodies to the European
Commission out of a total of over 10,000 such bodies in existence, rang-
ing from large multinationals such as SGS, Inchcape, and Bureau Vertitas
to small in-house testing facilities. Virtually all those approved were Eu-
ropean (Mcsserlin 1998).
In June 1997 the United States and the EU concluded an MRA that
covered conformity requirements in telecommunications equipment, in-
formation technology products, medical devices, and pharmaceuticals
and that was expected to address acceptance of test data, laboratory
accreditation, and final product certification. After a two-year period,
certifications performed anywhere by a facility recognized under the
MRA in the United States or Europe were to be accepted.



INTEGRATING DOMESTIC POLICIES   157
The agreements cover more than $40 billion of bilateral trade. The
MRA on telecommunications and information technology products
alone could save consumers and manufacturers approximately $1.4 bil-
lion, implying that the frictional costs abolished were equivalent to a 5
percent tax on the goods traded (Wilson 1999). Although this is a signif-
icant cost reduction, the MRAs are regarded as a second-best solution by
U.S. industry, which would prefer to rely much more heavily on supplier
self-certification instead of third-party conformity assessment.
MRAs have been a much less satisfactory route for developing and
transition countries seeking to reduce standards-related costs in the EU.
Their own standards are not acceptable in the EU, and they find it diffi-
cult to prove that their testing institutions are reliable. The Europe
Agreements, signed from 1992 on between the EU and the Central and
East European countries, recognized the possibility of concluding MRAs.
By 1995-96, however, the EU had decided to push instead for complete
harmonization, at least in part because it felt that recognizing transition
economy standards, some of which were suspected of protectionist in-
tent, would cause significant difficulties.3 Progress on complete harmo-
nization has been very slow.
One transition economy, the Czech Republic, cut through the impasse
with respect to its own imports by applying the principle of "one stan-
dard, one test, supplier's declaration of conformity" (Unter 1998). Under
this approach, goods that satisfy international standards and are accom-
panied by a supplier's declaration (based on in-house or third-party tests)
that the product conforms to the relevant standard may enter the coun-
try without additional testing.
Frictional Transactions Costs. A key aspect of many of the procedural
and administrative requirements ("red tape") that slow trade and move-
ment across frontiers is that they raise the border prices of goods and
services. This distinguishes these "real trade costs" from measures such
as import duties and quantitative import restrictions that raise the costs
of goods to consumers but do not affect border prices. Duties and quan-
titative restrictions generate revenue for the state or rents for those who
obtain the right to import. No rents or revenues are generated if a meas-
ure simply imposes additional costs by requiring firms to spend resources
that could otherwise be used productively. As a result, the domestic wel-
fare benefits of eliminating frictional barriers can be significantly higher
than those from the abolition of tariffs (see the appendix to chapter 2).
Any determination of the welfare implications of policy integration for
RIA members must consider the extent to which the regulatory barriers



158    REGIONAL INTEGRATION AND DEVELOPMENT
that are being reduced are frictional. It must also establish whether the re-
duction in frictional costs applies to all trade. This is difficult to do with
precision. Efforts to quantify the effects of differential standardization,
duplicative testing requirements, and excessive controls and inefficiencies
in customs clearance in the EU suggest that these barriers may easily give
rise to a "tax" of up to 2 percent of the value of the goods shipped. In a
developing country the figure can be a multiple of this; estimates of the av-
erage cost range up to 10 percent of the transaction value (WTO 1998).
In Egypt in the early 1990s customs red tape was a significant barrier to
trade; redundant testing and idiosyncratic standards alone imposed taxes
equivalent to between 5 and 90 percent of the value of specific shipments
of goods subject to mandatory standards. Reducing these frictional barri-
ers through an RIA with the EU would give Egypt an estimated welfare
gain equal to 4 percent of GDP, compared with no change if the FTA were
limited to abolishing tariffs (Hoekman and Konan 1999).
Maximizing the Benefits of Policy [ntegration through Nondiscrimi-
nation. As chapter 2 made clear, lowering tariffs on partner goods can
lead to welfare-reducing trade diversion, but if markets are competitive,
reducing purely frictional barriers within an RIA cannot reduce welfare.
If frictional barriers are reduced on a nondiscriminatory rather than pref-
erential (members-only) basis, the gains will be correspondingly larger
because all traders, regardless of location, will benefit from the cost
reductions. Many administrative simplification and harmonization
measures designed to facilitate customs clearance and trade can be, and
generally are, applied to nonmembers as well as members of an RIA. Of
course, nondiscrimination will not obtain if the required policy integra-
tion can only be achieved through formal agreements between govern-
ments to accept each other's regulatory regimes and apply common
administrative procedures or substantive requirements. In such cases the
benefits can be extended by allowing nonmembers to conclude parallel
agreements with RIA members-ideally, as a matter of course. Agree-
ments such as MRAs regarding health, safety, and prudential conformity
assessment systems can easily be stand-alone; they do not have to include
the preferential trade dimensions of the RIA.
Policy Integration to Increase Competition
In addition to reducing transactions costs, policy integration may benefit
countries by breaking down barriers to entry (access to markets) and re-
straining governments from using domestic regulatory policies to protect



INTEGRATING DOMESTIC POLICIES                  159
domestic firms. Both are routes to increased competition in domestic
markets and thus are almost bound to increase economic welfare, pro-
vided that they do not undermine other policy objectives. The challenge
is usually to determine whether the market access effect of a regulation is
necessary to achieve its underlying policy objectives; mechanisms for
deciding what is legitimate and for resolving disputes can be vital to the
stability of an RIA (box 6.3). These issues pervade many aspects of the
modern economy. Here, we illustrate them by reference to services, pub-
lic procurement, and contingent protection.
Services Liberalization. International services transactions frequently
require consumers and providers to be at the same place at the same
time.4 Market-access restrictions on services accordingly include not only
barriers to cross-border trade in services but also policies affecting the
physical entry of services producers into markets. Most countries restrict
the access of foreign services, services firms, and services workers to their
Box 6.3 When Is Nondiscrimination Discriminatory? The Economic Effects of
Recycling Requirements
A number of U.S. states, motivated by a desire to reduce the rate at which dumps and landfills
become exhausted, have passed mandatory recycling laws requiring that a minimum percentage of
the content of newsprint be recycled material. These laws hurt Canadian producers of pulp and
newsprint, who find it prohibitively expensive to import old newsprint to combine with the virgin
wood they routinely use. (Forestry products are a major Canadian export and a source of strong com-
parative advantage.) The result has been a significant reduction in these producers' sales in the
United States. During the penod 1989-92, the market share of Canadian producers in the United
States fell by 10 percent, from 56 to 50 percent (Vogel 1995: 229). The issue, however, has not been
brought to either the WTO or NAFTA for dispute settlement.
Similar regulations have been a source of market access disputes in the EU. A 1981 Danish
bottle-recycling law required that all beer and soft drinks be packaged in reusable containers and
that retailers take back all containers sold. Metal containers were banned. Other EU producers con-
tended that the law imposed discriminatory costs on them. The 1988 Cecchini report argued that
although the law was nondiscriminatory de jure, two-way transport costs for bottles were prohibitive
beyond a distance of 300 kilometers (Cecchini 1988: 60, cited in Vogel 1995: 84), and the European
Commission challenged the law, arguing that instruments less restrictive of trade could be used to
attain the government's objectives. The European Court of Justice, however, found the disposal and
reuse requirements legal and required only that Denmark accept metal containers, as long as
producers met the reuse and recycling requirements.
These examples illustrate that although domestic regulations can restrict market access, other
members may tolerate the situation if it is justified by overriding noneconomic objectives. But they
also suggest that RIA members may need to develop clear rules of the game and establish credible
dispute settlement systems to determine wlhen these rules have been violated.



160       REGIONAL INTEGRATION AND DEVELOPMENT
domestic markets. The restrictions are difficult to quantify, but they are
generally thought to be very high in comparison with tariffs on goods
(see table 6.1). Clearly, the gains from increasing competition in services
are likely to be massive.
The striking thing about services restrictions is their variety. Some-
times trade in services is simply prohibited; markets in domestic trans-
Table 6.1 Tariff Equivalents of Restrictions on Services Trade
(percent)
Average Tariff,
Economy or Region              Merchandise    Business and Financial Services  Construction
North Americab                      6 0                    8.2                    9 8
Western Europe                      6.0                    8 5                   18.3
Australia and New Zealand           5 0                    6.9                   24.4
Japan                               6 0                   19.7                   29 7
China                              18.0                   18.8                   40.9
Taiwan (China)                     n.a.                    2 6                    5.3
Other newly industrializing        n.a                     2.1                   10 3
countries
Indonesia                          13.0                    6.8                    9.6
Other Southeast Asia               10.0                    5.0                   17.7
India                              30.0                   13.1                   61.6
Other South Asiac                  25.0                   20.4                   46.3
Brazil                             15 0                   35.7                   57 2
Other Latin America                12.0                    4 7                   26.0
Turkey,                            13.0                   20 4                   46.3
Middle East and North Africa       20.0                    4.0                    9 5
Central and Eastern Europe and     10.0                   18.4                   51.9
Russia
South Africa                        6.0                   15.7                   421
Other Sub-Saharan Africa           n.a.                    0.3                   11.1
Rest of world (ROW)                n.a.                   20.4                   46.3
n.a. Not available.
Note: The basic methodology is to estimate a gravity equation vis-A-vis U.S. trade, using exports as the de-
pendent variable and per capita income, GDP, and a dummy for Western Hemisphere countries as regres-
sors. Hong Kong (China) and Singapore are used as a free-trade benchmark in the regressions, and devia-
tions from predicted imports relative to this benchmark are taken as an indication of barriers to trade.
These are backed out from a constant-elasticity import demand function as follows: TIT0 = (M1M0) *
(1/e), where T, is the power of the tariff equivalent, To-1, the free-trade benchmark, M,1Mo is the ratio of
actual to predicted imports (normalized relative to the free-trade benchmark ratio for Hong Kong and Sin-
gapore), and s is the demand elasticity (assumed to be -4).
a. Tariff averages are unweighted across all commodities and for the latest available year, which in many
cases is 1997 or 1998. Country coverage of regions is not comprehensive. The reported figures should be
regarded as indicative of the prevailing order of magnitude only
b. North American values involve assigning Canadian and Mexican values to the United States.
c Turkey and Other South Asia are not available separately in the U.S. data and have been assigned ROW
values.
Source: Hoekman (2000).



INTEGRATING DOMESTIC POLICIES    161
port and basic telecommunications services may be reserved exclusively
for domestic suppliers. Many countries also require that activities such as
legal, insurance, educational, surveying, and investment advisory servic-
es be performed by residents (who may be foreign nationals) or by citi-
zens. Even if there are no formal prohibitions, services suppliers in such
fields as law, accountancy, and rnedicine often must obtain local certifi-
cation or licensing. The standards are often set by professional bodies,
which may also be responsible for licensing procedures-and have a self-
interest in limiting competition. Their principal weapon is nonrecogni-
tion of foreign qualifications, dliplomas, and training programs. This
may be justified when training standards differ significantly among
countries, but more often, the allegedly key differences are invisible to
the disinterested eye.
In many services sectors, suppliers need access to distribution and
communications infrastructure, especially telecommunications net-
works. A dominant telecommunications carrier, whether public or pri-
vate, may discriminate among users of its network services by imposing
restrictions on the ability of new services providers to link to the network
or by forcing them to build infrastructure to reach interconnection
points. For example, discrimination in ancillary services (such as being
listed on computer reservation systems) can substantially reduce the
competitiveness of an airline. Limitations on advertising are commonly
used to reduce the ability of foreign insurance firms to compete, and re-
stricted distribution arrangements can effectively bar market access for
branded products. In all these cases, procompetitive regulatory interven-
tion is required to overcome the barriers-for domestic as well as foreign
suppliers.
There is substantial evidence that such restrictive policies reduce com-
petition in services industries and are very costly. Producer services, in
particular, play a crucial role in the development and growth prospects
of any nation; agricultural output can be lost because of poor transport
and storage facilities, and substandard communications networks can
raise the costs of doing business. Access to global networks in communi-
cations and transport and to competitive support services is vital for in-
ternational competitiveness in manufacturing. For example, many of the
cost-raising effects of Egypt's domestic policies are the result of regula-
tions that restrict competition in services and make Egyptian services in-
dustries inefficient. Maritime shipping is a monopoly, and estimates from
the mid-1990s show that rates then were 25 percent higher than those in
neighboring countries for equivalent routes; fees charged by the public
companies that provide port services for the hahdling and storage of
goods were 30 percent higher (M0ohieldin 1997). Policy integration that



162    REGIONAL INTEGRATION AND DEVELOPMENT
reduced average services prices by 15 percent could lead to estimated
welfare gains for Egypt of about 5 percent of GDP, depending on what
proportion of nontariff barriers affecting trade in goods is frictional and
on whether cost reductions are applied on a nondiscriminatory basis
(Hoekman and Konan 1999). Services sector production would double.
Liberalizing services can be a multilayered affair, and if it is to be suc-
cessful and have a significant effect on competition and welfare, all lay-
ers must be considered. Before Portugal's accession to the EU, entry into
its insurance markets was effectively prohibited. After accession, many
EU firms entered the market, but this had no effect on the conduct of in-
cumbent firms because prices remained regulated (Barros 1995).5 Simi-
larly, the U.S. federal government deregulated trucking in the 1970s and
1980s, but about two-thirds of truck shipments are intrastate, and many
states continue to regulate the sector. Daniel and Kleit (1995) estimate
that state-level entry restrictions raise average prices of truck shipments
by 20 to 30 percent above what would prevail if entry were unrestricted.
As with tariffs on traded inputs, the higher is the tariff equivalent of
distortionary regulatory policies for services, the higher is the tax on in-
dustries that use those services as inputs. Given the importance of servic-
es in the production process, it is important to take account of the extent
to which inappropriate regulatory regimes raise their costs. The exclu-
sion of liberalization commitments on services in the Euro-Mediter-
ranean FTAs implies that many industries in Mediterranean countries
will confront negative effective protection in the second half of the tran-
sition period to regional free trade. The tariff on their outputs will have
fallen to zero, but they will still be paying above world prices for their
services inputs. Extending liberalization to services would ease manufac-
turing's adjustment to the new competitive situation (Hoekman and
Djankov 1997).
Public Procurement A second area that is frequently almost devoid of
competition is government procurement, where public entities are often
permitted or required to discriminate in favor of domestic firms when
procuring goods and services. A well-known example is the U.S. "Buy
American" laws that apply to all government contracts.6 Many develop-
ing countries have legislation under which contracts can be awarded to a
foreign supplier only if the foreign supplier's bid is l5 or 20 percent
cheaper than the lowest bid by a domestic supplier. Discrimination can
take the form of outright bans, price preferences, local content rules, or
residency requirements. In principle, application of the national treat-
ment rule would go far toward eliminating procurement favoritism, but



INTEGRATING DOMESTIC POLICIES  163
national treatment may need to be complemented by some degree of pol-
icy integration if market access opportunities are to he effective. In the
EU, for example, common procedural requirements relating to calls for
tenders, minimum periods between publication and closure of bids, and
so forth have been imposed to ensure transparency and give potential
suppliers realistic opportunities to bid for contracts. Only a limited num-
ber of RIAs (the EU, NAFTA, and the CER) subject public procurement
to national treatment and incorporate additional procedural policy inte-
gration. Their rules are very similar to those in the WTO Agreement on
Government Procurement (Hoekman and Mavroidis 1997).
It is important to note that there are legitimate reasons for govern-
ments to buy goods and services preferentially from local suppliers.
Products may be effectively nontradable, or there may be problems of
asymmetric information that necessitate active monitoring of suppliers
by customers and therefore require geographic proximity (Evenett and
Hoekman 1999). In these cases, until the market failures can be ad-
dressed RIA-wide, discrimination may be optimal.
Contingent Protection. A third area in which increased competition is
vital is contingent protection. Trade policies in this category go beyond
tariffs and quotas to include antidumping duties, countervailing duties,
and "emergency protection" to address balance of payments disequilib-
ria or to guard an industry from "excessive" import competition. Con-
tingent protection-in particular, antidumping-has become a major
market access barrier in many jurisdictions.
Eliminating the possibility of contingent protection is of great impor-
tance for enterprises and consumiers, as the threat alone can be a barrier
to market access. It was this threat that moved Canada to seek an FTA
with the United States in the late 1980s and the EFTA countries to nego-
tiate the EEA agreement with the EU (Hindley and Messerlin 1993). A
problem is that governments which are unwilling to abolish antidumping
policy-which would be the optimal course from an economic perspec-
tive-will need to agree how much policy integration is required to make
abolition politically feasible. Thus, insofar as instruments of contingent
protection respond to industrial policies that are believed to distort
international competition, governments will need to agree on what alter-
native constraints are required on such policies. The most obvious indus-
trial policy that can give rise to competition concerns is subsidies. The
WTO Agreement on Subsidies and Countervailing Duties allows mem-
bers to impose duties on subsidized imports that materially injure
domestic competitors. Governments negotiating RIAs may therefore



164    REGIONAL INTEGRATION AND DEVELOPMENT
require disciplines on subsidy practices as a condition for abolishing
countervailing duties. This need not imply policy integration; members
can simply agree to refrain from using certain types of subsidies or to
limit their extent.
Much has been written about the absence of an economic rationale
for antidumping (Finger 1993a). The only argument that has potential
validity is that dumping can be predatory when a foreign firm or cartel
seeks to force domestic competitors out of the market by pricing below
cost with the intention of raising prices once the competition is gone. But
if there is predation, national antitrust laws should be able to address the
problem.7 This can be done unilaterally, without the need for interna-
tional agreement or for harmonization of competition regimes. If, how-
ever, there is a fear that foreign firms will be able to use anticompetitive
practices in their home markets to enhance their export competitiveness,
governments may condition the abolition of antidumping on agreement
regarding specific competition disciplines.
Spillovers
In addition to fostering competition by lowering transactions costs and
enhancing market access, policy integration can also improve welfare by
"internalizing" externalities-that is, by remedying the problem of one
country's actions impinging on another country other than via the price
mechanism. A distinction must be drawn between externalities associ-
ated with trade and investment liberalization and those with no con-
nection to trade policy, which should be the subject of cooperation quite
independent of any RIA. Even for the latter, however, the regular inter-
action between member governments that comes with an RIA may
facilitate communication and create trust such that agreements on
domestic policies to improve the joint allocation of resources become
possible.8
Effects of National Competition Policy on Trading Partners. Domes-
tic competition law and policy raise a combination of externality and
market access issues. Foreign antitrust policies may have a negative effect
on domestic welfare. A merger between foreign firms that is allowed by
the countries' antitrust authorities may increase export prices (Dixit
1984; Ordover and Willig 1986; Graham and Richardson 1997). Similar
issues arise with respect to the toleration or encouragement of export
cartels. An antitrust authority monitoring an industry that mostly ex-
ports may take a permissive attitude toward export cartels which allow



INTEGRATING DOMESTIC POLICIES                  165
the nation to improve its terms of trade. Analysts exploring the question
of whether the elimination of trade barriers gives rise to greater incen-
tives to use antitrust strategically have reached conflicting conclusions.
Horn and Levinsohn (1996) find that this outcome is not very likely even
in areas where it would seem to be most easily applied-merger policies
and export cartels. Bond (1997) suggests the opposite, both in theory
and on the basis of a study of U.S. merger standards in the late 19th cen-
tury. He concludes that a "race to the bottom" in policies emerged as
states competed to loosen merger policy in an attempt to attract busi-
nesses to register-and pay taxes-in their jurisdictions (box 6.4).
Social and Environmental Standards. Concerns that liberalization will
force countries to lower domestic standards have led to pressures on gov-
ernments to make social or environmental commitments when negotiat-
ing RIAs. As long as the national standards reflect the preferences and
natural conditions prevailing in a country, and as long as the social or en-
vironmental consequences of trade do not spill over from one country to
another, there is no reason why greater trade should induce trade part-
ners to change their laws. Altering domestic social or environmental
standards would entail a move away from what is considered beneficial
by the nation's citizens. Attempts to force partners to adopt "higher"
Box 6.4 Antitrust Law Competition by U.S. States In the 19th Century
In the United States in the late 19th and early 20th centuries, technological changes (railways and
the telegraph) led to a decline in transport costs and to Increased competiion, analogous to what
would happen in an RIA that abolished trade barriers. Federal law at the time focused on combating
the anticompetitive practices of large trusts engaged in interstate commerce, and the 1890 Sherman
Act's ban on price fixing gave firms an incentive to merge. State authorities were responsible for
approving mergers and granting corporate charters; states that were willing to have lax merger
standards could thus pass laws to encourage firms to merge and then incorporate in their jurisdic-
tions. The most liberal state, New Jersey, managed to attract more than half of the corporations with
a capitalization of $10 million or more. By 1902, New Jersey was able to pay off the entire state debt
from the tax it imposed on the capital stock of merging firms and to abolish property taxes. Other
states tried to follow New Jersey's lead, and a race to the bottom ensued. Eventually, 42 states
adopted similar laws, but with only limited success-once there, firms tended to stay in New Jersey.
Whether this race to the bottom was harmful is debatable, but it clearly illustrates how reduc-
tions in trade barriers can interact with other policies and how compefition between rules can have
marked effects on policy outcomes. Countries contemplating RIAs need to consider whether those
effects are likely to be benign (as the EU has assumed in areas where it has used obligatory mutual
recognition to pursue its Single Market Programme objectives) or malign.
Source: Bond (1997).



166      REGIONAL INTEGRATION AND DEVELOPMENT
standards that do not reflect their national preferences and conditions
should also be rejected, as they will result in inefficient resource alloca-
tion (Bhagwati and Srinivasan 1996). Gains from trade arise in large part
because countries differ; national social or environmental policies are
simply one determinant of these differences and do not constitute barri-
ers to trade or give rise to "unfair" trade. Of course, if pollution in one
country affects a neighbor, there will be gains from cooperation. The
vital issue in social and environmental policies is to ensure that there is
actually a real economic spillover to be addressed.9
There is little evidence of a race to the bottom in social and environ-
mental standards.'0 Thomas Holmes (1998) finds that U.S. manufactur-
ing firms are attracted to states which ban closed shops (a requirement
that all employees in a plant belong to a trade union) but that there has
been no tendency for states to compete in this dimension (box 6.5)."
Other experience suggests that, if anything, a "march to the top" is more
likely, driven in part by pressure from interest groups that seek to apply
uniform "high" standards in all RIA members. An example is the 1970
U.S. Clean Air Act, which allowed California to maintain stricter stan-
dards than the national norms embodied in the federal legislation. Over
time, the national norms converged to the higher California standard,
upon which California further tightened its norms. Other examples of the
"California effect"-the ratcheting upward of standards in competing
political jurisdictions (Vogel 1995: 259)-include EU standards on auto
emissions, packaging, and toxic chemicals. This process may easily be
detrimental to developing country members of North-South RIAs. An ex-
treme example is the plight of the former East Germany after unification,
when the former West German labor's insistence on equal wages and con-
ditions in both parts of the country led to massive job losses in the east.
Box 6.5 Labor Policies in U.S. States: Spillovers, but No "Race to the Bottom"
A recent empirical analysis of one aspect of the influence of U.S. states' labor policies on the loca-
tion of manufacturing (T. Holmes 1998) finds that the share of manufacturing in employment in
states with "pro-business" regulatory environments is one-third higher than in adjacent states with-
out such policies. This study Is noteworthy not only in indicating that state policies appear to matter
but also in suggesting that differences across states are stable through time-that there is no race
to the bottom. Despite its effects on location, the distribution of the policy studied (whether a state
has a law banning a requirement that all employees of a firm join a union) has not changed signifi-
cantly since 1958; in the past two decades only two states have passed such laws, and none have
repealed them.



INTEGRATING DOMESTIC POLICIES   167
Regional Free Trade Not Dependent on Policy Integration. The rheto-
ric of policymakers and their advisers often suggests that policy integra-
tion is necessary for attaining free trade. During the period leading to the
creation of the EEC, policymakers in many European countries argued
that some integration was required. Jelle Zijlstra, the minister of eco-
nomic affairs of the Netherlands, asserted in the early 1950s that credi-
ble tariff removal required "common policies on taxes, wages, prices and
employment policy" (Milward 1992: 188). Similarly, the Belgian govern-
ment believed that policy harmonization was required to equalize costs
and that without it a customs union would not be feasible, as countries
would impose new forms of protectionist policies. Thus, in the late
1940s the Belgian coal mining industry argued that a common market
could only be accepted if German wage and social security costs were
raised to Belgian levels.'2 French standards on such questions as equal
pay for men and women and the length of the workweek were initially
higher than in other EEC countries, and French officials persistently de-
manded policy harmonization in the social area as a precondition for
trade liberalization (Sapir 1995).
Such pressure to harmonize social standards generally stems, on the
one hand, from pressure by interest groups that fear the erosion of their
rents-hardly a valid rationale for policy integration-and, on the other,
from concern that domestic policies may be used to reimpose protection.
Addressing the latter often does not require policy integration; the na-
tional treatment principle will constrain most efforts to substitute do-
mestic policy for trade policy. The main exception concerns regulatory
regimes that block competition on a nondiscriminatory basis, as dis-
cussed above, in "Policy Integration to Increase Competition." Policy
integration has the potential to increase the welfare gains from RIA
membership, but there is no justification for making integration a condi-
tion for eliminating tariffs.
While governments may seek to agree on common regulatory princi-
ples to govern the behavior of public entities or restrict the use of
domestic policies, this is best done directly and should not be made
a precondition for trade liberalization. A linkage strategy that seeks to
use trade policy to induce policy reform in a partner country will be
costly, given the welfare losses from protection, and for small countries,
which have no market power or leverage, it will be ineffective. For
small countries, foreign econornic policies are best regarded as part of
their "environment"; they may not be ideal from the small country's
point of view, but this does not eliminate the gains that can be obtained
from trading.



168    REGIONAL INTEGRATION AND DEVELOPMENT
Compensation and Enforcement
Policy integration can be used as a mechanism to compensate members
for trade liberalization or to manage the integration of markets for "sen-
sitive" sectors where trade liberalization is simply not politically feasible.
It can also enhance the enforceability of free trade commitments. Thus,
the inclusion of policy integration initiatives in RIAs may be driven by
domestic political constraints that would otherwise impede international
cooperation in a policy area, or by a perception that the credibility of ef-
forts to liberalize trade and investment is enhanced by including policy
integration in an agreement.
Political and Credibility Gains from Cooperation on Domestic Poli-
cies. Policy integration may be an explicit RIA goal, with preferential
trade liberalization as a subsidiary goal designed to pay off powerful in-
terests that would otherwise oppose it. Conversely, policy integration
may be included in an RIA to bolster free trade politically. The EU's his-
tory suggests that Europe was not able to make much progress toward
internal trade liberalization until "it was discovered . . . that further
progress depended on . . . some policy of 'positive' integration . . . be-
cause the removal of discriminatory [trade] policies threatened to under-
mine just as many entrenched interests as [policy integration] would have
done" (Milward 1984: 421).
Conceptually, this type of situation is straightforward. Although there
are potential national gains from policy integration or trade reform,
there are likely to be some groups that lose from it or that are powerful
enough to insist on a payoff for accepting it. Environmental groups in
countries contemplating an RIA may insist that their concerns be ad-
dressed, or labor groups may insist that their interests be protected (as,
for example, in the case of NAFTA). Such favors, in the form of transfers
or exceptions, may be opposed by those who must pay for them, where-
as issue linkage-putting further (regulatory) issues on the table-might
increase the total gains from integration, assuage any losers, and mobi-
lize additional support for liberalization and integration.
In practice, however, identifying feasible or actual issue linkages is not
simple. National negotiating positions are driven by complex patterns of
lobbying and interest group politics, which are seldom apparent to out-
siders, and the explanations offered for the tradeoffs made between gov-
ernments are better viewed as part of the political process of selling the
outcome than as accurate accounts of motivations. But the nature of the
tradeoffs is clear in enough instances for us to be sure that this is an



INTEGRATING DOMESTIC POLICIES   169
important dimension of the negotiation of RIAs. In Europe, for example,
the formation of the ECSC and the creation of the Common Agricultur-
al Policy (CAP), both examples of deep integration, were clearly driven
by the perception that merely liberalizing trade policies in these sectors
was politically not feasible.
For a developing country, placing policy integration in an RIA context
may help overcome political opposition to reform in general by reducing
uncertainty, offering compensation, or making outside pressures for re-
form more effective. For example, Tunisia's decision to pursue a binding
reciprocal agreement with the EU was substantially driven by a view that
it would aid the general reform process (Mustafa Nabli, former minister
of planning of Tunisia, personal communication, 1998).
Policy Integration as an Aid to Enforcement of Trade Agreements.
Policy integration can help prevent backsliding on trade liberalization
commitments, and the threat of reversal of a trade liberalization may
help enforce policy integration commitments. Border policies can easily
be monitored without policy integration, but integration can reduce con-
flicts over the legitimacy of policies that have trade-impeding effects. In
addition to transparency and information, which can be aided by institu-
tions established to undertake policy integration, enforcement requires a
credible threat that sanctions will be imposed if a rule is violated. Policy
integration can help here by empowering institutions to enforce commit-
ments and make their decisions binding on RIA members. This can re-
duce uncertainty over enforcement and result in benefits by lessening the
need for hedging and "insurance."
It may be difficult or costly to retaliate against a violation of a negoti-
ated discipline within a single-subject agreement. For example, if one
member claims that its partner's domestic standards are inadequately en-
forced and imposes its own duplicative certification requirements, it
makes little sense for the partner to retaliate by doing the same. The
threat of a wider breakdown of cooperation may be more effective. An
important point is that policy integration may allow for credible threats
of noncooperation in areas of interest to a partner without threatening
trade relationships.
Policy integration can also help in enforcing trade liberalization com-
mitments by removing a government's ability to use domestic policy in-
struments as a substitute for trade policy. Similarly, mutual recognition
or harmonization of regulatory regimes may help reduce trading risks by
lowering (or, ideally, eliminating) the possibility of being confronted with
allegations of dumping or subsidization and by creating alternative ways



170    REGIONAL INTEGRATION AND DEVELOPMENT
of dealing with such disputes if they arise. The extent to which govern-
ments perceive a need to pursue policy integration as insurance against
opportunistic or nationalistic use of domestic policy instruments will de-
pend on many factors, including past use of specific policies, the level of
trust, the willingness to accept regulatory competition, and the type of
RIA involved. Where to draw the line is largely a political decision, and
RIAs differ markedly in their use of policy integration to constrain trade
and domestic policy.
The arguments just made do not imply that "anything goes" in policy
integration. They recognize the complexity and sui generis nature of RIA
formation. But they are simultaneously an injunction to the policymaker
to be explicit about issue linkages and cross-issue tradeoffs and how they
contribute to the general good. The EU's experience with the CAP illus-
trates that the economic cost of inappropriate deep integration can
be high.
Local versus Regional versus Global Cooperation
There are two dimensions to the geographic question: the set of countries
involved in establishing a norm, and the range of countries to which the
norm applies. On the former, there is a tradeoff between matching local
requirements and tastes and reaping the efficiency benefits of global stan-
dards. On the latter, nondiscrimination-that is, global application-is
always better, if it is feasible
Global Standards: The Preferred Route. Developing countries should
pursue outward-looking strategies based on global standards and best
practices to the greatest extent possible, rather than develop local or re-
gional norms. This not only saves resources that would be used in devel-
oping the standards but also reduces the chances of standards being used
as a means of segmenting markets and stifling competition. An addition-
al benefit is that such standards are typically relatively light; they usually
rely mainly on market forces and so help avoid laying the heavy hand of
officialdom on commercial activities. These injunctions are especially im-
portant if there are network externalities or economies of scale, for then
openness to the largest possible set of actors carries the greatest rewards.
Sometimes, formal agreements between states are needed to facilitate
trade, and an existing RIA provides a natural framework for negotiating
such agreements. But the rules that are adopted should be regional only
to the extent that this is unavoidable. The case for unavoidability applies
where, although the need for intercountry coordination is pressing,



INTEGRATING DOMESTIC POLICIES   171
global standards are not appropriate-perhaps because local tastes are
idiosyncratic-or have not yet been developed. If regional coordination
is required, it should, where possible, take the form of competition be-
tween regulatory regimes rather than harmonization, as this allows for
differences in preferences and economic conditions. If harmonization is
deemed necessary, it should be limited to the adoption of minimum
standards that comply with international norms, if these exist (see, for
example, Kanbur, Keen, and van Wijnbergen 1995).
It is sometimes argued that global fora are too unwieldy to use for de-
veloping standards. In fact, multilateral cooperation between states in
standard setting and prudential regulation predates the rise of regional-
ism, and it continues to be the clominant form of intergovernmental co-
operation because in many spheres the world is the optimal level for such
cooperation (box 6.6). Also, despite sometimes fiercely regional rhetoric,
many of the standards and norms embodied in RIAs are international;
for example, the EU has adopted the recommendations of the Bank for
International Settlements (BIS) on banking and relies on disciplines nego-
tiated in the WTO for government procurement. The more that interna-
tional norms are used by RIAs, the less likely it is that policy integration
will lead to detrimental outcomes. This is because these norms are set
through a process of consensus and tend to be minimum standards. RIAs
could help expand the use of such norms and become instruments
through which they can be enforced.
Less Discrimination Means More Benefits from Policy Integration.
Policy integration based on international standards is likely to have the
maximum effect on competition and hence on economic welfare within
an RIA, and it is less likely to be detrimental to nonmembers of RIAs.13
Even if policy integration is specific to an RIA, it may enhance global
welfare if it is applied on a nondiscriminatory basis to all traders. Mea-
sures to simplify customs procedures and documentary requirements re-
duce transactions costs on imports from all origins. Similarly, if in the
context of an RIA a government commits to a privatization program that
limits domestic subsidies or applies competition law more strictly, this
may benefit third countries as well as partners. Thus, nonmembers of the
RIA are likely to gain from openness, just as they do from lower MFN
tariffs. Conversely, regional policies applied discriminatorily are quite
likely to impose costs on foreigners, just as discriminatory tariffs can (see
chapter 8).
Mutual recognition or harmonization-based policy integration-in or
out of an RIA-can result in effective discrimination against outsiders.



172       REGIONAL, INTEGRATION AND DEVELOI'MENT
Box 6.6 Multilateral Policy Integration and the Enforcement Role of RIAs
More than 30 intergovernmental organizations emerged between 1860 and 1914. (The table lists a
number of them.) Most of them covered infrastructure, including postal services (1863), marine sig-
naling (1864), technical railway standards (1883), ocean telegraphy (1897), and aerial navigation
(1910). Multilateral institutions allowed the emergence of a Europe-wide market for industrial goods.
International interconnection norms agreed under the auspices of the Intemational Telegraph Union
eliminated the requirement that telegrams be printed out at each border post, walked across the
border, and retyped. The Universal Radiotelegraph Union aimed at preventing a global radio monop-
oly by requiring interconnection across different technologies.
Intemational railway unions promoted networks by standardizing rolling stock, allowing compa-
nies to use each other's rolling stock, and enforcing a single bill of lading so that one document could
be used for all trans-European shipments. This took place roughly a century before the EU's adoption
of a single customs document. All continental European countries except Russia and Spain adopted
the same rail gauge, drove on the left, and aligned signals, brakes, and timetables (Pollard 1974:
50-51). This standardization was largely driven by the private sector; business can often achieve co-
operation more readily than governments. The Rail Union (Central Office for International Railway
Transport) of 1890 played a significant role in dismantling protectionism in the late 19th century by
prohibiting transit duties on goods shipped by rail, and the Brussels Tariff Union made the remaining
restrictions transparent by publishing lists of tariffs in five languages.
Intergovernmental organizations proliferated after World War II with, for example, the formation
of the GATT (now the WTO), the United Nafions system, the International Civil Aviafion Organization
(ICAO), and the Organisation for European Economic Co-operafion, which became the OECD. By de-
veloping norms and practices to facilitate the expansion of international markets and manage con-
flicts between jurisdictions, these bodies help foster economic growth. The hallmark of such organi-
zations is that governments cooperate in developing standards, ranging from technical requirements
for maritime transport (the International Maritime Organization), to customs procedures (the World
Customs Organization), to labor rights and working conditions (the International Labour Organisa-
tion), to capital adequacy standards for banks (the Bank for Intemational Settlements). The resulting
standards are often adopted by RIAs.
RIAs differ from the global standards institutions by offenng enforcement mechanisms. Resolv-
ing interstate conflicts has always figured on the agenda of international conferences and organiza-
tions but has often been ineffective. In the early 20th century a number of proposals were made to
develop intemational instruments for the extension of binding arbitration to intellectual property, the
principle of equality of foreigners in taxation (national treatment), civil and commercial procedure,
customs tariffs, the right of foreigners to hold property, the regulation of companies, and claims for
damages. These were opposed by countries that sought to retain their national sovereignty (Murphy
1994:104). Of the bodies mentioned above, only the United Nations and the WTO have binding dis-
pute settlement mechanisms.
Trade-Related Intergovemmental Organizations Established before 1914
Objective                          Organization and date of creation
Infrastructure   Intemational Telegraph Union (1865)
and related    Universal Postal Union (1874)
"software"    Intemational Railway Congress Association (1884)
Central Office for International Railway Transport (1890)
Diplomatic Conference on Intemational Maritime Law (1905)
Universal Radiotelegraph Union (1906)
Permanent International Association of Road Congresses (International
Automobile Convention) (1909)
(Box continues on next page.)



INTEGRATING DOMESTIC POLICIES           173
Box 6.6 (continued)
Trade-Related Intergovemmental Organizations Established before 1914 (continued)
Standards      Intemational Bureau of Weights and Measures (1875)
Metric Union (1877)
International Bureau of Analytical Chemistry of Human and Animal Food (1912)
Intellectual   International Union for the Protection of Industrial Property (1883)
property     International Union for the Protection of Literary and Artistic Works (1886)
Trade          Brussels Tariff Union (International Union for the Publication of Customs
Tariffs) (1890)
Hague Conference on Private Intemational Law (1893)
Intemational Bureau of Commercial Statistics (1913)
Dispute        Permanent Court of Arbitration (1899)
settlement   Intemational Court of Prize (1907; never ratified)
Source: Adapted from Murphy (1994): 48-49.
What matters, then, is whether outsiders can be "recognized" in turn.
The principles of open access and conditional MFN can be applied to
third parties seeking to join the "club." If a country meets the minimum
conditions for membership in any particular product or standard, it
should be able to participate in a policy integration initiative."4
Introducing such a transitivity rule in the WTO could help expand
the reach of MRAs: if countries A and B and countries B and C each
had MRAs, A and C would automatically accept each other's norms
and products. Thus, subject to conditions ensuring that the original mu-
tual recognitions were genuine (for example, that if A could not agree
to the extension to C, it would have to suspend the agreement with B),
it would be possible to move toward broader, even global, standards
regimes. The more open RIAs are to efforts by nonmembers to partici-
pate in their policy integration dimensions, and the more they rely on
international standards, the less worrisome policy integration in RIAs
will be.
But participation in the most important dimension of MRAs-accept-
ance of conformity assessment-may be difficult for developing coun-
tries. Neither the United States nor the EU has concluded MRAs with
developing countries. Special efforts may be required to ensure that
regional policy integration does not perpetuate or increase this discrimi-
nation. One option would be to seek agreement from partner countries
that conformity assessment by internationally established certification
entities will be accepted and that greater reliance will be placed on the
private sector and the market mechanism (self-certification). It is of par-
ticular importance to keep policy integration free of the preferential



174    REGIONAL INTEGRATnON AND DEVELOPMENT
dimensions of the RIAs. Such contamination is not unknown; for exam-
ple, the EU tried to include restrictive preferential rules of origin in
MRAs with the United States, but U.S. industry put up successful resist-
ance (Wilson 1999). Developing countries may find it useful to seek
WTO disciplines that ban such attempts.
POLICY INTEGRATION TO DATE: MORE PROMISE
THAN REALITY
RIAs differ in the extent to which they have pursued policy integration.
Many embody only the minimum required to implement an FTA or a
customs union, while others have gone far toward full integration of do-
mestic policies. Determining the extent to which RIAs have a policy inte-
gration agenda can only be done case by case and must to some extent be
a subjective exercise. This section provides an overview of the status of
and prospects for policy integration in RIAs.
Coverage
The coverage of policy integration varies enormously across RIAs. On
balance, few RIAs have gone far down the policy integration track, and
most do not go much, if at all, beyond the WTO.
Cooperation on Product Standards Rarely Goes beyond WTO Rules.
Most RIAs that include disciplines on product standards use internation-
al standards as their norms (as do WTO standards agreements). Some
have adopted foreign standards unilaterally; for example, CARICOM
members generally use British standards (Page 1998).
Australia and New Zealand have gone farthest on standards-in some
respects, beyond the EU. A joint accreditation system for conformity as-
sessment was created in 1991 to harmonize auditing and certification
systems and criteria with international standards. In 1995 this was com-
plemented by an agreement to create the Australia New Zealand Food
Standards Council. The 1997 Trans-Tasman Mutual Recognition Agree-
ment, along EU lines, covers both goods and services. Goods sold legally
in one country can be sold in the other, and persons licensed to practice
an occupation in one country are automatically licensed in the other.5
In NAFTA, each party has agreed to accredit or otherwise recognize
testing and certification performed by partner country bodies. The
Group of Three and MERCOSUR provide for MRAs in this area but



INTEGRATING DOMESTIC POLICIES   175
have not yet made significant progress. Other RIA pacts do not address
certification and conformity assessment.
Services Liberalization: A Recent Phenomenon. Outside the EU, ef-
forts to liberalize international services transactions within RIAs became
prominent only in the late 1980s and early 1990s-in CUSTFA and,
later, NAFTA, and in the CER. RIAs have not led the WTO by much, if
at all; services were introduced in the GATT agenda in 1986. The EU
began to make substantial progress only in the mid-1980s, with the
launch of the Single Market Programme. Even though national treatment
applied formally to member state services providers, European services
markets in fact remained highly segmented until the late 1980s. Doctors,
pharmacists, lawyers, accountants, and other professionals had to retake
certification examinations to practice in a different country. Liberaliza-
tion of financial services was impeded because many members continued
to restrict capital movements. Much of the SMP aimed at integrating EU
services markets, and many of the directives issued by the European
Commission related to services industries. For example, in 1989 the Sec-
ond Banking Directive made home countries responsible for prudential
supervision (setting and enforcing liquidity and solvency standards) and
allowed credit institutions authorized in one state to establish branches
and provide banking services anywhere in the EU. Directives were also
developed for investment services, mutual funds, insurance, road and air
transport, telecommunications, professional (accounting, legal, and med-
ical) services, and mutual recognition of diplomas in pharmacy and other
"regulated" professions.
The NAFTA and the CER have comprehensive coverage of services
activities, using a negative list; that is, all services sectors are covered un-
less specifically exempted. A NAFTA Professional Services Annex sets
out procedures for developing standards for professional practice, re-
quires the abolition of citizenship and permanent residency requirements
in licensing and certifying professional service providers, and establishes
work programs to liberalize licensing for foreign legal consultants and
engineers.
The Andean Pact, the CACM, and the SADC have done little to liber-
alize services. In SACU and CARICOM certain services sectors have
been integrated more for historical than for deliberate policy reasons. In
CARICOM national treatment applies to banking, health, education,
tourism, and transport services, and many of these services are provided
jointly. The Group of Three is similar to NAFTA, but sectoral coverage is
narrower. (Transport remains the subject of negotiation.) In MERCO-



176    REGIONAI INTEGRATION AND DEVELOPMENT
SUR free circulation of services is a long-term objective, to be achieved
by 2007. Progress has been slow, however; members are still negotiating
a framework agreement. ASEAN members have until recently restricted
services liberalization to their commitments under the General Agree-
ment on Trade in Services (GATS). In 1997 they agreed in principle to
full liberalization (on a preferential basis) of most services by 2020. The
FTAs between the EU and Mediterranean countries do not include serv-
ices; those with Central and Eastern European countries do. Except in
the EU, either the multilateral GATS process is leading liberalization of
services or the GATS commitments of RIA members do not differ signif-
icantly from RIA commitments (Hoekmiiani and Sauve 1994; Page 1998).
Competition and Industrial Policy Disciplines. The EU imposes far-
reaching disciplines on member state regulatory policies to underpin the
realization of a single market. Common rules apply to subsidies, monop-
olies, government procurement practices, and antitrust. These are com-
plemented by detailed European legislation relating to the achievement
of the Internal Market. The goal of these competition provisions is to en-
sure a "level playing field" or "equal competitive opportunities" in Eu-
ropean markets for suppliers from all member states. All of the EU pro-
visions relating to competition and industrial policies are also found in
the EEA.
A commitment to apply common disciplines in areas such as antitrust,
subsidies, and state monopolies is also a central dimension of the EU-
centered RIAs. Increasingly, the EU appears to be requiring its partners
in Eastern and Central Europe and the Mediterranean to adopt its inter-
nal market rules and to introduce national legislation that is consistent
with its norms (independent of any consideration of trade effects). En-
forcement of competition rules is largely left to national bodies, and dis-
pute settlement provisions are largely political in nature, residing with
the Association Councils, which may make recommendations when dis-
putes arise but may not issue binding instructions.
The CER agreement between Australia and New Zealand is similar to
the EU in its coverage of competition and industrial policies. In 1988 a Pro-
tocol on Acceleration of Free Trade in Goods stipulated that nationals of
one state could be investigated by the competition authorities of the other
and could be required to respond to requests for information. Each nation
amended its antitrust legislation to extend it to Australian or New Zealand
firms with market power in either the national market or the combined
market. Courts are empowered to sit in either country; orders can be
served in either; and judgments by each country's courts or authorities are



INTEGRATING DOMESTC POLICIES  177
enforceable in the other. Unlike the situation in the EU, the application of
antitrust remedies remains strictly national. The CER also includes sub-
sidy disciplines stronger than those of the WTO. Industry-specific subsi-
dies are banned, and export subsidies were prohibited in 1987.
Other RIAs have less, or nothing, on competition policy, although
many RIAs have rules on subsidies. Export subsidies are frequently pro-
hibited, as in the Andean Pact, the CACM, CARICOM, MERCOSUR,
and the Group of Three. MERCOSUR members have begun harmoniz-
ing their competition laws and creating a mechanism for coordinated ac-
tion to prevent anticompetitive practices from affecting intra-area trade
(Rowat, Lubrano, and Porrata 1997). A December 1996 protocol pro-
hibits concerted practices that restrict or distort competition and affect
trade between member states. The protocol gives MERCOSUR institu-
tions the power to enforce these rules, although implementation remains
the responsibility of national competition agencies.'6 The 1996 Canada-
Chile FTA requires each member to adopt or maintain measures to pro-
scribe anticompetitive business conditions, and it also requires competi-
tion authorities to consult and ccooperate.
Persistence of Contingent Protection. A number of RIAs, including
the EU, the EEA, the CER, and the Canada-Chile FTA, ban the use of in-
struments of contingent protection such as antidumping and countervail-
ing duties. The CER abolished antidumping on trade between Australia
and New Zealand in 1990, following a decision that national competi-
tion regimes should be applied to such trade. As in the EU, elimination
of antidumping was linked to the transition path toward complete
free trade.
Antidumping remains applicable to intra-MERCOSUR trade.'7 There
is no explicit requirement or commitment that antidumping on intrabloc
trade will be eliminated, but the arguments for maintaining it on internal
trade have been subject to periodic inconclusive reviews.
The Canada-Chile FTA specifies that antidumping on products ceases
to be applicable on intra-FTA trade as tariffs on those products are elim-
inated or by January 1,2003, at the latest. The abolition does not extend
to countervailing duties. A committee has been set up to define subsidy
disciplines further and to eliminate the need for countervailing duties.
Elimination of antidumping occurred without agreement on the applica-
tion of common competition rules for trade. The explicit exception for
maintaining countervailing duties suggests that subsidy and industrial
policies were a greater cause for concern than private anticompetitive
practices.



178    REGIONAL INTEGRATION AND DEVELOPMENT
Antidumping and countervailing duties remain applicable to imports
from the Central and Eastern European and Mediterranean countries
that have signed FTAs with the EU and to the customs union between
the EU and Turkey (Togan 1997). Notwithstanding the agreement to
adopt EU-consistent competition disciplines, the Europe Agreements do
not specify that antidumping will be phased out or eliminated. The EU
insists that application of competition laws and principles is not
enough; what is necessary (but not sufficient) is that all of the Internal
Market directives also be applied."8 The EU RIAs illustrate that there is
no presumption that pursuing harmonization in competition policies
will lead to the elimination of contingent trade policies on intraregional
trade.
Given that the substantive case for antidumping duties is so weak in
the first place, it is difficult to identify economic conditions under which
RIAs should abolish them. A number of observations are pertinent, how-
ever:
1. Explicit linkages between abolishing antidumping and adopting
common antitrust rules are by no means the norm.
2. Antidumping may be used as a tool to extract concessions from
potential partners (as in the Europe and Euro-Mediterranean
Agreements).
3. If an RIA aims at a single market, there can be no role for
antidumping. Although common antitrust disciplines may be
adopted as part of deeper integration efforts, their dominant aim
is usually integration per se rather than abolition of antidumping
internally.
4. Much depends on the extent to which governments can be pre-
vented from intervening directly in domestic industries. This
seems to be one of the main concerns underlying the EU's unwill-
ingness to give up its right to impose antidumping duties on
imports from the European transition economies and the
Mediterranean countries. The fact that NAFTA did little to im-
pose disciplines on subsidies may be an important element in U.S.
unwillingness to give up antidumpiing. The Canada-Chile FTA is
also noteworthy in this connection in that countervailing duties
remain applicable.
Environmental and Social Policies. The EU (as is often the case when
it comes to policy integration) is in the vanguard of cooperation on envi-
ronmental policy. In part, this reflects efforts by "high standards" coun-
tries such as Denmark, Germany, and the Netherlands to export their



INTEGRATING DOMESTIC POLICIES  179
norms to all members. It has taken decades for the environment to
become an EU competence, and many years for the European Commis-
sion to begin to affect environmental policy in less "green" member
states.
Environmental regulations are much less prominent in other RIAs.
NAFTA includes a number of supplemental agreements, including one
on the environment, that are largely aimed not at harmonization but at
ensuring that national laws and regulations are enforced. A committee
can levy fines or recommend trade sanctions if a NAFTA signatory does
not apply its national law.
The EU has adopted a number of social policy directives. The Treaty
of Rome supports harmonization of social policies, although (except for
a requirement that men and women be treated equally) there are no
deadlines, targets, or criteria. 'rhe Social Chapter of the Maastricht
Treaty sets out a list of basic rights and freedoms for workers (and for
citizens in general), but these do not apply in all member countries. Most
other RIAs do not contain disciplines on social policies, or if they do,
they have never been implemented or developed. MERCOSUR has a
working group discussing the issue. NAFTA's North American Agree-
ment on Labor Cooperation provides a forum for exchanging informa-
tion and monitoring labor conditions, but, as with the environmental
side agreement, only the failure to enforce relevant national laws can be
challenged.
Depth of Integration: A Function of Political Objective
Most RIAs are commercial instruments-their primary objective is to fa-
cilitate trade. Over time, some RIAs have pursued elements of domestic
policy integration, driven by the potential for gains from cooperation in
particular areas and by concerns that governments might use domestic
policies as a substitute for trade policies. Such efforts are by their very
nature limited and partial. They are also paralleled, in both intent and
outcome, by initiatives in the WTO. It is useful to distinguish them from
initiatives in RIAs that have as their objective economic integration or
union, which are discussed in this subsection.
The RIA that has gone furthest in integrating domestic policies-the
EU-is an agreement between states seeking to achieve economic union.
Cooperation consequently extends far beyond liberalization of trade in
products and factors. The EU has created supranational institutions with
mandates to make policy in specific areas, the power to initiate rule-



180    REGIONAL INTEGRATION AND DEVELOPMENT
setting proposals, and a significant degree of autonomy-as manifested
in secretariats staffed by career professionals that are formally independ-
ent of member governments and have their own sources of funding.
Once adopted, the rules that are developed have direct effect in national
law, thus allowing citizens to invoke EU rules in national courts. In areas
where the EU has competence, rules are often set by majority vote. En-
forcement of EU rules is binding, with ultimate enforcement power vest-
ed in the supranational European Court of Justice.
Next to the EU, the CER has achieved the most in integrating do-
mestic policies. It has adopted elements from the EU-the Australia
New Zealand Food Standards Council works under majority voting
rules-but does not rely on supranational institutions. Instead, integra-
tion depends on national implementation of laws and regulations on a
nondiscriminatory basis; regular review and expansion of the agree-
ment; and extensive mechanisms for consultations and transparency.
The institutional structure is light. The CER is an important example
for developing countries because it illustrates that deep integration can
be achieved without the supranational structure found in the EU. But
this model may be difficult to emulate; the high degrees of similarity
and trust between the two partner countries seem to be a vital ingredi-
ent in its success.
Most other RIAs do not come close to the EU or CER levels of policy
integration. Most do not have supranational institutions responsible
for implementing their agreements but rely instead on secretariats,
staffed by delegations from member states, that lack initiation powers or
autonomy. Rule-making is based on consensus and unanimity, whether
formal or informal. Enforcement mechanisms are often diplomatic and
restricted to governments
In fact, the enforcement mechanisms of many RIAs reveal their funda-
mental "shallowness," relying as they do on either diplomatic conflict
resolution or nonbinding panel processes. WTO procedures are often
stronger than those in RIAs. Even where an RIA's panel-based dispute
resolution is stronger than that of the WTO, the latter plays a significant
role de facto. For example, the arbitration of investor-state disputes in
NAFTA is binding and is directly accessible by investors, whereas only
governments have access to the WTO-yet in many cases NAFTA mem-
bers have preferred to use WTO dispute settlement procedures. Canada
and the United States have generally used the WTO to deal with disputes
other than antidumping. Even for antidumping, where panel decisions
are binding in national law, the substantive requirements they impose are
those of the WTO.



INTEGRATING DOMESTIC POICIES   181
Prospects for Policy Integration in RIAs
Although, to date, little has been achieved in the way of policy integra-
tion in most RIAs, integration is under discussion in connection with a
number of existing and prospective agreements. Two major regional ini-
tiatives now under negotiation-the Asia-Pacific Economic Cooperation
(APEC) initiative and the Free Trade Area of the Americas (FTAA)-
illustrate the issues and challenges. Among them are ensuring that policy
integration efforts include the abolition of instruments of contingent pro-
tection; focusing on trade facilitation and the reduction of redundant red
tape; agreeing to binding and enforceable commitments; and applying as
many rules as possible on a nondiscriminatory basis.
APEC and the Goal of Nondiscriminatory Liberalization. APEC has
created a large number of bodies to explore possible coordination in
areas relevant to the organization's goal of free trade. Services, competi-
tion policy, investment regimes, standards and conformity assessment,
customs procedures, intellectual property rights, and government pro-
curement are all under discussion, although contingent protection is not.
The process for pursuing liberalization and integration is unique in that
it relies on unilateral "individual action plans" in each area that are to be
implemented on a nondiscriminatory basis. The plans submitted by the
18 members take up thousands of pages. Given that implementation of
commitments can take many years (the free trade objective is to be real-
ized by developed country members in 2010 and by developing country
members in 2020), many of the plans are no more than statements of in-
tent. Commitments are not legally enforceable, and countries are not
bound by any agreement. The focus instead is on developing nonbinding
principles relating to the application of domestic policies such as compe-
tition law and standards.
The nonbinding nature of APEC-based policy integration suggests
that the WTO may well be the primary vehicle for locking in whatever
unilateral reforms are implemented by members. APEC activities have
largely been instruments for communication, information exchange, and
technical assistance. Although this can be beneficial in providing focal
points and identifying best practices, insofar as policy integration
requires formal agreements between governments it would appear that
the APEC process is inherently limited to what can be achieved by apply-
ing the national treatment and MFN principles. These principles can give
rise to major economic benefits and should certainly be pursued, but pol-
icy integration that requires fornal instruments of cooperation such as



182    REGIONAL INTEGRATION AND DEVELOPMENT
MRAs must be based on binding, enforceable agreements so that viola-
tions can be contested. Since the objective of APEC members is to apply
liberalization and policy reforms on a nondiscriminatory basis, the WTO
is the natural forum in which to embed such agreements-in the process,
ensuring that they are open to all WTO members. The WTO is also the
best institution for locking in country-specific liberalization commit-
ments, as its dispute settlement mechanism is substantially stronger than
the diplomatic processes envisaged for APEC.
A weakness of the APEC discussions is the absence of serious efforts
to limit the reach of contingent protection. One of the strengths of the
initiative is the emphasis on policy integration-often based on identifi-
cation and unilateral adoption of best practices-with the aim of facili-
tating trade by reducing transactions costs associated with customs and
with testing and certification procedures. Another strength, at least ac-
cording to current plans, is that APEC will not have trade preferences for
intramember trade; the intention is that all its trade liberalization will be
MFN. Bergsten (1997), however, argues that APEC will need at least the
threat of preferences and discrimination if it is to prosper and to include
the United States. Thus, this strength is more of a challenge than a fact.
The FTAA: No Serious Negotiations as Yet. As with APEC, the FTAA
has a large number of working groups exploring possible modalities for
cooperation on domestic regulatory policies. Free intraregional trade in
merchandise is to be attained by 2005. It remains to be seen what
emerges as regards services liberalization and policy integration more
generally, for although there has been a great deal of preliminary and
procedural analysis, substantive negotiations have yet to begin. A likely
difference from APEC is that commitments and agreements will be bind-
ing and subject to enforcement procedures.
The negotiating agenda of the FTAA is very broad and extends
beyond what is currently covered by the WTO. In this respect there is
substantial potential for members to achieve deeper integration. A key
challenge-indeed, a litmus test-will be whether agreement can be
achieved to go significantly beyond the WTO by establishing free trade
and investment in services and the right of establishment as general prin-
ciples, and the extent to which initial liberalization commitments will go
farther than governments have committed to under the GATS. Another
indicator of the depth of integration will be whether contingent protec-
tion will be applicable to intra-FTAA trade. A working group on subsi-
dies, antidumping measures, and countervailing duties has been engaged
in information collection and exchange, with a particular emphasis on



INTEGRATING DOMESTIC POLICIES      183
eliminating remaining agricultural export subsidies and other trade-
distorting measures affecting agriculture. The terms of reference of the
group make no mention of developing recommendations for disciplining
the use of contingent protection.
CONCLUSION
This chapter has explored the role of policy coordination in developing
country RIAs, drawing mainly on EU experience, which offers by far the
most detailed example to hand. The conclusion that emerges is that al-
though policy integration offers scope for considerable gains, great polit-
ical and technical efforts are required to achieve it. Developing country
governments should not casually assume that policy integration will au-
tomatically generate big welfare gains: unless they invest heavily in it,
policy integration will evade them. We have also emphasized that gov-
ernments should exploit the many opportunities to reduce transactions
costs and market segmentation that multilateral organizations and uni-
lateral action offer.
NOTES
1. Schiff and Winters (2002a) discuss nonregulatory cooperation.
2. The principle of positive comity requires that a government body take into
account the interests of another country in the application of national law.
3. Poland, for example, is said to have maintained a "non-transparent system
of certification and product approvals that is not harmonized with international
standards, nor does it accept manufacturer self-certification. Due to the need to
gain separate certification, foreign exports meeting international standards might
not obtain Polish approval" (Wilson 1999: 37). One standards-related trade dis-
pute between the EU and Poland involved allegations that a Polish ban on im-
ports of gelatin on the basis of health concerns reflected a desire to protect local
producers (Financial Times, February 19, 1998).
4. This discussion draws on Hoekman and Braga (1997).
5. Indeed, firms may have entered Portugal's market precisely to take advan-
tage of the regulated prices.
6. See Greenwold and Cox (1993) and Francois, Palmeter, and Nelson (1997)
for descriptions and analyses of U.S. procurement practices.
7. Research suggests that predation is very much the exception, not the rule,
in antidumping cases. In more than 90 percent of actual antidumping actions, an
antitrust authority would not have intervened on grounds of competition, let
alone predation (Schone 1996; Messerlin 1997). The consensus among most



184     REGIONAI INTEGRATION AN) DEVELOPMENT
economists is that antidumping as practiced today has, de facto, nothing to do
with predation.
8. Among the examples of non-trade-related policy areas in which coopera-
tion can overcome problems created by externalities are shared natural re-
sources, such as the environment and watercourses, and common infrastructure
projects, such as power generation and transborder transport schemes (corri-
dors). Schiff and Winters (2002a) discuss these cases in greater detail.
9. Schiff and Winters (2002a) examine the issue of regional spillovers in areas
such as the environment, natural resources, infrastructure, and energy; and
whether RIAs can help in reaching a cooperative solution.
10. The focus here should be on policies, not outcomes. An RIA may lead to
industries' shifting production facilities to a partner country where environmen-
tal policies are less stringent, but as long as environmental policies in the partner
countries are appropriate for local conditions and national preferences, this is de-
sirable. Of course, for a variety of reasons, policies might not be optimal, and in
these cases shifting the location of production might be either costly or beneficial.
11. Holmes controls for the possibility that the data reflect the passage of
closed-shop legislation as a reaction to a concentration of industry, rather than
vice versa, by focusing on the border areas of each statc. Along cithcr side of thc
border, other conditions affecting location are likely to be very similar, while the
amount of industry in the border areas is sufficiently small not to reflect the im-
portance of industry to the state's overall level of activity.
12. The discussion of proposals for a European customs union in the early
1950s included virtually every question that came to be addressed in the Maas-
tricht treaty: a common European currency, monetary policy, freedom of labor,
mutual recognition of professional qualifications, a common company law, a free
capital market, and common workplace and product safety standards (Milward
1992: 191).
13. These statements must be qualified if there is imperfect competition in the
affected sectors, for then the RIA may be able to benefit at the expense of non-
members by its choice of standard.
14. In this context, the "club" is not the RIA; it is defined relative to one stan-
dard or product, not the whole range of products covered by the RIA. If the lat-
ter definition applied, the proposal would merely amount to requiring RIAs to be
open to any country that can negotiate suitable entry conditions, which, as we
argue in chapter 8, is not a satisfactory or realistic way of achieving nondiscrim-
ination in the world economy.
15. This MRA was originally concluded among the states of Australia in
1992; essentially, what was done was to extend it to New Zealand.
16. Since MERCOSUR institutions operate by consensus, a national competi-
tion authority will have the option of refusing to implement a decision. The pro-
tocol must be ratified by the parliaments of member states before it becomes ef-
fective. Uruguay and Paraguay have yet to create the necessary institutions.



INTEGRATING DOMESTIC POLICIES      185
17. Antidumping has been used actively by some members of MERCOSUR.
Argentina initiated 33 antidumping cases on imports from Brazil during the peri-
od 1992-96, making Brazil the number one target for such actions (Tavares de
Araujo and Tineo 1998).
18. The EU is requiring its future members to adapt national competition
rules to EU standards more strictly than has been the norm for existing member
states (P. Holmes 1996: 5). Pittman (1998) documents that the Central and East-
ern European countries have already gone far down the harmonization road.






CHAPTER 7
Regional Integration as Politics
Countries sometimes form trade blocs for noneconomic reasons, such as
national security, peace, and assistance in developing political and social
institutions. These are public goods that are unlikely to be adequately
provided in the absence of some form of intervention, such as an RIA.
Political objectives can be important for RIAs-sometimes overwhelm-
ingly so-but it is still desirable that they be achieved efficiently and that
policymakers pay heed to their economic cost. However, no economic
analysis of an RIA is complete without taking the potential welfare ben-
efits of political objectives into account. This chapter examines some of
the political objectives of regionalism, discusses their economic implica-
tions, and assesses whether trade preferences are necessary for achieving
these goals. It is not a review of the political science of regionalism but,
rather, an assessment of how political objectives should shape our eco-
nomic appraisal of regionalism.
Political benefits such as peace and security can sometimes swamp the
simple material considerations that usually determine economic policy.
Moreover, since such benefits are typically shared by only a limited
number of countries-usually, neighbors-it makes sense to seek them
on a regional basis rather than multilaterally. Thus, this issue forms a rel-
atively more important part of the analysis of RIAs than of some other
international issues.
The following section explains how, under some circumstances, the
formation of an RIA may be an effective way of dealing with security
tensions between neighboring countries. The argument is, essentially,
that mutual trade fosters peace between countries and regionalism
187



188    REGIONAI INTEGRATION AND DEVELOPMENT
fosters trade. We show that for an RIA designed to enhance security, the
optimum external tariff (on imports from nonmembers) declines over
time and as integration deepens.
We next briefly discuss the use of RIAs to reduce and manage pres-
sures for migration. Although such objectives have been frankly admit-
ted to in practice-for example, in NAFTA and in the Europe
Agreements between the EU and the countries of Central and Eastern
Europe-their effectiveness in this regard is far from certain.
Joining an RIA with democratic countries can help a developing coun-
try achieve or uphold democracy if the RIA imposes on its members
"club rules" such as democracy and civil rights. As is discussed in the
third section of the chapter, this assistance is likely to be more effective if
the other members have large economies; larger partners are generally
able to impose greater costs on (or withdraw greater benefits from) re-
calcitrants than can smaller ones. The size of the developing country and
its proximity to its democratic partners also matter. A partner country is
likely to be far more concerned about possible spillovers from events in a
nearby, sizable developing country than about those in a small, distant
one or resulting from "a quarrel in a faraway country between people of
whom we know nothing."'
The final section examines the connection between economic integra-
tion and political integration-in particular, whether increased regional
integration is likely to weaken or strengthen the nation-state. It argues that
resource pooling and collective action can increase the effectiveness of the
state in small or even medium-size nations in dealing with economic prob-
lems such as pollution, guarding against third-country security threats,
and enhancing international influence by lowering negotiation costs and
increasing bargaining power in dealings with the rest of the world. Coop-
eration of this kind, however, does not usually require trade preferences.
REGIONAL INTEGRATION AS A MEANS OF REDUCING
FRICTIONS BETWEEN ANTAGONISTIC NEIGHBORS
The idea that trade can be an important force for creating and maintain-
ing peaceful relations between countries dates at least from the 18th cen-
tury (see box 7.1) and is probably a good deal older. Since forming an
RIA almost always increases trade between the partners, there seems to
be a good chance that the pacific effects of trade would extend to this
particular form of trade relationship.2 In fact, this is true only when the
RIA is between relatively evenly balanced partners whose governments
genuinely wish to improve security and when the benefits from the RIA
are distributed in a relatively equal way.



REGIONAL INTEGRATION AS POLITICS               189
Box 7.1 Trade and Peace:The Political Lineup Follows the Economic Lineup
The notion of using international trade to defuse tension and bring nations together dates back at
least to the publication in 1795 of Perpetual Peace by Immanuel Kant (1992: 157). In the 19th cen-
tury British politician Richard Cobden persistently advocated that Britain trade freely with her neigh-
bors to convince them of the advantages of free trade and to lock them more fully into the commu-
nity of nations. Cordell Hull, U.S. secretary of state from 1933 to 1944, and one of the architects of
the postwar international trading order, advocated this view throughout his public life. He declared
that "if we could increase commercial exchanges among nations over lowered trade and tariff barri-
ers and remove unnatural obstructions to trade, we would go a long way toward eliminating war it-
self" (Hull 1948: 84). Of World War II, he wrote, "Yes, war did come, despite the trade agreements.
But it is a fact that war did not break out between the United States and any country with which we
had been able to negotiate a trade agreement.... With very few exceptions, the countries with which
we signed trade agreements joined together in resisting the Axis. The political lineup followed the
economic lineup" (365).
Diplomatic considerations were at the heart of the 1860 Anglo-French (Cobden-Chevalier) com-
mercial treaty (Irwin 1993: 95,96). France was worried about offending protectionist interests, while
Britain was somewhat reluctant to pursue an agreement that would violate its policy of unilateral
free trade. Nevertheless, both governments saw a commercial treaty as a way of defusing tensions
and improving diplomatic relations and decided to sign it. In 1889 the Italian economist Vilfredo
Pareto argued that customs unions were "a means to better political relations and eventual pacifica-
fion" (Machlup 1977: 41).
As early as 1943, French statesman Jean Monnet, one of the founding fathers of the European
Economic Community (EEC), wrote, "There will be no peace in Europe if the states reconstitute them-
selves on a basis of national sovereignty with its policies of prestige and economic protection ... the
constitution of large armies will again be necessary . . . Europe will once again be recreated in
fear ... unless the States of Europe join in a Federation or a 'European entity' that results in a com-
mon economic unit" (Monnet, "Notes de reflexion," August 5,1943).
Monnet and compatriot Robert Schuman, the other architect of the EEC, explicitly argued that a
goal of the European Coal and Steel Community (ECSC), the precursor of the EEC, was to make Fran-
co-German war not only "unthinkable, but materially impossible" (Swann 1992: 6). The preamble to
the 1951 Paris Treaty establishing the ECSC "resolved to substitute for age-old rivalries the merging
of their essential interests, [and] to create, by establishing an economic community, the basis for a
broader and deeper community among peoples long divided by bloody conflicts." Walter Hallstein,
who had been a president of the EC Commission, later put it succinctly: "We are not in business at
all, we are in politics" (Swann 1992: ix).
Another potential approach to achieving peace in Europe would have been to deal with defense
matters directly. But by 1956, such attempts as the European Political and Defence Communities had
failed. Monnet and Schuman understood that, given the extent of mistrust after the war, the indirect
approach of economic integration was the only one likely to be acceptable to all parties at the time.
Trade as a Promoter of Peace
Increasing international trade might improve security in three related
ways:
1.   More trade means greater economic interdependence between the
countries involved. This increases the stake each country has in
the welfare of its neighbor and makes war more costly. It also in-



190    REGIONAL INTEGRATION AND DEVELOPMENT
creases the number of people who have an economic interest in
peaceable relations and so helps strengthen political pressures
against going to war.
2. More trade means more interaction between the peoples and
governments of the two countries, more familiarity with the
neighbor's goods and services, and greater understanding of their
cultural, political, and social institutions. All this is likely to in-
crease trust.
3. Secure trading relations will reduce the likelihood of war by in-
creasing security of access to the partner's supplies of strategic
raw materials and reducing the threat of trade embargo. This ar-
gument is especially important in a world of high trade barriers
where access to other sources is difficult-the situation in Europe
around 1950, when Jean Monnet and other French leaders were
concerned that the German coal barons in the Ruhr would have
too much control over French industry (Duchene 1994). The
post-World War I solution of direct control of German resources
had been discredited and was not an option. Monnet instead
pushed for the creation of the ECSC so that Germany could re-
build its industry without being a threat to France. This sectoral
approach to security was later expanded to the integration of gen-
eral trade relations, with the creation of the EEC.
Economists have examined some of the implications of uncertain ac-
cess to strategic raw materials. Arad and Hillman (1979) show how fear
of being cut off from foreign sources of defense equipment can cause
countries to overinvest in their own defense industries. Hillman and
Long (1983) discuss the optimal exploitation of a mineral resource if the
alternative foreign supply is uncertain. In both cases an RIA or a global
institution that ensured partner supplies would be both politically and
economically advantageous.
A fourth possible strategic benefit of trade is that the greater trust gen-
erated by increased trade may, in time, pay a peace dividend as defense
spending falls.
In all these cases one can envisage a virtuous cycle whereby increasing
trade permits closer integration, first through more thorough trade liber-
alization and then through policy or deep integration that binds the par-
ties closer together and facilitates yet further growth in trade, and so on
(see box 7.2, below).
Although direct evidence of the effect of trade on the likelihood of
conflict between any pair of countries is limited, numerous studies have
confirmed Chan's (1984) conclusion that conflict between countries is



REGIONAL INTEGRATION AS POLITICS                191
Box 7.2 ModelingTrade and Security Externalities
If increased trade between two countries reduces tensions between them because of increased trust
and economic dependency, what can economic analysis say about the type of intervention that can
capture this security externality? Schiff and Winters (1998) examine this Issue within a formal three-
country model, with two small, antagonistic countries and a large "rest of the world."
The model shows that, absent security externalities, nondiscriminatory free trade is the optimal
policy for the two small countries. If, however, increased mutual imports provide additional security
for small countries, it is worth subsidizing them by lowering their price relative to other goods. This
could be accomplished directly, but, given the fiscal objections to subsidies and their liability to cap-
ture, it is more frequently done by creating an RIA that taxes imports from the rest of the world, rais-
es their prices, and therefore lowers the relative prices of intrabloc imports. The arrangement is op-
timal (welfare is maximized) when external tariffs are set so that the marginal benefit from
security-including a "peace dividend" from reduced defense expenditures-equals the marginal
cost of trade diversion. Under these circumstances, the net welfare impact of forming an RIA is pos-
itive and Is not ambiguous, as it would be in the absence of security effects.
Strictly, to be equivalent to an import subsidy, the RIA should be accompanied by taxes on do-
mestic sales to raise their prices relative to member imports, but even if this is not done, there will
be security benefits from the trade preferences (although then the optimal tariffs are lower). The
case for an RIA is strengthened if security is related to both imports from and exports to the poten-
tially antagonistic partner, and there may be yet further benefits because in this case full optimiza-
tion requires that countries cooperate.
Over time, the optimal tariff on imports from nonmembers is likely to fall in such a security-
inspired RIA. As antagonism between the two countries diminishes, due to increased intrabloc trade,
the subsidy on intrabloc trade (or the optimal external tariff) falls. Thus, if security is the main moti-
vation for forming the RIA, the RlA's external trade policy should become increasingly open over time.
This is precisely what took place in the EC/EU: average tariffs on manufacturing products fell from
about 13 percent in 1958 to about 3 percent after the Uruguay Round.
Are trade preferences the only way to obtain security benefits? Since deep integration can also
lower trading costs and increase trade flows, it can enhance security even in the absence of trade
preferences (and if external trade barriers are present, their optimal level falls following deep inte-
gration). But it seems likely that countries will engage In deep integration-which implies giving up
a degree of sovereignty-only with countries they already trust. If trust is initially low, trade prefer-
ences may be the only available instrument until the degree of trust has increased. This, in fact,
increases the warranted level of the external tariff in the early stages of an RIA, for preferences not
only boost trade directly but also, by creating conditions for future deep Integration, promise further
security and economic gains later.
less prevalent if both are democratic. Polachek (1992, 1996) explains
this finding through the effect of democracy on international trade. He
estimates that democratic countries trade more with each other than do
other countries and, using detailed data from the Conflict and Peace Data
Bank, finds trade to have a significant and negative impact on conflict.
Polachek estimates that a 6 percent increase in trade lowers his measure
of conflict by about 1 percent. A key feature of these results is that



192    REGIONAL INTEGRATION AND DEVELOPMENT
Polachek tested for causality and found that an increase in trade between
partners caused a reduction in conflict but that reduced conflict did not
increase trade.3
We are not saying that trade always promotes peace or that trade is
sufficient for peace. Clearly, trade partners do fight, and sometimes over
trade issues, as discussed below. On the whole, however, there is persua-
sive evidence that trade will generally tend to foster peaceable, if not
friendly, relations between countries.
RIAs as Promoters of Peace
So far, we have been discussing the effect on conflict of trade in general-
implicitly, nondiscriminatory trade. It is widely held, however, that dis-
criminatory trade has the same effect. Although this argument seems
highly plausible under certain conditions, it is not always the case.
Political scientists have researched the use of trade diplomacy within
a regional context and have concluded that trade relations, including
RIAs and, especially, deeper arrangements, might assist political rela-
tions between member countries by developing means for avoidance
and management of intramural conflict. The negotiations between lead-
ers of neighboring countries that are required to form and operate an
RIA tend to generate trust between them. This helps them identify with
each other, understand each others' problems, and interpret each others'
actions. Trade talks allow political or economic elites to form coalitions
for subsequent collaboration and consensual action.4 Wallace (1994: 4)
argues that the "most striking phenomenon of formal European inte-
gration has been the interpenetration of national administrations, with
ministers and officials from different governments in close and continu-
ous contact."
The main motivation for creating the ECSC in 1951 and the EEC in
1957 was to reduce the threat of war in Europe, especially war between
France and Germany. Similar motives are found in the creation of
ASEAN-to reduce tensions between Indonesia and Malaysia (De Rosa
1995)-and of APEC and the CACM, which include potential political
or military opponents (Page 1998). Anwar (1994), in assessing ASEAN's
role as promoter of regional peace, notes that intraregional conflicts oc-
curred among the five founding members before ASEAN was founded
but not afterward. Srinivasan (1994: 7) argues that greater economic
interdependence among South Asian countries would help defuse ten-
sions among them, stating that "promoting freer movement of goods,



REGIONAL INTEGRATION AS POLITICS    193
services, people and capital in the region might also facilitate the resolu-
tion of political and territorial disputes."
Security also seems to have played an important role in the Southern
Cone of South America. The Argentine and Brazilian militaries long jus-
tified their claims to resources for defense partly on the potential threats
from each other. In the mid-1980s the two countries signed nuclear co-
operation agreements and economic agreements covering steel and auto-
mobiles in the expectation that this would help reduce tensions between
them by curtailing the power of the military and strengthening their frag-
ile democracies. The creation of MERCOSUR in 1991 reinforced this
process and bound smaller neighbors into it.
Rubens Ricupero, secretary-general of the United Nations Conference
on Trade and Development (UNCTAD) and former finance minister of
Brazil, confirms the importance of MERCOSUR's security aspects:
Both countries were emerging from a period of military govern-
ments during which considerable tension had characterized the bilat-
eral relationship, centered on a long-standing controversy about
competing hydroelectric projects in international rivers of the Plata
Basin. Both militaries had also continued to pursue their secret nu-
clear programs. It was essential to start with agreements in the eco-
nomic area in order to create a more positive external environment
that rendered it possible to contain the military nuclear programs
and to replace rivalry by integration. This effort was developed
along successive stages and eventually led to signature by the two
governments of Brazil and Argentina. (Personal communication,
1998)
Thus, as with the ECSC and the EC, the indirect path to enhancing se-
curity through economic integration was deemed an essential first step.
Africa provides what may be a related example. In 1986, 11 of 15
members of ECOWAS ratified a mutual defense protocol that authorizes
military intervention by the community in conflicts between members or
if conflict in a member country is instigated from outside and is likely to
endanger peace and security in the entire community (Oxford Analytica
Daily Brief, September 5, 1997). One interpretation of this initiative is
that by creating institutions spanning the linguistic divide-an almost
unique achievement in West Africa-ECOWAS enabled neighbors to de-
velop cooperative behavior that eventually allowed them to address mu-
tual security concerns. In fact, in this view, ECOWAS has gone farther
than either MERCOSUR or the EU, for it uses an actual defense pact to
bolster security, while the others use only economic integration.



194    REGIONAL INTEGRATION AND DEVELOPMENT
In cases where security is an issue and is amenable to trade-related
policy, creating an RIA may be the optimal approach (see box 7.2).
Under such circumstances, we can also infer that the optimal level of
protection that RIA members maintain against imports from nonmem-
bers will fall over time as trade grows and also following policy inte-
gration. This last observation can be used as a test for identifying coun-
tries' unobservable motivations for creating an RIA: if security were
the main motive, tariffs would fall, and therefore if tariffs do not fall,
we know that the efficient pursuit of security was not the main objec-
tive of the RIA, regardless of the rhetoric surrounding its creation. This
observation can also serve as a policy prescription for "security-
inspired" RIAs.
RIAs Are Not Always Effective Routes to Peace
In the EU and MERCOSUR, integration helped enhance security by in-
ternalizing security externalities associated with intrabloc trade. The
main reason for the success of these RIAs in achieving this goal is that the
members were actually looking for arrangements to solve a security
problem, not an economic one, and that defense pacts were not feasible,
given low trust at the time. Thus, their objectives were political rather
than economic, and member countries structured the RIAs in such a way
as to attain those objectives. Among other things, economic gains and
losses were shared in ways that member countries perceived as fair.5
Many other RIAs, however, are motivated by economics. In these cases
an asymmetric distribution of benefits and costs may result in frictions
among member countries. In other words, the pursuit of economic gain
may result in security losses if the gains for one member come essentially
at the expense of other members.6
Thus, the finding that trade enhances security does not allow us to
conclude that policies which promote trade within a region will always
improve the prospects for regional peace. Indeed, they may have precise-
ly the opposite effect. Policy-induced integration promotes trade at a
price. The tariff preferences that induce regional trade can create power-
ful income transfers within the region and can lead to the concentration
of industry in a single location. The countries or regions which lose in-
come or industry can be so resentful that separatist movements arise and
the overall risk of conflict is increased. In such cases, disintegration, as
happened with East and West Pakistan, may be the outcome.
A clear example of how integration can trigger conflict was the Amer-
ican Civil War. The United States constituted a customs union in which



REGIONAL INTEGRATION AS POLITICS  195
the North produced manufactures that were sold in the South, while the
southern states produced cotton that was exported to Europe. Tariffs
nearly triggered civil war in 1828, when Congress, dominated by north-
ern interests, sharply raised the U.S. import duty on manufactures. The
effect of what was known in the South as the "Tariff of Abominations"
was to increase the price that northern manufacturers could charge in the
South, generating a massive income transfer from the South to the
North. South Carolina refused to collect the duties and threatened to se-
cede unless the tariff was rescinded. The federal government sent in
troops, but Congress backed down before fighting developed. In 1860
northern interests tried again, and this time Congress would not back
down. This, perhaps as much as slavery, was the issue that led the south-
ern states to try to quit the Union (Adams 1993).
Another example is the East African Community (EAC). Here, Kenya
was the equivalent of the northern states in the American case. Tanzania
and Uganda complained about the income transfers that the common ex-
ternal tariff on manufactures created. They also feared that there would
be an increasing agglomeration of manufacturing in Nairobi, which had
a head start on industrialization compared with the smaller industrial
centers of Dar es Salaam in Tanzania and Jinga in Uganda. Arguments
about compensation for the income transfers led to the collapse of the
EAC, the closing of borders, and the confiscation of Community assets
in 1978. This atmosphere of hostility contributed to conflict between
Tanzania and Uganda in 1979.
In the light of these examples, in which the trade policy used to pro-
mote regional integration was so unfair that it actually worsened intrare-
gional security, the success of the EEC looks even more remarkable.
Among the favorable factors was, first of all, a genuine desire for securi-
ty, perhaps traceable to Europe's extraordinarily bloody history, which
saw three Franco-German conflicts in a century, and to the failure of a
number of other efforts at integration such as the European Political and
Defence Communities. It also helped that the key players, France and
Germany, were relatively evenly balanced. Had they not been, an alter-
native solution to the security problem-hegemonic domination-would
have been more likely than the reliance on mutual benefits that charac-
terizes the RIA route.
Furthermore, reflecting the genuine wish for reconciliation, the Euro-
pean Community has always pursued regional integration in ways that
avoid transfers large enough to trigger conflict. This was partly a matter
of negotiating style and partly of design (Winters 1997b). The style was
consensual: negotiators were always looking for compromise and concil-



196    REGIONAI INTEGRATION AND DEVELOPMENT
iation. When a country signaled that a Community policy would cause it
major political or economic problems, it was accommodated, either by
being offered compensation, as with the British budget rebate negotiated
by Prime Minister Margaret Thatcher, or through a very gradual adjust-
ment process, as with application of the rules on labor mobility to new
members Portugal and Spain. The key design feature was that the Com-
munity's external tariffs were generally low and declining, so that the in-
come transfers arising from exploitation of consumers in one nation by
producers in another were relatively small.
The one exception was agriculture, which has been highly protected,
has generated large income transfers between countries, and has been a
source of some internal political conflict. In grand terms, however, even
agriculture was part of the inspired peace bargain. The Common Agri-
cultural Policy arose because France wanted access to German markets
at the high prices that German farmers also desired. The conflict over
agriculture has not often been between France and Germany but between
these two countries and other members, especially the United Kingdom.
The economic costs of forming a security-enhancing RIA arc impor-
tant, even if the objective is as apparently noneconomic as securing
peace. Before deciding to form such an RIA, policymakers should be
convinced that trade will enhance trust significantly-that contact will
help, not harm, general relations, and that it will create no new frictions.
They should also ask whether an RIA is the most efficient mechanism for
internalizing security externalities. Should they choose to form an RIA
for security reasons, they should ensure that tariffs are not set higher
than is absolutely necessary to capture the security externalities that
exist, and they need to realize that these tariffs should decline over time
and following deeper measures of integration.
RIAs AND SOCIAL AND POLITICAL PRESSURES: POTENTIALLY
HELPFUL, BUT NOT A PANACEA
The formation of an RIA is sometimes seen as a means of preventing or
reducing the spread of civil disturbances or civil war from neighboring
nations or of controlling migratory flows. The EU has been concerned
with such threats from North Africa, and this has been one motivation
behind the Euro-Mediterranean Agreements. The hope is that these
agreements, including their associated aid protocols, will improve the
economic situation in the North African partner countries and help con-
tain the problems. Similarly, both the United States and Mexico have



REGIONAL INTEGRATION AS POLITICS  197
been concerned about the possibility of occasional social strife and vio-
lence in Mexico spreading northward, and both hoped that NAFTA, by
facilitating access to U.S. markets, would help improve Mexico's eco-
nomic situation and reduce social tensions. Implicit in this hope is that
NAFTA will increase Mexican economic welfare.
Closely related is the manifest desire of rich countries to stem large-
scale migration from poorer countries even if it does not immediately
threaten political and social stability. This too has played a role in the
formation of RIAs, including NAFTA and the EU's Europe Agreements
with the countries of Central and Eastern Europe (OECD 1995). In the
case of NAFTA, Presidents Salinas and (George H. W.) Bush argued that
helping Mexico export more goods would help it export fewer people,
thereby reducing migration pressures. The recent Euro-Mediterranean
agreements also provide evidence of such motivation.
Whether RIAs do help solve migration problems depends on whether
trade and migration are complements or substitutes. Standard trade the-
ory holds that they are substitutes, so that increased trade integration is
likely to reduce income or wage differentials and decrease labor migra-
tion flows. More recent analysis, however, and some empirical results
have shown that North-South trade and migration may well be comple-
ments, so that integration may not lower migration, especially of un-
skilled labor.
Four main arguments lead in this direction. First, it may plausibly be
argued that migration entails a fixed cost and that developing country
capital markets are highly imperfect (L6pez and Schiff 1998). If so, very
poor people may not be able to afford migration, and a policy that in-
creased their incomes could relax their capital constraint and allow them
to move. Second, the costs of migration decrease as more information
about the destination country becomes available. Since, for the reasons
discussed above, this seems a likely effect of an RIA, migration to the
new partners could increase. Third, even if an RIA is welfare improving
overall-which, we have argued, cannot be taken for granted-it may
not benefit unskilled workers, who are the real betes noires of the poten-
tial countries of immigration. For example, since the mid-1980s, un-
skilled workers have fared poorly in Mexico, with declines in their real
income of 10 to 15 percent (Hanson and Harrison 1999). To the extent
that this deterioration is attributable to NAFTA, the incentives for these
people to emigrate, if not their ability to do so, is increased by integra-
tion. Finally, the changes in the production structure induced by an RIA
will cause some people to migrate internally within member countries,
and evidence suggests that once people have been shaken loose from



198    REGIONAL INTEGRATION AND DEVELOPMENT
their "homelands," their propensities to migrate internationally increase.
That is, having once uprooted and moved from, say, rural areas to Mex-
ico City or to areas bordering the United States, it is a small step to move
farther, to the United States itself (Sewastynowicz 1986; Morrison and
Zabin 1994).
Migration remains an issue even within common markets in which
free mobility of labor is ostensibly an objective. The EU set a 13-year
transition period before Portugal and Spain were permitted completely
free access to other members' labor markets. Similarly, it is widely ac-
cepted that fear of migration is one of the reasons behind the reluctance
of the EU to consider Turkish membership seriously. Even among long-
standing members of the EU, migration has not been made easy, and
labor mobility is much lower in Europe than in the United States (Blan-
chard and Katz 1992). In part, this is cultural, but it also reflects multiple
policy frictions, such as pension transferability, housing systems, and
health provision, that make effective migration complicated or worse.
THE ROLE OF REGIONAL INTEGRATION IN STRENGTHENING
DEMOCRACY AND POLITICAL INSTITUTIONS
RIAs can be useful tools for improving political institutions. Trade blocs
with strong club rules can help anchor democratic reforms in member
countries. Membership in an RIA can increase the likelihood of achiev-
ing or upholding democracy, especially if the bloc includes large demo-
cratic developed countries. Newer or less politically advanced countries
may gain from joining an RIA that includes at least one large developed
country if accession is part of a strategy of pursuing political, economic,
or social reforms (or preventing backsliding), when such moves would
not be feasible without the conditionality embodied in the RIA's club
rules. Those rules often include democracy and human rights.
MERCOSUR put its-at the time, informal-democracy rule into
practice in April 1996, when the commander of Paraguay's armed forces
was said to be contemplating a military coup. The bloc's four presidents
(with backing from the United States and the Organization of American
States) reportedly quelled the rumored coup with a strong joint state-
ment that democracy was a condition of membership in the bloc. Two
months later MERCOSUR amended its charter to formally exclude any
country that "abandons the full exercise of republican institutions."7 In
forming FTAs with MERCOSUR, Bolivia and Chile accepted democracy
as a condition for membership (see box 7.3).



REGIONAL INTEGRATION AS POLITICS                  199
Box 7.3 Democracy in MERCOSUR
The Presidential Declaration on the Democratic Commitment in MERCOSUR, signed in San Luis, Ar-
gentina, on June 25,1996, made democracy a condition of membership for the four member coun-
tries, Argentina, Brazil, Paraguay, and Uruguay. This condition was extended to the FTAs between
MERCOSUR and Bolivia and Chile in the Protocol of Ushuaia, signed on July 24,1998.
How credible is the 'democratic commitment"? Are member countries likely to act on it and
punish countries that deviate from democratic principles? One indication of the credibility of the
commitment is provided by the specifics of the Protocol of Ushuaia, which sets forth the full validity
of democratic institutions as an indispensable condition for the existence of the MERCOSUR agree-
ments and establishes procedures for consulting on violations. If there is rupture of the democratic
order in one state, the other countries are to consult among themselves.and with the affected state.
If these consultations prove ineffective, the other countries will decide on the nature and extent of
the measures to be applied. The sanctions listed range from suspension of the right to participate in
the organs of the various agreements up to suspension from MERCOSUR. These measures are to be
adopted by consensus and communicated to the affected state, which does not participate in the de-
cisionmaking process. The sanctions are to end once it has been verified that democratic order has
been fully reestablished.
MERCOSUR and the associated countries have reduced ambiguity about the "democratic com-
mitment" to a minimum. Punishment seems very likely, and the market of MERCOSUR plus Bolivia
and Chile is important enough to the member countries that they are likely to consider the threat of
retaliation seriously. Thus, the threat is likely to be effective.
It is generally thought that the framers of MERCOSUR borrowed the democracy idea and lan-
guage from the 1957 Treaty of Rome that established the EEC. But although it seems understood
that only democracies are eligible for (new) membership, the Treaty of Rome does not mention
democracy (although it does mention peace and liberty).
Why is there a strong statement on democracy in MERCOSUR and not in the EEC? All EEC mem-
ber countries had been democratic for more than a decade in 1957, and the United States acted as
a sort of guarantor. MERCOSUR member countries, by contrast, were emerging from long periods of
military rule, there was no extemal guarantee, and preserving and consolidating fragile democracies
was an important objective of the governments involved.
The EEC had no formal democracy requirement, although by conven-
tion and practice it was understood from the mid-1960s that such a con-
dition existed, at least for new members. Bhalla and Bhalla (1997: 159)
argue that it was generally understood that "the acceptance of the poor-
er economies of Greece, Spain and Portugal was motivated largely by the
desire to help these restored democracies remain democratic by bolster-
ing them politically and economically." Similarly, the Europe Agreements
with accession candidates in Central and Eastern Europe and the Baltic
area are designed to "facilitate" the transition countries' "full integration
into the community of democratic nations" (title 1, art. 2). Latvia, a can-
didate for EU accession, is reviewing its citizenship policies for its Russ-
ian minority to meet EU concerns about human rights (Washington Post,



200    REGIONAL INTEGRATION AND DEVELOPMENT
July 24, 1998). The EU agreements with Mediterranean countries also
include respect for human rights and the rule of law, as does the Cotonou
Agreement between the EU and the ACP countries. In the 1992 Treaty on
European Union, explicit reference is made to democracy, although not
in any operational form.
Conditions regarding democracy and human rights will be truly effec-
tive only if the penalties for violating them are severe and their enforce-
ment credible. As with economic policies (chapter 4), it is difficult to pin
down exactly what inspires credibility, but for developing countries an
explicit statement of the club rule seems necessary, and an explicit and
plausible plan for its enforcement highly desirable (see box 7.3). In the
absence of such conditions, it is difficult to see how the objective will be
achieved-and, in particular, how it will be achieved without the need to
resort to the sort of explicit political pressure that MERCOSUR had to
exert in the case of Paraguay.
The enforceability of club rules depends both on the value of belonging
to the bloc and on the credibility of the threat of action. First, new mem-
bers who, in joining a large bloc, obtain significant benefits-including ac-
cess to a large market and greater bargaining power with the rest of the
world-are unlikely to break the rules (or backslide) and so risk losing
these benefits. Admittedly, parties that pose threats to democracy may not
be moved by such considerations, but ordinary economic agents will be,
and so the condition will increase the difficulty of building support for in-
surrection. Second, the credibility of the threat of action is likely to be
greater if violation of club rules by new members entails a large cost to the
other members. That cost may be direct and economic or (perhaps more
likely) indirect and political, such as a demonstration effect.
The effectiveness of club rules is likely to vary with the nature of the
membership. RIAs between small low-income countries, which typically
trade very little with each other, are less able to impose significant costs
on recalcitrants by ejecting them. Choosing large, important partners im-
proves the chances that club rules will be enforced. Because a country is
likely to be more concerned about the social and political conditions of a
nearby than of a far-off country, enforcement of club rules is likely to be
more effective in RIAs between neighbors than in those between distant
partners.
Two other circumstances will affect the credibility of enforcement.
First, it seems unlikely that nondemocracies, or countries where democ-
racy is very fragile, will prove stern disciplinarians even if a democracy
rule exists. Second, the enforcer has to see enforcement as an important
issue in itself. It is also important that there be no competing issues that
are more vital to the enforcer. While the Cold War was in progress, West-



REGIONAL INTEGRATION AS POLITICS  201
ern powers were far more concerned to ensure that client countries were
securely anticommunist than with the details of their governance struc-
tures. Imposing democracy rules was just not an element of policy.
We have noted the importance of proximity. The question of distance
is relevant to one current debate. The EU's Cotonou Agreement with the
ACP states has the goal of eventually replacing the Lom6 Conventions
with a series of regional economic partnership agreements (EPAs) with
groups of ACP countries (McQueen 1998). It includes, as an "essential
element," the developing partners' respect for human rights and democ-
racy, as well as injunctions to manage their economies properly (Council
of the European Union 1998). Some commentators (for example, Win-
ters 2001; Schiff and Winters 2002a, 2002b) argue that the EU's interest
in enforcing good economic policy in the ACP countries, let alone
"good" political practice, is very doubtful. The ACP countries are too
small and distant to affect any EU interest materially, and the various EU
members frequently have different views on any particular case. In addi-
tion, disciplining former colonies for pursuing economic or political
policies that are not approved in European capitals looks like an interna-
tional public relations nightmare. On this view, EPAs will generate no
additional credibility for the ACP countries. A contrary view, at least on
economic credibility, is set forth by Collier and others (1997), who see
the EU as the stern external agent of restraint that developing countries
require in order to convince the world that they are reliable and will suc-
ceed economically. The argument is essentially that the EU is so impor-
tant to the ACP countries that they would never flout it. Thus, the EU
can do its partners some good at almost no cost to itself.
An interesting feature of the club rules-democracy argument for RIAs
is that it is one case where there is clearly no multilateral substitute. Mul-
tilateral trade arrangements cannot propose and enforce these sorts of
rules, which are simply not part of the mandate of organizations such as
the WTO. Other international organizations such as the United Nations
or regional and multilateral development institutions may be able to per-
suade member countries to abide by some rules, but so far, these do not
include democracy or other constraints on political regimes. It would ap-
pear that positive spillover effects in the political arena are only possible
in large regional arrangements that include club rules.
REGIONALISM AND THE NATION-STATE
Joining an RIA necessarily requires surrendering some immediate control
over policymaking and losing some political autonomy, if only over



202        REGIONAL INTEGRATION AND DEVELOPMENT
tariffs on partners' exports. (So, of course, does membership in the
WTO.) Some RIAs, however, go deeper than this and create institutions
for joint decisionmaking. For example, as the EU's integration has deep-
ened, decisionmaking has increasingly moved away from national capi-
tals to Brussels, and much of the current debate is shaped by the belief
that some form of political unification must eventually follow the cre-
ation of an integrated economic unit (box 7.4).8 But such integration
need not result in the suppression of the nation as an organizational
framework or in the loss of effective sovereignty. On the contrary, by
pooling sovereignty, members of an RIA may be able to preserve and en-
large it and thus strengthen the concept of national identity and integrity.
Nation-states can strengthen themselves by creating a united front
against external pressures or by joining forces in international negotia-
tions. Setting      aside   considerations      of coalition      building    and    policy
spillovers, however, such cooperation does not strictly require an RIA in
the sense of trade preferences.
Box 7.4 Trade Preferences Do Not Inevitably Lead to Political Integration
Although few customs unions aim explicitly at political union, many have gone far beyond simple
preferential trading arrangements to achieve a degree of political integration. Economic integration
has often been a precursor to and facilitator of closer political association (as in the EEC) or even
state formaton, as in the ZolIverein, founded in 1834, which contributed to the creation in 1871 of
the German nation-state. Similarly, Moldavia and Wallachia formed a customs union in 1847 before
uniting as Romania in 1878.
But economic integration is not needed for politcal integration. In the past-and today in some
developing countries-nations had internal trade barriers, which meant that they were politically,
but not economically, integrated. In the Middle Ages customs collectors were frequently stationed
not just at the boundaries but in the interior of political units-at important market centers, junctions
of trade routes, or mountain passes. England and Scotland became united under a single monarch in
1603, but attempts to reach agreement on commercial union failed until the Act of Union in 1707
(Irwin 1993). Although politically unified under the king for centuries, France remained divided-
even after several reforms-by 1,600 internal tolls and tariffs until the French Revolution enabled
their abolition in 1790. Each Swiss canton retained tariff autonomy unbl 1848. More recently, Hong
Kong (China) and mainland China have achieved political unity without full economic integration, as
internal trade barriers and dual customs jurisdictions persist.
Not only is economic integration not necessary for political integration but it can also decrease
regional political unity. Common external tariffs or rules of origins may operate as irritants and as a
stimulant to separatist movements. Such was the case with Western Australia in the Australian Com-
monwealth, the Prairie Provinces in Canada, and the North and the South in the antebellum United
States.



REGIONAL INTEGRATION AS POLITICS  203
Regional Integration to Deal with Outside Threats and
Regional Hegemons
The normal approach to external security threats is for countries to form
alliances independent of any trade preferences. It is possible, however, to
start with a trade pact, based on "hopes that economic union between the
weak would ripen into political union, and that by the political union of
the weak a power might be established strong enough to defend against
aggression from outside" (Viner L950: 92). Nations that feared being ab-
sorbed forcibly by larger states have united to forestall such coercion.
Thus, the Austrian emperor proposed (but eventually aborted) an eco-
nomic union with Spain and Bavaria as a defensive scheme against France
in 1665 (Viner 1950: 93). More recently, the Gulf Cooperation Council
(GCC) was created in 1981 partly in response to the potential threat of re-
gional powers such as the Islamic Republic of Iran and Iraq (Schiff and
Winters 1998), and ASEAN was partially motivated by a perceived need
to stem the threatened spread of Communism in Southeast Asia. A major
motive of Central and Eastern European countries in applying for mem-
bership to the EU is as protection from a perceived threat from Russia.
The Southern African Development Coordination Conference
(SADCC) was initially formed in 1980 to provide a united front against,
and reduce dependence on, South Africa. After apartheid ended, South
Africa was invited to join the group, now the Southern African Develop-
ment Community (SADC). The difference was that whereas SADCC
involved cooperation on trade matters in general but not mutual
trade preferences per se, SADC is developing a trade protocol based on
preferences.
Regional Integration and Negotiations with the Outside World
Regional cooperation (which may, but need not, involve trade prefer-
ences) can strengthen the voices of small nations. These countries often
face severe disadvantages in dealing with the rest of the world because of
their low bargaining power and high negotiation costs. Bilateral and
multilateral negotiations often require substantial financial resources,
time, and expert knowledge, which are limited in small countries. As the
world has become more integrated and the number of issues to be dealt
with in the international arena has grown, the incentive for small coun-
tries to cooperate with their neighbors has grown, as well.
Small countries can substantially reduce their negotiation costs and at
the same time increase their market and negotiating power by pooling



204    REGIONAL INTEGRATION AND DEVELOPMENT
their resources and acting together to articulate shared interests. This is
more likely to come about (Andriamananjara and Schiff 2001):
*   If their interests are similar (so that intrabloc negotiation costs
are low)
*   If the cost of international negotiations is high (creating greater
incentives to cooperate), and
*   If a large number of issues needs to be dealt with (which both in-
creases the incentives and makes it easier to construct packages in
which every party can gain).
Establishing a regional grouping typically involves "logrolling": "I'll
vote for your issue if you vote for mine." By trading support for each
other's preferred issues, countries can get more than they could obtain
unilaterally (Andriamananjara and Schiff 2001). Even so, such coalition
formation is neither easy nor common. Members will usually have to
sacrifice some preferred positions even before the international negotia-
tion process begins as the coalition settles on priorities. These steps can
be politically difficult, especially if the group is large and the countries
differ widely. The coalition also needs to devise ways of responding to of-
fers and setbacks in negotiations, for it is quite certain that they will not
achieve all that they hope for initially. Setting up a secretariat and for-
mulating suitable institutional rules for decisionmaking may help in
these processes but require significant time and resources up front.
Relatively shallow but highly successful examples of cooperation of
this kind are the Scandinavian and ASEAN groups in the WTO. These
blocs pool their resources to attend meetings, providing regular briefings
for each other. If they agree on an issue (and they put some effort into
discovering beforehand whether they do or not), the representatives may
speak for the group; if not, individual countries look after their own in-
terests.
At a deeper level, small Caribbean nations increased their bargaining
power by establishing CARICOM in 1973 to pool their negotiation re-
sources and formulate common policy stances (see box 7.5). This al-
lowed the member states to reduce their negotiation costs and exert
greater influence outside the region than would have been possible had
they acted independently. The region acquired bargaining power at the
very highest level of North-South politics. Representatives of CARICOM
countries took the lead in formulating and articulating the positions of
the ACP group in negotiating the Lome Conventions. By pooling their
support, the CARICOM nations succeeded in getting their nationals
elected to key international positions such as Commonwealth secretary-
general and ACP secretary-general. In the process, they ensured that the



REGIONAL INTEGRATION AS POLITICS                   205
Box 7.5 CARICOM and International Negotiations
All Caribbean states face problems of political and economic viability because of their extremely
small size. CARICOM, created in 1973, has as members Antigua and Barbuda, The Bahamas, Barba-
dos, Belize, Dominica, Grenada, Jamaica, Montserrat, St. Kitts and Nevis, St. Lucia, St. Vincent and
the Grenadines, and Trinidad and Tobago. The members have populations ranging from 10,000 to
2.55 million and GDPs of from $0.24 billion to $6.3 billion.
CARICOM was intended to give the region a more powerful voice and presence to defend its in-
terests in intemational affairs (Byron 1994; IADB 1995). The group has three main areas of activity:
economic integration; cooperation in noneconomic areas (health, education, and transport) and In
operation of common services; and coordination of the foreign policies of the member states. CARl-
COM has been particularly active in negotiating preferential access to European and North American
markets, negotiating for consistent and remunerative commodity prices, obtaining larger flows of
concessionary finance for the region, and raising the region's profile in multilateral institutions
(Byron 1994). It has been involved in negotiations between the ACP countries and the EU and with
the GATT/WT0, the proposed Free Trade Area of the Americas, the United Nations Conference on
Trade and Development (UNCTAD), and the United Nations Convention on the Law of the Sea (UNC-
LOS). It has taken part in commissions or joint councils with Canada, Cuba, Japan, Mexico, the Unit-
ed States, the Organization of American States, and the Group of Three. A single microstate would
not have had the human, physical, or financial capacity to conduct fruifful negotiations unilaterally in
so many areas.
region's interests in commodity trade and development cooperation were
taken into account. They also consolidated multilateral links with other
parts of the developing world, established themselves as full participants
in the activities of the United Nations (U.N.), despite that organization's
earlier ambivalence on the issue of microstate membership, and focused
on getting U.N. organs to address the development needs of small island
developing states. Finally, they succeeded in collectively negotiating a
whole range of preferential market access agreements-for example,
CARIBCAN with Canada, the Caribbean Basin Initiative with the Unit-
ed States, and the Lome Conventions and (along with other developing
nations) the generalized system of preferences (GSP) with the EU. De-
spite its relatively limited trade and investment impact, CARICOM has
been successful in serving as a political instrument in joint negotiations
on trade and investment with larger countries and regional trading blocs.
The existence of a visible regional and supranational authority may
attract more foreign assistance (or even foreign direct investment) be-
cause it is easier for the donor community to deal with the group as an
entity than with each country individually. Regarding the SADCC, Inotai
(1991) observes: "More recently, common activities emerged in order to
attract higher volumes of external financial resources. By 1988, SADCC



206     REGIONAL INTEGRATION AND DEVELOPMENT
could ensure external financing for 20 industrial projects, and is now
working on getting additional resources for 11 more projects." The
SADCC's successor, the SADC, now acts as a regional coordinating
mechanism with the donor community. For instance, in February 1996
the U.S. Department of Commerce signed a memorandum of under-
standing with the SADC that outlines six areas for cooperation in ad-
vancing commercial development in southern Africa.
As for other RIAs, in ASEAN Japanese aid has played an important
role in assisting regional industrial projects, including automobile assem-
bly and parts production (Bhalla and Bhalla 1997). Similarly, following
its advocacy of regionalism as a principle, the European Union has ac-
tively assisted subregional integration-for example, in Central Europe,
the Mediterranean, and the CACM.
Finally, the objective of strengthening negotiating and bargaining
power is not confined to the formation of blocs by small countries.
Whalley (1998b) argues that this idea was shared by the countries in-
volved in the creation of the EEC in the late 1950s, which felt that to-
gether they would have much greater leverage in negotiations with the
United States than they would individually. Similar arguments were
made in the United Kingdom in favor of joining the EEC. Whalley asserts
that the goal of increasing negotiating power, especially vis-a-vis
NAFTA, was also present in the formation of MERCOSUR.
Having discussed the appeal and efficacy of RIAs in the political
arena, we turn to the issue of market power as part of the discussion,
in chapter 8, of the effects of regional integration on global trade liberal-
ization.
NOTES
1. British Prime Minister Neville Chamberlain's infamous phrase about the
German invasion of Czechoslovakia in 1938 (Taylor 1979).
2. Recall that the worry about RIAs is not that they reduce trade but that the
extra trade they bring about reduces economic welfare.
3. Although these tests are informative, it should be noted that they relv on a
limited, temporal notion of causality: A causes B if observing the current and
past outcomes of A helps predict the current value of B.
4. Others take the view that friendly nations or allies, rather than enemies,
tend to form RIAs-that the order of causation is from friendship to regional in-
tegration. Mansfield (1993) argues that countries will lower barriers only with
allies (not with adversaries) because increased trade will raise incomes, which
can enable higher defense expenditures, which in turn only makes sense if the



REGIONAL INTEGRATION AS POLITICS      207
countries are allies. The fact that an RIA might be as likely to reduce income as
to increase it counts against this view as a general theory.
5. One cause of perceptions of unfairness can clearly be short-to-medium-
term fluctuations in exchange rates. As noted in chapter 1, for reasons of space
we decided not to deal with macroeconomics and to limit the analysis to real
phenomena.
6. Of course, tension arising from the redistribution of rents is not specific to
RIAs. For instance, Schiff (1998) argues that if moving all the way to free trade
results in such a redistribution of rents across ethnic groups that it raises tensions
among those groups (destroys social capital), it may be harmful economically.
7. Presidential Declaration on the Democratic Commitment in MERCOSUR,
San Luis, Argentina, June 25, 1996; Talbott (1996); "Survey on MERCOSUR,"
The Economist, October 12, 1996.
8. As president of the European Commission, Jacques Delors stated that one
consequence of the Single European Act would be that four-fifths of those deci-
sions then taken in national capitals would eventually be made in Brussels (Mil-
ward 1992: 2). The increasing number of lobbying groups setting up in Brussels
seems to support the view of the EU's increasing importance in decisionmaking.






CHAPTER 8
Trade Blocs and the Rest of the World
RIAs are, by nature, exclusive clubs. Every country in the world is ex-
cluded from nearly every RIA in the world, and every RIA excludes near-
ly every country. Discrimination against excluded countries is real and
causes significant trade diversion, for which we present new empirical
evidence in this chapter. Trade diversion can harm excluded countries (in
shorthand, the rest of the world, ROW), particularly where they face
large trade blocs. The extent of the harm depends on how much the RIA
diverts trade and on the structure of the ROW economy in the sectors in
which trading patterns are affected. As a rule, trade blocs harm non-
members least if they are less trade diverting, liberalize their external
trade, and boost global competition by increasing member efficiency and
growth.
The other main issue in the intersection between RIAs and the rest of
the world is whether RIAs are stepping stones toward globally freer
trade, or millstones around the neck of progress toward that goal. Many
arguments have been advanced for the benign view-that regionalism
stimulates global trade negotiations, that it makes negotiations simpler,
that "open regionalism" liberalizes trade, or that blocs can advance far-
ther and faster than global negotiations. In truth, however, the world of
multiple trade blocs is still too new to allow a definitive answer to this
question, and much of the evidence shows that advances in multilateral
trade negotiations have led, not followed, the formation of trade blocs.
Moreover, there are strong analytical reasons to suggest that regionalism
is more likely to undermine than to support full free trade and that it
may increase the chances of trade wars.
209



210    REGIONAL INTEGRATION AND DEVELOPMENT
The WTO's rules and practices on regionalism impose some discipline
on the worst kind of trade blocs but cannot, in the end, prevent members
from creating RIAs that harm themselves or others. Moreover, the rules
do not fully apply to developing countries and are not enforced very ac-
tively anyway. We cannot identify rule changes that are both feasible and
desirable, but we do urge that current rules be clarified and enforced. We
also suggest that more attention be paid to assessing the economic effects
rather than just the legal standing of proposed RIAs.
TRADE DISCRIMINATION: STILL SIGNIFICANT
Trade discrimination against excluded countries commonly causes signif-
icant trade diversion. Although this diversion is not inevitably harmful to
the rest of the world, it will be injurious under two common circum-
stances: when nonmembers tax their international trade (for example, by
imposing tariffs), and when nonmembers' export prices fall as a result of
falling demand. Merely examining the value of ROW exports to the RIA
is not sufficient to identify harm; one really needs to consider the evolu-
tion of ROW imports and the prices paid for them relative to the prices
of exports-the terms of trade. In addition, the difference between the
value of a unit of exports and the resources required to produce it have
to be examined.
One might think that after five decades of tariff cuts and two decades
of reform of nontariff barriers, trade restrictions would be so low as to
hardly matter. But although tariffs of 30 percent are less distortionary
than tariffs of 60 percent, trade barriers are still high enough to impart a
significant bias to international trade. Preferential reductions in these
tariffs-through RIAs-will further distort trade and probably impose
costs; even a discriminatory tariff of 5 percent can have significant effects
on import sourcing if goods are highly substitutable. If discrimination
extends to areas such as standards testing and enforcement, or to public
procurement, trade diversion is likely to be even more significant.
The degree of discrimination arising from an RIA is related to the
height of the "normal," nonpreferential trade barrier-the most-
favored-nation (MFN) barrier-and to the proportion of trading part-
ners or of trade covered by MFN status.' Only about 14 percent of EU
members' imports pay tariffs.2 These imports, which are clearly discrim-
inated against, include some goods from developing countries outside the
ACP group and from non-European industrial countries. Even for manu-



TRADE BLOCS AND THE REST OF THE WORLD                211
factures, the least restricted sector, tariffs range up to 22 percent for
motor vehicles, 18.2 percent for footwear, and 13.4 for clothing. Tariff
peaks facing nonpreferred exporters are even higher in the United
States-up to 25 percent for motor vehicles, 57 percent for footwear, and
35.3 percent for apparel-and in Japan (10.2, 48.8, and 17.8 percent, re-
spectively, for these goods).
Developing country tariffs also remain high enough to distort the
tradeoff between home and imported supplies and between preferred
and nonpreferred sources of imports. Table 8.1 presents average and
maximum     tariffs for selected developing countries and the common ex-
ternal tariffs of some current or planned customs unions. These data
show significant discrimination against suppliers who do not receive ex-
emptions.
Table 8.1 Tariff Averages and Peaks in Selected Developing Countries
and Customs Unions
(percent)
Country or customs union and date of tariff data  Average              Maximum
Brazil (1997)                                    12 2                      70
Mexico (1997)                                    13.4                     260
Venezuela (1997)                                 12.0                      35
Kenya (1994)                                     35.1                      62
Senegal (1996)                                   27.0                      124
India (1997)                                     30.1                      260
Indonesia (1996)                                 13.0                      200
Thailand (1996)                                  22.8                      100
Andean Pact (CET)                                12.8a
MERCOSUR (GET, 2006)                             12.0                      63
SACU (1997)                                       8.7                      78
UDEAC (CET, 1995)                                18.6                      30
UEMOA (1996)                                     13.2b                     -
- Not available.
Note: CET, common external tariff; MERCOSUR, Common Market of the South; SACU, Southern African Cus-
toms Union; UDEAC, Union douanibre et economique de l'Afrique Centrale (renamed CEMAC); UEMOA,
West African Economic and Monetary Union.
a. Weighted tariff from Echavarria (1998).
b. The detailed CET is still under negotiation. The reported figure is an estimate from International Mone-
tary Fund (IMF) sources.



212    REGIONAL INTEGRATION AND DEVELOPMENT
TRADE DIVERSION AND EXCLUDED COUNTRIES
The cost of trade diversion has been a recurring theme of this book, usu-
ally from the viewpoint of partner countries. Here we examine whether
trade diversion could be large enough to harm excluded countries seri-
ously. Strictly, trade diversion refers only to the social cost incurred by
the importing countries as high-cost supplies displace low-cost ones. But
nonmember countries can also lose welfare if their exports are displaced,
which can happen not only because of policy distortions but also because
member costs have fallen as a result of, say, efficiency-enhancing deep in-
tegration. In this section (and here only), we interpret diversion as in-
cluding these latter effects, although, as argued briefly below, such effects
should not be subject to international control or redress.
Evidence of Trade Diversion
It used to be fashionable to argue that RIAs did not actually raise trade
barriers and caused only very limited trade diversion. ln the most studied
case, that of the EEC and its associated arrangements, trade diversion in
manufactures was generally held to be slight (Balassa 1974; Truman
1975; Winters 1987; Sapir 1992). More recent evidence, however, sug-
gests that diversion can be significant even when regional integration is
accompanied by external liberalization.
Bayoumi and Eichengreen (1997) find that the formation of the EEC
reduced the annual growth of member trade with other developed coun-
tries by 1.7 percentage points, with the main attenuation occurring over
1959-61, just as preferences were starting to bite. Cumulating the de-
cline in growth over 1957-73, and noting that total EEC imports from
the rest of the world were $83.1 billion in 1973, puts lost ROW exports
at $24 billion in that year. The formation of EFTA had similar, if smaller,
effects. Frankel and Wei (1998) find that by 1990 trade diversion had
largely erased the EEC's tendency to trade unusually heavily with the rest
of the world, while Sapir (1997) finds that over 1960-72, "EFTA exports
to the EC suffer[ed] from their nonpreferential status," as did other Eu-
ropean nonmember countries' exports in later periods. Our own analy-
sis, which focuses on 1980-96, finds evidence of trade diversion from the
deepening of the EC and EFTA and possibly from the formation of
NAFTA and MERCOSUR (Yeats 1998; Soloaga and Winters 2001; see
also the discussion in chapter 2).
NAFTA is too recent, and its experience too confounded by the 1994
devaluation of the peso, to allow the firm identification of trade diver-



TRADE BLOCS ANI) THE REST OF THE WORLD   213
sion, but there is indicative evidence at the sectoral level. Mexico in-
creased tariffs on non-NAFTA imports of clothing from 20 to 35 percent
in March 1995, just as it was reducing those on NAFTA imports. U.S; ex-
ports to Mexico increased in value by 47 percent between 1994 and
1996, while those from the rest of the world fell by 66 percent. The ex-
planation for a large part of this change is probably not the tariff per se
but the combination of the devaluation, which made assembly in Mexico
very competitive, and NAFTA's rules of origin, which strongly encour-
aged Mexican manufacturers to use U.S. clothing parts. In the U.S. mar-
ket, imports of clothing and finished textiles from Mexico increased by 91
percent and those from Canada by 93 percent over 1993-96 as these im-
ports were exempted from higher tariffs and from import quotas. Mean-
while, imports from Asia fell. The U.S. International Trade Commission
(USITC 1997) views these changes as evidence of trade diversion.
Loss of Exports to Trade Diversion
Previous chapters have shown that trade diversion is directly harmful to
RIA members that suffer it. Diversion also has an immediate and direct
effect on the exports of the rest of the world: they fall. This is frequently
taken as sufficient evidence of harm, for in the traditions of trade diplo-
macy and the GATT, exports are "good," and imports are "bad." If we
are interested in economic welfare, however, we cannot draw this con-
clusion so readily. Indeed, if everything else, including imports, were held
constant, a reduction in a country's exports would improve its economic
welfare because the goods-or the resources used to produce them-
could be redirected to the domestic market.
In fact, of course, we cannot hold everything else constant: the ROW's
loss of exports reduces its ability to buy imports. The losses its con-
sumers face as they cut back on imports must be balanced against their
gains from consuming the resources that were to be exported. There
are three situations in which these components will not be perfectly off-
setting:
1. Because of market distortions, a dollar spent on imports may con-
fer more welfare than can be produced by diverting to local con-
sumption the resources required to produce a dollar's worth of
exports.
2. Lower demand from within the RIA may drive down the price of
ROW exports, so that the rest of the world loses purchasing
power and welfare because each remaining unit of exports buys
fewer imports. It is the change in the relative prices of ROW ex-



214     REGIONAL INTFFGRATION AND DEVELOPMENT
ports and imports-the terms of trade-that matters, not the loss
of exports per se, and it is this that should be the focus of investi-
gation.3
3. If the loss of exports is nonmarginal, some of the benefits of spe-
cialization itself will be lost.
If a unit of exports generates more welfare than would alternative uses
of the resources taken to produce it, losing exports becomes costly. For
each unit of exports lost, real income will fall by the difference between
the value of the exports-that is, the imports that they buy-and their
value in the domestic economy.4 This will happen if exporting generates
supernormal profits because export markets are imperfectly competitive.
Those profits will be lost on any trade that is diverted and cannot be re-
placed by alternative sales at the same price. Industries with economies
of scale are in a similar position. If the creation of an RIA causes indus-
try in the rest of the world to lose scale economies, the cost of all its out-
put increases, imposing costs on other consumers of its output and
reducing its profit margins. This is the story behind predictions (for ex-
ample, by Haaland and Norman 1992) that EFTA would lose signifi-
cantly from the EU's Single Market Programme.
A major potential wedge between the value and the cost of exports is
export taxes. Under perfect competition, the price of a unit of exports
equals the returns to the factors used in producing it (including entrepre-
neurship), plus any taxes imposed. If the export is lost and not replaced,
only the former is recouped. Explicit export taxes are relatively rare
these days, but implicit taxes abound.5 Most important, import taxes
(tariffs) actually tax exports.6 If an RIA cuts an excluded country's total
exports, this eventually implies a decrease in imports, and if these im-
ports are worth more than they cost (because they pay taxes), welfare is
lost as they fall. Similar arguments apply to other implicit export taxes,
such as excise taxes on inputs, excessive fees for international communi-
cations services, and an overvalued currency. With average tariffs ex-
ceeding 10 percent in most developing countries, real income losses
equivalent to over 10 percent of the value of the diverted trade will be
common. Trade diversion can also cause losses if exports generate posi-
tive externalities. Most relevant, probably, are exports of manufactures,
which many observers believe have spillover effects through their role in
training managers and workers, increasing marketing experience and
reputation within markets, and improving technological know-how in
general. The evidence in favor of such externalities is mixed (Aitken,
Hanson, and Harrison 1997), but if they do exist, exports are worth
more than the revenue they generate, and their loss is socially harmful.



TRADE BLOCS AND THE REST OF THE WORLD     215
A related argument views production and employment as valuable in
themselves and holds that trade diversion reduces them. To be true, this
requires both that the diverted exports cannot be replaced by other ex-
ports and that the resources released as total exports fall cannot be reem-
ployed. Under these circumstances, trade diversion causes losses, but one
questions why the resources cannot be employed elsewhere. Setting aside
transitional unemployment as workers seek new jobs, which is real
enough but limited in duration, it is not clear why aggregate employment
depends on the level of exports. And if it does, is the problem the trade
diversion caused by the RIA, or labor market rigidities in the rest of the
world?
The Effect of Large RIAs on Nonmembers' Terms of Trade
The effect of an RIA on the prices at which ROW firms can sell their
products depends largely on its size. Small RIAs will rarely matter, as
they almost never affect the prices at which trade occurs. But some RlAs,
such as the EU or, potentially, the FTAA, are large enough to affect world
prices. Their behavior has implications for all the players in the market
(positive for buyers if the price falls, and negative for sellers), whether or
not they deal with the RIA itself. The significance of price changes is that
they affect not only marginal trade but also the whole volume of existing
trade (see chapter 2). If a shock means that exporters have to drop their
prices to sell the last 1 percent of exports to member countries, this is a
small misfortune if only that 1 percent carries the lower price. But if ex-
porters cannot discriminate between buyers, they have to drop their price
on all sales-a pure loss on the first 99 percent.
It is more common, perhaps, for goods to be differentiated by place of
production and for different mnarkets to be segmented. Here, exporters
face downward-sloping demand curves in each of their markets: they
have to reduce prices to sell more units. Lower demand from the RIA
market exerts downward pressure on the prices of their sales in the RIA,
leading to exports being reoriented toward other markets and to re-
sources being switched to producing other goods. If the RIA takes a large
proportion of the output of the affected goods, and if the goods account
for a large share of total output, the price reductions could be significant
for suppliers in the rest of the world. The effect will be greatest for these
suppliers' sales in the RIA and next greatest in other markets for the af-
fected exports, but it will be felt in markets for all other goods. Whichev-
er of these applies, factors of production in the exporting country earn
less; that is, their income falls.



216      REGIONAL INTEGRATION AND DEVELOPMENT
Despite being central to the theoretical literature for more than a centu-
ry, the terms-of-trade effects of commercial policies have been almost en-
tirely neglected by empirical economists. They have, however, been identi-
fied by analysts using computable general equilibrium (CGE) models-for
example, Gasiorek, Smith, and Venables (1992) and Scollay and Gilbert
(2001). The latter model many of the proposed RIAs in the Asia-Pacific re-
gion and find that nonmembers typically lose, as shown in table 8.2.
Turning to genuine empirical work exploring actual outcomes, among
20th century publications we cannot identify even one empirical ex post
study of regional integration that focuses on1 price effects.7 Recent re-
search by the World Bank has started to fill this lacuna, as described
next.
Livestock Trade in South America. An example of "large market" ef-
fects with small export volumes is sales of live cattle in South America
(Gupta and Schiff 1997). Cattle are not easily transported within devel-
oping countries and are subject to rigorous veterinary regulations in
most developed countries. Consequently, Latin America represents a nat-
ural regional market. In 1966-68, before the formation of the Andean
Pact in 1969, Peru imported mainly from Argentina but also from
Colombia (table 8.3) and accounted for about 30 percent of Argentine
cattle exports. By 1970, the situation was reversed, with the RIA member
supplier, Colombia, displacing the nonmember, Argentina.
Although quantities do not necessarily indicate welfare effects, these
shifts also led to price changes. Before the formation of the Andean Pact,
Table 8.2 Estimates of Potential Welfare Effects of Selected Asia-Pacific RiAs
(percentage of real GDP)
Welfare effect  Welfare effect  Welfare effect
RIA                             for members   for nonmembers   for world
FTAA                                0.08          -0.02          0.01
APEC preferences                    0 58          -0.12           0 27
Japan, Republic of Korea, and China  0.50         -0.03          0 09
Japan and Chile                    -0.03           0.00          -0 01
Japan and Canada                   -0.02          -0.01         -0 01
AFTA and CER (Australia and New Zealand)  0.44    -0 01           0.00
Note: AFTA, ASEAN Free Trade Area; APEC, Asia-Pacific Economic Cooperation; ASEAN, Association of
Southeast Asian Nations; CER, Closer Economic Relations agreement; FTAA, Free Trade Area of the Ameri-
cas.
Source: Scollay and Gilbert (2001).



TRADE BLOCS AND THE REST OF THE WORLD           217
Table 8.3 Peru's Cattle Imports before and after Formation of the Andean Pact
Item               1966-68 ("before")  1970-72 ("after")
Imports (thousands of metric tons)
From Argentina                     32               1
From Colombia                      7                27
Premium on exports to Peru (percent)
By Argentina                       7                4
By Colombia                       -3                2
Source: Gupta and Schiff (1997).
Argentine exporters' prices were 7 percent higher per unit in Peru than
elsewhere. After that, the premium fell to 4 percent. For Colombia, the
change in premiums on sales to Peru was in the opposite direction. In all,
the loss to Argentina amounted to perhaps $700,000. This is not huge,
even at 1966 prices (unless you are an Argentine cattle exporter), but it
illustrates that even a small RIA can have negative effects on neighbors in
segmented markets.
Brazil and MERCOSUR. Brazil's membership in MERCOSUR has
been accompanied by a significant decline in the relative prices of im-
ports from nonpartner countries (Chang and Winters 2002). Figure 8.1
reports the relative price of Brazil's imports from Argentina versus prices
of imports from the United States, averaged (unweighted) over the 323
products that Brazil imported from both countries in every year during
the period 1990-96. Although both price averages fell, presumably be-
cause of macroeconomic conditions in Brazil, U.S. prices fell by much
more, as MERCOSUR came into operation. Formal econometric esti-
Figure 8.1 Average Prices of Brazil's Imports from Argentina Relative to Prices of Im-
ports from the United States, 323 Commodities, 1990-96
2 0  - -  -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   -   - _
O  O   15  --  -  -  -  -_ 
E    e_  10  -- - - - -
E E     0
r!    _ 5 __ ____       -_   _   ____      __ -i_---
-10   -_ _ _ _- _ _- _ _- _ _- _ _- _ _- _ _- _ _ _ _- _ _- _ _-  
1990  1991  1992  1993  1994  1995   1996
Year
Source: Chang and Winters (2002).



218     REGIONAL INTEGRATION AND DEVELOPMENT
mates suggest that these changes in relative price are substantially the re-
sults of differences in tariffs on the two suppliers. Similar results appear
for comparisons between Argentine and developing country export
prices (for Chile, Korea, and Mexico), although the sets of goods export-
ed by Argentina and each of the other exporters are much smaller, mak-
ing the estimates less precise.
Theory suggests that MERCOSUR will increase Argentine (member)
prices and decrease U.S. (nonmember) prices. The results above are con-
sistent with this view but do not prove it. The researchers therefore also
compared nonmembers' export prices to the Brazilian (MERCOSUR)
market with those to non-MERCOSUR markets. If exporters can charge
different prices in different markets, export prices to non-MERCOSUR
markets (which are numerous and are collectively much more important
than MERCOSUR markets) can be taken as a norm. Changes in the rel-
ative prices of exports to Brazil and elsewhere then reflect factors specif-
ic to Brazil, including MERCOSUR membership. Figure 8.2 plots the
prices of U.S. exports to Brazil and to the non-MERCOSUR world aver-
aged over the set of products exported to both markets in all years. It
clearly demonstrates that whereas non-MERCOSUR prices rose, those
to Brazil declined sharply as MERCOSUR preferences were phased in.
Figure 8.3 reports the same information for Korean exporters.
The figures are startling, but of course, other things may have
happened to Brazil besides its entrance into MERCOSUR. Chang and
Figure 8.2 Average Prices of U.S. Exports to Brazil Relative to Prices of U.S. Exports to
the Rest of the World, 1,356 Commodities, 1991-96
15 -- - - - - - - - - - - - - - - - - -
~a0
flc
o o., 0
-10-   ___      ___ 
-15 -_--_--_--_--_____-__-__-__--__-__-__-__- 
1991   1992   1993   1994   1995   1996
_..To Brazil -4To Rest of World
Source: Chang and Winters (2002).



TRADE BLOCS AND THE REST OF THE WORLD       219
Figure 8.3 Average Prices of Exports from the Republic of Korea to Brazil Relative to
Prices of Korea's Exports to the Rest of the World, 99 Commodities, 1990-96
10
5 
EC -5      ---          ---- -      -A- - - -
10
2)
1990     1 0   13 3   1'38  199   19395  19936
-+ To Brazil *  To Rest of World
Source: Chang and Winters (2002).
Winters find, however, that even after allowing for costs, exchange rates,
inflation, and MFN tariffs, the relative prices of nonmember exports to
Brazil and elsewhere are significantly related to Brazil's tariffs on MER-
COSUR members. For products that Argentina also exports to Brazil,
U.S. suppliers appeared to reduce their export prices to Brazil relative to
non-MERCOSUR markets by slightly over one-fifth of the reduction in
internal tariffs, and other exporters reduced their prices a bit more.
Schiff and Chang (forthcoming) find similar but stronger price effects.
Using the same data, they distinguish between products that Argentina
exports to Brazil and those it does not. For the former, they confirm
Chang and Winters' results qualitatively, using somewhat different crite-
ria to define market presence. For the latter-goods that Argentina does
not currently supply-they distinguish between goods that Argentina
does not export at all and goods that it exports to markets other than
Brazil and for which Argentina thus represents "potential competition"
to U.S. exporters in Brazil's marlcet. In the first case MERCOSUR has no
effect on U.S. export prices, whereas in the second Schiff and Chang find
results similar to those of Chang and Winters. It is worth noting that the
price changes recorded here benefit Brazil and, almost certainly, Argenti-
na at the expense of excluded countries. Whether the gains outweigh the
losses globally depends on details of the model that the papers do not ex-
plore directly-specifically, the nature of costs and competition. The the-



220        REGIONAL INTEGRATION AND DEVELOPMENT
oretical model used to motivate the empirical estimates would suggest
overall gains, but other underpinnings for the same estimates suggest the
contrary. In other words, even the threat of entry by the preferred part-
ner imposes costs on nonmembers. Both studies find that Brazil's MFN
liberalization raises relative prices of excluded countries in Brazil's mar-
ket. Chang and Winters, however, find these effects much weaker than
the adverse preferential ones.
EFTA and the EC. Finally, box 8.1 illustrates, using the case of EFTA
and the EC, the main problems that small economies can have if their
principal market joins or deepens an RIA.
One of the main fears of excluded countries is that a newly formled or
deepened RIA will absorb investment that they might otherwise have re-
ceived. Regional integration frequently leads to temporary investment
booms within an RIA, and if there is a finite set of investment opportu-
Box 8.1 Trade Diversion and Investment Switching in Europe
In the late 1980s the five continental EFTA countries were heavily dependent on the EC market. On
average, they exported 35 percent of their GDP, and 53 percent of their exports went to the EC. Thus,
roughly 18 percent of their income was generated directly from sales to the EC.
The EC's Single Market Programme (SMP) was designed to integrate EC markets and, by increasing
competition and exploiting economies of scale, raise the efficiency of EC producers. This was a direct
threat to EFTA producers, especially in the sectors where the EC market was most fragmented and
imperfectly competitive and so offered the most room for improvement in the skill-intensive engi-
neering sectors. Because these sectors were also imperfectly competitive in EFTA, they were the
ones that generated the largest labor rents and profits in EFTA members. Estimates (probably rather
conservative) suggested that EFTA output in some of these sectors would fall by over 5 percent and
gross domestic product (GDP) by nearly 0.5 percent as the SMP was implemented (Haaland and Nor-
man 1992).
Shifts in investment around this time made these predictions look only too plausible. More or less as
the SMP was announced, every EFTA country suffered a decline in FDI inflows and an increase in
outflows. These changes reversed themselves only as countries embraced the European Economic
Area (EEA) or full accession to the EC (Baldwin, Forslid, and Haaland 1996).
Although analysts reckoned that many excluded countries would benefit from the SMP as EC de-
mand for their exports increased and EC export prices fell (Gasiorek, Smith, and Venables 1992), in
EFTA these benefits were swamped by production losses. The solution for the EFTA countries was to
join the SMP so that they too could benefit from the large, more efficient market. It was predicted
that this move would turn the GDP loss of 0.5 percent into a gain of nearly 3 percent. Because the
countries did join the EEA and eventually sought full accession, the costs of their being excluded
could not be observed. Yet the fact that they did move closer to the EC, and the observed shifts in in-
vestment flows, suggest that the threats were perceived to be serious (Haaland and Norman 1992).



TRADE BLOCS ANI) THE REST OF THE WORLI)  221
nities or a finite volume of investable funds, it is not unreasonable to fear
some switching between recipients. We have not, however, found hard
evidence that investment or FDI inflows have fallen in excluded countries
as a result of RIA creation.
The effect of the EC's Single Market Programme on EFTA provides
probably the strongest test of losses attributable to investment switching.
The expected shock was large, and since investment is forward looking,
actual investment is likely to respond to these expectations. FDI fell in
every EFTA country after the SMP was announced and recovered only as
these countries signed onto the EEA or committed to accession to the EU
(Baldwin, Forslid, and Haaland 1996). Unfortunately, although the
study by Baldwin, Forslid, and Haaland is suggestive, it offered no for-
mal ex post analysis of the situation, and so it is possible that the de-
crease in FDI reflected cyclical factors rather than a structural effect.
THE ROAD TO MULTILATERALISM: ARE RIAs STEPPING
STONES, OR MILLSTONES?
In this section, we shift the focus from the immediate consequences of re-
gionalism for nonmember welfare to whether regionalism sets up forces
that encourage or discourage evolution toward globally freer trade. Given
the difficulties in understanding the dynamics of any reform, it is hardly
surprising that there are as yet no categorical answers (see box 8.2). Chap-
ter 3 showed that RIAs can be set up in ways that promote liberalism in-
ternally. The systemic aspects considered here are less optimistically bal-
anced. Most of the plausible systemic arguments suggest that regionalism
will tend to undermine multilateralism, and those authors who do not
come to this conclusion nevertheless tend toward the implication that even
if regionalism helps reduce levels of protection, it does so in ways that in-
crease the chances of fragmenting the world economy into warring blocs.
The stakes in the bet as to whether RIAs are stepping stones to multi-
lateral trade liberalization or millstones around its neck are huge.' Open-
ing trade and increasing competition have been part of virtually every
sustained economic growth experience, and the unprecedented postwar
growth of world output and income was clearly allied to these factors.
Moreover, systemic effects affect everybody, even bystanders. This is par-
ticularly important for small- andl medium-size economies, which depend
heavily on international trade and are the principal beneficiaries of an
orderly and nondiscriminatory trading regime.



222        REGIONAI INTEGRATION AND DEVELOPMENT
8.2 Stepping Stones, or Millstones? A Summary
At first blush, regionalism looks like an important step on the road to global free trade. The recent ex-
plosion of RIAs (see figure 1.1, in chapter 1) has, after all, gone hand-in-hand with the greatest mul-
tilateral trade liberalization in history. But despite the large number of RlAs in the world, most are too
recent, too incomplete, or too idiosyncratic to support conclusions about a world of many RIAs, and
there is little empirical evidence about the effects of regionalism on the world trading system.
Even though regionalism and nondiscriminatory trade liberalizatfon have progressed together over
the past decade, especially in Latin America, the little evidence we have suggests that multilateral
opening has led the way and trade blocs have followed. Many countries reformed their trade well be-
fore they joined RlAs: Argentina, Brazil, Mexico, and Turkey are salient examples (Foroutan 1998).
Many members of RIAs in Africa, and some in Lafin America, have not reformed. And many countries
have reformed without joining RlAs, among them Chile in the 1970s and 1980s, Indonesia, and
Korea.
There is also some question as to the degree of openness that trade blocs foster. Relatively few RIAs
advance significantly beyond the WTO obligations their members have taken on in difficult areas
such as agriculture, sernices, and controlling anfidumping actions.
The perception that the existence of trade blocs makes global liberalization easier to achieve be-
cause It reduces the number of negotiating parties is also open to question. Blocs may well find it
just as difficult to achieve internal agreement, and their combined size will make it easier for them to
resist global pressures to liberalize, as well as give them the power to force others to liberalize. We
need only recall the difficulty the EU had in negofiating in previous GATT rounds and its long obsti-
nacy about opening up its agricultural markets.
Although economists have built many theoretical models of RIAs and
the world trading system, few have generated testable predictions about
observable phenomena, and still less about measurable ones. Moreover,
even if there were testable predictions, there is very little to test them on.
Few RIAs have been sufficiently effective and long-lived to warrant
drawing empirical conclusions from their experience, and a world of
multiple RIAs is a recent development. This has two implications. First,
although we need to make all the use of empirical information that we
can, ultimately we have to rely heavily on a priori reasoning. Second, the
question about regionalism and multilateral trade concerns the future,
not the past. The four decades ending in 1986 saw remarkable progress
in liberalizing trade in industrial goods multilaterally, but there was little
effective regionalism. In the past decade and a half, in addition to con-
tinuing progress on industrial goods, a start has been made on agricul-
ture and services, and a huge reversal has occurred in developing country
rhetoric about and practice of trade policy. This has been accompanied
by a growth of regionalism but, at least in a temporal sense, could not
have been caused by it.' The question for this section is precisely what



TRADE BLOCS AND THE REST OF THE WORLD    223
this burst of regionalism, and any continuation of it, imply for future
multilateral trade liberalization.'0
Regionalism affects the progress of multilateral trade liberalization by
several routes: by altering the internal incentives for trade liberalization,
by affecting the way in which RIA members interact, and by changing
the interactions between RIAs and the rest of the world. Chapter 3 dealt
with the first two, noting, among other things, the danger of producer
lobbies driving RIAs toward trade diversion and the beneficial dynamic
that trade-creating RIAs might have in reducing protection. Here we ex-
amine the interactions between members and nonmembers, discussing
specifically domino regionalism, multilateral trade negotiations, and
"open regionalism."
Essentially, there are three arguments. The first is that the behavior of
the major blocs-the EU, NAFTA, and APEC-will affect the multilater-
al system in ways that are basically exogenous for developing countries.
All that the latter can do is to seek to influence that behavior, perhaps via
international institutions, and to prepare for its consequences. The sec-
ond is that RIAs (including those involving developing countries) may af-
fect the behavior of developing countries themselves, altering their own
propensities for nondiscriminatory liberalization and their willingness to
support and protect the multilateral system. If enough developing coun-
tries are affected, this too will have systemic effects, in which case devel-
oping countries must themselves take some direct responsibility for the
consequences of regionalism. The third is that regionalism could affect
the processes of multilateralism as well as the outcomes (see "Negotiat-
ing Power of RIAs," below).
It might seem that asking whether regionalism could reverse multilat-
eral trade liberalization is merelv an intellectual indulgence. After all, do
not the WTO rules governing RIAs expressly forbid RIAs to increase
trade barriers? But this view is too simplistic, for several reasons.
For many developing countries, there is a wide gulf between their ac-
tual (applied) tariffs and the maxima committed to in their formal bind-
ings in the WTO. For example, when Mexico nearly doubled tariffs on
503 import items from non-NAFTA sources in 1995, it did so without vi-
olating any bindings. WTO rules are ambiguous and poorly enforced. A
determined government can make trade policy more restrictive in ways
more or less immune to WTO disciplines-say, through antidumping ac-
tions or health regulations. In a world edging toward general trade liber-
alization, any deceleration in that process is equivalent to an increase in
protection relative to the original path.



224     REGIONAL INTEGRATION AND DEVELOPMENT
Multilateralism as a Process
This chapter is mainly concerned with outcomes: progress toward multi-
lateral free trade. But one also needs to ask about multilateral processes
and behavior-the "ism" part. "Multilateralism" is a much-used but lit-
tle-defined, term. In one attempt at precision, Ruggie (1992) sees it as a
deep organizing principle of postwar international life, with three defin-
ing characteristics:
*   Indivisibility-the system is a whole in that the actions of one
party affect all parties and each party acknowledges its allegiance
to the whole.
*   Generalized rules of conduct-interactions between parties are
governed by widely recognized general principles rather than ad
hoc or particularistic interests.
*   Diffuse reciprocity-all parties expect to gain from the system but
do not demand precise reciprocity in every separate transaction.
The translation of these principles into concrete form in the interna-
tional trading system is straightforward. The system is indivisible in that
it permits an extremely dense and far-reaching network of trade links
and intergovernmental contacts, and it is viewed as having an existence
separate from all the individual trade links between participants. More-
over, its separate existence is seen as valuable. The trading system's most
obvious generalized norm is nondiscrimination, one element of which
(MFN) immediately and automatically extends bilateral agreements to
all members. Reciprocity is diffuse in that governments do accept indi-
vidual actions that appear nor to be in their immediate interests, but it
is generally accepted that, overall, every country has to gain from the
system.
Under these criteria, regionalism clearly undermines multilateralism,
as it defies MFN. It is also divisive, since it exacerbates tendencies for
parties to focus more strongly on some links than on others. One argu-
ment in defense of regionalism is that it is subject to generalized rules
that all have agreed and that indivisibility should not preclude some links
being stronger than others. The first point merely recognizes that the
world trading system is imperfectly multilateral; the second, in practice,
depends on whether governments shift their focus from general to partic-
ular trade relations. Since RIAs certainly shift the focus to some degree,
regionalism does corrode the multilateralism of the global trading sys-
tem, if only mildly at present.
Because the multilateralism of the trading system is part of a broader
multilateral order encompassing areas such as security, money, and the



TRADE BLOCS AND THE REST OF THE WORLD    225
environment, it is conceivable that regionalism could have a wider effect
on multilateralism. At present this does not seem to be a problem, but if
trading frictions between, or about, trading blocs grow more strident,
more than just trade could be at stake, as was seen in the breakup of the
EAC in the 1970s.
Negotiating Power of RIAs
Many RIAs explicitly aim to increase the negotiating power of their
members on the world scene. In customs unions, which have a common
trade policy, the scope for enhancing negotiating power by coordinating
the positions of several countries is obvious. Coordination in trade nego-
tiations is open to any set of countries, of course, but having a formal
RIA makes it easier and more credible. We argue that this coordination
could slow multilateral liberalization or even lead to increases in protec-
tion.
In 1991 Paul Krugman famously suggested that the worst number of
(equal-sized) RIAs for world welfare was three. With fewer, larger blocs,
more trade is tariff free (which is beneficial). There is, however, more
trade diversion, and blocs have greater bargaining power and so raise
their tariffs against each other (both of which are harmful). Obviously, a
single bloc is best because that means global free trade, but for small
numbers of blocs, diversion and competitive tariff setting outweigh the
benefits of intrabloc free trade.
Krugman's result has turned out to be very fragile. He analyzed only
customs unions, but most RIAs are FTAs, in which members maintain
their own external tariff regimes. FTA members have two good reasons
for lowering their tariffs on nonmembers as the FTA expands: to reduce
trade diversion, and to compete to capture tariff revenue and boost their
competitiveness by reducing tariffs relative to their partners (see the dis-
cussion in chapter 3).
Furthermore, countries interact more or less continuously on trade is-
sues, and each interaction influences those that come after it. This opens
the possibility that current cooperation can be maintained by the threat
of future punishment. In this context, a trade agreement is sustainable if,
for each party, the value of the stream of benefits expected to arise from
keeping the agreement exceeds that of the stream arising from breaking
it. Breaking the agreement would entail choosing a policy that maximizes
immediate welfare but with a period of punishment ensuing. The trade-
off depends on three sets of factors:



226     REGIONAL INTEGRATION AND DEVELOPMENT
*   The rate at which the future is discounted. Cheating (raising one's
tariff) means high immediate payoffs followed by lower payoffs
during the punishment phase. The higher the rate of discount
(that is, the lower the relative weight placed on the future), the
more likely a country is to cheat.
*   The probabilities that cheating will be punished (and how, and for
how long) and that cooperation will be rewarded (that the other
party will not cheat and that the agreement will not be overturned
by some exogenous shock).
*   Trading arrangements, such as RIAs, that affect the volumes and pat-
terns of trade. The lower is the tariff in a trade agreement, the greater
are the (immediate) benefits of cheating (raising one's own tariffs uni-
laterally) and the stronger (more costly) is the punishment needed to
make the agreement sustainable. Stronger punishment allows more
cooperative behavior-that is, lower tariffs in the agreement.
This framework allows us to ask directly and simply whether, by
changing the various incentives, regionalism fosters lower or higher tar-
iffs on the rest of the world.
Regionalism and Tariff Levels
Bond and Syropoulos (1996) suggest that, starting from worldwide free
trade, introducing RIAs and allowing them to expand creates two coun-
tervailing forces: the incentive to cheat grows, but so does the welfare
loss in the resulting trade war. In their model, the former dominates, with
the result that it is more difficult to maintain free trade in a bloc-ridden
world and that the minimum tariff which can be supported by this kind
of cooperation increases as bloc size increases. In short, regionalism in-
creases the pressures for protectionism.
The discount rate is crucial in the operation of these so-called trigger
strategies because it trades off the immediate benefits of cheating against
the eventual costs of a trade war. This raises the question of the time scale
over which these games are played. In terms of individual tariffs and tar-
iff wars, the period required for retaliation is so short that there are hard-
ly any gains to cheating. Thus, discipline seems virtually complete, and
the model suggests that nothing much affects the cooperative outcome.
It is, however, more plausible to view regimes as the instruments, the
GATT rounds as the natural periodicity, and policies such as the zeal with
which antidumping policies are applied and the use of health and techni-
cal regulations as the weapons. Then the periods required to recognize de-
fection and retaliate become much longer. The important effect of region-
alism is not on the "tactics" of trade policy but on the "strategy." RIAs



TRADE BLOCS AND THE REST OF THE WORLD                    227
Box 8.3 Regionalism and Protection
Theory cannot determine whether RIAs will increase their protection against nonmembers. Foroutan
(1998) attempts to settle the matter through direct empirical observation of the behavior of blocs.
While most empirical work necessarily considers particular case studies, her study takes a cross-
developing country approach.
Because nearly every country is In or is discussing an RIA, it is difficult to devise a comparator group
for members of RlAs. Foroutan exploits the fact that in the past not every country was a member, and
not every bloc affected trade significantly. She classifies RiAs according to whether they are 'effec-
tive," defined as having a material effect on the share of intrabloc trade in the group's total trade.
Comparing members of effective and ineffective blocs sheds some light on the consequences of
RIAs for external trade policy.
The attribution is necessarily rough and ready, but effective RIAs among developing countries (up to
1995) are defined as including the CACM (1960-75 and again since 1990), the Andean Pact (since
1990), MERCOSUR, UEMOA, and SACU. Among the Individual countries affected by their RIA mem-
berships are Cameroon, Israel, Kenya, Mexico, and Zimbabwe. Foroutan then compares these devel-
oping countries with those that were not In effective RIAs, using three dimensions: average applied
tariffs and nontariff barrier (NTB) coverage; Uruguay Round concessions; and openness. The sam-
ples differ across the three exercises because of data availability; the table below shows results for
average tariffs and for openness.
Average applied tariffs and NTB coverage
The Latin American RIAs now have some of the lowest average tariffs and NTB coverage among de-
veloping countries and have achieved the greatest liberalization of any group since the mid-1 980s.
Except for Chile, the small "non-RIA" group has made much less progress. In Africa, until 1994, nei-
ther RIA nor non-RIA countries had displayed much tariff liberalizabon (NTB data are not available);
the latest available data yield a mean average tariff that is almost the same for the two groups.
South Asia has liberalized but remains highly protected, East Asia has always been relatively liberal
but now has higher average tariffs than reforming Latin America, and the Middle East and North
Africa region shows no reform and fairly high average protection. The most liberal group in the study
is that consisting of members of North-South RIAs: Israel, Mexico, and Turkey.
Uruguay Round concessions
Here, the only feasible comparison is between the Labn America RIA group and all non-RIA coun-
tries. The RIA group cut its bound tariff by more and bound more of its tariffs in the Uruguay Round
than did the non-RIA group, but it also completed the round with significantly higher bound tariffs.
Openness
The measure of openness used was (nonfuel imports + exports)/GDP, and the non-RIA group dis-
played the greatest average increases in openness between 1980-84 and 1990-94. The Latin
America-RIA countries show some increase, but not to the levels achieved in the 1970s.
Strong trade reform appears to be mostly associated with RIA membership. Is it possible to conclude
that trade reform is a result of RIA membership? There are five reasons for withholding that judg-
ment at present.
1. A more detailed examinafion shows that much of the trade reform preceded RIA membership,
as happened in Argentina, Brazil, Mexico, and Turkey.
2. At the country level, many RIA members have not reformed (examples are members of the
CACM in its early period and African RIA members), whereas plenty of non-RIA countries, including
(Continued on next page.)



228        REGIONAI INTEGRATION AND DEVELOPMENT
Box 8.3 (continued)
Chile, Indonesia, and Korea, have done so. General tendencies notwithstanding, reform clearly
involves much more than RIA membership alone.
3. Many of our hypotheses about the effects of RIAs on protection operate only over fairly long
time periods, whereas the data cover mostly rather recent integration.
4. The general results depend very heavily on the Latin American experience. But since other
forces may well have been at work in that region, we must be cautious about attributing too much to
regionalism.
5. The results on actual openness tend in the opposite direction from those on policies.
Foroutan's results certainly refute the simple hypothesis that RIAs necessarily and immediately lead
to protectionism, and they are consistent with the idea that regionalism helps lock in previous MFN
liberalization. As with so much of this debate, the jury is not so much "still out" as "still listening"
while sufficient evidence is accumulated. Better than waiting for the evidence to come in, however,
would be for RIAs to arrange their policies and institutions to ensure that they actually deliver their
liberal promise rather than the opposite.
Average Tariffs in Developing Country RIA Members and Other Countries
(percent)
Average tariff                     Openness,
Non-RIA                           Non-RIA
RIA Countries     Countries       RIA Countries     Countries
Group          1981-   1990-   1981-    1990-   1980-    1990-   1980-    1990-
85      94       85      94       84      94       84      94
Countries in North-   20.6     10.2                      19       22
South RIAs
Latin America         31.0     13.8    25.8    15.4      26      30       174     107
Sub-Saharan Africa    35.6     31.0    30.7    24.4      47      45       40      48
South Asia                             59.9    41 8                       22      30
East Asia                              20.5    15.0                       43      68
Middle East and North                  26.3    26.6                       40      43
Africa
Note: Blanks denote not applicable.
l (Exports + nonfuel imports)/GDR
Source: Foroutan (1998).
will tend to reduce the incentive to take a world view of the broad trends
of trade policy because intrabloc trade comes to be seen as a substitute for
trade with the rest of the world. The danger is that countries in RIAs will
be less willing to sustain liberal regimes. Box 8.3, which explores the issue
empirically, suggests not that regionalism is accompanied by an immedi-
ate descent into protectionism or a retreat from it but, rather, that multi-
lateral liberalization has tended to lead regional liberalization.
Bond, Syropoulos, and Winters (2001) use a similar framework to ex-
amine the deepening of an existing regional arrangement. They observe



TRADE BLOCS AND THE REST OF THE WORLD    229
that as the bloc deepens, its trade with the rest of the world tends to de-
cline. The excluded countries then find the original agreement unattrac-
tive and will initiate tariff increases unless the bloc reduces its tariffs. The
bloc will almost certainly prefer lowering its external tariff to getting into
a trade war, and so a new, lower tariff equilibrium is usually feasible. It is
important to note, however, that at this new equilibrium the rest of the
world could be worse off than it was before the deepening. It cannot pre-
vent the bloc from deepening its integration, and even though it finds the
new equilibrium the best alternative from among the new set of feasible
outcomes, this does not imply anything relative to the starting point.
Has Regionalism Spurred Multilateral Negotiations?
Many commentators argue that the creation of the EEC in 1957 led to
these kinds of endogenous tariff adjustment. Thus, for example, they
suggest that the formation of the EEC led directly to the Dillon and
Kennedy Rounds of GATT negotiations as the United States sought to
mitigate the EEC's potential for diverting trade (Lawrence 1991; Sapir
1993; WTO 1995). Although this scenario is perfectly conceivable, the
argument is not straightforward.
1. It seems unlikely that multilateral negotiations would have ceased
completely had the EEC not been created, especially given the
global reach of the United States during the 1 960s. Thus, at most,
the EEC affected the timing and extent, but not the existence, of
the rounds.
2. Agriculture played an important role in the formation of the EEC,
and the EEC was probably more successful in resisting that sec-
tor's liberalization in the multilateral trade negotiations than its
members would have been individually. As a result, future liberal-
ization probably became more, not less, difficult.
3. Suppose that the hypothesis were true and that the creation of the
EEC did lead to negotiations. The logic of the argument is essen-
tially coercive: EEC members did something that their trading
partners considered harmful and then offered to mitigate it in re-
turn for concessions. Coercion may be warranted, and the out-
come may have been beneficial, but this is a dangerous game. It
depends critically on the willingness of the partners to fold, by ne-
gotiating, rather than to fight by raising tariffs, and to respond
multilaterally rather than regionally. In economists' models such
as we have just explored, we can work out the incentive to fold,
but in the real world it is not so easy.



230    REGIONAL INTEGRATION AND DEVELOPMENT
4. Even if coercion worked for the EEC, it probably would not for
smaller RIAs of developing countries.
It has also been argued that regionalism was behind the Tokyo Round.
Winham (1986) reports that both the first EEC enlargement (including
free trade with EFTA) and the restrictiveness of the CAP were factors in
the U.S. desire for a round. The former observation seems no more com-
pelling than those about the creation of the EEC, while the latter is dis-
tinctly two-edged: to attribute the Tokyo Round to regionalism requires,
first, that the CAP induced negotiations and, second, that regionalism in-
creased trade restrictions in agriculture. For this to be advantageous in
its net effect on multilateral progress again requires a negotiating struc-
ture in which might and countervailing power are the critical forces be-
hind liberalization.
Finally, consider the Uruguay Round. The WTO (1995) says, "there is
little doubt that . . . the spread of regionalism [was a] major factor in
eliciting the concessions needed to conclude" the round. There was, in-
deed, a perception that the failure of the round would lead to regional
fragmentation. This almost certainly encouraged the spread of "defen-
sive" regionalism during the early 1990s, but whether that trend pres-
sured the two major parties in the round to agree is not clear. After all,
they were the prime "regionalists," and they would certainly not have
been the principal casualties of fragmentation. Some senior EU nego-
tiators have said that the 1993 Seattle APEC summit induced the EU
finally to concede on agriculture and to conclude the Uruguay Round
(Bergsten 1997). Again, this may be true, but there are strong counter-
arguments. For example, APEC was not advertised as a discriminatory
RIA, and any discrimination would have been far in the future. Further-
more, the principal necessary condition for the EU to complete the round
was agricultural reform, which was initiated in 1990 and completed in
1992 (Hathaway and Ingco 1996).
These arguments do not inform developing country trade policy di-
rectly: no single developing country exerts enough leverage to affect
global trade talks. Developing countries, however, have a vital interest in
the world trading system and may be able to influence developed country
behavior indirectly via diplomacy or WTO rules-especially if they act
together. One alarming possibility is that regionalism might undermine
U.S. or EU willingness to participate actively in the multilateral system."
Over the past three decades the United States and the EU have been
major players, monitoring both smaller countries' policies and each
other's. A loss of interest by either would upset this delicate balance and
reduce the WTO's overall effectiveness.



TRADE BLOCS AND THE REST OF THE WORLD   231
Domino Regionalism
Above, we implicitly assumed that nonmember countries could respond
to an RIA only through MFN negotiations. A second response, however,
is to join the RIA or create a new one-what has been termed "domino
regionalism" (Baldwin 1995, 1997). The idea is that one act of regional
integration can stimulate the next because the larger a bloc is, the greater
the costs to excluded countries of not belonging to it.12 Baldwin (1995)
coined the phrase to describe the process by which, after three decades of
resistance, three Scandinavian countries decided in the late 1980s to seek
EU membership. Although these countries were still uncomfortable with
the EU politically, the economic pressures from the Single Market Pro-
gramme were overwhelming, and as one Scandinavian country joined,
the pressures on the next increased. Similarly, Canada sought to turn
U.S.-Mexican trade talks into an expansion of CUSFTA, which eventual-
ly became NAFTA; several Latin American and Caribbean countries
later sought accession to NAFTA; Bolivia and Chile have FTA associa-
tions with MERCOSUR; Mediterranean and Eastern European countries
are racing to conclude association agreements with the EU; and a num-
ber of late entrants are seeking membership in the Cross-Border Initiative
in Africa. And when multilateral progress loses momentum, domino
regionalism receives a further boost. For example, in the wake of the
largely sterile Seattle WTO ministerial conference in 1999, Singapore, a
former paragon of nondiscrimination, sought agreements with New
Zealand and the United States.
The spread of regionalism is not evidence of its virtue. In a regional-
ized world a country may be better off inside than outside an RIA, but
this tells us nothing about whether it prefers a regionalized to a nonre-
gionalized world. To give a graphic illustration, if there is gang warfare
in your neighborhood, it may be best to belong to a gang but that does
not make gangs a good thing.
A further problem with the view that domino effects necessarily render
regionalism benign is that it takes two to tango. Even if excluded coun-
tries seek access to an RIA, the existing members may not wish to let them
in, at least not without significant down payments (see Andriamananjara
and Schiff 2001). RIAs tend to turn the terms of trade against nonmem-
bers, so the optimal bloc size, looking from the inside out, is smaller than
the whole world: there needs to be somebody outside to exploit.
Even voluntary regionalism can make everyone worse off. A simple il-
lustrative model of such effects is shown in figure 8.4. Frankel, Stein, and
Wei (1997) divide a world of many countries into four continents, with



232     REGIONAI. INTEGRATION AND DEVELOPMENT
Figure 8.4 Domino Regionalism: Changes in Welfare as RIAs Form
2 -
1     /     
2                         0\                         World
--     Inside
the RIA
co -2                                                Outside
3           /             the RIA
-4-
0        1       2        3        4
Number of Blocks
Source: Frankel, Stein, and Wel (1997).
zero trading costs between countries within the same continent and positive
costs between continents. At the start, each continent has an MFN trade
policy. Any one continent can then improve its welfare by forming an FTA:
overseas producers will have to lower their prices to mitigate their loss of
competitiveness and will, as a result, suffer either a decline in income or the
loss of part of the FTA market. Next, a second continent benefits by creat-
ing an RIA, converting a loss of welfare into gain, and then the third does
the same, converting larger losses into smaller ones. Even the fourth conti-
nent gains by creating an RIA, although by then all continents are worse off
than under MFN policies. World welfare falls at every stage, but no conti-
nent has the incentive to undo the regionalism individually."3
Could such processes lead all the way to global free trade? Almost cer-
tainly not, because insiders benefit by turning the terms of trade against
outsiders and so seek to prevent unlimited expansion of their blocs. If
several blocs of roughly equal size formed, it is possible (but not certain)
that they would subsequently negotiate with each other to achieve glob-
al free trade. In most models, however, it is more likely that multiple
blocs will be of different sizes and that the final steps toward free trade
will be vetoed by the larger blocs (Nordstrom 1995; Campa and Soren-
son 1996). Only if RIAs were somehow obliged to accept any country
that wished to join would evolution to global free trade be likely, and
even then, in the course of achieving it, countries might suffer lower wel-
fare than they had under MFN tariffs. Thus, the speed and security of the
convergence to global free trade would be critical considerations to take
into account in advocating such "open access," even if it were feasible-
which it is not (see box 8.4).14 Insider resistance to expansion certainly



TRADE BLOCS AND THE REST OF THE WORLD                   233
Box 8.4 Is Open Access the Key to Benign Regionalism?
Both in theoretical models (for example, Yi 1996) and in policy discussions (Bhagwati 1991; Serra
and others 1997), and as long ago as the preparatory meetings for the London Monetary and Eco-
nomic Conference of 1933, it has been suggested that the key to ensuring that RIAs lead to multilat-
eralism is "open access."That is, any country willing to abide by an RIA's rules should be guaranteed
admission. To date no RIA has offered such unconditional access. Most have restricted geographic
domains, and even within these domains it is existing members who determine whether applicants
meet the conditions. A simple rule of, say, internal free trade or "national treatment for Investors"
may be objectively assessable, although even then the transition period needs to be negotiated, but
anything more-dispute settlement, excluded sectors, a common antidumping policy, and so
on-certainly requires negotiation and threatens candidates with delay and veto.
Where an RIA involves few conditions (as is the case with the Cross-Border Initiative in eastern and
southern Africa, which does not rigorously enforce either internal preferences or external tariff har-
monization), there is little incentive to exclude geographically eligible countries. Indeed, there is per-
haps some incentive to include them, in an attempt to enhance the bloc's negotiating power in world
trade talks and vis-3-vis international institutions. The result is expansion to include a wide range of
different economies.
NAFTA, which has as its principal rules completely free trade in goods and open investment, might
on that account have automatic entry, but it does not Accession is not restricted geographically but
is "subject to terms and conditions as may be agreed . . ." This resistance, however, probably has
more to do with avoiding adjustment in the United States than with the need to negotiate issues such
as quotas for professional migration or dispute settlement. MERCOSUR is happy to accept new
members, but because fairly deep integrabon is planned, detailed negotiation is required. Associa-
tion with MERCOSUR, the status chosen by Chile, is easier but does not offer full integration, and
even it required several years of talks.
The EU stands ready to sign association agreements with many neighboring countries and with all of
the middle-income members of the Africa, Caribbean, and Pacific group-but only on its own terms
concerning such issues as rules of origin, excluded sectors, and the use of antidumping duties. Full
membership in the EU is anything but open access. The United Kingdom had to apply three times; it
will take Turkey at least 30 years to get in; and there is no bmetable for countries such as Ukraine
and Georgia. Negotiations are tortuous even when accession is agreed in principle. The EU's White
Paper on Eastern European Accession is hundreds of pages long; each candidate faces a formidable
list of demands and requirements prior to membership and is subject to annual reports on its
progress. In several cases these countries are required to adopt policies from which some existing
members are exempt-for example, the social chapter of the Maastricht Treaty.
In all the cases listed here, accession may be better for nonmembers than staying outside and suf-
fering discrimination, but it may not be better than MFN trade. Thus, since the expansion of RlAs all
the way to global free trade is far from ensured, one cannot necessarily view successful accessions
as stepping stones to multilateralism. Moreover, negotiated accession can lead to asymmetric
agreements in which benefits to developing country candidates are reduced and may be appropriat-
ed by existing members through side conditions on issues such as the environment, labor regula-
tions, and rules of origin. Since the more complex aspects of RIAs, especially those with budgetary
implications, must be negotiated, access can never be automatic and unconditional. Hence, it seems
rather naive to believe that in practice the wro could write or enforce general rules for open access.
Although relatively open access-as in APEC, which has the slogan of "open regionalism"-might
seem less threatening and exploitative than closed access (as in NAFTA), one cannot rely on this
route to deliver benign regionalism.



234      REGIONAL INTEGRATION AND DEVELOPMENT
occurs in the real world, although probably less for consciously exploita-
tive reasons than out of a general resistance to change and a fear of budg-
etary consequences. Consider, for example, how long it took the EU even
to admit the possibility of Turkish accession, or how tortuous the process
of accession has become for the countries of Central and Eastern Europe.
Similarly, NAFTA has rejected overtures from many countries, and
APEC had a moratorium on accession from 1993 to 1996.
Most formal analysis of domino regionalism assumes countries of
equal size that are identical in every respect except for the goods in which
they have comparative advantage. Thus, the only significant characteris-
tic of an RIA is its size, rather than which countries are members, and
there is no issue about how to share the benefits of integration. Once we
break free of this assumption, the compatibility of different partners and
struggles over distribution start to matter, and they greatly complicate
the process and analysis of RIA enlargement.
Given the different interests of excluded and member countries, what
does the evidence show about the domino effect? Table 8.4 provides a
summary. The "Strong expansion" column lists nine RIAs that have ex-
panded over time while maintaining the nature of the RIA or strengthen-
ing it. Among these are the EU; APEC; CUSFTA, in its expansion to
NAFTA; and other RIAs in Sub-Saharan Africa, Asia, Central and East-
ern Europe, and the Caribbean. The "Intermediate expansion" list of
RIAs that have expanded through a weaker form of integration contains
two customs unions, the EU and MERCOSUR, that expanded by form-
ing FTAs-in the case of the EU, with Mediterranean countries and with
Central and Eastern European countries and for MERCOSUR, with
Table 8.4 Domino Regionalism: Strong, Intermediate, and No Expansion
Intermediate
Strong expansion                   expansion        No expansion
EU (formerly EC): 1957, Belgium, France, Germany, Italy, Euro-Mediterranean  Andean Pact, Arab
Luxembourg, Netherlands, 1973, Denmark, Ireland,  Agreement. FTAs  Common Market,
United Kingdom; 1981, Greece; 1986, Portugal, Spain;  between EU and  CACM, CBI, CEPGL,
1995, Austria, Finland, Sweden; inclusion of some of  Mediterranean  COMESA, EAC,
the Central and Eastern European countries and some  countries  ECOWAS, G3, GCC,
Mediterranean island countries planned                       10C, SMRC, SACU
FTAs between EU and
CUSFTA: 1989, Canada, United States; 1994, expansion  Mexico, EU and
to NAFTA through inclusion of Mexico; possible ex-  Chile, and so on
pansion to FTAA
FTAs between MER-
APEC: 1989, Australia, Brunei Darussalam, Canada, In-  COSUR and Chile,
donesia, Japan, Republic of Korea, Malaysia, New  and between MER-
Zealand, Philippines, Singapore, Thailand, United  COSUR and Bolivia



TRADE BLOCS AND THE REST OF THE WORLD                     235
Table 8.4 (continued)
Intermediate
Strong expansion                        expansion         No expansion
States; 1991, China, Hong Kong (China), Taiwan
(China); 1993, Mexico, Papua New Guinea; 1994,
Chile; 1998, Peru, Russian Federation, Vietnam
CARICOM: 1973, Antigua and Barbuda, Barbados, Ja-
maica, St. Kitts and Nevis, Trinidad and Tobago; 1974,
Belize, Dominica, Grenada, Montserrat, St. Lucia, St
Vincent and the Grenadines; 1983, The Bahamas (part
of the Caribbean Community but not of the Common
Market)
CEMAC (originally UDEAC): 1966, Cameroon, Central
African Republic, Chad, Republic of Congo, Gabon;
1989, Equatorial Guinea
SADC: 1980, Angola, Botswana, Lesotho, Malawi,
Mozambique, Swaziland, Tanzania, Zambia, Zimbab-
we; 1990, Namibia; 1994, South Africa; 1995, Mauri-
tius; 1998, Democratic Republic of the Congo, Sey-
chelles
UEMOA: 1994, Benin, Burkina Faso, C6te divoire, Mali,
Niger, Senegal, Togo; 1997, Guinea-Bissau
CEFTA 1993, Czech Republic, Hungary, Poland, Slovak
Republic; 1996, Slovenia; 1997, Romania; 1999, Bul-
garia
AFTA: 1992, Indonesia, Malaysia, Philippines, Singapore,
Thailand; 1994, Brunei Darussalam; 1995, Vietnam;
1997, Lao People's Democratic Republic, Myanmar;
1999, Cambodia
Note: AFTA, ASEAN Free Trade Area; APEC, Asia-Pacific Economic Cooperation; ASEAN, Association of
SoutheastAsian Nations; CACM, Central American Common Market; CARICOM, Caribbean Community and
Common Market; CBI, Cross-Border Initiative (Africa); CEFTA, Central European Free Trade Area; CEMAC,
Economic and Monetary Community of Central Africa; CEPGL, Economic Community of the Countries of
the Great Lakes; COMESA, Common Market for Eastern and Southern Africa; CUSFTA, Canada-United
States Free Trade Agreement; EAC, East African Cooperation; EC, European Community; ECOWAS, Eco-
nomic Community of West African States; EU, European Union; FTA, free trade area; FTAA, Free Trade Area
of the Americas; G3, Group of Three (Colombia, Mexico, and Venezuela); GCC, Gulf Cooperation Council;
IOC, Indian Ocean Commission; MERCOSUR, Common Market of the South; NAFTA, North American Free
Trade Agreement; SAARC, South Asian Association for Regional Cooperation; SACU, Southern African Cus-
toms Union; SADC, Southern African Development Community; UDEAC, Union douanere et economique de
l'Afrique Centrale; UEMOA, West African Economic and Monetary Union.
Source: Authors' compilation.



236    REGIONAL INTEGRATION AND DEVELOPMENT
Chile and Bolivia. Note that MERCOSUR differs from other RIAs in
that its members wanted to expand by accepting Chile as a member but
Chile preferred to maintain an independent trade policy and form an
FTA with MERCOSUR. Finally, there are 13 RIAs in the "No expan-
sion" column.
Although the number of RIAs that have expanded is approximately
the same as the number of those that have not, when the size of RIAs is
considered, the expanding RIAs strongly dominate. This is exactly what
theory would predict: it is the economic power of the bloc-to stimulate
efficiency at home and to discriminate against outsiders-that encour-
ages entry. Finally, we reemplhasize that through expansion of RIAs a
domino effect is not necessarily a good thing.
Regionalism as Insurance
The main incentive for domino regionalism is to avoid being outside a
bloc when nearly everyone else is inside, either because one's terms of
trade would suffer or out of fear of a trade war that would close markets
that were not contractually bound open. This "insurance motive" was
seen in the spread of RIAs in the 1920s and 1930s (Eichengreen and
Frankel 1995). Throughout that period, France pursued an active re-
gional policy toward its own colonies and in Eastern Europe as a count-
er to British and German influence. As Germany reasserted itself, it
adopted regional means, starting with a proposed customs union with
Austria in 1931, which the other powers blocked only by exercising their
powers over the German financial system. Germany then gradually built
an even tighter web of regional trading arrangements (Irwin 1993).
Britain had granted some preferences to its colonies from 1919 on, and
in 1932 it deepened and widened these as it introduced higher tariffs on
other partners. The United States, which had increased tariffs strongly
for essentially domestic reasons under the 1929 Smoot-Hawley Act, tried
to recapture bilateral markets in the Reciprocal Trade Agreements Act of
1934. The lesson of this period is that regionalism grew up remarkably
quickly to fracture a relatively evenhanded, if somewhat sclerotic, trad-
ing regime.
The process of breaking into regional blocs is potentially explosive
(Oye 1992). Not only is there an incentive for each country to join a bloc
even if the result is that eventually everyone is worse off, but there is an
incentive to join early. The costs of remaining outside escalate as the bloc
grows, and if existing members use accession negotiations to extract an
"entrance fee," that also rises (see box 8.4).'5



TRADE BLOC(S AND THE REST OF THE WORLD                      237
The term "insurance motive" highlights another possible problem
with developing countries' seeking to defend their market access by sign-
ing RIAs. The more uncertain the world, the higher are insurance premi-
ums, and the costs of errors are lower if one is insured. In other words,
large powers may gain from saber rattling-by maintaining tariffs or
hostile antidumping regimes--while small countries are deciding
whether to join them, and after joining, the small countries will be less
concerned to preserve a global system than previously. As saber rattling
is effective only if there is some chance of coercion, this makes regional-
ism of this kind look quite hostile to multilateralism.16 Box 8.5 explores
Box 8.5 Insurance Policies
Signing an RIA may offer assured access to partner markets in two senses: first, through avoidance
of day-to-day harassment from antidumping or countervailing duties and other administrative
means of protection and, second, as a haven if total trade war breaks out. Mercifully, the latter is a
very low-probability event at present, and it is difficult to forecast what form it would take. Some re-
cent results from Whalley (1 998a), however, give a feel for the orders of magnitude at stake.
Whalley uses a CGE model of the world economy disaggregated into seven countries or regions. Six
of these potentially fight trade wars, while the other area (the rest of the world) remains completely
passive. Trade wars have different outcomes according to which countries or regions have combined
to form customs unions. Using actual 1986 values as a base, Whalley explores trade wars in which
each country or bloc fixes its tariffs to maximize its own welfare, taking rivals' tariffs as given. (Of
course, this ignores WTO strictures on raising tariffs, but the insurance is intended for just such a
breakdown.) The only constraint is that if an RIA exists before war breaks out, it remains in operation
afterward. The table presents the implications of four possible configurations of customs unions
measured relative to 1986 actual values.
The precise numerical results should not be taken seriously, for they are subject to a host of uncer-
tainties, but the broad pattern is informative. In trade wars large economies (such as the United
States and the EC) suffer least or even gain. Others suffer heavily, especially if they are highly de-
pendent on a large bloc that becomes very restrictive, as is the case with Canada and Other Western
Europe in the table. Entering the bloc basically solves the problem: compare Canada in columns 1
and 2, Mexico in columns 2 and 3, and Other Western Europe In columns 3 and 4.As one small coun-
try protects itself, the burdens on others increase, as happens with Japan and the rest of the world.
With losses of this size, the incentives to seek accession if the danger of a trade war increases are
huge. The Other Western Europe group has a turnaround of 43.6 percent of GDP from joining a North
American-European customs union. Even a small perceived increase in the probability of a trade war
would be sufficient to persuade it to bear trade diversion and other costs of entry to avoid the worse
outcome.
The table refers to customs unions, and hence the blocs coordinate their external trade policies and
can exploit, in so small a model, a great deal of market power to raise their tariffs. If the RIAs were
FTAs, the costs of being excluded would be rather lower and would not rise monotonically as the FTA
expanded (because a large FTA tends to reduce its external tariffs to lessen the costs of trade diver-
sion). The basic idea that being outside is very costly holds true, however.
(Box continued on next page.)



238     REGIONAL INTEGRATION AND DEVELOPMENT
Box 8.5 (continued)
Losses of Economic Welfare from Trade Wars, for Economies Inside and Outside
Customs Unions
(equivalent variations as a percentage of GDP)         NorthAmerican-
No customs   Canada-US   NorthAmerican  European
Trade War with:   union     customs union  customs union  customs union
United States         1.2          0.5         -0.4          0.5
Canada              -25.5          0.9          0.4        -0.7
Mexico                           -8.5           0.l        -0.3
Japan                             -5.2        -5.2         -5.4
EC (12 members)     -8.5           3.4          3.4          2.5
Other Western Europe  -5.2       -33.1        -33.5         10.1
RestofWorld         -10.6        -10.9        -11.1        -13.4
World total         -6.0          -5.8         -5.8         -6.2
Source: Whalley (1998a).
the high incentives for "insurance regionalism" in the current world.
Such problems do not look very serious at present, for developed coun-
tries have many other nonexploitative relationships with developing
countries and may not find "insurance premiums" worth pursuing in this
way. But insurance is for worst-case scenarios, and our point is that de-
veloping countries should not complacently assume that regionalism will
preserve them if a crisis arises.
One can observe the same sort of phenomena in today's world, albeit
in a much attenuated form. One of the reasons that New Zealand Prime
Minister Helen Clark cited for her country's seeking a regional arrange-
ment with Singapore was a fear of being left outside as the world split up
into blocs, into none of which New Zealand fitted naturally. This fear
was arguably stimulated by the stasis in APEC (which accounts for a
very large proportion of New Zealand's trade) and "the failure of world
leaders to agree on an agenda for a new wave of trade liberalization"
(Far Eastern Economic Review, August 17, 2000).
RIAs as Negotiating Partners: Do They Promote Free Trade?
If RIAs make trade negotiations easier, perhaps they can help the world
evolve toward freer trade. Coordinated coalitions may, as noted above,
have greater negotiating power than their members individually, and
such coalitions may facilitate progress just by reducing the number of
players represented in a negotiation (Krugman 1993; Kahler 1995). But
this result is not guaranteed. For example, a negotiation between one
dominant partner and a competitive fringe of small countries might be



TRADE BLOCS AND THE REST OF THE WORLD   239
easier and proceed farther than if the fringe coalesced into a significant
counterforce, but if the blocs are genuinely unified, it is probably reason-
able to expect negotiations to be easier with fewer partners.
Unfortunately, this condition is rarely met, so any gains from having
fewer players in the last stage of a negotiation are offset by the complex-
ity of agreeing joint positions in the first phase. This tradeoff is examined
in Andriamananjara and Schiff (2001). The difficulties of achieving a Eu-
ropean position on agriculture and cultural protection in the Uruguay
Round are well known, and formulating EEC positions in the Tokyo
Round proved complex (Winham 1986). Moreover, two-stage negotia-
tions need not be more liberal than one-stage ones (Basevi, Delbono, and
Mariotti 1994). To be sure, Germany and the United Kingdom pressured
France to agree to the agricultural deal in the Uruguay Round, but they
had to make potentially trade-restricting concessions on "commercial
defense instruments" (antidumping) to clinch the deal. The negotiating
power of African countries would not be greatly enhanced by coopera-
tion, and the benefits are not likely to outweigh the costs of combining
their different interests into a single negotiating position (Wang and Win-
ters 1998).
The customs unions that attend the next round of global trade talks
will need to establish procedures for determining their negotiating posi-
tions. SACU's previous practice of delegating all responsibility to South
Africa begins to look less tenable as divisions emerge between members,
and MERCOSUR has yet to devise really robust internal decisionmaking
capacity. Thus, at least in the foreseeable future, RlAs do not seem likely
to facilitate even traditional trade negotiations. Moreover, if RIAs are
being extended or new ones are being created, the process can swamp de-
veloping countries' limited negotiating capacity (box 8.6).
As the WTO has extended its reach, it has embraced subjects in which
most central customs union authorities have no mandate to negotiate.
Mixing national and customs union responsibilities seems unlikely to
simplify matters, and it is not realistic to expect member countries to sur-
render sovereignty on sensitive issues to regional bodies just because
trade negotiations are in train.
Do RIAs Make It Easier to Tackle Tough Issues?
It is frequently claimed that a strength of the regional approach to liber-
alization is that handling tough cases is easier (Kahler 1995). That is,
there are areas in which regional liberalization or harmonization be-
tween like-minded countries is feasible when multilateral progress is not.



240        REGIONAI. INTEGRATION AND DEVELOPMENT
Box 8.6 The Overburdened Negotiator
Reserves of administratve skill, political capital, and imagination are finite; if they are devoted to an
RIA, they are not available for multilateral objectves. These arguments were advanced-somewhat
implausibly-to explain EU and U.S. behavior during the Uruguay Round, but they must be several
bmes more important for developing countries.
Negotiating an RIA, especially with a major power that has its own objectives, will absorb a huge
proportion of the policymaking skills of a developing country. Such skills are typically so scarce that
many developing countries have the same negotiator dealing with both the EU in Brussels and the
WTO in Geneva. In fact, the governments of a number of smaller European countries have asked the
WTO to postpone its trade policy review exercises for their countries because they are wholly ab-
sorbed in negotiating association agreements with the EU.
Moreover, flying to Brussels or Washington to undertake regional negotiations is altogether more
gratifying than working quietly at home to reduce arbitrary tariff peaks or improve customs adminis-
tration. This is one reason why policymakers and private sector groups are prepared to grasp the
painful nettles of reform that RIA negotiations call for and to stick to the agreements they make. But
the process can be harmful if the RIA does not deliver benefits commensurate with the opportunity
costs of negotiation.
The importance of developing country capacity constraints is illustrated in the recent discussion of
multilateral trade policy, where lack of capacity has become a prominent theme. For example, the
WTO's Doha Ministerial Declaration, a 5,000-word document, contains 19 references to capacity
building and 21 to technical assistance for developing countries (see Winters 2002).
This seems most likely for activities that are highly restricted (such as
agriculture, trade subject to antidumping measures, and some services)
and for areas that are highly technical or sensitive (standards, competi-
tion policy, and services regulation). For example, NAFTA and the
Group of Three have tackled investment; Brazil has accepted free trade in
information technology goods within MERCOSUR but refused to sign
the global Information Technology Agreement (ITA) in 1997; and Chile
and Canada have eschewed antidumping actions on mutual trade but not
in relation to third countries.'7
Until recently, however, even RIAs of developed countries, let alone
those between developing ones, had not advanced much farther with lib-
eralization     than   the multilateral system         (Hoekman        and   Leidy    1993).
Agriculture frequently remained restricted (for example, in EFTA); trans-
port, culture, and other "sensitive" services were excluded (CUSFTA);
and government procurement was ignored de facto, if not de jure (EEC).
The EU, especially in its Single Market Programme, has advanced be-
yond the GATT on a broad front, but this took 30 years to initiate and is,
to date, unique. More recently, there have been other advances, such as
NAFTA's plans to ultimately liberalize agriculture and procurement. But
overall, RIAs have not led multilateral liberalization to the extent that is
sometimes supposed.



TRADE BLOCS AND THE, REST OF THE WORLD  241
There is also the question of whether regionalism is actually more ef-
fective than multilateralism in liberalizing deeply. Has global liberaliza-
tion actually been ruled out, or, if RIAs were not an option, would a lit-
tle more time and effort yield global progress? Having got the ball
rolling, will RIAs later slow it down, for the reasons discussed above?
On a prescriptive note, to the extent that RIAs are justified by their abil-
ity to open up otherwise closed sectors, it is important to ensure that the
subsequent switch from regionalism to the multilateral track can be man-
aged effectively. The necessity for, and the means of achieving, this
switch should be written into the initial terms of RIAs.
An extension of the "tough cases" argument is that RIAs help develop
blueprints for subsequent multilateral negotiations (Bergsten 1996;
Lawrence 1996). For example, the EU pioneered "bulk" mutual recogni-
tion for industrial standards and services harmonization, and NAFTA's
investment chapter may inform a multilateral negotiation (if there is
one). On the other hand, the EEC suggested the CAP as a model for agri-
culture in the Kennedy Round. (See Preeg 1970: 152.)
But the major powers could also seek to use RIAs to reinforce their
initial positions in future multilateral negotiations. If they have greater
politico-economic power within their own regions than in the world in
general, it is easy to imagine their building up coalitions for their own
policies before taking issues into a multilateral round. Arguably, the
United States used approaches by potential partners in the Americas and
Asia to broaden the negotiating agenda for its relations with Europe
(Ostry 1998), while Europe did the same with the EEA and its Europe
Agreements.
The benefits of developing regional blueprints depend heavily on
whether they are liberalizing (Bhagwati 1993) and on whether they are
otherwise well suited to developing countries' needs and capacities.
Major powers already use the carrot of access to the generalized system
of preferences (GSP) to promote environmental and labor conditions in
developing countries; the EU looks for action in such areas and on intel-
lectual property in the Europe Agreements; and the United States has
used NAFTA as a tool for enforcing Mexican labor and environmental
standards. By negotiating singly with the major powers, developing
countries are essentially placed in competition with each other and lose a
good deal of their (small) reserves of negotiating power. The deals they
can achieve could be much less favorable than those that might emerge
from multilateral talks under the WTO, and they may be less open and
liberal, as well.
Moreover, there are dangers in such tactics. First, even if the major
powers' aspirations are inherently desirable, building up rival teams can



242    REGIONAL INTEGRATION AND DEVELOPMENT
make final negotiations more, rather than less, difficult. Second, when
only one regional bloc is advocating a policy, other countries might suffi-
ciently resent the pressure to adopt it that they pull back. De facto rejec-
tion by developing countries of the OECD draft Multilateral Agreement
on Investment in 1998 contained at least elements of this reaction. Third,
the time it takes to build regional coalitions can delay multilateral talks.
Fourth, coalitions rooted in formal RIAs are here to stay: multilateral
processes may fail, but the blocs remain. This is quite different from a ne-
gotiating coalition, which dissolves if it fails to gain its objectives.
To be sure, progress is required in "new" areas such as standards, but,
as was noted in chapter 6, it is frequently better pursued independent of
tariff preferences. Thus, while we may well learn from RIAs how to tack-
le particular aspects of liberalization, this is not a convincing reason for
pursuing regionalism per se.
Open Regionalism: Little More Than a Slogan
Open regionalism was the idea of the 1990s. Crafted to describe APEC's
original aspirations and to convey their complete consistency with multi-
lateral objectives, the accolade has been applied to many blocs at some
point in their history. Unfortunately, it is difficult to pin down exactly
what "open" means.
Before trying to clothe the concept of "open regionalism" with mean-
ing, we should remove its fig leaves. First, although there is a presump-
tion that greater openness makes RIAs more benign, nothing can
guarantee that they do not hurt the rest of the world. Second, the main
reason for pursuing "open regionalism" is to benefit the members them-
selves, but none of the definitions that have been proposed absolutely
guarantees such an outcome.
Srinivasan (1998) has called "open regionalism" an oxymoron, and
that is surely true at the limit of its range: a perfectly open economy
could not discriminate in the way that regionalism in the trade sense re-
quires. Once we come in from the edge, however, it does make sense to
ask whether some RIAs are more open than others.
Writing about APEC, the group for which the term was coined, Berg-
sten (1997) offers five definitions of "open regionalism":
1. "Open access," whereby any country willing to abide by an RIA's
rules may join that RIA. Bergsten argues that, strictly interpreted,
this is not realistic because RIAs have restricted geographic do-
mains, but easy access for countries within their domains would
be one acceptable definition. Bergsten defines the EU as open in
this respect.



TRADE BLOCS AND THE REST OF THE WORLD   243
2. Unconditional MFN (or concerted unilateralism). This was the
definition of early APEC advocates, who saw the coalition as a
means of encouraging countries to liberalize together and so
provide for each other some of the terms of trade and political-
economy benefits of a full GATT round. Although some members
of APEC still adhere to this aspiration, Bergsten is doubtless cor-
rect in saying that it is (quite unacceptable in the United States,
where reciprocity is seen as an essential part of liberalization.'" It
would, of course, also not be regionalism as defined in this book.
3. Conditional MFN, whereby intra-APEC liberalization would be
extended to any country that reciprocated. Given APEC's size, no
country would be likely to reject the offer of an FTA, Bergsten
says, but it might take time for others to come on board, and in
the interim there would be trade discrimination. There might also
be resentment at APEC's making take-it-or-leave-it offers. Condi-
tional MFN is very similar to open access but requires an
operational definition of reciprocity. Bergsten suggests using con-
ditional MFN as a threat and indeed says APEC has already im-
plicitly done so-as a means of obtaining (nearly) global deals
such as the ITA, which was initiated by APEC in 1996.
4. Global liberalization through traditional unilateral and multilat-
eral trade liberalization on an MFN basis. It is possible that APEC
could lead a movement to remove all trade barriers by its own in-
ternal free trade deadline of 2010 for developed countries and
2020 for developing countries.
5. Trade facilitation, under which APEC countries would reduce
border frictions and pursue policy integration, but with a focus on
elements that operate multilaterally.
APEC has yet really to decide among these alternatives, as there has
not yet been any "APEC liberalization." Members have certainly not in-
troduced any discriminatory trade policies (with the minor exception of
the APEC business visa), but neither have they moved beyond imple-
menting their Uruguay Round obligations and, for developing members,
their own unilateral reforms.
For the future, Bergsten recommends a combination of points 2
through 5 to make clear that APEC's objective is global liberalization,
that it will achieve internal liberalization by 2020, that it (or at least its
larger members) will discriminate in a WTO-consistent fashion if other
countries do not reciprocate, and that it welcomes proposals by other
countries on sectors and timetables for liberalization (in other words,
that APEC wishes to negotiate, r ather then dictate terms for, its extension
to the rest of the world). Although perhaps more positive in its approach



244    REGIONAL INTEGRATION AND DEVELOPMENT
and timetable, this manifesto is not very different from the EEC's 1957
statement that "by establishing a customs union between themselves,
member states aim to contribute, in the common interest, to the harmo-
nious development of world trade, the progressive abolition of restric-
tions on international trade and the lowering of customs barriers"
(Treaty of Rome, art. 110, sec. 1).
Other commentators have taken up the cry of "open regionalism."
Echavarria (1998) sees "open regionalism" in the Andean Pact in the
sense that it lowered tariffs relative to the 1980s. Brazil's foreign minis-
ter, Luiz Felipe Lampreia, speaking to the U.N. General Assembly in Sep-
tember 1996, proclaimed "open regionalism" in MERCOSUR because
of the tariff reductions and the group's enlargement. The U.S. Council of
Economic Advisers defined U.S. policy as laying the foundation for
"open, overlapping plurilateral trade agreements as stepping stones to
global free trade" (CEA 1995), while Pelkmans and Brenton (1999) see
recent EU policy as "open regionalism" in the sense of "GATT compati-
bility" and as having gone "some way" toward "open regionalism" in
the nonpreferential sense. Former WTO director-general Renato Rug-
giero declared, "The answer [to marrying regionalism and multilateral-
ism] is . . . 'open regionalism,"' which he defined as "in practice . . . the
MFN principle" (quoted in Srinivasan 1998).
Openness is an important component of development and a valid ob-
jective for all developing countries; if countries feel they must have re-
gionalism, let it be genuinely open. But "open regionalism" is a slogan
rather than an analytical term. It is defined in so many different ways
that it conveys no information about an RIA other than that its members
are embarrassed to be thought of as protectionist. In particular, a claim
of "open regionalism" should confer on a bloc no presumption of eco-
nomic value and no immunity from analysis to see if it is acting in either
its own or in other countries' interests.
REGIONALISM AND THE WTO
The analysis thus far suggests that international policy toward regional-
ism should aim at:
* Encouraging RIAs to achieve trade creation and avoid trade di-
version, both for the sake of members and to minimize harm to
excluded countries;
* Permitting deep integration, including nation building, between
members;



TRADE BLOCS AND THE REST OF THE WORLD   245
* Preserving the effects of previous liberalizations and providing
credibility for any liberalizations that form part of the RIA; and
* Supporting a liberalizing dynamic within member countries and
in the world trading system as a whole.
The instrument we have for international policy on trade blocs is the
WTO. This section explores how it manages regionalism and whether its
rules could be reformed to help it do better.'9
GATT, and All That
RIAs are an officially sanctioned, but conditional, exception to the
GATT's rules on nondiscrimination. The conditions imposed on RIA for-
mation doubtless constrain and mold the pattern of regionalism in the
world, but they are not adequate, and are not well enough enforced, to
ensure that regionalism is economically beneficial for either its members
or excluded countries. And although improved adherence to current
rules is desirable, no feasible reform can guarantee that only beneficial
RIAs are created. The responsibility for good outcomes falls on govern-
ments themselves; apart from a complete ban, they cannot tie their own
or each other's hands sufficiently tightly in the WTO to preclude the pos-
sibility of harmful RIAs being formed.
The world trading system works pragmatically and consensually. The
GATT was created in 1947 as a temporary body to assist countries in
trade liberalization. Its role was to codify and record a series of tariff re-
ductions that its members wished to make and to provide a structure to
give credibility to those reductions. To discourage the reversal or nullifi-
cation of tariff cuts, it restricted the use of policies that impose duties on
trade on an ad hoc basis (such as antidumping duties and emergency pro-
tection) and of equivalent policies, such as internal taxes on imports. It
also defined some important mechanics for the trading system-for ex-
ample, the valuation of trade for customs purposes.
A key concept of the GATT-indeed, the cornerstone of the present
world trading system-is nondiscrimination between different sources of
the same imported good. Nondiscrimination is achieved by requiring
members to accord each other MFN treatment except in specified cir-
cumstances. With an assurance of nondiscrimination, when A negotiates
a reduction in one of B's tariffs, A knows that the commercial value of its
effort will not be undermined by B's then offering C an even lower tariff.
This, in turn, makes A more willing to "buy" the concession by reducing
one of its own tariffs on B and so encourages trade liberalization.



246    REGIONAL INTEGRATION AND DEVELOPMENT
Over 50 years of operation, the GATT operated in a low-key, consen-
sual, member-driven fashion. It did not "adjudicate" trade disputes;
rather, it had a dispute settlement process that was less concerned with
law than with solving disputes in a way that preserved the previously ne-
gotiated balance of benefits, maintained consensus, and allowed the lib-
eralizing bandwagon to continue to roll. The WTO, which was created
in 1995 to oversee the GATT and certain other agreements, is more le-
galistic but still focuses on pragmatic and mutually acceptable solutions
to problems. The WTO administers a set of rules for behavior, not a set
of outcomes; it is concerned with members' meeting agreed obligations
and being accorded agreed rights rather than with economic outcomes
per se. The WTO/GATT has undoubtedly been a force for economic
good, but its role has not been defined in those terms.
The GATT traditionally did not intrude into domestic politics. It had
no ability to force member countries to liberalize if they did not wish to,
and it was extremely light-handed in its requirements about the shape of
domestic legislation. The WTO is more far-reaching. Its greater breadth
and its "single undertaking"-under which members must subscribe to
(virtually) all its rules instead of treating some as optional extras, as had
been the case-constrains governments more tightly. Nevertheless, the
WTO, like the GATT, can be effective only if it proceeds more or less by
consensus.
Given this background, the WTO can enhance the economic well-
being of developing countries in four ways:
1. If sufficient members wish, it can organize periodic rounds of
tariff negotiations that offer opportunities and incentives to mem-
bers to reduce their barriers to trade.
2. It provides guidelines for domestic policy-directly in some cases,
but, more importantly, indirectly, by shaping the terms of the de-
bate. Governments resisting pressures to protect particular groups
are immeasurably strengthened if they can point to prohibitions
in the WTO agreements. Similarly, lobbyists are strengthened if
the WTO explicitly permits the policies they want, even if only
conditionally. "After 145 countries have debated and confirmed
the legitimacy of policy X," they say, "what right has the govern-
ment to deny us its benefits?" (The answer is, "almost every
right," for the WTO defines minimum standards of behavior, not
norms-but it is a hard case to make.)20
3. It can protect the rights of WTO members against certain rules vi-
olations by other members. It cannot, however, guarantee mem-
bers against harm.21



TRADE BLOCS AND THE REST OF THE WORLD    247
4. It provides a forum and a mechanism for governments to manage
the spillovers from members' trade policies to their partners. Bag-
well and Staiger (1998) show how the GATT traditions of reci-
procity and nondiscrirmination combine to minimize such
spillovers-reciprocity, because two members offering mutual tar-
iff concessions are likely to generate (at least partially) offsetting
terms-of-trade effects, and nondiscrimination, because it prevents
aggressive coalitions from forming.
These four potential benefits provide a yardstick for assessing the
WTO's current rules about RIAs and exploring whether they can be im-
proved.
The Rules for RIAs: Useful, but Not Infallible
Article XXIV of the GATT specifies the conditions under which coun-
tries may violate the MFN clause by forming RIAs. It imposes three prin-
cipal restrictions (see the annex to this volume). An RIA:
*   Must not, "on the whole," raise protection against excluded
countries;
*   Must reduce internal tariffs to zero and remove "other restrictive
regulations of commerce" other than those permitted by other
GATT articles; and
*   Must cover "substantially all trade."
The GATT's logic is essentially mercantilist-stressing the rights of
trading partners to market access-rather than economic, with a focus
on the economic costs and benefits of policy. From the mercantilist per-
spective, the first two conditions make sense. The rule against increasing
protection against excluded countries preserves the sanctity of tariff
bindings by ensuring that forming an RIA does not provide a wholesale
way of dissolving previous bindings. It is supplemented by the require-
ment that compensation is due to individual partners for tariff increases
induced by the RIA if other reductions to keep the average constant do
not maintain a fair balance of concessions. Together with the 1994
Uruguay Round Understanding on the Interpretation of Article XXIV,
which deals with how to measure tariff barriers for RIAs, these provi-
sions offer reasonable assurances abou,t the barriers facing nonmembers.
The condition on reducing internal tariffs to zero helps defend the
MFN clause by making it subject to an "all-or-nothing" exception. If
countries were free to negotiate different levels of preference with each
trading partner, binding and nondiscrimination would be fatally under-
mined. No member could be sure that it would receive the benefits it ex-



248     REGIONAL INTEGRATION AND DEVELOPMENT
pected from negotiating and reciprocating a partner's tariff reduction.
Also, if a customs union is a first step toward nation building, it is inap-
propriate that an international trade treaty should stand in the way of
such progress. Thus, internal free trade, such as is (usually) achieved
within a single country, would seem to be an acceptable derogation of
MFN, whereas preferences would not be. The third condition reinforces
this interpretation by requiring a serious degree of commitment to an
RIA in terms of sectoral coverage.
Article XXIV also makes sense when viewed as a guide to economic
policy. The requirement not to raise the average level of protection
against excluded countries' exports not only honors those countries'
market access rights but also closes one otherwise available route to in-
creased protectionism. This is desirable on any account, but in the con-
text of RIAs, the danger and costs of trade diversion will be greater if
members can increase their external tariffs. The second and third condi-
tions-no internal tariffs, and substantial coverage-are important in
heading off pressure to use tariffs to fine-tune political favoritism toward
domestic industries or partner countries; they help prevent governments
from restricting RIAs to swaps of trade-diverting concessions and so
avoiding politically more painful trade creation. These rules essentially
require a serious commitment to integration of member markets as a
condition for proceeding.
Article XXIV is generally an aid to better RIAs, but it is certainly not
sufficient for good economic policy. Even if the conditions were applied
without exception, they would not preclude harmful RIAs. RIAs that are
wholly GATT-compatible can nevertheless be predominantly trade
diverting; excluded countries can suffer terms-of-trade declines; protec-
tion can increase; and institutions can arise that make liberal policies less
likely.
Moreover, there are major difficulties in interpreting the WTO condi-
tions on regionalism. Despite the Uruguay Round Understanding, there
is no agreement about what "substantially all trade" means-even about
whether it refers to the proportion of actual trade covered or to the in-
clusion of all major sectors of the economy. Similarly, the treatment of
nontariff barriers in assessing the overall level of trade restriction is not
defined, nor is that of rules of origin. The requirement that "other re-
strictive regulations of commerce" be removed between members is am-
biguously worded: several exceptions to this requirement are identified
explicitly, but other barriers, including antidumping duties and emer-
gency protection, are not. Complete integration between members of an
RIA would abolish all these barriers, and so their continuation-as, for



TRADE BLOCS AND THE REST OF THE WORLD   249
example, in NAFTA and the Euro-Med Agreements-suggests an unwill-
ingness to proceed very far in that direction.
Perhaps because of its ambiguities, Article XXIV has been notoriously
weakly enforced. RIAs have to be notified to the GATT, and until 1996
each was then reviewed by an ad hoc working party to see if it was in
conformity with Article XXIV. According to the WTO (1995: 16), of 69
working parties reporting up to and including 1994, only 6 were able to
agree that an RIA met the requirements of Article XXIV, and only 2 of
the 6 RIAs-CARICOM and the customs union of the Czech and Slovak
Republics-are still operative. Ihe remaining working parties did not
conclude that the agreements were not in conformity; they merely left the
matter undetermined.
This agnosticism is essentially the product of the GATT's consensual
nature. The first major test of Article XXIV was the Treaty of Rome,
which established the EEC. The political pressure to permit it was enor-
mous: EEC countries would almost certainly have put the EEC before
the GATT in the event of conflict, and the United States strongly sup-
ported the treaty. The treaty, however, clearly violated Article XXIV, and
so the only feasible solution was not to push the review to conclusion.22
Given a start like this-along with the EEC's willingness to support more
or less any RIA in the GATT, the need for working parties to reach con-
sensus, and the GATT's inability to make an adverse determination with-
out the acquiescence of the party at fault-it is hardly surprising that
future reviews proved little more demanding.23 Nor have matters
improved with the establishment of a single Committee on Regional
Trading Agreements (CRTA, discussed below) to conduct the reviews.
The inability to rule on conformity does not mean that the rules have had
no effect, for we do not know the extent to which they have influenced
the structure of RIAs that have come forward or which potential
arrangements they have deterred.24 It is not an encouraging record, how-
ever, either from the point of view of enforcing current rules or from that
of rewriting the rules to increase their ability to distinguish good from
bad RIAs.
Finally, Articles XXIV.10 and XXV of the GATT can be used to grant
waivers to make otherwise inadmissible policies GATT-legal. This was
done for the ECSC (1952) and the United States-Canada Auto Pact
(1965). Under the WTO, waivers are still feasible but are time limited.
Weaker Rules for Developing Countries. As if all this were not
enough, a further complication for developing countries is the Enabling
Clause of 1979, which significantly relaxes the conditions for creating



250    REGIONAL INTEGRATION AND DEVELOPMENT
RIAs that include only developing countries. It drops the conditions on
the coverage of trade and allows developing countries to reduce tariffs
on mutual trade in any way they wish and to employ nontariff measures
"in accordance with criteria which may be prescribed by the" GATT
members. It then supplements the first condition with the nonopera-
tional requirement that the RIA not constitute a barrier to MFN tariff re-
ductions or cause "undue difficulties" for other contracting parties.
In practice, developing countries have had virtual carte blanche.
Twelve preferential arrangements have been notified under the Enabling
Clause, including the LAIA, ASEAN, and the GCC. Internal preferences
of 25 and 50 percent figured in ASEAN's trading plans, in many of the
arrangements concluded under the LAIA, and in the GCC. There is little
sign that internal preferences have undermined MFN agreements with
other trading partners, but until recently, these countries did not make
many MFN agreements. Indeed, until the late 1980s, the Latin American
and African countries' frequent use of regional arrangements and their
weak participation in the multilateral rounds might suggest a substitu-
tion of one form of liberalization for another. More worrying were the
sectoral agreements that abounded in Latin America. Nogues and Quin-
tanilla (1993) argue that there is little doubt that the 17 agreements that
Argentina and Brazil had signed by 1986 generated significant trade di-
version.
The Enabling Clause further dilutes the weak discipline that Article
XXIV imposes. Even if Article XXIV does not actually stop many harm-
ful practices, it does at least avoid automatically giving them the
respectability of legal cover. Thus, while the GATT knowingly and will-
ingly permitted LAFTA (1960) and the initial notification of ASEAN
(1977) to violate Article XXIV (Finger 1993b), at least it required con-
tinuing consultation with partners and left open the possibility of chal-
lenge in the dispute settlement process. The Enabling Clause offers more
cover in various areas and thus erodes even this discipline.25
Loopholes in Rules on Services. Article XXIV and the Enabling
Clause of the GATT refer to trade in goods. The equivalent for services is
Article V of the GATS, which is modeled closely on them (see the annex
to this volume). The requirement not to raise barriers to third countries
is rather tighter: it is applied sector by sector rather than "on the whole,"
and third-country suppliers already engaged in "substantive business" in
an RIA's territory before the RIA is concluded must receive RIA treat-
ment. The "substantially all trade" ambiguity is only slightly abated,
with an explicit note that "substantially" is to be "understood in terms



TRADE BLOCS AND THE REST OF THE WORI-1)  251
of number of sectors, volume of trade and modes of supply." For covered
sectors, "substantially all discrimination" is to be removed, but since the
requirement is defined as consisting of "elimination of existing discrimi-
natory measures, and/or prohibition of new or more discriminatory
measures" (emphasis added), it need amount to very little.26 Developing
countries receive "flexibility" on "substantially all discrimination" and
exemption from the need to give RIA treatment to third-country firms
with "substantial business" in member countries.
Proposals for Improving the Rules on RIAs: "Feasible" and
"Desirable" Do Not Overlap
The WTO rules on RIAs are not exactly broken, but they are creaky, and
it is worth asking what might be done about them. We focus here on
their economic content and benefits and on the feasibility of reform. We
also require that any reforms be clear and operational, removing ambi-
guity rather than adding to it. Feasibility, which depends on the attitudes
of member countries and the need to remain within the GATT's basic
pragmatic and consensual framework, seems to us more of a constraint
than devising economically sensible rules.27 Few countries within RIAs
appear to seek tighter discipline, least of all the EU and the United States.
The EU is intent on enlargement-extending its Mediterranean agree-
ments and implementing the Cotonou Agreement-while the United
States is heavily engaged in the FTAA negotiations. Neither the EU nor
the United States submitted proposals to the 1999 Seattle ministerial
conference of the WTO, and while they accepted the issue as part of the
Doha Development Agenda, neither has sought to advance the topic.28
An RIA that does not reduce external barriers almost inevitably caus-
es trade diversion. One counter to this would be to require RIA members
to liberalize, both to reduce diversion and to induce external trade cre-
ation with nonmembers. Finger (1993b) views these reductions as a
"price" to be negotiated to persuade nonmembers to forgo their MFN
rights. How far the parties are prepared to go in a negotiation, however,
is determined by the prevailing rules and enforcement mechanisms,
which define the outcome should negotiations fail. Unfortunately, these
provisions currently give nonmembers almost no negotiating power.
Other authors have therefore made more concrete proposals.
Greater Liberalization by RIAs. Bhagwati (1993) suggests a require-
ment that for each tariff heading, a customs union's common external
tariff be bound at the minimum tariff for that heading among all mem-



252     REGIONAl. INTEGRATION AND DEVELOPMENT
bers. This does not guarantee the elimination of trade diversion (suppose
that the tariffs of three members were 98, 99, and 100 percent), but it
would clearly reduce it. It would impose a high (mercantilist) price on
RIA formation that only "serious" integrators would pay, and it would,
overall, be trade liberalizing. As a reform it is admirably clear, and if fea-
sible, it would be desirable economically. Its demanding nature, however,
makes it very unlikely to succeed in the present circumstances.
Serra and others (1997) propose that members of FTAs be required to
bind their tariffs at the actual applied rates on the eve of the R1A. Apart
from what this might do to pre-FTA applied rates, the suggestion is ran-
dom in its liberalizing effect, which reduces its moral force. Bhagwati, by
contrast, would just ban FTAs. Like his proposal on customs unions, this
is consistent with seeking to restrict RIAs to those that are committed to
far-reaching integration, but it too faces severe feasibility constraints, es-
pecially since some FTAs proceed quite far in other directions and have
become the more popular form of regional integration.
Conditions for Rules of Origin. Tied up with the FTA question is that
of rules of origin. Serra and others (1997) suggest requiring that they be
no more restrictive than before the RIA. This is laudable, but, in practice
it is difficult to determine, ad hoc in nature, open to manipulation, and
potentially very complex in the face of technological changes. A better
course would be a requirement precluding the manipulation of such rules
for protectionist purposes-requiring, for example, that countries adhere
to a single set of rules of origin agreed internationally, or that a country's
preferential rules be the same as its nonpreferential ones. R. Wonnacott
(1996) suggests a number of milder reforms in this direction: that rules
of origin be banned where tariffs differ between members by less than,
say, 2 percentage points; that for each commodity they be banned for the
FTA member with the lowest tariff, and so on. These rules might be ac-
ceptable but would only scratch the surface.
Exports-The Wrong Criterion. One proposal envisions mounting ex
post reviews to determine whether nonmember exports have fallen since
an RIA was created and, if they have, demanding changes in policies
(McMillan 1993). Although it is frequently taken seriously (see, for in-
stance, Frankel, Stein, and Wei 1997), the proposal is wrong in virtually
every respect. Exports are the wrong criterion (welfare is related to im-
ports, not exports); quantitative targets are the wrong way to formulate
trade policy (they create more rents and market power than do tariffs);
economic modeling is still too imprecise to permit identification of caus-



TRADE BLOCS AND THE REST OF THE WORLD   253
es with any credibility; and ex post adjustment after five years is no basis
for the policy predictability sought by investors.
More Vigorous Enforcement of Existing Rules. Two arguably more
feasible proposals entail enforcing current rules more rigorously. Even
they, however, currently encounter fierce opposition and will require
great political commitment by many WTO members to be implemented.
To be acceptable to the major powers, they will certainly need to be ac-
companied by a grandfather clause to assure current RIAs that they will
not be undermined by new interpretations.
1. A precise definition and enforcement of "substantially all trade"
would be a useful innovation. A quantitative indicator would be
clear, but it would need to be high, given that the kinds of trade
restrictions countries wish to maintain typically constrain existing
trade quite fiercely already. The frequently cited 80 percent,
which dates from consideration of the Treaty of Rome, is not ad-
equate. Even 90 percent, which seems to inform current EU-
MERCOSUR talks, is not indicative of serious integrationary in-
tent. We would advocate 95 percent after 10 years and 98 percent
after 15.
2. Similarly, a more constraining view of "other restrictive regula-
tions of commerce"-to ensure that these regulations cover the ef-
fects of rules of origin on excluded countries and that obvious
barriers such as safeguards actions and antidumping duties are
abolished internally-would be useful. The latter requirement
would increase the degree of trade creation, as safeguard and an-
tidumping policies are explicitly aimed at preserving domestic
output levels. Thus, these requirements would raise the political
bar for "serious" regionalism.
More Radical Reforms. There are three major proposals for creating a
"liberal dynamic":
1. Time limits. Srinivasan (1998) proposes that RIAs be permitted
only temporarily; all RIA concessions would have to be extended
to all countries within, say, five years. This is effectively a ban on
RIAs and certainly precludes any gains they might offer in terms
of deep integration or nation building. It is not a serious con-
tender.
2.  "Open access." Beginning at least with the U.S. submission to the
preparatory committee of the 1933 London Monetary and Eco-
nomic Conference, scholars and policymakers have argued that



254    REGIONAL INTEGRATION AND DEVELOPMENT
requiring RIAs to admit any country willing to accept their rules
both reduces the RIAs' adverse effects on excluded countries and
establishes a liberal dynamic (Viner 1950). While this may be true
if admission can be guaranteed, virtually every RIA extant has ge-
ographic restrictions on membership and also has features that re-
quire negotiation and so vitiate the promise of "open access" (see
box 8.4).
3. Prohibition. Bagwell and Staiger (1998) assert that FTAs neces-
sarily undermine the ability of the WTO to address the spillovers
from one country's trade policy to another's welfare through re-
ciprocal tariff negotiations. This might lead to fewer such negoti-
ations or to worse outcomes in those that occur. The implication
is that, as in Bhagwati's view, FTAs should be banned. (The case
against customs unions in this dimension is weaker, especially if
the partners have "similar" trade policy objectives.)
Procedural Fixes: The CRTA. The Committee on Regional Trading
Agreements (CRTA), which reports to the WTO's General Council, was
established in February 1996 to increase the transparency, efficiency, and
consistency of the WTO's treatment of RIAs. It was seen as a means of
ensuring more rigorous review of new RIAs, since a single group would
review all of them, using the same criteria and with more searching noti-
fication and information requirements. The CRTA would also undertake
periodic reviews of existing RIAs and could resolve some of the systemic
issues that remained after the Uruguay Round. The more thorough re-
view was seen as a route to better compliance with WTO requirements,
while the consideration of conceptual issues was a step toward refining
the rules and codifying them more precisely.
Unfortunately, the CRTA has still not got into its stride after six years.
Its assessments of particular cases have been stymied by the lack of clear
systemic rules, and its discussion of rules has been stalemated on exactly
the same "substantially all" and "other regulations" issues as bedeviled
the Uruguay Round discussions. By December 2000, the CRTA had initi-
ated consideration of 86 RIAs (including 32 inherited from previous
working parties). It had completed factual analysis of 62 of these but had
been unable "to conclude any examination referred to it" (report by the
chairman of the CRTA to the WTO General Council, December 2000).
The CRTA is the key to improved near-term management of RIAs by
the WTO. Although RIAs are open to dispute if third countries feel ag-
grieved (at least until the CRTA has formally certified their WTO-con-
formity), the rules on "serious" intent to integrate are not really vulnera-



TRADE BLOCS AND THE REST OF THE WORLD   255
ble to the dispute process. Members do not normally bring internal dis-
putes to the WTO, and third countries are unlikely to press RIAs to in-
clude more sectors if they expect this extension to increase trade diver-
sion.29 Similarly, why would a third country seek to free RIA members
from the threat of each other's antidumping legislation? Hence if issues
concerning the coverage of sectors or the depth of integration are missed
at the outset, they are missed forever.
Recently, however, an RIA did figure in the WTO's dispute settlement
process. As part of the EU-Turkey customs union, Turkey imposed quo-
tas on imports of certain textiles and clothing so that the EU's quotas
(under the Multifibre Arrangement) would not be undermined by de-
flected trade. India complained and won a ruling against the quotas on
the grounds that they were not fundamental to the customs union. The
dispute panel stated that alternative means should be found to meet
Turkey's objectives. These might include levying tariffs (for which India
could demand compensation), phasing out the EU restrictions, or using
rules of origin to control the movement of third-country goods from
Turkey to the EU. The ruling suggests a willingness to subordinate Arti-
cle XXIV to other parts of the GATT, which could have significant im-
plications for new customs unions or for expansions.
The WTO's rules on regionalism are far from perfect, but it is impos-
sible to devise rules that reliably sort beneficial from harmful RIAs. If
that process is to be followed at: all, it will have to be done case by case
by the CRTA. But the CRTA faces a serious timing problem. Unless
agreements are submitted to the WTO early in the process of negotia-
tion-when they would be very provisional-reviews will generally be
too late to influence their initial form or to affect public debate. If the re-
views then call for changes that would upset carefully negotiated com-
promises, they will be resisted and resented by members, which is bad
news for a consensual organization. Thus, considerable political courage
will be called for to enforce CRTA findings until the requirements are
sufficiently understood and respected by members to be met ab initio.
This process of review and response would be aided by detailed econom-
ic studies of RIAs, extending well beyond the legalities of GATF Article
XXIV and GATS Article V. But both immediately and ultimately, the re-
sponsibility for good RIAs lies with governments themselves.
Conclusion: Rules Are Not the Answer
Rules are not the answer, or even a large part of the answer, for ensuring
that regionalism will be beneficial. Rather, the need is for understanding.



256     REGIONAL INTEGRATION AND DEVELOPMENT
In the end, subject only to imperfect constraints on the effects on exclud-
ed countries, governments will act in what they believe to be their best
interests, and rules can help illuminate these interests by informing the
parties to the debate about received opinion and norms. But on topics
that are as case specific and sui generis as RIAs, advocates will always be
able to find plausible exceptions to norms, and no set of rules seems like-
ly to be very effective at distinguishing beneficial from harmful RlAs.
Thus, it seems unlikely that great efforts to rewrite the WTO's rules on
RIAs will be adequately rewarded.
None of the fundamental reforms to GATT Article XXIV and GATS
Article V proposed to date seems likely to command enough political
support to make any progress, and none of the minor ones seems worth
the effort. Accordingly, we offer the following recommendations:
*   Extend the disciplines of Article XXIV and Article V to RIAs be-
tween developing countries.
*   Enforce these disciplines rigorously in the CRTA, especially those
on the coverage and depth of liberalization.
*   Use the dispute settlement procedure to enforce third countries'
right not to face increases in protection, whether the increases are
direct, or indirect through the use of tools such as rules of origin.
Even these steps will require major expenditure of political capital by
many countries. We therefore suggest that continuing studies of particu-
lar RIAs be conducted, not just with respect to Article XXIV and Article
V disciplines but also in terms of the RIAs' overall effects on economic
performance. This will help policymakers and the public understand
what they can and cannot expect from regional integration.
NOTES
1. The term "most-favored-nation" arises from the long-standing trade diplo-
macy practice of guaranteeing a partner treatment at least as good as is granted
to the most favored partner. If this extends to all partners, all are treated equally
in a nondiscriminatory fashion. Universal MFN is the cornerstone of the GATT,
the treaty that has governed trade relations since 1947. But the term "most fa-
vored" is now ironic. Almost no country pays tariffs above MFN levels, and
most pay less via the generalized system of preferences (GSP), other unilateral
preferences, or formal RIAs. Only 10 countries pay MFN tariffs on all their ex-
ports to the EU. All others, except for the Democratic People's Republic of
Korea, which pays above MFN levels, receive some sort of preference (Winters
2000).



TRADE BLOCS AND THE REST OF THE WORLD         257
2. The 14 percent figure is what is left after allowing for imports from EU
members, RIA partners, trade under the GSP, and goods with zero MFN tariffs.
This is about 40 percent of the EU's imports from non-EU sources (Sapir 1998).
3. It is also possible that by Improving efficiency or generating economies of scale,
the RIA will drive down the prices of its exports (ROW imports). Then the rest of the
world benefits from the RIA, as Gasiorek, Smith, and Venables (1992) predicted for
the EU's Single Market Programme and Flores (1997) for MERCOSUR.
4. This result, of course, refers to total exports; if exports diverted from the
RIA can be sold elsewhere at the same price, no damage is done.
S. In any case, one feels uneasy criticizing an RIA because it reduces an ex-
porter's ability to levy taxes on members' imports.
6. Import taxes allow domestic producers to raise their prices and thus bid re-
sources away from exporters. Alternatively, trade may be thought of as the
process of swapping exports for imports: if the government requires 10 percent
of the imports to be turned over to it as tax, the effect is the same as if it required
10 percent of the exports to be turned over. Trade is discouraged, and both ex-
ports and imports fall.
7. The common approach to measuring the effects of RIAs on nonmembers is
to consider changes in ROW exports; this, as we have argued above, is at best an
indirect approach to measuring the welfare of the rest of the world.
8. Loosely, multilateral trade liberalization means that nearly all countries re-
duce barriers on imports from nearly all partners.
9. Some theorists have argued, albeit in very particular models, that regional-
ism has sprung from the success of multilateralism. For example, Ethier (1998)
assumes that as the world economy liberalizes, the incentive for any individual
developing country to liberalize also increases. He sees RIAs as the preferred
modality for this liberalization because, in comparison with nondiscriminatory
opening, they appear to allow greater confidence that FDI inflows will follow re-
form. Freund (1998) argues that as MFN tariffs come down, RIAs become easier
to sustain (in a sense defined in the section "Regionalism and Tariff Levels" in
this chapter), and trade diversion becomes less costly relative to the benefits of
increased competition.
10. On a broad historical canvas, of course, regionalism, which goes back to
at least 1815, predates multilateralism. But the current trading system based on
the GATT arose in 1947 from other forces and preceded significant regional
arrangements between independent powers by at least several years.
11. Levy (1995) shows theoretically how acquiring fringes of FTA partners
could weaken EU and U.S. interest in mutual negotiations.
12. The concept, although not the name, goes back at least to Hufbauer
(1989).
13. Frankel, Stein, and Wei's model is not at all robust; small asymmetries in
size or changes in parameter values can change the results. It is adequate, never-
theless, for demonstrating formally the interaction between RIAs.



258     REGIONAL INTEGRATION AND DEVELOPMENT
14. In November 2001, at the 14th Pacific Economic Co-operation Council
meeting in Hong Kong, Jim Sutton, trade negotiations minister of New Zealand
and a champion of open regionalism, said it was inconceivable that any bloc
would grant totally automatic right of entry to countries willing to abide by its
rules.
15. In the 1930s it was not only terms of trade disadvantages that worried
politicians but also simple mercantilist effects, as countries sought to reduce im-
ports competitively. Suppose that the United Kingdom aimed to reduce its im-
ports by £1 million. Under MFN rules, each of n equal partners might expect to
lose £(1/n) million in exports. But if m partners were exempt because they had an
RIA with Britain, each remaining partner would face losses of £1/(n - m) mil-
lion. If the "entrance fee" absorbed nearly all the benefit of accession, the fee
would increase from 1/n for the first to 1/(n - 1) for the second, and so on. The
incentive to be first is obvious.
16. For any individual facing given risks, having insurance is better than not,
but abolishing the risk and the insurance it necessitates would be better.
17. Oye (1992) argues that the 1930s also fit this pattern. He asserts that re-
gional arrangements such as U.S. bilateral arrangements under the Reciprocal
Trade Agreements Act were politically feasible because they almost guaranteed
export expansion in partner markets in return for import liberalization. In this
way they started to relax restrictions that were immune to multilateral efforts.
18. It now seems likely that Japan would also be unwilling to pursue unilater-
al approaches (Scollay and Gilbert 2001).
19. The rules of international commerce are embodied in three main agree-
ments: the General Agreement on Tariffs and Trade (GATT), the General Agree-
ment on Trade in Services (GATS), and the Agreement on Trade-Related Aspects
of Intellectual Property Rights (TRIPs). They are administered by the WTO, two
of whose principal tools are the Trade Policy Review Mechanism (TPRM) and
the Understanding on Dispute Settlement. The WTO has 145 members. The
principal nonmember economies, which are seeking accession, are Russia, Saudi
Arabia, and Ukraine; the countries of Eastern Europe and Central Asia constitute
the other main group of current candidates.
20. The parallel is that the law permits me to smoke, subject to some condi-
tions. That does not make smoking good for me, nor does it help me stop my
kids from smoking.
21. For example, if a country is harmed by another's breaking into its export
markets, there is-properly-no redress under the GATT.
22. For example, in reviewing the treaty, the head of the GATT secretariat ex-
pressed "the view, with which he thought there was no disagreement, that the in-
cidence of the common tariff was higher than that of the rates actually applied by
the member states at the time of entry into force of the Treaty of Rome" (GATT
document C/M/8, para. 6; cited in GATT 1994: 750).
23. Under GATT procedures, finding a party in violation of its obligations re-



TRADE BLOCS AND THE REST OF THE WORLD        259
quired unanimity (and therefore the consent of the violating party). This is not
true of the WTO.
24. Within the GATT, there was a feeling that the article had influenced the
structure of U.S.-Canadian and U.S.-Israeli agreements (personal communication
from a senior GATT staff member, 1999). We can also identify cases in which
WTO rules, or their equivalent, have prevented RIAs. For example, in 1932 the
United Kingdom and the United States refused to waive their MFN rights, pre-
venting the implementation of the Ouchy Convention, a forerunner of the
Benelux customs union (Viner 1950). Similarly, negotiators of the draft Multilat-
eral Agreement on Investment (MAI) found no way of preventing some conces-
sions on services among members from also applying to nonmembers via the
GATS MFN clause, and they therefore held back such concessions. GATS Article
V permits regional arrangements, but the MAI was far too narrowly defined to
qualify as one.
25. There is also an unresolved dispute about whether Article XXIV can be
applied to an arrangement notifiecl under the Enabling Clause, as the United
States demanded with respect to MERCOSUR.
26. Together, these requirements seem to impose no discipline on the sectors
that are excluded from the RIA, although these sectors may still be covered by
the members' GATS obligations.
27. Schott (1989) offered an early analysis of these sorts of points.
28. WTO document JOB (99)/4797/Rev.2.
29. The United States and its partners still use the WTO dispute settlement
procedure, but there has never been a formal GATT dispute between the EU and
any country with which it has a formal RIA.






CHAPTER 9
Rules of Thumb for Regionalism
Experience shows that sound international trade policy is a fundamental
requirement for economic development. Under almost any circum-
stances, sound policy includes maintenance of open borders for trade in
goods and services and for foreign direct investment. This openness is
best achieved by reducing barriers to trade and investment evenhandedly
on all partners, through nonpreferential trade liberalization. Experience
shows, too, that unilateral trade liberalization is desirable whether or not
trading partners reduce their barriers in return. But political pressures
can push governments toward regionalism (preferential trade liberaliza-
tion), and so in some circumstances RIAs may be appropriate solutions
to national policy needs. For example, RIAs may confer credibility on
policy regimes, help to solve political problems, or increase competition.
Both when RIAs are desirable for some positive purpose and when
they arise for essentially political reasons, there are strong returns to en-
suring that they are set up with economic efficiency in mind. Even if their
objective is ostensibly noneconomic (say, enhancing national security), it
is important to minimize the economic price paid for such gains. Unfor-
tunately, it is impossible to devise universal policy guidelines for applica-
tion to every single RIA because so much depends on the circumstances
of each case. We can, however, derive a number of concrete and fairly ro-
bust rules of thumb that have wide application and provide a useful
benchmark against which to measure specific plans. In what follows,
these rules of thumb are listed under the eight central messages that
emerge from the analysis in the foregoing chapters.
261



262    REGIONAL INTFGRATION AND DEVELOPMENT
Message 1: Use RlAs as a way of fostering competition.
If a developing country government wishes to participate in a RIA, it
should use it as a procompetitive instrument. Governments should ask
whether the RIA will foster greater competition in domestic markets and
should focus on incorporating provisions that will help achieve this
outcome.
To increase the likelihood that this criterion is satisfied:
1. Minimize the extent of discrimination against nonmembers by
using the opportunity provided by the RIA to lower protection
against the rest of the world-unilaterally, if necessary. This is of
particular importance for small countries.
2. Resist pressure from sectoral lobbies to favor trade-diverting
RIAs. Minimize exceptions to the coverage of regional free trade,
and include services.
3. Go beyond the abolition of formal trade barriers and reduce the
invisible" barriers associated with the enforcement of regulatory
and procedural requirements.
4. Include formal commitments granting the right of establishment
and national treatment to foreign direct investors, including those
in services markets.
Message 2: North-South dominates South-South.
Not all partners are equal. RlAs with high-income countries are more
likely to generate significant economic gains than are those with poorer
ones.
1. Developing country governments that want to take part in RIAs
should generally seek to form them with large, rich neighboring
countries or blocs. By virtue of their size and (usually) efficiency,
such partners are more likely to increase domestic competition
and to be better sources of trade and of FDI-related technol-
ogy transfers. Provided that the partners are genuinely inter-
ested in the performance of the developing country, richer part-
ners will generate more policy and political credibility than
poorer ones.
2. If rich partners are already providing largely duty-free access, a
North-South RIA may give rise to large transfers to northern pro-
ducers, which will directly reduce the developing partner's wel-
fare. Care should be taken to ensure that corresponding benefits



RUILES OF THUMB FOR REGIONALISM   263
such as credibility, financial transfers, and access to the partner's
more sensitive markets are locked in before committing to the
RIA.
3. To lessen the size of any resource transfers, governments should
reduce the average level of tariffs and other trade barriers that are
maintained against nonmembers.
4. RIAs between small economies are likely to be trade diverting. If
they are pursued, particular attention should be paid to securing
the benefits of creating larger "home" (regional) markets by in-
creasing intrabloc competition. Attention should also be given to
lowering protection against the rest of the world in order to ob-
tain the benefits of greater extrabloc competition.
5. RIAs between poor economies could cause divergence in income
between partners, which could harm some partners (typically, the
less developed and poorer ones) absolutely and heighten political
tensions between members. These effects could be mitigated by
reducing trade barriers against nonmembers and by including
provisions for compensation for tariff revenue losses.
Message 3: Credibility gains require explicitness.
RIAs can enhance the credibility of economic and political reform pro-
grams, but generally only if they explicitly include provisions and mech-
anisms that directly affect the policies of interest.
Governments engaging in RIAs are advised to:
1. Fully specify the transition path to regional free trade (what will
be reduced or eliminated, and when) and minimize the probabili-
ty of reversals of liberalization by prohibiting the application of
instruments of contingent protection.
2. Sign RIAs with partners large enough to enforce agreements and
close enough to wish to do so, and incorporate explicit language
in the RIA that requires action on the part of partner countries in
instances of noncompliance or backsliding.
3. Include binding dispute settlement mechanisms that are not
diplomatic in nature or conditional on foreign policy considera-
tions.
4. Keep in mind that economic credibility will be achieved only if the
RIA represents good policy in its own right. No RIA will stimu-
late investment if the funcdamentals, such as secure property rights
and good macroeconomic policies, are not in place.



264     REGIONAL INTEGRATION AND DEVELOPMENT
Message 4: Only efficient RIAs are likely to help politically.
RIAs can help solve political problems; but if they are economically
wasteful or divisive, they could have opposite effects.
1. Trade preferences may be useful for achieving political objectives
such as national security, but the optimal level of such preferences
declines over time.
2. Cooperation with others can help reinforce national sovereignty
where individual governments are too small to address problems
alone. This can apply both to technical matters and to pooling na-
tional efforts in international negotiating or regulatory fora. Be-
cause the optimal groupings for cooperation and pursuit of
shared interests will vary by issue, one cannot assume that every
RIA will automatically deliver these benefits.
Message 5: Regional cooperation does not generally require trade
preferences.
The existence of widespread intercountry spillovers calls for coopera-
tion between developing countries in areas other than trade policy,
such as regulatory reform and infrastructure provision. Usually,
however, these goals should be pursued independent of trade discrimti-
nation.
1. Governments should adopt good practices unilaterally through
the simplification of administrative requirements and recognition
of foreign regulatory regimes where this does not conflict with the
ability to achieve public interest objectives. Although the politics
may be easier if several countries reform in concert, strong re-
formers should avoid being tied indissolubly to slower ones.
2. If regional product standards or regulations for producers are re-
quired to offset market failures, international norms should be
adopted if these exist or can be developed. If regional norms are
necessary, they should be nondiscriminatory and open to all coun-
tries, wherever possible.
3. Where harmonization is necessary, partners should define mini-
mum, rather than universal, standards within an RIA and should
ensure that regulations are appropriate to all members' levels of
development and administrative capabilities.
4. Trade preferences might aid policy integration or regional cooper-
ation by putting more issues on the table or by providing an insti-
tutional framework for cooperation. Multilateral negotiations or



RULES OF THUMB FOR REGIONALISM   265
unilateral adoption of international norms may, however, be more
effective.
Message 6: Beware of transactions costs in operating RL4s.
Governments should consider carefully the transactions and implemen-
tation costs associated with different types of RIAs.
1. FTAs require costly rules of origin. Their costs can be reduced by
adopting the same rules as apply to nonpreferential trade and by
using the "change in tariff heading" approach instead of criteria
based on local content or value added.
2. Customs unions may be preferable to FTAs, but only if internal
borders are actually abolished (frequently, they remain) and the
common external tariff is low.
3. Customs unions' institutions for setting common trade policy
should be designed to reduce protectionist biases by, for example,
ensuring that a single political body is responsible for both the
gains and the losses generated by trade barriers.
Message 7: RlAs may have positive or negative fiscal implications.
The fiscal dimensions of RIAs are important for countries where trade
taxes generate a significant share of government revenue.
1. In economies with less-developed tax systems, trade taxes can
offer a relatively effective and uncorrupt means of raising rev-
enue. Trade taxes are not, however, a permanent solution to rev-
enue needs. The RIA should be used as an opportunity to develop
and reform domestic tax structures. Policymakers should ensure
that reasonably nondistortionary domestic taxes can compensate
for any decline in tariff revenues.
2. Customs unions may also ease collection costs by reducing the
scope for smuggling.
Message 8: Do not rely on the W7/0 to ensure that RIAs are
beneficial.
Countries should not rely on the WTO to ensure that RIAs are beneficial
to members and outsiders. The WTO forbids some destructive forms
of regionalism, but its main contribution toward constraining the poten-
tially negative implications of regionalism for nonmembers is as an
instrument for pursuing global liberalization on an MFN basis.



266    REGIONAL INTEGRATION AND DEVELOPMENT
1. Governments should monitor what happens after an RIA is im-
plemented and use this information as the basis for notification
and negotiation in the WTO. Surveillance efforts should include
price as well as quantity effects on a commodity-by-commodity
basis.
2. RIAs-the country's own and those of others-should be assessed
on the basis of national development objectives and criteria, not
according to whether they satisfy the relevant WTO articles.
3. The WTO review procedures and dispute settlement process
should be more vigorously used to enforce the existing rules and
contest the negative effects of protectionist dimensions of RIAs,
including rules of origin and the application of contingent pro-
tection.
4. The differences between the WTO's rules for developing and de-
veloped country RIAs should be eliminated. In particular, the
rules requiring nearly complete commodity coverage and no in-
creases in average protection levels should apply to all countries.



Appendix. Selected WTO Provisions on
Regional Integration Arrangements
GATT ARTICLE XXIV
4. The contracting parties . . . also recognize that the purpose of a
customs union or of a free trade area should be to facilitate trade
between the constituent territories and not to raise barriers to
trade.
5. (a) With respect to a customs union ... the duties and other reg-
ulations of commerce imposed at the institution . . . shall not
on the whole be higher or more restrictive than the general in-
cidence of the duties and regulations of commerce applicable
in the constituent territories prior to the formation of such
union ...
(b) With respect to a free-trade area . . . the duties and other reg-
ulations of commerce maintained in each of the constituent
territories and applicable at the formation of such free-trade
area . . . shall not be higher or more restrictive than the corre-
sponding duties and other regulations of commerce existing in
the same constituent territories prior to the formation of the
free-trade area ...
(c) Any interim agreement ... shall include a plan and schedule
for the formation of such a customs union or of such a free-
trade area within a reasonable length of time.
7. (a) Any contracting party deciding to enter into a customs
union or a free-trade area, . . . shall promptly notify the
CONTRACTING PARTIES and shall make available to them such
information ...
8. (a) A customs union shall be understood to mean the substitution
of a single customs territory for two or more customs territo-
ries, so that: (i) duties and other restrictive regulations of
commerce (except, where necessary, those permitted under
Articles XI, XII, XIII, XIV, XV and XX) are eliminated with
respect to . . . substantially all the trade in products originat-
ing in such territories . ..
8. (b) A free trade area shall be understood to mean a group of two
or more customs territories in which the duties and other re-
strictive regulations of commerce (except, where necessary,
those permitted under Articles XI, XII, XIII, XIV, XV and XX)
267



268    REGIONAL INTEGRATION AND DEVELOPMENT
are eliminated on substantially all the trade between the con-
stituent territories in products originating in such territories.
THE ENABLING CLAUSE (THE DECISION ON DIFFERENTIAL
AND MORE FAVORABLE TREATMENT, RECIPROCITY AND
FULLER PARTICIPATION OF DEVELOPING COUNTRIES)
1. Notwithstanding the provisions of Article I of the General Agree-
ment, contracting parties may accord differential and more favor-
able treatment to developing countries, without according such
treatment to other contracting parties.
2. The provisions of paragraph 1 apply to the following: . . .
(c) Regional or global arrangements entered into amongst less-
developed contracting parties for the mutual reduction or elimi-
nation of tariffs and, in accordance with criteria or conditions
which may be prescribed by the CONTRACTING PARTIES, for the mu-
tual reduction or elimination of non-tariff measures, on products
imported from one another; . . .
THE URUGUAY ROUND UNDERSTANDING ON THE
INTERPRETATION OF ARTICLE XXIV
2. The evaluation . . . of the duties and other regulations of
commerce . . . shall . . . be based upon an overall assessment of
weighted average tariff rates and of customs duties collected....
For this purpose, the duties and charges to be taken into consider-
ation shall be the applied rates of duty. It is recognized that for the
purpose of the overall assessment of the incidence of other regula-
tions of commerce for which quantification and aggregation are
difficult, the examination of individual measures, regulations,
products covered and trade flows affected may be required.
3. The "reasonable length of time" referred to in Article XXIV 5(c)
should exceed 10 years only in exceptional cases....
GATS ARTICLE V
1. This Agreement shall not prevent any of its Members from being
a party to or entering into an agreement liberalizing trade in



APPENDIX     269
services between or among the parties to such an agreement, pro-
vided that such an agreement:
(a) has substantial sectoral coverage,' and
(b) provides for the absence or elimination of substantially all dis-
crimination, in the sense of Article XVII, between or among
the parties, in the sectors covered under subparagraph (a),...
3.  (a) Where developing countries are parties to an agreement of the
type referred to in paragraph 1, flexibility shall be provided
for regarding the conditions set out in paragraph 1, particu-
larly with reference to subparagraph (b) thereof, in ac-
cordance with the level of development of the countries
concerned, both overall and in individual sectors and
subsectors ...
4. Any agreement referred to in paragraph 1 shall be designed to fa-
cilitate trade between the parties to the agreement and shall not in
respect of any Member outside the agreement raise the overall
level of barriers to trade in services within the respective sectors
or subsectors compared to the level applicable prior to such an
agreement.
NOTES
1. [The following note is in the original GATS text.] This condition is under-
stood in terms of number of sectors, volume of trade affected and modes of sup-
ply. In order to meet this condition, agreements should not provide for the a pri-
ori exclusion of any mode of supply.






Bibliography
The word processed denotes informally reproduced works that may not
be commonly available through libraries. Web addresses for works
available online as full texts are provided, if known, at the end of the
entry. Abstracts of CEPR Discussion Papers may be accessed at
<http://www.cepr.org/pubs/new- dps/dp_papers.htm>.
Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. "The Colonial
Origins of Comparative Development: An Empirical Investigation." Ameri-
can Economic Review 91 (5): 1369-1401.
Adams, Charles. 1993. For Good and Evil: The Impact of Taxes on the Course
of Civilization. Lanham, Md.: Madison Books.
Aitken, Brian J., and Ann E. Harrison. 1999. "Do Domestic Firms Benefit from
Direct Foreign Investment? Evidence from Venezuela." American Economic
Review 89 (3): 605-18.
Aitken, Brian J., Gordon H. Hanson, and Ann E. Harrison. 1997. "Spillovers,
Foreign Investment, and Export Behavior." Journal of International Econom-
ics 43 (1-2, August): 103-32.
Aitken, N. D. 1973. "The Effect ol the EEC and EFTA on European Trade: A
Temporal Cross-Section Analysis." American Economic Review 63 (5):
881-92.
Aitken, N. D., and W. R. Lowry. 1973. "A Cross-Sectional Study of the Effects of
LAFTA and CACM on Latin American Trade." Journal of Common Market
Studies 11 (4, June): 326-36.
Aitken, N. D., and R. S. Obutelewicz. 1976. "A Cross-Sectional Study of EEC
Trade with the Association of African Countries." Review of Economics and
Statistics 58 (4, November): 425-33.
Alburo, Florian A., Carlos C. Bautista, and Maria Socorro H. Gochoco. 1992.
"Pacific Direct Investment Flows into ASEAN." ASEAN Economic Bulletin
(March).
Alogoskoufis, G. 1995. "The Two Faces of Janus: Institutions, Policy Regimes
and Macroeconomic Performance in Greece." Economic Policy: A European
Forum (April): 147-84.
Amjadi, Azita, and L. Alan Winters. 1999. "Transport Costs and 'Natural' Inte-
gration in Mercosur." Journal of Economic Integration 14 (4, December):
497-521.
Anderson, Kym, and Hete Norheim. 1993. "History, Geography and Regional
Economic Integration." In Kym Anderson and Richard Blackhurst, eds.,
271



272     REGIONAI INTEGRATION AND D.VELOPMENT
Regional Integration and the Global Trading System, 19-51. New York: St.
Martin's Press.
Anderson, Svein S., and Kjell A. Eliassen. 1993. Making Policy in Europe: The
Eutropeification of National Policy-Making. London: Sage Publications.
Andriamananjara, Soamiely, and Maurice Schiff. 2001. "Regional Cooperation
among Microstates." Revieiv of International Economics 9 (1, February):
42-51.
Anwar, Dewi Fortuna. 1994. Indonesia in ASEAN: Foreign Policy and Regional-
isnii. New York: St. Martin's Press.
Arad, Ruth W., and Arye L. Hlillman. 1979. "Embargo Threat, Learning and De-
parture from Comparative Advantage." journal of International Economics 9
(May): 265-75.
Bagwell, Kyle, and Robert W. Staiger. 1998. "Will Preferential Trading Arrange-
ments Undermine the Multilateral Trading Svstem?" Economic Journal 108
(July): 1162-82.
. 1999. "An Economic Theory of GATT." American Economtic Re-
view 89 (1, March): 215-48.
Balassa, Bela. 1974. "Trade Creation and Trade Diversion in the European
Common Market: An Appraisal of the Evidence." Manchester School of Eco-
inomic and Social Studies 42: 93-135.
1989. "Tariff Reductions and Trade in Manufactures among the
Industrial Countries." In Bela Balassa, Comparative Advanitage, Trade Policy
and Economic Development. 13 t-39. ILondon: Harvester Wheatsheaf.
Balasubramanyam, Venkataraman N., Mohammed A. Salisu, and David Saps-
ford. 1996. "Foreign Direct Investment and Growth in EP and IS Countries."
Econonic Jouriial 106 (January): 92-105.
Baldwin, Richard E. 1989. "The Growth Effects of 1992." Economic Policy: A
European Forum 4 (October): 248-81.
. 1992. "Measurable Dynamic Gains from Trade." Journal of Politi-
cal Econonmy 100 (1, February): 162-74.
1993. "On the Measurement of Dynamic Effects of Integration."
Empirica-Austrian Economic Papers 20 (2): 129-45.
. 1994. Toivards an Integrated Europe. London: Centre for
Economic Policy Research.
1995. "A Domino Theorv of Regionalism." In Richard E. Baldwin,
Pertti Haaparanta, and Jaakko Kiander, eds., Expandintg Membership of the
European Union. Cambridge, U.K.: Cambridge University Press.
1997. "The Causes of Regionalism." World Econoniy 20 (7):
865-88.



BIBLIOGRAPHY     273
Baldwin, Richard E., and Rikard Forslid. 1996. "Trade Liberalization and
Endogenous Growth: A q-Theory Approach." NBER Working Paper 5549.
National Bureau of Economic Research, Cambridge, Mass. Available at
<http://papers.nber.org/papers/w5549>.
Baldwin, Richard E., and Elena Seghezza. 1996. "Testing for Trade-Induced
Investment-Led Growth." CEPR Discussion Paper 1331. Centre for Econom-
ic Policy Research, London.
Baldwin, Richard E., and Anthony J. Venables. 1995. "Regional Economic Inte-
gration." In Gene Grossman and Kenneth Rogoff, eds., Handbook of Inter-
national Economics, vol. 3: 1597-1644. Amsterdam: North-Holland.
Baldwin, Richard E., Rikard Forslid, and Jan 1. Haaland. 1996. "Investment Cre-
ation and Diversion in Europe." World Economy 19 (6, November): 635-59.
Baldwin, Richard E., Joseph F. Francois, and Richard Portes. 1997. "The Costs
and Benefits of Eastern Enlargement: The Impact on1 the EU and Central Eu-
rope." Economic Policy: A European Forum 24 (April): 125-76.
Bandyopadhyay, Subhayu, and Howard J. Wall. 1999. "Customs Union or Free
Trade Area? The Role of Political Asymmetries." Review of International
Economics 7 (4): 665-72.
Barrett, Scott. 1994 "Conflict and Cooperation in Managing Interniational Wa-
ter Resources." Policy Research Working Paper 1303. World Bank, Policy
Research Department, Public Economics Division, Washington, D.C.
Processed.
Barros, Pedro. 1995. "Conduct Effects of Gradual Entry Liberalization in Insur-
ance." Journal of Regulatory Economics 8: 45-60.
Basevi, Georgio, Flavio Delbono, and Mario Mariotti. 1994. "Bargaining with a
Composite Player: An Application to the Uruguay Round of GATT Negotia-
tions." Journal of International and Comparative Economics 3: 161-74.
Bayoumi, Tamim, and Barry Eichengreen. 1997. "Is Regionalism Simply a Diver-
sion? Evidence from the Evolution of the EC and EFTA." In Taketoshi Ito and
Anne 0. Krueger, eds., Regionalism versus Multilateral Trade Arrangements.
NBER-East Asia Seminar on Economics, vol. 6. Chicago: University of
Chicago Press.
Ben-David, Dan. 1993. "Equalizing Exchange: Trade Liberalization and Eco-
nomic Convergence." Quarterly Journal of Economics 108 (3, August):
653-79.
. 1994. "Income Disparity among Countries and the Effects of Freer
Trade." In Luigi L. Pasinetti and Robert M. Solow, eds., Economic Growth
and the Structure of Long-Term Development, 45-64. London: Macmillan.
. 1996. "Trade and Convergence among Countries." Journal of
International Economics 40: 297-98.



274     REGIONAL INTEGRATION AND DEVELOPMENT
. 1998. "Convergence Clubs and Subsistence Economies." Joturnal of
Development Economics 55: 155-57.
Bergsten, C. Fred. 1996. "Globalizing Free Trade." Foreign Affairs 75
(May-June): 105-20.
. 1997. "Open Regionalism." World Economy 20: 545-65.
Bernal, Richard L. 1998. The Integration of Small Econonites in the Free Trade
Area of the Americas. Policy Papers on the Americas 9. CSIS Americas Pro-
gram. Washington, D.C.: Center for Strategic and International Studies.
Bhagwati, Jagdlsh. 1987. "VERs and Quid Pro Quo DFI and VIEs: Political-
Economy-Theoretic Analysis." International Economic journal 1: 1-4.
1991. The World Trading System at Risk. Princeton, N.J.: Princeton
University Press.
. 1993. "Regionalism and Multilateralism: An Overview." In Jaime
de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integra-
tion, 22-57. New York: Cambridge University Press.
Bhagwati, Jagdish, and Arvind Panagariya. 1996. "Preferential Trading Areas
and Multilateralism: Strangers, Friends, or Foes?" In Jagdish N. Bhagwati
and Arvind Panagariya, eds., The Economics of Preferential Trading Agree-
ments. Washington, D.C.: AEI Press.
Bhagwati, Jagdish, and T. N. Srinivasan. 1996. "Does Environmental Diversity
Detract from the Case for Free Trade?" In jagdish Bhagwati and Robert E.
Hudec, eds., Fair Trade and Harmonization: Prerequisites for Free Trade?
Cambridge, Mass.: MIT Press.
. 1999. "Outward-Orientation and Development: Are Revisionists
Right?" Yale University Economic Growth Center Discussion Paper 806.
New Haven, Conn.
Bhalla, A. S., and P. Bhalla. 1997. Regional Blocs: Btiilding Blocks or Stumbling
Blocks? New York: St. Martin's Press.
Bilal, Sanoussi. 1998. "Why Regionalisn May Increase the Demand for Trade
Protection." Journal of Economic Integration 13: 30-61.
Blanchard, Oliver Jean, and Lawrence F. Katz. 1992. "Regional Evolutions."
Brookings Papers on Econontic Activity 1: 1-61.
Blomstr6m, Magnus, and Ari Kokko. 1997. "Regional Integration and Foreign
Direct Investment: A Conceptual Framework and Three Cases." Policy
Research Working Paper 1750. World Bank, International Trade
Division, Washington, D.C. Available at <http://econ.worldbank.org/re-
source.php?type=5>.
. 1998. "Foreign Investment as a Vehicle for International Tech-
nology Transfer." In Giorgio Barba Navaretti, Partha Dasgupta, Karl G.



BIBLIOGRAPHY     275
Maler, and Domenico Siniscalco, eds., Creation and Transfer of Knowledge:
Institutions and Incentives. Heidelberg and Berlin: Springer Verlag.
Bond, Eric W. 1997. "Competition Policy in Customs Unions: Theory and an Ex-
ample from U.S. History." Pennsylvania State University, University Park.
Processed.
Bond, Eric W., and Constantinos Syropoulos. 1996. "The Size of Trading Blocs:
Market Power and World Welfare Effects." Journal of International Econom-
ics 40: 411-37.
Bond, Eric W., Constantinos Syropoulos, and L. Alan Winters. 2001. "Deepen-
ing of Regional Integration and Multilateral Trade Agreements." Journal of
International Economics 53 (2, April): 335-61.
Brada, Josef C., and Jose A. Mendez. 1988. "An Estimate of the Dynamic Effects
of Economic Integration." Review of Economic and Statistics 70 (February):
163-68.
Byron, Jessica. 1994. "CARICOM in the Post-Cold War Era: Regional Solutions
or Continued Regional Contradictions?" Working Paper Series 178. Institute
of Social Studies, The Hague.
Campa, J. M., and T. L. Sorenson. 1996. "Are Trade Blocs Conducive to Free
Trade?" Scandinavian Journal of Economics 98 (2): 263-73.
Catinat, Michael, and Alexander Italianer. 1988. "Completing the Internal Mar-
ket: Primary Microeconomic Effects and Their Implementation in Macro-
econometric Models." Commission of the European Communities, Brussels.
CEA (Council of Economic Advisers). 1995. Council of Economic Advisers An-
nual Report. Washington, D.C.
CEC (Commission of the European Communities). 1988. "The Economics of
1992." European Economy 35. Luxembourg.
. 1996. European Economy. Luxembourg.
Cecchini, Paolo, with Michael Catinat and Alexis Jacquemin. 1988. The Euro-
pean Challenge, 1992: The Benefits of a Single Market. Aldershot, Hants,
U.K.; Brookfield, Vt.: Gower.
Chan, Steve. 1984. "Mirror, Mirror on the Wall . . . Are the Freer Countries
More Pacific?" Journal of Conflict Resolution 28: 617-48.
Chang, Won, and L. Alan Winters. 2002. "How Regional Blocs Affect Excluded
Countries: The Price Effects of MERCOSUR." American Economic Review
92 (4, September): 889-904.
Coe, David T., and Elhanan Helpman. 1995. "International R&D Spillovers."
European Economic Review 39: 859-87.
Coe, David T., and Alexander W. F[offmaister. 1999. "Are There International
R&D Spillovers among Randomly Matched Trade Partners? A Response to



276     REGIONAL INTEGRATION AND DEVELOPMENT
Keller." IMF Working Paper WP/99/18. International Monetary Fund, Wash-
ington, D.C. Available at <http://www.imf.org/external/pubind.htm>.
Coe, David T., Elhanan Helpman, and Alexander W. Hoffmaister. 1997. "North-
South R&D Spillovers." Economic Journal 107 (January): 134-49.
Collier, Paul. 1996. "The Role of the African State in Building Agencies of
Restraint." In Mats Lundahl and Benno J. Ndulu, eds., New Directions tn
Development Economics: Growth, Environmental Concerns, and Govern-
ment in the 1990s, 282-98. Studies in Development Economics, vol. 3. Lon-
don and New York: Routledge.
Collier, Paul, Patrick Guillaumont, Sylviane Guillaumont, and Jan Willem Gun-
ning. 1997. "The Future of Lome: Europe's Role." World Economy 20 (3):
285-306.
Cooper, C. A., and B. E Massell. 1965. "Toward a General Theory of Customs
Unions for Developing Countries." Journal of Political Economy 73: 256-83.
Cordoba, J. 1996. "Rules of Origin in Free Trade Agreements." World Bank,
Latin America and the Caribbean Department, Washington, D.C. Processed.
Council of the European Union. 1998. "Negotiating Directives for the Negotia-
tion of a Development Partnership Agreement with the ACP Countries." In-
formation Note 10017/98. European Union, Brussels.
Daniel, Timothy P., and Andrew N. Kleit. 1995. "Disentangling Regulatory Pol-
icy: The Effects of State Regulations on Trucking Rates." Journal of Regula-
tory Economics 8: 267-84.
Dee, Philippa, Alexis Hardin, and Michael Schuele. 1998. "APEC Early Volun-
tary Sectoral Liberalization." Productivity Commission Staff Research Paper.
Auslnfo, Canberra.
de Melo, Jaime, Claudio Montenegro, and Arvind Panagarlya. 1993. "L'intgra-
tion regionale hier et aujourd'hui." Revue d'Economie du Developpement 1
(1-2, February): 1-49. Also published as "Regional Integration, Old and
New." Policy Research Working Paper 985. World Bank, Country Economics
Department, Washington, D.C., 1992, processed.
de Melo, Jaime, Arvind Panagariya, and Dani Rodrik. 1993. "Regional Integra-
tion: An Analytical and Empirical Overview." In Jaime de Melo and Arvind
Panagariya, eds., New Dimensions in Regional Integration. New York: Cam-
bridge University Press.
DeRosa, Dean A. 1995. Regional Trading Arrangements Among Developing
Countries: The ASEAN Example. Research Report 103. International Food
Policy Research Institute.
Dickens, W. T., and L. F. Katz. 1987. "Inter-Industry Wage Differences and
Industry Characteristics." In Kevin Lang and Jonathan S. Levin, eds., Unem-



BIBLIOGRAPHY     277
ployment and the Structure of Labor Markets, 48-89. Oxford, U.K.:
Blackwell.
Di Leva, Charles. 1998. "International Environmental Law and Development."
Georgetown International Environmental Law Review 10 (2): 501-49.
Dinar, Ariel, and Aaron Wolf. 1997. "Economic and Political Considerations in
Regional Cooperation Models." Agricultural and Resource Economics
Review 26 (1, April): 7-22.
Dissou, Yasid. 2002. "Dynamic Effects in Senegal of Regional Trade." Review of
International Economics 10 (1): 177-99.
Dixit, Avinash. 1984. "International Trade Policies for Oligopolistic Industries."
Economic Journal 94: 1-16.
. 1988. "Anti-Dumping and Countervailing Duties under Oligop-
oly." European Economic Review 32: 55-68.
Djankov, Simeon, and Bernard Hoekman. 1997. "Trade Reorientation and Pro-
ductivity Growth in Bulgarian Enterprises." Policy Research Working Paper
1707. World Bank, International Economics Department, International Trade
Division, Washington, D.C. Available at <http://econ.worldbank.org/re-
source.php?type=5>.
. 2000. "Foreign Investment and Productivity Growth in Czech En-
terprises." World Bank Economic Review 14 (1): 49-64.
Dollar, David. 1992. "Outward-Oriented Developing Economies Really Do
Grow More Rapidly: Evidence from 95 LDCs, 1976-1985." Economic De-
velopment and Cultural Change 40 (April): 523-44.
Doyle, Michael. 1986. Empires. Ithaca, N.Y.: Cornell University Press.
Duchene, Francois. 1994. Jean Monnet: The First Statesman of Interdependence.
New York: Norton.
Echavarria, Juan Jose. 1997. "Trade Flows in the Andean Pact Countries:
Unilateral Liberalization or Regional Preferences?" In Shahid Javed Burki,
Guillermo E. Perry, and Sara Calvo, eds., "Trade: Towards Open Regional-
ism. Annual World Bank Conference on Development in Latin America and
the Caribbean 1997." World Bank, Washington, D.C.
. 1998. "Trade Flows in the Andean Countries: Unilateral Liberaliza-
tion or Regional Preferences?" Organization of American States, Trade Unit,
Washington, D.C. Processed.
Edwards, Sebastian. 1993. "Openness, Trade Liberalization, and Growth in De-
veloping Countries." Journal of Economic Literature 31 (3, September):
1358-93.
. 1998. "Openness, Productivity and Growth: What Do We Really
Know?" EconomicJournal 108 (447, March): 383-98.



278     REGIONAL INTEGRATION AND DEVELOIPMENT
Eichengreen, Barry, and J. Frankel. 1995. "Economic Regionalism: Evidence
from Two 20th Century Episodes." North American Journal of Economics
and Finance 6: 89-106.
Emerson, Michael. 1988. The Economics of 1992: The E.C. Commissiol's As-
sessment of the Economic Effects of Completinig the Internal Market. Oxford,
U.K.: Oxford University Press.
Engerman, Stanley L., and Kenneth L. Sokoloff. 1997. "Factor Endowments, In1-
stitutions, and Differential Paths of Growth among New World Economies."
In Stephen Haber, ed., How Latin America Fell Behind, 260-304. Stanford,
Calif.: Stanford University Press.
Ethier, W. J. 1998. "Regionalism in a Multilateral World." Journal of Political
Economy 106 (6, December): 1214-45.
European Commission. 1996. "Economic Evaluation of the Internal Market."
Ezuropean Economy. Reports and Studies 4. Luxembourg.
. 1997. The Single Market Review. Subseries Ill: The Dismantling of
Barriers. Vol. 3: Custonis and Fiscal Formnalities at Frontiers. Luxembourg.
.1998. The Single Market Review. Subseries IV: Im?pact on Trade and
Investme7zt. Vol. 4 (1): Foreign Direct Investnient. Luxembourg.
Evenett, Simon J., and Bernard Hoekman. 1999. "Government Procurement Dis-
crimination and International Trade." World Bank, Development Research
Group, Trade, Washington, D.C. Processed.
Fernandez, Raquel, and Jonathan Portes. 1998. "Returns to Regionalism: An
Analysis of Nontraditional Gains from Regional Trade Agreements." World
Bank Economic Review 12 (2, May): 197-220.
Finger, J. Michael, ed. 1993a. Antidunipin7g: How It Works and Who Gets Hurt.
Ann Arbor: University of Michigan Press.
I. 1993b. "GATT's Influence on Regional Arrangements." In Jaime de
Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration.
New York: Cambridge University Press.
Flores, R. 1997. "The Gains from MERCOSUR: A General Equilibrium, Imper-
fect Competition Evaluation." Journal of Policy Modeling 19 (1, February):
1-18.
Foroutan, Faezeh. 1996. "Turkey, 1976-85: Foreign Trade, Industrial Productiv-
ity, and Competition." In Mark J. Roberts and James R. Tybout, Industrial
Evolution in Developing Coun?tries: Micro Patterns of Turnover, Productivity,
and Market Structure, 314-37. New York: Oxford University Press.
. 1998. "Does Membership in a Regional Preferential Trade Arrange-
ment Make a Country More or Less Protectionist?" World Economny 21
(May): 305-36.



BIBLIOGRAPHY     279
Francois, Joseph F. 1997. "External Bindings and the Credibility of Reform." In
Ahmed Galal and Bernard Hoekman, eds., Regional Partners in Global Mar-
kets: Limits and Possibilities of the Euro-Med Agreements. London: Centre
for Economic Policy Research.
Francois, Joseph F., and C. R. Shiells, eds. 1994. Modeling Trade Policy: Applied
General Equilibrium Assessments in North American Free Trade. Cambridge,
U.K.: Cambridge University Press.
Francois, Joseph F., David Palmeter, and Douglas Nelson. 1997. "Government
Procurement in the U.S.: A Post-Uruguay Round Analysis." In Bernard Hoek-
man and Petros Mavroidis, eds., Law and Policy in Public Purchasing: The
WTO Agreement on Government Procurement. Ann Arbor: University of
Michigan Press.
Frankel, Jeffrey A., ed. 1998. The Regionalization of the World Economy. Chica-
go: University of Chicago Press.
Frankel, Jeffrey A., and David Romer. 1999. "Does Trade Cause Growth?"
American Economic Review 89 (3, June): 379-99.
Frankel, Jeffrey A., and Andrew K. Rose. 2000. "Estimating the Effect of Cur-
rency Unions on Trade and Output." NBER Working Paper 7857. National
Bureau of Economic Research, Cambridge, Mass. Available at <http://
papers.nber.org/papers/w7857>.
Frankel, Jeffrey A., and Shang-Jin Wei. 1998. "Regionalization of World Trade
and Currencies: Economics and Politics." In Jeffrey A. Frankel, ed., The
Regionalization of the World Economy. Chicago: University of Chicago Press.
Frankel, Jeffrey A., Ernesto Stein, and Shang-Jin Wei. 1997. Regional Trading
Blocs in the World Economic System. Washington, D.C.: Institute for Interna-
tional Economics.
Freund, Caroline L. 1998. "Multilateralism and the Endogenous Formation of
PTAs." International Finance Discussion Paper 614. Board of Governors of
the Federal Reserve System, Washington, D.C.
Fujita, Masahisa, Paul Krugman, and Anthony J. Venables. 1999. The Spatial
Economy: Cities, Regions, and International Trade. Cambridge, Mass.: MIT
Press.
Fukase, Emiko, and Will Martin. 2000. Free Trade Membership as a Stepping
Stone to Development: The Case of ASEAN. World Bank Discussion Paper
421. Washington, D.C.
Fuleihan, Basil. 1997. "The Impact of EU Association on Lebanon's Fiscal Posi-
tion." In Wassim Shahin and Kamal Shehadi, eds., Pathways to Integration:
Lebanon and the Euro-Mediterranean Partnership, ch. 6. Beirut: Lebanese
Center for Policy Studies.



280     REGIONAL INTEGRATION AND DEVELOPMENT
Gasiorek, Michael, Alasdair Smith, and Anthony J. Venables. 1992. "'1992':
Trade and Welfare-A General Equilibrium Model." In L. Alan Winters, ed.,
Trade Flows and Trade Policy after 1992. Cambridge, U.K.: Cambridge Uni-
versity Press.
Gatsios, Konstantine, and Larry Karp. 1991. "Delegation Games in Customs
Unions." Review of Economic Studies 58 (2): 391-97.
. 1995. "Delegation in a General Equilibrium Model of Customs
Unions." European Economiiic Review 39 (2): 319-33.
GATT (General Agreement on Tariffs and Trade). 1994. Analytical Index: Guide
to GATT Law and Practice. 6th ed. Geneva.
Graham, Edward M., and J. David Richardson, eds. 1997. Global Competition
Policy. Washington, D.C.: Institute for International Economics.
Greenwold, S., and A. Cox. 1993. "The Legal and Structural Obstacles to Free
Trade in the United States Procurement Market." Public Procurement Latv
Review 2: 237-52.
Grossman, Gene M., and Elhanan Helpman. 1995. "The Politics of Free-Trade
Agreements." American Economic Review 85 (4): 667-90.
Gupta, Anju, and Maurice Schiff. 1997. "Outsiders and Regional Trade Agree-
ments among Small Countries." Policy Research Working Paper 1847. World
Bank, Development Research Group, Washington, D.C. Available at
<http://econ.worldbank.org/resource.php?type=5>.
Haaland, Jan I., and Victor D. Norman. 1992. "Global Production Effects of Eu-
ropean Integration." In L. Alan Winters, ed., Trade Flows and Trade Policy
after 1992. Cambridge, U.K.: Cambridge University Press.
Haddad, Mona, and Ann Harrison. 1993. "Are There Positive Spillovers from
Direct Foreign Investment? Evidence from Panel Data for Morocco." lournal
of Development Economics 42 (1, October): 51-74.
Hall, R. E., and C. 1. Jones. 1999. "Why Do Some Countries Produce So Much
More Output per Worker Than Others?" Quarterly Journal of Economics
114 (1): 83-116.
Hamilton, C. B. 1988. "Restrictiveness and International Transmission of the
'New' Protectionism." In Robert E. Baldwin, Carl B. Hamilton, and Andre
Sapir, eds., Issues in U.S.-EC Trade Relations. Chicago: University of Chicago
Press.
Hansen, Niles M. 1969. "Regional Development and the Rural Poor." Joumral of
Human Resources 4 (2, spring): 205-14.
Hanson, Gordon H. 1996. "Economic Integration, Intraindustry Trade and
Frontier Regions." European Economic Review 40 (April): 941-49.
. 1997. "Increasing Returns, Trade and the Regional StructLre of
Wages." Economic Journal 107: 113-33.



BIBLIOGRAIPHY    281
1998. "Regional Adjustment to Trade Liberalisation." Regional Sci-
ence and Urban Economics 28: 419-44.
Hanson, Gordon H., and Ann Harrison. 1999. "Trade Liberalisation and Wage
Inequality in Mexico." Industrial and Labor Relations Review 52 (2):
271-88.
Harrison, Ann. 1994. "Productivity, Imperfect Competition and Trade Reform:
Theory and Evidence." Journal of International Economics 36 (1-2, Febru-
ary): 53-73.
- 1996. "Determinants and Effects of Direct Foreign Investment in
C6te d'lvoire, Morocco, and Venezuela." In Mark J. Roberts and James R.
Tybout, eds., Industrial Evolution in Developing Countries: Micro Patterns of
Turnover, Productivity, and Market Structure, 163-86. New York: Oxford
University Press.
Harrison, Glenn W., Thomas E Rutherford, and David G. Tarr. 1996. "Increased
Competition and Completion of the Market in the European Union: Static
and Steady State Effects." Journal of Economic Integration 11 (3, September):
332-65.
. 1997. "Trade Policy Options for Chile: A Quantitative Evaluation."
Policy Research Working Paper 1783. World Bank, Development Research
Group, Trade, Washington, D.C. Available at <http://econ.worldbank.org/
resource.php?type=5>.
. 2002. "Trade Policy Options for Chile: The Importance of Market
Access." World Bank Economic Review 16 (1): 49-79.
Hathaway, Dale E., and Merlinda D. lngco. 1996. "Agricultural Liberalisation
and the Uruguay Round." In Will Martin and L. Alan Winters, eds., The
Uruguay Round and the Developing Countries. Cambridge, U.K.: Cambridge
University Press.
Hathaway, Oona A. 1998. "Positive Feedback: The Impact of Trade Liberaliza-
tion on Industry Demands for Protection." International Organization 52
(summer): 575-612.
Henderson, David. 1999. The MAI Affair: A Story and Its Lessons. London:
Royal Institute for International Affairs.
Henrekson, Magnus, Johan Torstensson, and Rasha Torstensson. 1997.
"Growth Effects of European Integration." European Economic Review 41
(8, August): 1537-57.
Herin, Jan. 1986. "Rules of Origin and Differences between Tariff Levels in
EFTA and in the EC." EFTA Occasional Paper 13. Geneva.
Hillman, Arye L., and Ngo Van Long. 1983. "Pricing and Depletion of an Ex-
haustible Resource When There Is Anticipation of Trade Disruption." Quar-
terly Journal of Economics 98 (May): 215-33.



282     REGIONAL INTEGRATION AND DEVELOPMENT
Hindley, Brian, and Patrick Messerlin. 1993. "Guarantees of Market Access and
Regionalism." In Kym Anderson and Richard Blackhurst, eds., Regional Inte-
gration and the Global Trading System. New York: St. Martin's Press.
Hinojosa-Ojeda, Ra6il A., Jeffrey D. Lewis, and Sherman Robinson. 1995. "Re-
gional Integration Options for Central America and the Caribbean after
NAFTA." North American Journal of Economics and Finance 6 (2, fall):
121-48.
Hirschman, Albert 0. 1958. The Strategy of Economic Development. New
Haven, Conn.: Yale University Press.
Hoekman, Bernard. 2000. "The Next Round of Services Negotiations: Identify-
ing Priorities and Options." Federal Reserve Bank of St. Louis Review 82 (4,
July-August): 31-48.
Hoekman, Bernard, and Carlos A. Primo Braga. 1997 "Protection and Trade in
Services: A Survey." Open Economies Review 8: 285-308.
Hoekman, Bernard, and Simeon Dlankov. 1997. "Effective Protection and In-
vestment Incentives in Egypt and Jordan: Implications of Free Trade with Eu-
rope." World Development 25: 281-91.
Hoekman, Bernard, and D. Konan. 1999. "Deep Integration, Nondiscrimina-
tion, and Euro-Mediterranean Free Trade." In Jurgen von Hagen and Mika
Widgren, eds., Regionalism in Europe. Boston: Kluwer Academic.
Hoekman, Bernard, and Michael P. Leidy. 1993. "Holes and Loopholes in Inte-
gration Agreements: History and Prospects." In Kym Anderson and Richard
Blackhurst, eds., Regional Integration and the Global Trading System,
218-45. New York: St. Martin's Press.
Hoekman, Bernard, and Petros Mavroidis, eds. 1997. Law and Policy in Public
Purchasing: The WTO Agreement on Government Procurement. Ann Arbor:
University of Michigan Press.
Hoekman, Bernard, and Pierre Sauve. 1994. "Regional and Multilateral Liberal-
ization of Trade in Services: Complements or Substitutes?" Journal of Com-
mon Market Studies 32: 283-317.
Holmes, Peter. 1996. "Competition Policy and Integration: Levelling or Tilting
the Playing Field." Global Economic Institutions Working Paper 21. London:
Centre for Economic Policy Research.
Holmes, Peter, and Alasdair Smith. 1997. "Dynamic Effects of Regional Integra-
tion." In Regionalism and Development: Report of the European Comnis-
sion and World Bank Seminar (June 2, 1997). European Commission Studies
Series 1. Brussels.
Holmes, Thomas. 1998. "The Effect of State Policies on the Location of Manu-
facturing: Evidence from State Borders." Journal of Political Economy 106:
667-705.



BIBLIOGRAPHY     283
Horn, Henrik, and James Levinsohri. 1996. "Merger Policies and Trade Liberal-
ization." NBER Working Paper 6077. National Bureau of Economic
Research, Cambridge, Mass. Available at <http://papers.nber.org/papers/
w6077>.
Horn, Henrik, Harald Lang, and Stefan Lundgren. 1995. "Managerial Effort In-
centives, X-Inefficiency and International Trade." European Economic Re-
view 39 (1, January): 1 7-38.
Hsieh, Chang-Tai. 2001. "Comment on 'Trade Policy and Economic Growth: A
Skeptic's Guide to the Cross-National Evidence,' by Rodriguez and Rodrik."
In Ben S. Bernanke and Kenneth Rogoff, eds., NBER Macroeconomics Annu-
al 2000, 325-30. Cambridge, Mass.: MIT Press.
Hufbauer, G. C. 1989. The Free Trade Debate: Reports of the Twentieth Centu-
ry Fund Task Force on the Future of American Trade Policy. New York: Pri-
ority Press.
Hull, Cordell. 1948. The Memoirs of Cordell Hull. New York: Macmillan.
Hunter, Linda, James R. Markusen, and Thomas F. Rutherford. 1992. "U.S.-
Mexico Free Trade and the North American Auto Industry: Effects on the
Spatial Organisation of Production of Finished Autos." World Economy 15
(1, January): 65-8 1.
IADB (Inter-American Development Bank). 1995. "Economic Integration in the
Americas: Periodic Note on Integration." Washington, D.C. Available at
<http://www.iadb.orgfint/intpub/nota/homepage.htm>.
Inotai, Andras. 1991. "Regional Integration among Developing Countries, Re-
visited." Policy Research Working Paper 643. World Bank, Country Econom-
ics Department, Washington, D.C. Processed.
Irwin, Douglas A. 1993. "Multilateral and Bilateral Trade Policies in the World
Trading System: A Historical Perspective." In Jaime de Melo and Arvind
Panagariya, eds., New Dimensions in Regional Integration. New York: Cam-
bridge University Press.
Jacquemin, Alexis, and Andr6 Sapir. 1991. "Competition and Imports in the Eu-
ropean Market." In L. Alan Winters and Anthony J. Venables, eds., European
Integration: Trade and Industry. Cambridge, U.K.: Cambridge University
Press.
Johnson, H. G. 1965. "An Economic Theory of Protectionism: Tariff Bargaining
and the Formation of Customs Unions." Journal of Political Economy 73:
256-83.
Jones, Charles 1. 2001. "Comment on 'Trade Policy and Economic Growth: A
Skeptic's Guide to the Cross-National Evidence,' by Rodriguez and Rodrik."
In Ben S. Bernanke and Kenneth Rogoff, eds., NBER Macroeconomics Annu-
al 2000, 330-38. Cambridge, Mass.: MIT Press.



284     REGIONAL INTEGRATION AND DEVELOPMENT
Jones, Maldwyn A. 1983. The Limtits of Liberty. East Kilbride, Scotland:
Thomson.
Kahler, Miles. 1995. International Inistitutions and the Political Economy of In-
tegration. Washington, D.C.: Brookings Institution Press.
Kanbur, R., M. Keen, and S. van Wijnbergen. 1995. "Industrial Competitiveness,
Environmental Regulation and Direct Foreign Investment." In Ian Goldin and
L. Alan Winters, eds., The Economics of Sustainable Development, 289-302.
Cambridge, U.K.: Cambridge University Press.
Kant, Immanuel. 11795] 1992. Perpetual Peace: A Philosophical Essay. Bristol,
U.K.: Thoemmes.
Karras, Georgios. 1997. "Economic Integration and Convergence: Lessons from
Asia, Europe and Latin America." Journal of Economic Integration 12 (4):
419-32.
Kawai, Masahiro, and Shujiro Urata. 1996. "Trade Imbalances and Japanese
Foreign Direct Investment: Bilateral and Triangular Issues." Discussion Paper
Series F-52. Institute of Social Science, University of Tokyo.
Keller, Wolfgang. 1998. "Are International R&D Spillovers Trade-Related? Ana-
lyzing Spillovers among Randomly Matched Trade Partners." European Eco-
nomic Review 42 (8, September): 1469-81.
. 2002. "Geographic Localization of International Technology Dlffu-
sion." American Econoniic Revietw 92 (1, March): 120-42.
Kemp, M. C., and H. Wan, Jr. 1976. "An Elementary Proposition Concerning
the Formation of Customs Unions." Journal of International Economics 6:
95-97.
Kokko, Arn. 1996. "Productivity Spillovers from Competition between Local
Firms and Foreign Affiliates." Journal of International Development 8
(4, July-August): 517-30.
Konan, Denise E., anid Keith E. Maskus. 2000. "Joint Trade Liberalization and
Tax Reform in a Small Open Economiiy: The Case of Egypt." Journal of
Development Economics 61 (2): 365-92.
Krishna, Pravin. 1998. "Regionalism and Multilateralism: A Political Economy
Approach." QuarterlyJournal of Economics 113 (1, February): 227-51.
. Forthcoming a. "Are Regional Trading Partners 'Natural'?" Jourinal
of Political Econiomy.
. Forthcoming b. "Comment on 'Regionalism   and Economic
Development' by Anthony Venables." In R. Devlin, A. Estevadeordal, and
S. Evenett, eds., Regional Iuitegration and Trade in the Developmnent
Agenda. Brookings Institution and Inter-American Development Bank Con-
ference, May 31-June 1, 2001. Washington, D.C.: Brookings Institution
Press.



BIBLIOGRAPHY     285
Krishna, Pravin, and Devashish Mitra. 1998. "Trade Liberalization, Market
Discipline and Productivity Growth: New Evidence from India." Journal of
Developmenzt Economics 56 (2, August): 447-62.
Krueger, Anne 0. 1997. "Political Economy of Mexico's Entry into NAFTA." In
Takatoshi Ito and Anne 0. Krueger, eds., Regionalism versus Multilateral
Trade Arrangements. Chicago: University of Chicago Press.
Krugman, Paul R. 1991. "Is Bilateralism Bad?" In Elhanan Helpman and Assaf
Razin, eds., International Trade and Trade Policy. Cambridge, Mass.: MIT
Press.
. 1993. "Regionalism versus Multilateralism: Analytical Notes." In
Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional In-
tegration, 58-78. New York: Cambridge University Press.
Krugman, Paul R., and Raul Livas Elizondo. 1996. "Trade Policy and the Third
World Metropolis." Journal of Development Economics 49: 137-50.
Krugman, Paul, and Anthony J. Veriables. 1990. "Integration and the Competi-
tiveness of Peripheral Industry." CEPR Discussion Paper 363. Centre for Eco-
nomic Policy Research, London.
. 1996. "Integration, Specialization, and Adjustment." European
Economic Review 40 (3-5): 959-67.
Kugler, Maurice. 2001. "Sectoral Diffusion of Spillovers from Foreign Direct In-
vestment." Department of Economics Working Paper. University of
Southampton, U.K.
Lawrence, Robert Z. 1991. "Emerging Regional Arrangements: Building Blocs
or Stumbling Blocks?" In Richard O'Brien, ed., Finance and the International
Economy 5: The AMEX Bank Review Prize Essays, 23-3S. New York:
Oxford University Press.
. 1996. Regionalism, Multilateralism, and Deeper Integration. Wash-
ington, D.C.: Brookings Institution Press.
Lawrence, Robert Z., and Robert E. Litan. 1990. "The World Trading System
after the Uruguay Round." Boston University International Law Journal 8 (2,
fall): 247-76.
Lecomte, Henri Bernard Solignac. 1998. "Options for Future ACP-EU Trade Re-
lations." ECDPM Working Paper 60. European Centre for Development Pol-
icy Management, Maastricht, the Netherlands.
Levinsohn, James. 1993. "Testing the Imports-as-Market-Discipline Hypothe-
sis." Journal of International Economics 35 (1-2, August): 1-22.
Levy, Philip. 1995. "Free Trade Agreements and Inter-Bloc Tariffs." January.
Yale University, New Haven, Conn. Processed.
. 1997. "A Political Economy Analysis of Free-Trade Agreemenits."
American Economic Review 87 (4): 506-19.



286     REGIONAL INTEGRATION ANI) DEVELOPMENT
L6pez, Ram6n E., and Maurice Schiff. 1998. "Migration and the Skill-
Composition of the Labor Force: The Impact of Trade Liberalization in
LDCs." Canadian Journal of Economics 31 (2): 318-36.
Lucas, Robert E. 1988. "On the Mechanics of Economic Development." journal
of Monetary Economics 22 (July): 3-42.
Lumenga-Neso, Olivier, Marcelo Olarreaga, and Maurice Schiff. 2001. "On
'Indirect' Trade-Related Research and Development Spillovers." Policy
Research Working Paper 2580. World Bank, Development Research Group,
Trade, Washington, D.C. Available at <http://econ.worldbank.org/resource.
php ?type = 5>.
Machlup, Fritz. 1977. A History of Thouight on Economic Integration. New
York: Columbia University Press.
Magee, Stephen P., and Hak-loh Lee. 1997. "Tariff Creation and Tariff Diversion
in a Customs Union: The Endogenous External Tariff of the EEC,
1968-1983." Nota Di Lavoro 38 (97): 1-39. Fondazione Eni Enrico Mattei,
Milan.
Mansfield, Edward D. 1993. "Effects of International l'olitics on Regionalism
in International Trade." In Kym Anderson and Richard Blackhurst, eds.,
Regional Integration and the Global Trading Systein. New York: St. Martin's
Press.
Marshall, Alfred. 1920. Principles of Economics. London: Macmillan.
Martin, Will. 1996. "Assessing the Implications for Lebanon of Free Trade with
the European Union." World Bank, Washington, D.C. Processed.
McKibbin, Warwick J. 1994. "Dynamic Adjustment to Regional Integration: Eu-
rope 1992 and NAFTA." Journal of the Japanese and International Economy
8: 422-53.
McMillan, John. 1993. "Does Regional Integration Foster Open Trade? Eco-
nomic Theory and GATT's Article XXIV." In Kym Anderson and Richard
Blackhurst, eds., Regional Integration and the Global Trading System,
292-310. New York: St. Martin's Press.
McQueen, Matthew. 1998. "Lome versus Free Trade Agreements: The Dilemma
Facing the ACP Countries." World Economy 21 (4, June): 421-44.
Messerlin, Patrick A. 1983. "Bureaucracies and the Political Economy of Protec-
tion: Reflections of a Contineital European." Weltwvirtschaftliches Archiv
117: 468-96.
. 1997. "Competition Policy and Anti-Dumping Reform." In Jeffrey
J. Schott, ed., The World Trading System: Challenges Ahead. Washington,
D.C.: Institute for International Economics.
. 1998. "Technical Regulations and Industry Standards in the EU."
World Bank, Washington, D.C. Processed.



BIBLIOGRAPHY     287
2001. Measuring the Costs of Protection in Europe: European
Commercial Policy in the 2000s. Washington, D.C.: Institute for International
Economics.
Midelfart Knarvik, Karen Helene, Henry G. Overman, Stephen Redding, and
Anthony J. Venables. 1999. "The Location of Industry in Europe." Economic
Papers 142. European Commission, Brussels.
Milward, A. S. 1984. The Reconstruction of Europe, 1945-51. Berkeley:
University of California Press.
. 1992. The European Rescue of the Nation-State. Berkeley: Universi-
ty of California Press.
Milward, Alan S., and S. B. Saul. 1973. The Economic Development of Conti-
nental Europe, 1780-1870. Totowa, N.J.: Rowman and Littlefield.
Mohieldin, M. 1997. "The Egypt-EIJ Partnership Agreement and Liberalization
of Services." In Ahmed Galal and Bernard Hoekman, eds., Regional Partners
in Global Markets: Limits and Possibilities of the Euro-Med Agreements.
London: Centre for Economic Policy Research.
Morrison, A. R., and C. A. Zabin. 1994. "Two-Step Mexican Migration to the
United States: The Role of Mexican Export Agriculture." Tulane University,
Department of Economics, New Orleans, La. Processed.
Motta, Massimo, and George Norman. 1996. "Does Economic Integration
Cause Foreign Direct Investment?" International Economic Review 37 (4,
November): 757-83.
Moukarbel, Iskandar. 1997. "'Impact of the Free Trade Agreement with the Eu-
ropean Union Countries on the Lebanese Economy." In Wassim Shahin and
Kamal Shehadi, eds., Pathways to Integration: Lebanon and the Euro-
Mediterranean Partnership, ch. 7. Beirut: Lebanese Center for Policy Studies.
Murphy, Craig. 1994. International Organization and Industrial Change: Glob-
al Governance since 1850. New York: Oxford University Press.
Neven, Damien J. 1996. "Regulatory Reform and the Internal Market." In Or-
ganisation for Economic Co-operation and Development, Regulatory Reform
and International Market Openness. Paris.
Ng, Francis, and L. Alan Winters. 1998. "Tariff Reform in Senegal: A Simple
Numerical Analysis." World Bank, Development Research Department,
Washington, D.C.
Nishimizu, Mieko, and John M. Page, Jr. 1982. "Total Factor Productivity
Growth, Technological Progress, and Technical Efficiency Change: Dimen-
sions of Productivity Change in Yugoslavia, 1965-78." Economic Journal 92
(368, December): 920-36.
Nogues, Julio, and R. Quintanilla. [993. "Latin America's Integration and the
Multilateral Trading System." In Jaime de Melo and Arvind Panagariya, eds.,



288     REGIONAL INTEGRATION AND DEVELOPMENT
New Dimensions in Regional Integration, 278-313. New York: Cambridge
University Press.
Nordstrom, H. 1995. "Customs Unions, Regional Trading Blocs and Welfare."
In Richard Baldwin, Pertti Haaparanta, and Jaakko Kiander, eds., Expanding
Membership of the European Union, 54-78. Cambridge, U.K.: Cambridge
University Press.
North, Douglass C. 1990. Institutions, Institutional Change and Economic
Performance. Cambridge, U.K.: Cambridge University Press.
OECD (Organisation for Economic Co-operation and Development). 1995.
Regional Integration and the Multilateral Trading System: Synergy and Diver-
gence. Paris.
Olarreaga, Marcelo, and Isidro Soloaga. 1998. "Endogenous Tariff Formation:
The Case of Mercosur." World Bank Ecoiormic Review 12 (2, May):
297-320.
Olson, Mancur, Jr. 1996. "Big Bills Left on the Sidewalk: Why Some Nations Are
Rich and Others Poor." Journal of Economic Perspectives 10 (2): 3-24.
Ordover, J. A., and R. D. Willig. 1986. "Perspectives on Mergers and World
Competition." In Ronald E. Grieson, ed., Antitrust and Regulation. Lexing-
ton, Mass.: Lexington Books.
Ostry, Sylvia. 1998. Presentation to the First Academic Colloquium of the Amer-
icas, San Jose, Costa Rica.
Oye, Kenneth. 1992. Economic Discrimination and Political Exchange: World
Political Economy in the 1930s and 1980s. Princeton, N.J.: Princeton Univer-
sity Press.
Page, Sheila. 1996. "Intensity Measures for Regional Groups." Prepared for the
European Association of Development Research and Training Institutes
(EADI) Ninth General Conference, September 11-14, Vienna.
. 1998. "Regions and Developing Countries." Overseas Development
Institute, London. Processed.
Palmeter, N. David. 1993. "Pacific Regional Trade Liberalization and Rules of
Origin." Journal of World Trade 27 (5, October): 49-62.
Panagariya, Arvind. 1997. "The Meade Model of Preferential Trading: History,
Analytics and Policy Implications." In Benjamin J. Cohen, ed., International
Trade and Finance: New Frontiers for Research, 57-88. Cambridge, U.K.:
Cambridge University Press.
Panagariya, Arvind, and R. Findlay. 1996. "A Political-Economy Analysis of
Free-Trade Areas and Customs Unions." In Robert C. Feenstra, Gene M.
Grossman, and Douglas A. Irwin, eds., The Political Economy of Trade
Policy: Papers in Honor of Jagdish Bhagwati. Cambridge, Mass.; London:
MIT Press. Reprinted 1999 in Jagdish Bhagwati, Pravin Krishna, and Arvind



BIBLIOGRAPHY     289
Panagariya, eds., Trading Blocs: Alternative Approaches to Analyzing Prefer-
ential Trade Agreements, 335-56. Cambridge, Mass.: MIT Press.
Panagariya, Arvind, and Pravin ](rishna. 2002. "On Necessarily Welfare-
Enhancing Free Trade Areas." Journal of International Economics 57 (2):
353-67.
Pelkmans, J. 1990. "Regulation and the Single Market: An Economic Perspec-
tive." In Horst Siebert, ed., The Completion of the Internal Market. Tuibin-
gen, Germany: J. C. B. Mohr.
Pelkmans, J., and P. Brenton. 1999. "Bilateral Trade Agreement with the EU:
Driving Forces and Effects." In Olga Memedovic, Arie Kuyvenhoven, and
Willem T. M. Molle, eds., Multilateralism and Regionalism in the
Post-Uruguay Round Era: What Role for the EU? 87-120. Boston, Mass.:
Kluwer Academic.
Perroni, Carlo, and John Whalley. 1994. "The New Regionalism: Trade Liber-
alization or Insurance?" NBER Working Paper 4626. National Bureau of
Economic Research, Cambridge, Mass. Available at <http://papers.nber.org/
papers/w4626>.
Pittman, Russell. 1998. "Competition Law in Central and Eastern Europe: Five
Years Later." Antitrust Bulletin 43 (1): 174-228.
Polachek, Solomon W. 1992. "Conflict and Trade: An Economics Approach to
Political Interactions." In Walter Isard and Charles H. Anderton, eds., Eco-
nomics of Arms Reduction and the Peace Process: Contributions from Peace
Economics and Peace Science, 89-120. Amsterdam: North-Holland.
. 1996. "Why Democracies Cooperate More and Fight Less: The Re-
lationship between International Trade and Cooperation." Review of Interna-
tional Economics S (3): 295-309.
Pollard, Sidney. 1974. European E.conomic Integration, 1815-1970. London:
Thames and Hudson.
Porter, Michael. 1998. On Competition. Cambridge, Mass.: Harvard Business
School Press.
Preeg, Ernest H. 1970. Traders and Diplomats; An Analysis of the Kennedy
Round of Negotiations under the General Agreement on Tariffs and Trade.
Washington, D.C.: Brookings Institution.
Puga, Diego, and Anthony J. Venables. 1997. "Preferential Trading Arrange-
ments and Industrial Location." Journal of International Economics 43 (3-4):
347-68.
. 1998. "Trading Arrangements and Industrial Development." World
Bank Economic Review 12 (2): 221-50.
Richardson, Martin. 1994. "Why a Free Trade Area? The Tariff Also Rises."
Economics and Politics 6 (1): 79-96.



290     REGIONAL INTEGRATION AND DEVELOPMENT
. 1995. "Tariff Revenue Competition in a Free Trade Area." Euro-
pean Economic Review 39: 1429-37.
Ricupero, Rubens. 1998. "What Policy Makers Should Know about Regional-
ism." Keynote address, World Bank conference on regionalism, Geneva, May.
Roberts, Mark J., and James R. Tybout. 1996. Industrial Evolution in Develop-
ing Countries: Micro Patterns of Turnover, Productivity, and Market Struc-
ture. New York: Oxford University Press.
Robinson, Sherman, and Karen Thierfelder. 1999. "Trade Liberalization and Re-
gional Integration: The Search for Large Numbers." TMD Discussion Paper
34. International Food Policy Research Institute, Trade and Macroeconomics
Division, Washington, D.C.
Robson, Peter. 1987. Integration, Developpenient et Equite. Paris: Economica.
. 1998. Econonmics of International Integration. 4th ed. London:
Routledge.
Rodriguez, Francisco, and Dani Rodrik. 2001. "Trade Policy and Economic
Growth: A Skeptic's Guide to the Cross-National Evidence." In Ben S.
Bernanke and Kenneth Rogoff, eds., NBER Macroeconomics Annual 2000.
Cambridge, Mass.: MIT Press.
Rodrik, Dani. 1988. "Imperfect Competition, Scale Economies, and Trade Policy
in Developing Countries." In Trade Policy Issues and Empirical Analysis,
109-37. National Bureau of Economic Research Conference Report series.
Chicago: University of Chicago Press.
Roemer, J. E. -1977. "The Effect of Sphere of Influence and Economic Distance
on the Commodity Composition of Trade in Manufactures." Review of Eco-
nomic Statistics 59: 318-27.
Romer, Paul M. 1986. "Increasing Returns and Long-Run Growth." Journal of
Political Economy 94 (5): 1002-37.
. 1990. "Endogenous Technical Change." Journal of Political
Economy 98: S71-S102.
Rowat, Malcolm, Michele Lubrano, and Rafael Porrata. 1997. Competition Pol-
icy and MERCOSUR. World Bank Technical Paper 385. Washington, D.C.
Ruggie, John G. 1992. "Multlateralism: The Anatomy of an Institution." Inter-
national Organization 45 (3, summer): 561-98.
Rutherford, Thomas E, Elisabet E. Rutstrom, and David Tarr. 2000. "A Free
Trade Agreement between the European Union and a Representative Arab
Mediterranean Country: A Quantitative Assessment." In Bernard Hoekman
and Jamel Zarrouk, eds., Catching Up with the Competition: Trade Opportu-
nities and Challenges for Arab Countries. Ann Arbor: University of Michigan
Press.



BIBLIOGRAPHY     291
Sachs, Jeffrey D., and Andrew Warner. 1995a. "Economic Reform and the
Process of Global Integration." Brookings Papers on Economic Activity 1:
1-95.
. 1995b. "Natural Resource Abundance and Economic Growth."
NBER Working Paper 5398. National Bureau of Economic Research, Cam-
bridge, Mass. Available at <http://papers.nber.org/papers/w5398>.
Sapir, Andre. 1992. "Regional Integration in Europe." Economic Journal 102
(415): 1491-506.
. 1993. "The European Community: A Case of Successful Integra-
tion? A Comment." In Jaime de Melo and Arvind Panagariya, eds., New Di-
mensions in Regional Integration. New York: Cambridge University Press.
. 1995. "The Interaction between Labour Standards and Internation-
al Trade Policy." World Economy 18 (6): 791-803.
. 1997. "Domino Effects in West European Trade, 1960-92." CEPR
Discussion Paper 1576. Centre for Economic Policy Research, London.
. 1998. "The Political Economy of EC Regionalism." European Eco-
nomic Review 42: 717-32.
Scharpf, Fritz W. 1988. "The Joint Decision Trap: Lessons from German Feder-
alism and European Integration." Public Administration 66 (autumn):
239-78.
Schattschneider, E. E. 1935. Politics, Pressures and the Tariff. New York: Arno
Press.
Schiff, Maurice. 1997. "Small is Beautiful: Preferential Trade Agreements and
the Impact of Country Size, Market Share, and Smuggling." Journal of Eco-
nomic Integration 12 (3, September): 359-87.
. 1998. "Ethnic Diversity and Economic Reform in Sub-Saharan
Africa." Journal of African Economies 7 (3): 348-61.
. 2001. "Will the Real Natural Trading Partner Please Stand up?"
Journal of Economic Integration 16 (2, June): 245-62.
Schiff, Maurice, and Won Chang. Forthcoming. "Market Presence, Contesta-
bility, and the Terms-of-Trade Effect of Regional Integration." Journal of In-
ternational Economics.
Schiff, Maurice, and Yanling Wang. 2002a. "Education, Governance, and Trade-
Related Technology Spillovers in Latin America." World Bank, Development
Research Group, Trade, Washington, D.C. Processed.
. 2002b. "Regional Integration and Trade-Related Technology Diffu-
sion: The Case of NAFTA." World Bank, Development Research Group,
Trade, Washington, D.C. Processed.



292     REGIONAL INTEGRATION AND DEVELOPMENT
. 2002c. "Technology Diffusion and Productivity Gains: Mexico and
Poland's Trade with CUSFTA and the F.U." World Bank, Washington, D.C.
Schiff, Maurice, and L. Alan Winters. 1998. "Regional Integration as Diplo-
macy." World Bank Economic Review 12 (2, May): 271-95.
. 2002a. "Regional Cooperation and the Role of International Orga-
nizations and Regional Integration." Policy Research Working Paper 2872.
World Bank, Development Research Group, Trade, Washington, D.C. Avail-
able at <http://econ.worldbank.org/resource.php?type=5>.
. 2002b. "Regionalism and Development: The Implications of World
Bank Research for ACP and Latin American Countries." Journal of World
Trade 36 (3, June): 479-99.
Schiff, Maurice, Yanling Wang, and Marcelo Olarreaga. 2002. "Trade-Related
Technology Diffusion and the Dynamics of North-South and South-South In-
tegration." Policy Research Working Paper 2861. World Bank, Development
Research Group, Trade, Washington, D.C. Available at <http://econ.world-
bank.org/resource.php?type =5>.
Schone, Rainer. 1996. "Alternatives to Anti-Dumping from an Antitrust Perspec-
tive." Ph.D. dissertation. University of St. Gallen, Switzerland.
Schott, Jeffrey, ed. 1989. Free Trade Areas and U.S. Trade Policy. Washington,
D.C.: Institute for International Economics.
Scollay, Robert, and John P. Gilbert. 2001. New Regional Trading Arrangements
in the Asia Pacific? Policy Analyses in International Economics 64. Washing-
ton, D.C.: Institute for International Economics.
Serra, Jaime, Jose C6rdoba, Gene Grossman, Carla Hills, John Jackson, Julius
Katz, Pedro Noyola, and Michael Wilson. 1997. Reflections on Regionalism:
Report of the Study Group on International Trade. New York: Carnegie En-
dowment for International Peace.
Sewastynowicz, James. 1986. "Two-Step Migration and Upward Mobility on the
Frontier: The Safety Valve Effect in Pejibaye, Costa Rica." Economilc Devel-
opment and Cultural Change 34 (July): 731-53.
Shepsle, Kenneth A., and Barry A. Weingast. 1981. "Political Preferences for the
Pork Barrel: A Generalization." American Journal of Political Science 25
(February): 96-1i1.
Smarzynska, Beata K. 2000. "Technological Leadership and Foreign Investors'
Choice of Entry Mode." Policy Research Working Paper 2314. World Bank,
Development Research Group, Trade, Washington, D.C. Available at
<http://econ.worldbank.org/resource.php?type=5>.
.2002. "Composition of Foreign Direct Investment and Protection of
Intellectual Property Rights: Evidence from Transition Economles." Policy
Research Working Paper 2786. World Bank, Development Research Group,



BIBLIOGRAPHY     293
Trade, Washington, D.C. Available at <http://econ.worldbank.org/resource.
php?type=5>.
Smith, Alasdair, and Anthony J. Venables. 1988. "Completing the Internal Mar-
ket in the European Community: Some Industry Simulations." European Eco-
nomic Review 32: 1501-25.
Soloaga, Isidro, and L. Alan Winters. 2001. "Regionalism in the Nineties: What
Effect on Trade?" Policy Research Working Paper 2156. World Bank,
Washington, D.C. Available at <http://econ.worldbank.org/resource.php?
type=5>. Also issued as CEPR Discussion Paper 2183, Centre for Economic
Policy Research, London, 1999.
Solow, Robert M. 1956. "A Contribution to the Theory of Economic Growth."
Quarterly Journal of Economics 70 (1): 65-94.
Srinivasan, T. N. 1994. "Regional Trading Arrangements and Beyond: Exploring
Some Options for South Asia: Theory, Empirics and Policy." Report IDP-142.
World Bank, South Asia Region, Washington, D.C.
. 1997. "The Common External Tariff of a Customs Union: Alterna-
tive Approaches." Japan and the World Economy 9: 447-65.
. 1998. "Regionalism and the WTO: Is Nondiscrimination Passe?" In
Anne 0. Krueger, ed., The WTO as an International Organization. Chicago:
University of Chicago Press.
Stephenson, S. 1996. "The Economic Impact of Rules of Origin in the Asia-
Pacific Region." Prepared for the Eighth Trade Policy Forum of the Pacific
Economic Co-operation Council, Seoul, September.
Summers, Lawrence H. 1991. "Regionalism and the World Trading System." In
Policy Implications of Trade and Currency Zones: A Symposium Sponsored
by the Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, August
22-24, 1991, 295-302. Kansas City, Mo.: Federal Reserve Bank of Kansas
City.
Swann, Dennis. 1992. The Economics of the Common Market. Har-
mondsworth, Middlesex, U.K.: Penguin.
Talbott, Strobe. 1996. "Democracy and the National Interest." Foreign Affairs
75: 47-63.
Tavares de Araujo, Jose, and Luis Tineo. 1998. "The Harmonization of Compe-
tition Policies among Mercosur Countries." Antitrust Bulletin 43 (1): 45-70.
Taylor, Telford. 1979. Munich: The Price of Peace. New York: Vintage.
Telieur, E. 1998. "Free Trade: Does South Africa Gain?" Trade and Industry
Monitor 6: 1-6. Trade and Industry Secretariat, South Africa.
Tinbergen, Jan. 1954. International Economic Integration. Amsterdam: Elsevier.
Togan, Subidey. 1997. "Opening Up the Turkish Economy in the Context of the
Customs Union with EU." Journal of Economic Integration 12: 157-79.



294     REGIONAI. INTEGRATION AND DEVELOPMENT
Torrent, Ramon. 2002. "Regional Integratiol Instruments and Dimensions: An
Analytical Framework." In Robert Devlin and Antoni Estevadeordal, eds.,
New Frontiers in Trade and Regional Integration, ch. 8. Washington, D.C.:
Inter-American Development Bank.
Truman, E. M. 1975. "The Effects of European Economic Integration on the
Production and Trade of Manufactured Products." In Bela Balassa, ed., Euro-
pean Economic Integration, 3-40. Amsterdam: North Holland.
Tybout, James. 2000. "Manufacturing Firms in Developing Countries: How Well
Do They Do, and Why?" Journal of Economic Literature 38 (1, March):
11-44.
Tybout, James R., and M. Daniel Westbrook. 1995. "Trade Liberalization and
the Dimensions of Efficiency Change in Mexican Manufacturing Industries."
Journal of International Economics 39 ( 1-2, August): 53-78.
Tybout, James, Jaime de Melo, and Vittorio Corbo. 1991. "The Effects of Trade
Reforms on Scale and Technical Efficiency: New Evidence from Chile." Jour-
nal of International Economics 31 (3-4, November): 231-50.
Unter, Brian. 1998. "Maximizing Custom Benefits-A Global Model for Regula-
tory Reform." Prepared for the U.S.-China Standards, Testing, and Certifica-
tion Workshop, Information Technology Industry Council, Washington,
D.C., February 17-18.
USITC (U.S. International Trade Commission). 1997. The Impact of the North
American Free Trade Agreement on the U.S. Economy and Industries: A
Three-Year Review. USITC Publicationi 3045. Washington, D.C.
Vamvakidis, Athanasios. 1998. "Regional Integration and Economic Growth."
World Bank Economic Review 12 (2, May): 251-70.
1999. "Regional Trade Agreements or Broad Liberalisation: Which
Path Leads to Faster Growth? Time Series Evidence." IMF Staff Papers 46 ( I,
March). Available at <http://www.imf.org/external/pubind.htm>.
Venables, A. 2000. "Les accords d'int6gration r6gionale: facteurs de convergence
ou de divergence?" (Regional integration agreements: Forces for convergence
or divergence?). Revue d'Economie du Developpement (June): 227-46.
. 2002. "Winners and Losers from Regional Integration Agree-
ments." London School of Economics, Department of Economics. Processed.
Verdoorn, P. J. 1960. "The Intra-Bloc Trade of Benelux." In E. A. G. Robinson,
ed., Economic Consequences of the Size of Nations. London: Macmillan.
Viner, Jacob. 1950. The Customs Union Issue. New York: Carnegie Endowment
for International Peace.
Vogel, David. 1995. Trading Up: Consiimer and Enlviron,nental Regulation in a
Global Economy. Cambridge, Mass.: Harvard University Press.



BIBLIOGRAPHY     295
Wallace, William. 1994. Regional Integration: The West European Experience.
Washington, D.C.: Brookings Institution Press.
Wang, Z. K., and L. Alan Winters. 1998. "Africa's Role in Multilateral Trade
Negotiations: Past and Future." Journal of African Economies 7: 1-33.
Whalley, John. 1998a. "The Interface between Environmental and Trade Poli-
cies." In Alan M. Rugman and John J. Kiron, with Julie A. Soloway, eds.,
Trade and the Environmenit: Economic, Legal and Policy Perspectives, 72-81.
Elgar Reference Collection. International Library of Critical Writings in Eco-
nomics, vol. 87. Cheltenham, U.lC.: Elgar.
. 1998b. "Why Do Countries Seek Regional Trade Agreements?" In
Jeffrey A. Frankel, ed., The Regionalization of the World Economy, ch. 3.
Chicago: University of Chicago Press.
Willmore, L. N. 1974. "The Pattern of Trade and Specialisation in the Central
American Common Market." Journal of Economic Studies 1: 113-34.
1976. "Trade Creation, Trade Diversion and Effective Protection in
the Central American Common Market." Journal of Development Studies 12:
396-14.
Wilson, John S. 1996. "Eliminating Barriers to Trade in Telecommunications and
Information Technology Goods and Services: Next Steps in Multilateral and
Regional Liberalization Efforts." In Organisation for Economlc Co-operation
and Development, Regulatory Reform and International Market Openness.
Paris.
. 1999. "Product Standards and International Trade." World Bank,
Washington, D.C. Available at <wwwl.worldbank.org/wbiep/trade/Stan-
dards/papers.htm>.
Winham, Gilbert R. 1986. International Trade and the Tokyo Round Negotia-
tion. Princeton, N.J.: Princeton Lrniversity Press.
Winters, L. Alan. 1987. "Britain in Europe: A Survey of Quantitative Trade Stud-
ies." Journal of Common Market Studies 25: 315-35.
. 1992a. "Integration, Trade Policy and European Footwear Trade."
In L. Alan Winters, ed., Trade Flows and Trade Policy after "1992,"
175-209. Cambridge, U.K.: Cambridge University Press.
, ed. 1992b. Trade Flows and Trade Policy after "1992." Cambriclge,
U.K.: Cambridge University Press.
1993. "The European Community: A Case of Successful Integra-
tion." In Jaime de Melo and Arvind Panagariya, eds., New Dimensions in
Regional Integration. New York: Cambridge University Press.
. 1994. "The EC and Protection: The Political Economy." European
Economic Review 38: 596-603.



296     REGIONAL INTEGRATION AND DEVELOPMENT
. 1997a. "Lebanon's Euro-Mediterranean Agreement: Possible Dy-
namic Benefits." In Wassim Shahin and Kamal Shehadi, eds., Pathways to In-
tegration: Lebanon and the Euro-Mediterranean Partnership, ch. 3. Beirut:
Lebanese Center for Policy Studies.
. 1997b. "What Can European Experience Teach Developing Coun-
tries about Integration?" World Economy 20: 889-912.
. 2000. "The EU's Preferential Trade Agreements: Objectives and
Outcomes." In Pitou van Dijck and Gerrit Faber, eds., The External
Economic Dimension of the European Union, 195-222. The Hague: Kluwer
Law International.
. 2001. "Post-Lome Trading Arrangements: The Multilateral Alter-
native." In Jurgen von Hagen and Mika Widgren, eds., Regionalism in Eu-
rope. Boston: Kluwer Academic.
. 2002. "Doha and the World Poverty Targets." Prepared for Annual
Bank Conference on Development Economics, World Bank, Washington,
D.C., April 25, 2002. Available at <http://ccon.worldbank.org/view.php?
type=S&confid=2579&id=14983>.
Winters, L. Alan, and Won Chang. 2000. "Regional Integration and Import
Prices: An Empirical Investigation." Journal of International Economnics 51
(2): 363-77.
Wolff, Edward N., and M. Ishaq Nadiri. 1993. "Spillover Effects, Linkage Struc-
ture, and Research and Development." Structural Change and Economic
Dynamics 4 (December): 315-31.
Wonnacott, Paul, and Mark Lutz. 1989. "Is There a Case For Free Trade Areas?"
In Jeffrey Schott, ed., Free Trade Areas and U.S. Trade Policy, 59-84. Wash-
ington, D.C.: Institute for International Economics.
Wonnacott, Paul, and Ronald Wonnacott. 1981. "Is Unilateral Tariff Reduction
Preferable to a Customs Union? The Curious Case of the Missing Foreign Tar-
iffs." American Economic Review 71 (4): 704-14.
Wonnacott, Ronald J. 1996. "Free Trade Agreements: For Better or Worse?"
American Economic Review 86 (2, May): 62-66.
World Bank. 2000. Trade Blocs. Policy Research Report. New York: Oxford
University Press.
. Various years. World Development Indicators. Washington, D.C.
WTO (World Trade Organization). 1995. Regionalism anid the World Trading
System. Geneva: WTO Secretariat.
Annual Report, 1996. Geneva.
1998. "WTO Trade Facilitation Symposium, 9-10 March 1998,
Report by the Secretariat." Document G/C/W/115. Geneva.



BIBLIOGRAPHY     297
Yannopoulos, George N. 1990. "Foreign Direct Investment and European Inte-
gration: The Evidence from the Formative Years of the European
Community." Journal of Common Market Studies 28 (3, March): 235-59.
Yeats, Alexander J. 1998. "Does Mercosur's Trade Performance Raise Concerns
about the Effects of Regional Trade Arrangements?" World Bank Economic
Review 12 (1, January): 1-28.
Yi, Sang-Seung. 1996. "Endogenous Formation of Customs Unions under Imper-
fect Competition: Open Regionalism Is Good." Journal of International
Economics 41 (August): 153-77.






Index
A                                     buyer-seller linkages, 138
acceptance and recognition of         centralizing vs. decentralizing
foreign regulatory regimes,         forces, 137-1 38
152-153, 154                    FDI, 142
ACP countries, see African,           human capital, 137
Caribbean, and Pacific (ACP)    investment, 103, 115-117
countries                       knowledge spillovers, 137
"additive regionalism" (multiple      levels of trade barriers, 143-144
RIA memberships), 15-16,        modeling effects of RIAs, 139
75-78                           North-South vs. South-South
adjustment costs (of RIAs), 83            agreements, 139-141,
African, Caribbean, and Pacific           142-143
(ACP) countries                 trade diversion and intermember
bargaining powers, RIAs                 distribution, 141-144
strengthening, 204              trade liberalization affecting,
EU agreements, 112, 200, 201,           138-141
233, 251                      Albania, 29
African RIAs, 6, see also specific  AMU, see Arab-Maghreb Union
RIAs                          Andean Pact, 6
agglomeration and location of       closed regionalism, 29n3
industries, 140-141             competition policy, 177
domino regionalism, 234             domino regionalism, 234
multiple agreements, 75, 76, 78     excluded countries, trade
negotiating partners, RIAs as,          discrimination against, 211
239                             gravity models, 41, 45
selected list of, 27-28             increased trade between members,
AFTA, see Association of Southeast        32
Asian Nations (ASEAN) and       investment stimulus, 106, 115,1 17
ASEAN FTA                       livestock trade, 216
agglomeration and location of         lobbying and external trade
industries, 18-19, 124, 137         policy, 88
299



300     INDEX
Andean Pact (continued)               open regionalism, 242, 243
member countries, 27                peace and security, 192
multiple memberships, 78            policy integration, 181-182
natural trading partners fallacy,   private sector involvement, 89
66, 67                          RIA, status of, 1, 71
open regionalism, 244               welfare effects, estimated potential
public investment planning, failure     of, 216
of, 121n1                     Association of Southeast Asian
regionalism and protectionism, 227      Nations (ASEAN) and
services liberalization, 175            ASEAN FTA (AFTA), 6
trade effects of RIAs between       bargaining powers, RIAs
developing countries, 59            strengthening, 204, 206
Angola, 28                            convergence of per capita
antidumping rules, 163-164,               incomes, 128
177-178, 183n7, 185nl7,         domino regionalism, 235
240                             estimated potential welfare effects,
Antigua and Barbuda, 27                   216
anti-monde, see counterfactual (anti-  FDI, 118
monde) calculations             gravity models, 41, 44, 45
antitrust policy integration,         increased trade between members,
164-165, 176-177                    32
APEC, see Asia-Pacific Economic       member countries, 29
Cooperation                     natural trading partners fallacy,
Arab Common Market, 29, 234               66, 67
Arab Countries' Agreement on          openness associated with, increase
Investment agreement and            in, 38
Free Movement of Arab           peace and security, 192, 203
Capital, 105                    political integration and trade
Arab Free Trade Agreement, 32             integration, 203
Arab-Maghreb Union (AMU), 76          PTA, 29
Argentina                             services liberalization, 176
Brazilian imports from, 217-219     tax revenues, 94
Chile's bilateral agreement with, 75  trade effects of RIAs between
economies of scale, 52                  developing countries, 59
livestock trade, 216-217          Australia
peace and security, 193             CER, see Closer Economic
reform, RIAs leading to, 227            Relations agreement
RIAs, selected list of, 27          product standards, 174
Article XXIV of GATT, 247-251, 256    reform requirements in
Asian RIAs, 6, 29, 234, see also          agreements with New
specific RIAs                       Zealand, 110
Asia-Pacific Economic Cooperation     RIAs, selected list of, 27
(APEC)                          services trade restrictions, 160
domino regionalism, 234             Trans-Tasman MRA, 174
insurance motive, 238               Trans-Tasman Travel
investment policy, 106                  Arrangement, 150
level of integration, 20          Austria
member countries, 27                historical background to RIAs, 4
multilateralism and globalization,  insurance motive, 236
223, 230                        political integration and trade
neighborhood RIAs, exception to,        integration, 203
71                              RIAs, selected list of, 26



INDEX     301
B                                   Burkina Faso, 28
The Bahamas, 27                     Burundi, 27, 28
Bahrain, 29
balance of payments disequilibria,  C
163                           CACM, see Central American
Bangladesh, 29                            Common Market
Bank for International Settlements  Cambodia, 29, 94
(BIS), 171                    Cameroon, 28
Barbados, 27                        Canada
bargaining, see negotiation           antidumping rules, 240
Belgium, 5, 26, 167, 259n24           Article XXIV of GATT, 259n24
Belize, 27                            CARICOM market access
Benelux customs union, 5, 259n24          agreements, 205
Benin, 28                             Chile, agreement with, 71, 75,
Bhutan, 29                                177
bilateral investment treaties (BITs),  competition policy, 177
17, 104-105                     contingent protection, 163, 177
BIS, see Bank for International       domino regionalism, 231
Settlements                     estimated potential welfare effects
BITs, see bilateral investment treaties   of RIAs, 216
Bolivia                               insurance motive, 237, 238
Chile's bilateral agreement with,   NAFTA, see North American Free
75                                  Trade Agreement
democracy, fostering, 198           policy integration, 152
domino regionalism, 231, 236        recycling requirements, economic
multiple agreements, 75                 effects of, 159
multiple memberships, 78            RIAs, selected list of, 26, 27
public investment planning, failure  services trade restrictions, 159,
of, 121nl                           160
RIAs, selected list of, 27        Canada-United States Free Trade
border formalities, 155                   Agreement (CUSFTA), 2-3, 6,
Botswana, 28, 78                          26
Brazil                                difficult issues, ability to handle,
antidumping rules, 177, 185n17          240
economies of scale, 52              domino regionalism, 231, 234
excluded countries, trade           national trading partners fallacy,
discrimination against, 211         66-67
external trade policy, 85, 89       policy integration, 174
FDI, 120, 135                     CAP (Common Agricultural Policy),
information technology                  169,170, 196, 241
negotiations, 240             Cape Verde, 28
open regionalism, 244             capital, rate of return on, 114
peace and security, 193           Caribbean
reform, RIAs leading to, 227        domino regionalism, 231, 234
RIAs, selected list of, 27          RIAs, 27, 205
services trade restrictions, 160  Caribbean Community and
terms of trade of non-                  Common Market
MERCOSUR countries,                 (CARICOM)
217-220                         Article XXIV of GATT, enforcing,
Britain, see United Kingdom               249
Brunei Darussalam, 27, 29             bargaining powers, RIAs
Bulgaria, 26, 28, 29                      strengthening, 204-205



302    INDEX
CARICOM (continued)                 Central Europe
closed regionalism, 29n3            agreements with EU, see Europe
competition policy, 177                 Agreements
domino regionalism, 235             domino regionalism, 234
increased trade between members,    political integration and trade
32                                 integration, 203
openness associated with, increase  services trade restrictionis, 160
in, 37, 38                    Central European Free Trade Area
services liberalization, 175           (CEFTA)
trade effects of RIAs betweeni      domino regionalism, 235
developing countries, 59        increased trade between members,
CBI, see Cross-Border Initiative          32
(Africa)                        member countries, 28, 29
CEAO, see Economic Community of       natural trading partners fallacy,
West Africa                         66, 67
CEFTA, see Central European Free    CEPGL, see Economic Community
Trade Area)                         of the Countries of the Great
CEMAC, see Economic and                   Lakes
Monetary Community of         CER, see Closer Economic Relations
Central Africa; see also            Agreement (CER)
Union douaniere et            CETs, see customs unions
6conomique de l'Afrique       CGE, see computable general
Centrale (UDEAC)                    equilibriun
Central African Republic, 28        Chad, 28, 98
Central American Common Market      Chile
(CACM), 6                       antidumping rules, 240
closed regionalism, 29n3            Argentine export prices compared,
competition policy, 177                 218
disproportionate shares of tariff-  Canada, agreement with, 71, 75,
free trade, fiscal                  177
compensation for, 96-97         competition policy, 177
domino regionalism, 234             contingent protection, 177
economies of scale, 54              democracy, fostering, 198
gravity models, 43, 44, 45          domino regionalism, 231, 236
increased trade between members,    estimated potential welfare effects
32                                  of RIAs on, 216
member countries, 27                EU agreement with, 71, 130
natural trading partners fallacy,   investment incentives, 115
66, 67                          MERCOSUR, agreement with, 75,
openness associated with, increase      95, 233, 236
in, 38                          Mexico, agreement with, 71, 75
peace and security, 192             multiple RIA memberships
public investment planning, failure    ("additive regionalism"),
of, 121nl                           75-77
regionalism and protectionism,      reform, RIAs leading to, 228
227                             RIAs, selected list of, 27
services liberalization, 175        tax revenues, 95
subregional integration assistance  China, 27, 29, 160, 216
from EU, 206                  CILSS, see Permanent Interstate
trade effects of RIAs between           Committee on Drought
developing countries, 59            Control in the Sahel



INDEX     303
closed vs. open models of           Common Market of the South
regionalism, 242-244                (Mercado Comiun del Sur-
benign regionalism, open access as     MERCOSUR), 6
key to, 232-234, 242,           antidumping rules, 177, 185nl7
253-254                         bargaining power, RIAs
definitions of open regionalism,        increasing, 206
242-243                         Bolivia's FTA with, 75
examples of, 29n3                   Chile's bilateral agreement with,
growth and openness, 37, 125,           75, 95, 233, 236
127, 130, 135-137               competition policy, 177, 184n16
major move from closed to open      democracy, fostering, 198, 199,
models, xi, 1-2                     200
measuring degree of openness, 37    domino regionalism, 231,
regionalism and protectionism,          234-236
227                             economies of scale, S2
TFP and openness, 129               EU FTA, 130
usefulness of concept, 24           excluded countries, trade
Closer Economic Relations                discrimination against, 211
agreement (CER)                 external trade policies of customs
estimated potential welfare effects,    unions, 84, 89
216                             FDI, 118, 120, 121, 135
intra- and extrabloc trade, 41      gravity models, 41-45
policy integration, 147             increased trade between members,
competition provisions,               32
176-177                        information technology
contingent protection, 177            negotiations, 240
depth of integration, 180         investment stimuli, 105-106, 118,
public procurement, 163               120
services liberalization, 175      level of integration, 20
CMEA, see Committee for Mutual       lobbying and pressure groups, 89,
Economic Assistance                 91
Cobden, Richard, 189                  member countries, 27
Colombia, 27, 75, 78, 216-217         natural trading partners fallacy,
colonies, former colonies, and            66, 67
colonialism, 4, 5, 108, 236     negotiating partners, RIAs as,
COMESA, see Common Market for            239
Eastern and Southern Africa     neighborhood RIAs, 72
Committee for Mutual Economic         open regionalism, 27, 38, 233,
Assistance                          244
(CMEA/COMECON), 29,             peace and security, 193
116                             product standards, 174
Committee on Regional Trading         regionalism and protectionism,
Agreements (CRTA), 249,             227
254-255                         services liberalization, 175-176
Common Market for Eastern and         social standards, 179
Southern Africa (COMESA),      tax revenues, 95
6                               trade diversion and excluded
domino regionalism, 234                 countries, 212, 217-220
investment policy, 106              trade effects of RIAs between
member states, 28                       developing countries, 59,
overlapping blocs, 76, 78               65



304     INDEX
common markets, selected list of,  Cuba, 29
26-29                         CUSFTA, see Canada-United States
Communaute economique des pays           Free Trade Agreement
des grands lacs, see Economic  customs unions, see also specific
Community of the Countries         unions
of the Great Lakes              external trade policies, 84-87
Communaut6 6conomique et             FTAs vs., 16, 78-82
monetaire d'Afrique Centrale,   GATT provisions, 26
see Economic and Monetary       historical background to RIAs, 4,
Community of Central               5
Africa; see also Union         insurance motive, 237, 238
douaniere et 6conomique de     lobbying, 16, 88-93
l'Afrique Centrale             lower trading costs and greater
Comoros, 27, 28                          integration offered by, 81-82
comparative advantage, 69-71         neighboring countries, between,
competition, see procompetitive          72
effects                         selected list of, 26-29
computable general equilibrium     Czech Republic, 26, 28, 29, 75,
(CGE) models, 47-49, 52,           157
53, 216
concerted unilateralism, 243       D
Congo, Democratic Republic of, 28  deep integration, xi, 2, 20, 179-180
Congo, Republic of, 28               credibility and, 110
contingent protection, 163-164,      defining, 151
177-178                        high economic cost of
convergence of per capita incomes,       inappropriate arrangements,
127-130                            170
coordination of domestic policies,   political feasibility, 169
151-152                        WTO policy, 244
Costa Rica, 27, 54, 67, 75         deflection of trade, 79-81
C6te d'lvoire, 28, 97, 104         democracy, fostering, 188, 191,
Cotonou Agreement, 112, 200, 201,        198-201
251                           Denmark, 26, 150, 178
counterfactual (anti-monde)        developing countries, see also more
calculations, 14, 37, 40           specific topics
countervailing duties, 163,          burden of negotiating RIAs, 240
177-178                        excluded countries, trade
creation of trade, see trade creation    discrimination against, 211
credibility                          reasons for emphasis on, xi, 4-5
democratic enforcement actions,    WTO's means of assisting,
200-201                            246-247
RIAs providing, 107-113,         diffuse reciprocity as characteristic
168-169, 263                       of multilateralism, 224
Cross-Border Initiative (Africa)   discrimination in trade, see
(CBI), 6                           nondiscrimination principle
domino regionalism, 231, 234     disproportionate shares of tariff-free
member states, 27                      trade, fiscal compensation for,
open access, 233                       96-98
overlapping blocs, 76            dispute settlement procedures, 105,
cross-border shopping, 98-99             107, 255, 258nl9
CRTA, see Committee on Regional    diversion of trade, see trade
Trading Agreements                 diversion



INDEX     305
Djibouti, 28                        Economic Community of the
Doha Development Agenda, 150,            Countries of the Great Lakes
251                                (Communaute economique
domestic policy integration, see         des pays des grands lacs-
policy integration                  CEPGL), 28, 29n3, 234
Dominica, 27                        Economic Community of West
domino regionalism, 23, 231-236          Africa (CEAO)
closed regionalism, 29n3
E                                     disproportionate shares of tariff-
EAC, see East African Cooperation        free trade, fiscal
East African Community, collapse         compensation for, 97, 98
of, 6, 142, 195, 225            economies of scale, 54
East African Cooperation (EAC), 6     increased trade between members,
closed regionalism, 29n3                32
disproportionate shares of tariff-  investment and growth, 116-117
free trade, fiscal              natural trading partners fallacy,
compensation for, 98                66, 67
domino regionalism, 234             neighborhood RIAs, 72
investment and growth, 116          openness associated with, increase
member countries, 26                   in, 37
natural trading partners fallacy,   trade effects of RIAs between
66, 67                              developing countries, 59
neighborhood RIAs, 72               UEMOA, becoming, see West
overlapping blocs, 76, 78              African Economic and
public investment planning, failure     Monetary Union
of, 121nl                     Economic Community of West
East Asian Economic Caucus, 41           African States (ECOWAS),
Eastern Europe                           28, 29n3, 76, 193, 234
domino regionalism, 231, 234      economic effects of RIAs, 13-14,
EU agreements, see Europe               31-61, 265
agreements                      anti-monde (counterfactual)
multiple agreements, 75                 calculations, 14, 37, 40
political integration and trade     CGE models, 47-49, 52, 53
integration, 203                gravity model, 40-46
services trade restrictions, 160    growth of trade, 36-39
EC, see European Community (EC),      imperfectly competitive markets,
European Economic                   46-49
Community (EEC), and           increased trade between members
European Union (EU)                 of RIAs, 32-33
ECCAS, see Economic Community        intraindustry trade, 46-47
of Central African States       markets, large vs. small, 50-54
Economic and Monetary                 nondiscriminatory liberalizations,
Community of Central                effects compared with,
Africa (Communaute                  40-46, 63, 73
economique et mon6taire         policy integration, 154-155
d'Afrique Centrale-             political and economic
CEMAC), 28, 76,140, 235;            integration, connection
see also Union douaniere et         between, 188, 201-206
economique de l'Afrique         procompetitive effects of RIAs,
Centrale (UDEAC)                    14, 31-32
Economic Community of Central         product differentiation, 46-47,
African States (ECCAS), 76          71



306    INDEX
economic effects of RIAs (continued)  Euro-Mediterranean Agreements, see
trade creation, 14, 31, 33-36           Mediterranean-EU
trade diversion, 14, 23, 31,            agreements
33-36                         Europe Agreements, 3, 5, 26
economic partnership agreements       accession issues, 234
between EU and ACP              contingent protection, 178
countries, 112, 201             democracy, fostering, 199
economies of scale, 50-54             difficult issues, ability to handle,
ECOWAS, see Economic                     241
Community of West African       domestic policy reforms covered
States                              by, 110
ECSC, see European Coal and Steel     domino regionalism, 231
Community                       migration controls, 197
Ecuador, 27, 75, 78                   MRAs, 157
education and productivity growth,    political integration and trade
130                                integration, 203
EEA, see European Economic Area       subregional integration assistance,
EEC, see European Community               206
(EC), European Economic       European Coal and Steel
Community (EEC), and                Community (ECSC), 5, 189,
European Union (EU)                 192, 193
Egypt, Arab Republic of, 28, 95,    European Community (EC),
158, 161-162                        European Economic
El Salvador, 27, 75, 96-97, 121nl         Community (EEC), and
emergency protection, 163                 European Union (EU)
Enabling Clause, 26, 249-250          ACP countries, agreements with,
endogenous (self-generated)               112, 200, 201, 233, 251
growth theory, 125, 126,        agglomeration and location of
132-133                            industries, 141
enforcement                           Article XXIV of GATT, enforcing,
Article XXIV of GATE and other          249
WTO rules, 249, 253             bargaining power, RIAs
democratic requirements of RIAs,        increasing, 206
200-201                         CARICOM market access
policy integration, enforcement of      agreements, 205
free trade commitments via,     CBI (Africa) and, 27
168, 169-170, 172-173           Central Europe, agreements with,
environmental standards and policy       see Europe Agreements
integration, 159, 165-166,      Chile's bilateral agreement with,
168, 178-179                        75, 130
EPAs between EU and ACP               Common Agricultural Policy
countries, see economic             (CAP), 169, 170, 196, 241
partnership agreements          competition provisions, 176,
between EU and ACP                  185n18
countries                       convergence of per capita
Equatorial Guinea, 28                     incomes, 128
Estonia, 26, 67, 75                   democracy, fostering, 199-201
Ethiopia, 28                          difficult issues, ability to handle,
EU, see European Community (EC),          240, 241
European Economic               disproportionate shares of tariff-
Community (EEC), and                free trade, fiscal
European Union (EU)                 compensation for, 97



INDEX     307
domino regionalism, 231, 234         contingent protection, 163,
Eastern Europe, agreements with,       177-178
see Europe Agreements            depth of integration, 179-180
economies of scale, 51-52            environmental standards, 178
Egypt, FTA with, 95                  frictional transactions costs,
enlargement, focus on, 251             158
excluded countries, trade            harmonization, 152, 154
discrimination against,          international norms, adoption
210-211                            of, 171
external trade policy, 84, 85-86,    investment, 150
89                               MRAs, 149, 153, 154,
FDI, 119-120                           156-157,173-174
gravity models, 40-45                political and credibility gains,
growth and industrial location,        168-169
19, 124                          product standards, 155-157,
harmonization, 81, 98-99, 152,         174
154                              public procurement, 163
Iceland, agreement with, 110         regulatory norms, adapting,
insurance motive, 237, 238             148
investment, 106, 110, 116,           services trade restrictions, 160,
119-120, 150                       162, 173-174
Lebanon, FTA with, 97, 131           social standards, 167, 179
lobbying, 89, 92-93, 207n8         political integration, 202, 203,
Mediterranean policy, see              205, 206, 207n8
Mediterranean-EU               predominance in RIAs, 5, 12
agreements                      reform requirements in certain
member countries, 26                   agreements, 110
MERCOSUR FTA, 130                  RIAs, selected list of, 26-27
migration controls, 197, 198       rules of origin and trade
MRAs, 149, 153, 154, 156-157,          deflection, concerns for FTAs,
173-174                            80
multilateralism and globalization,  services trade restrictions, 160,
223                                162, 173-174
multilateral negotiations and      South Africa, agreement with, 71,
regionalism, 229-230               97
multiple agreements, 75            subregional integration assistance,
natural trading partners fallacy,      206
66, 67                          tax harmonization and
negotiating partners, RIAs as,         competition, 98-99
239                            tax revenues, 94, 95
nontariff barriers, harmonizing,   trade deflection, 81
81                             trade diversion and excluded
North-South agreements, 3              countries, 212, 214, 215,
open access, 233, 234                  220-221
open regionalism, 244              Treaty of Rome, 249
peace and security, 189, 192-193,  Turkey and, 81-82, 178, 234,
195-196                            253
Poland, agreement with, 80       European Economic Area (EEA), 5,
policy integration, 20, 147, 151       26, 177, 241
border formalities, 155         European Economic Community
competition provisions, 176,         (EEC), see European
185n18                             Community (EC), European



308     INDEX
EEC (continued)                     F
Econonmic Community (EEC),    FDI, see foreign direct investment
and European Union (EU)       Finlanid, 26, 150
European Free Trade Association     fiscal implications of RIAs, see
(EFTA)                              economic effects of RIAs
border formalities, 155           foreign direct investment (FDI),
contingent protection, 163              17-18, 101-102, 117, see
difficult issues, ability to handle,    also investment
240                             agglomeration and location of
gravity models, 40-45                   industries, 142
growth and investment, 116          BITs, 105
investment, 116, 220-221            convergence of per capita
rules of origin and trade deflection    incomes, 127
concerns for FTAs, 80           evidence of positive effects of
trade diversion and excluded            RIAs on, 119-121
countries, 212, 214, 220-221    IPR protections, 134-135
European RIAs                         knowledge capital and human
selected list of, 28-29                 capital, 127
European Single Market Programme      knowledge spillovers, 133-135
(SMP), 5, 41, 110, 115-116,     managed industrialization, 104
175                             motives for, 117-119
difficult issues, ability to handle,  trade-oriented explanation of,
240                                 122n8
domino regionalism, 231           France
investment switching, 220-221       historical background to RIAs, 4
trade diversion and excluded        insuranice motive, 236
countries, 214                   negotiating partners, RIAs as,
European Union (EU), see European         239
Community (EC), European        peace and security, 189, 190, 192,
Economic Communitv (EEC),           195
and European Union (EU)         policy integration and social
exchange rate policies, 4                 standards, 167
excluded countries, 23-25, 209-210    RIAs, selected list of, 26
export losses, 213-215, 252-253     sponsorship of former colonies,
multilateralism, 221-244, see also      108
multilateralism and           free trade agreements (FTAs), see
globalization                       also specific FTAs
tariff levels and RIAs, 226-229     binding of tariffs at actual applied
trade discrimination against,           rates, 252
210-211                         customs unions vs., 16, 78-82
trade diversion, 23, 212-221, see   domino regionalism, 232
also trade diversion            external trade policies, 83-84
WTO's policy toward regionalism,    GATT provisions, 26
244-256, see also World          historical background to RIAs, 5
Trade Organization              lobbying, 16
export losses by excluded countries,  rules of origin, 79-81, 252
213-215, 252-253                selected list of, 26-29
export taxes, 214                     trade deflection, 79-81
external trade policies, 82-93      Free Trade Area of the Americas
customs unions, 84-87                   (FTAA)
FTAs, 83-84                         education and productivity
lobbying, 16, 88-93                     growth, 130



INDEX     309
estimated potential welfare effects,  policy integration, 166, 167, 178
216                             RIAs, selected list of, 26, 29
lobbying and external trade         social standards, 166, 167
policy, 88, 89                Ghana, 28, 104
policy integration, 182-183       globalization, see multilateralism
private sector involvement, 89          and globalization
trade diversion and excluded      government planning projects,
countries, 215                      failure of, 103-104, 121n1
U.S. negotiations with, 251       government procurement, 162-163,
free trade vs. RIAs, see                  171
multilateralism and           gravity model, 40-46
globalization                 Great Britain, see United Kingdom
frictional transactions costs,      Great Depression, 4
157-158                       Greece, 19, 26, 111-112, 199
FTAA, see Free Trade Area of the    Grenada, 27
Americas                      Group of Three (G3), 6
FTAs, see free trade agreements       competition policy, 177
domino regionalism, 234
G                                     investment, 240
G3, see Group of Three                investment policy, 106
Gabon, 28                             member countries, 6
The Gambia, 28, 73                    multiple memberships, 78
GATT, see General Agreement on        natural trading partners fallacy,
Tariffs and Trade                   66, 67
GCC, see Gulf Cooperation Council     neighborhood RIAs, exception to,
General Agreement on Tariffs and          71
Trade (GATT), 245-247           product standards, 174
agreements notified to, increasing  growth, 18-19, 123-124
number of, 1                    agglomeration of industries, see
Article XXIV of GATT as rules           agglomeration and location
for RIAs, 247-251, 256              of industries
Enabling Clause, 26, 249-250        convergence of per capita
EU partnership predominance, 5          incomes, 127-130
MFN status, 245, 247, 259n24        economics of growth and RIAs,
multilateral negotiations and           36-39
regionalism, 229-230            education, 130
regionalism and protectionism,      endogenous (self-generated)
226, 227                            growth theory, 125, 126,
RIAs, selected list of, 26              132-133
services liberalization, 175        evidence of RIAs stimulating
succession of WTO to GATT, 29n:L        growth, 135-137
generalized norms as characteristic   FDI and knowledge spillovers,
of multilateralism, 224             133-135
Germany                               goal of trade policy, growth as,
agglomeration and location of           124-125
industries, 141                 human capital, 125, 126-127
environmental standards, 178        investment, 103, 115-117
FDI, 119                            knowledge, see knowledge capital
historical background to RIAs, 4        and spillovers
insurance motive, 236               location of industries, see
negotiating partners, RIAs as, 239      agglomeration and location
peace and security, 190, 192, 195       of industries



310    INDEX
growth (continued)                 Hoekman, Bernard, 147
North-South vs. South-South       Honduras, 27, 75, 98
agreements, 18-19, 123-125    Hong Kong (China), 27
convergence and spillovers,     "hub and spoke" regionalism,
127-130                            78
FDI and knowledge spillovers,   Hull, Cordell, 189
133-135                       human capital
human capital, 125, 126-127      agglomeration of location of
knowledge capital and                industries, 137
spillovers, 125-127,            growth, 125, 126-127
129-135                         migration controls, 196-198
openness and growth, 135-137      policy integration on social
TFP, 127, 128-131                    standards, 165-166, 167,
wrong partners harming                168, 178-179
growth, 130-1 33                Trans-Tasman Travel
open regionalism and, 37, 125,         Arrangement, 150
127, 130, 135-137             Hungary, 26, 28, 29, 67, 75
public planning projects, failure
of, 103-104, 121nl            I
specialization within region,     Iceland, 26, 67, 110, 150
103-104                      ICSID, see International Centre for
spillovers, 125, 127-135               Settlement of Investment
TFP, 127, 128-131                      Disputes
traditional (neoclassical) growth  IGAD, see Intergovernmental
theory, 125                        Authority on Development
wrong partners harming growth,    IMF, see International Monetary
130-133                            Fund
Guatemala, 27, 54, 75, 97          imperfectly competitive markets,
Guinea, 28                               46-49
Guinea-Bissau, 28                  import-substitution creed, 5,
Gulf Cooperation Council (GCC)           103-104
domino regionalism, 234           income convergence, 127-130
gravity models, 44, 45            India, 29, 160, 211
increased trade between members,  Indian Ocean Commission (IOC),
32                                 28, 76, 88, 234
member countries, 29             indirect trade deflection, 81
natural trading partners fallacy,  indivisibility as characteristic of
66, 67                             multilateralism, 224
openness associated with, increase  Indonesia
in, 37, 38                      excluded countries, trade
political integration and trade        discrimination against,
integration, 203                   211
trade effects of RIAs between      peace and security, 192
developing countries, 59       reform, RIAs leading to, 228
RIAs, selected list of, 27, 29
H                                    services trade restrictions, 160
harmonization                      industrialization
national standards and             agglomeration and location, see
regulations, 152, 153, 154,        agglomeration and location
171                                of industries
nontariff barriers in EU, 81       managed, 103-104
tax harmonization, 98-99           RIAs favoring, 65-66



INDEX     311
insurance motive, 236-238            stimulus to investment, RIAs as,
integration of domestic policies, see    113-117
policy integration              trade diversion and excluded
intellectual property rights (IPR)       countries, 220-221
protections, 134-135           treatment of investment in current
Intergovernmental Authority on           RIAs, 105-106
Development (IGAD), 76        IOC, see Indian Ocean Commission
intergovernmental organizations,   IPR, see intellectual property rights)
trade-related, 1 72-1 73,          protections
258n19, see also World Trade  Iran, Islamic Republic of, 203
Organization (WTO)            Iraq, 203
International Centre for Settlement  Ireland, 26
of Investment Disputes       Israel, 71, 259n24
(ICSID), 105                  Iay ,2
International Monetary Fund (IMF),  aly,
111
international norms, adoption of,
170-174                      Jamaica, 27
interventionism, 5                 Japan
intraindustry trade, 46-47           ASEAN aid, 206
investment, 17-18, 101-1 03          estimated potential welfare effects
BITs, 17, 104-105                      of RIAs on, 216
capital, rate of return on, 114    FDI, 117-119, 120
credibility provided by RIAs,      FTAs, 6
107-113                        insurance motive, 238
FDI, see foreign direct investment  RIAs, selected list of, 27
general policy reforms more           selecte hstrof, 27
important than RIAs,           services trade restrictions, 160
102-103
growth and industrial location,  K
103, 115-117                 Kant, Immanuel, 189
lower transaction costs, RIAs     Kenya
leading to, 114                 agglomeration and location of
multilateral agreements, 106-107,      industries, 142
North-South agreements         collapse of East African
North-South agreements
FDI, 119, 120-121                    Community, 121nl, 195
growth, effect on, 116           comparative advantage, 69-71
reform credibility, 108,         excluded countries, trade
111-112                            discrimination against,
South-South agreements, vs.,         211
17, 102, 116-117               multiple memberships, 78
stimulus to invest, 114, 115     public investment planning,
policies on, 101-102, 103-107          failure of, 121nl
policy integration, 150            RIAs, selected list of, 27, 28
public planning prolects, failure  knowledge capital and spillovers,
of, 103-104, 121nl                 126-127
reform of bad policy and           agglomeration and location of
incentives for good policy,        industries, 137
RIAs providing, 108-112        endogenous (self-generated)
South-South agreements, 17, 102,       growth theory, 125, 126,
116-117                            132-133



312     INDEX
knowledge capital (continued)         policy integration and special
FDI, 133-135                            interest groups, 168
North-South vs. South-South         political integration in EU, 207n8
agreements, 125-127,            pressure groups and
129-135                             MERCOSUR, 91
testing for effects of, 144nl       trade diversion, leading to, 88-90
TFP and R&D industries,           location of industries, see
129-131                             agglomeration and location
Korea, Republic of                        of industries
Argentine export prices compared,  logrolling, 204
218                           Luxembourg, 5, 26, 259n24
Brazilian export prices, 219
estinated potential welfare effects  M
of RIAs, 216                  rnacroeconomic policies, 4, 102
FDI, 120                          Madagascar, 27, 28, 88
FTAs, 6                           Maghreb groups, North Africa, 6
reform, RIAs leading to, 228        AMU, 76
RIAs, selected list of, 27          Euro-Maghrcb, see
Kuwait, 29                                Mediterranean-EU
agreements
L                                   Malawi, 27, 28
labor market, see human capital     Malaysia, 27, 29, 192
LAFTA, see Latin American Free      Maldives, 29
Trade Area                    Mali, 28
LAIA, see Latin American            managed industrialization,
Integration Association             103-104
Lao People's Democratic Republic,   market access, 65, 164, 205, 233,
29                                  234
Latin American RIAs, see also       markets, large vs. small, 50-54
specific RIAs                 Mashraq groups, North Africa, 6
domino regionalism, 231           Mauritania, 28
livestock trade, effect of large  Mauritius, 27, 28
RIAs on, 216-217              Mediterranean-EU agreements, 5, 6
RIAs in, 6, 27                      Article XXIV of GATT, enforcing,
Latin American Free Trade Area            249
(LAFTA), 29n3, 44,116, 128      democracy, fostering, 200
Latin American Integration            domino regionalism, 231, 234
Association (LAIA), 27, 45,     FTAs over customs unions, 74
106                             member countries, 26
Latvia, 26, 67, 75, 199               peace and security, 196
Lebanon, 97, 131                      service liberalization, 162
Lesotho, 28                           subregional integration assistance,
liberalization                            206
services, 159-162, 175-176        membership of RIA, choosing,
trade, see trade liberalization         64-74
Liberia, 28                           comparative advantage, 69-71
Liechtenstein, 26                     natural trading partners fallacy,
Lithuania, 26, 67                         66-69
lobbying, 16, 88-93                   neighboring countries, 71-73
expenditures, 92-93                 North-South vs. South-South
new environment created by RIAs         agreements, 73-74
for, 90-93                      policy integration, 170-174



INDEX     313
Mercado Comin del Sur, see            knowledge spillovers, 132, 133
Common Market of the            lobbying, 90
South                           multilateralisimi vs. regionalism,
MERCOSUR, see Common Market               224
of the South                    North-South agreements
Mexico                                    compared to unilateral
Argentine export prices compared,       liberalization under, 73
218                             open access vs., 233
Chile, agreement with, 71, 75       open regionalism and, 243
domino regionalism, 231             policy integration, 148, 171-173,
economies of scale, 52                  181-182
EU agreement with, 71               reasons for favoring RIAs over
excluded countries, trade               free trade with, 65
discrimination against, 211     reduction of trade diversion and,
FDI, 119, 120, 134                      251
growth and industrial location,     trade discrimiination and, 2-10
18, 123, 134                  Mozambique, 28
insurance motive, 238             MRAs, see mutual recognition
knowledge spillovers, 134               agreements
migration controls, 197, 198      multilateralism and globalizationi,
multiple agreements, 75                 63-66,167, 221-223
NAFTA, see North American Free      context of RIAs in, 3
Trade Agreement                 difficult issues, use of RIAs to
natural trading partners fallacy,       handle, 239-242
68                              diffuse reciprocity, 224
peace and security, 196-197         domino regionalism, 23,
reform, RIAs leading to, 227            231-236
RIAs, selected list of, 26, 27      excluded countries, 221-223
services trade restrictions, 160    generalized norms, 224
United States, ties to, 108         indivisibility, 224
MFN, see most-favored-nation          insurance, regionalism as,
(MFN) status                        236-238
Middle Eastern RIAs, 29, see also     intergovernmental organizations,
specific RIAs                       172-173, 258nl9
migration controls, RIAs as, 188,     investment agreements, 106-107,
196-198                             113
Mongolia, 29                          negotiating powers, 225-226,
Monnet, Jean, 189, 190                    238-239
Montserrat, 27                        open regionalism, 242-244
Morocco, 26, 67, 134                  policy integration, 170-174
most-favored-nation (MFN) status,     processes, behavior, and
256nl                               characteristics, 224-225
APEC, 181-182                       promotion of free trade via RIAs,
conditional and unconditional           238-239
MFN, 243                        reduction of trade barriers,
domino regionalism, 232                 138-140
excluded countries' response to     regionalism as developing from
RlAs, 231                           success of, 257n9
export prices, effect on, 219, 220  tariff levels and RIAs, 226-229
GATT, 245, 247, 259n24              WTO, 223
investment strategies, 104, 105,  multiple RIA memberships, see
106                                 "additive regionalism"



314     INDEX
mutual recognition agreements       Nicaragua, 27, 75
(MRAs), 153                   Niger, 28
harmonization and, 153, 154       Nigeria, 28
nondiscrimination principle,      nondiscrimination principle, 5, 23,
171-174                             33-34
product standards, 156-157, 174     effect of nondiscriminatory
stand-alone provisions, 149, 158        liberalizations, 40-46, 63, 73
Myanmar, 29                           excluded countries, trade
discrimination against,
N                                         210-211
NAFTA, see North American Free        MRAs, 171-174
Trade Agreement                 policy integration, 158, 1 71-174
Namibia, 27, 28, 78                   recycling requiremenits, economic
natural trading partners fallacy,         effects of, 159
66-69                           regional cooperation, trade
negotiation                               discrimination not required
burden of negotiating RIAs, 240         for, 264-265
difficult issues, handling, 239-242  nontariff barrier (NTB) coverage,
multilateral negotiations and           227
regionalism, 225-226,         North Africa, 6, 26, 76, 196, see
238-239                             also Mediterranean-EU
strengthening bargaining powers         agreements
of nations and regions,       North American Free Trade
203-206, 225-226                    Agreement (NAFTA), 2-3, 6,
trade negotiation process, effect of    26
RIAs on, 24                     agglomeration and location of
neighborhood RIAs, 71-73, 108,            industries, 143, 144
130                             antidumping, 178
neoclassical (traditional) growth     Article XXIV of GATT, enforcing,
theory, 125                         249
Nepal, 29                             bargaining power, RIAs
The Netherlands                           increasing, 206
Benelux customs union, 5,           Chile, projected advantages from
295n24                              joining, 75
environmental standards, 178        credibility of Mexican policy
policy integration, 167, 178            reforms, 109-110, 11t
RiAs, selected list of, 26          CUSFTA, extension to Mexico of,
social standards, 167                   2-3, 6, 26
New Zealand                           difficult issues, ability to handle,
CER, see Closer Economic                240, 241
Relations Agreement             disproportionate shares of tariff-
insuranice motive, 238                  free trade, fiscal
product standards, 174                  compenisation for, 97
reform requirements in              domino regionalism, 231, 234
agreements with Australia,      economies of scale, 52
110                             environmental standards, 179
RIAs, selected list of, 27          external tariffs, effect on, 84
services trade restrictions, 160    FDI, 101, 119, 135
Singapore, 231, 238                 gravity models, 41, 43, 44
Trans-Tasman MRA, 174               growth and industrial location, 18
Trans-Tasman Travel                 investment, 106, 240
Arrangement, 150                labor group interests, 168



INDEX     315
lobbyists' comments on, 88         MERCOSUR not competitive
Mexico's gains from, 74, 130,          with, 135
131, 143                       policy integration, role in, 172
migration controls, 197            Sub-Saharan Africa, trade with,
multilateralism and globalization,     78
223                            TFP levels in OECD countries,
natural trading partners fallacy,      128-130
66, 67, 68                    Organization of American States,
open access, 233, 234                  198
peace and security, 197          origin, rules of, 79-81, 252
policy integration, 150, 178     Ouchy Convention, 259n24
product standards, 174
public procurement, 163          P
rules of origin and trade        Pakistan, 29
deflection, 79-80             Panama, 75
services liberalization, 173     Papua New Guinea, 27
social standards, 179            Paraguay, 27, 84nl6, 198, 200
trade diversion and excluded     peace and security, 187-188
countries, 212-213             conflict triggered by integration,
North-North agreements, 66               194-196
North-South agreements, xi, 2-3,     historical background to trade as
262-263                            source of, 189
agglomeration and location of      insurance motive, 237
industries, 139-141, 142-143   modeling externalities, 191
comparative advantage, 70          political integration and trade
growth and, vs. South-South            integration, 203
agreements, see growth         RIAs and promotion of, 192-194,
investment, see investment             206n4
membership of RIA, choosing,       spread of war and unrest, effect of
73-74                              RIAs on, 196-197
natural trading partners fallacy,  trade as promoter of peace,
66                                 189-192
preference for, 15               Permanent Interstate Committee on
R&D, 129-130                           Drought Control in the Sahel
selected list of, 26-27                (CILSS), 67
tariff revenue transfers, 36     Peru, 27, 75, 78, 216-217
Norway, 26, 67, 150                Philippines, 27, 29
NTB, see nontariff barrier         Poland
growth and industrial location,
O                                        18, 123
OECD, see Organisation for           multiple agreements, 75
Economic Co-operation and       natural trading partners fallacy,
Development                        67
Oman, 29                             product standards, 183n3
open regionalism, see closed vs. open  RIAs, selected list of, 26, 28, 29
models of regionalism           rules of origin and trade deflection
Organisation for Economic Co-            concerns for FTAs, 80
operation and Development     policy integration, 19-21,
(OECD)                             147-149
investment, 106-107, 242           alternative routes, 148
knowledge capital, 132             antitrust policies, 164-165,
lobbying, 93                           176-177



316     INDEX
policy integration (continued)         services liberalizationi, 159-162,
border formalities, 155                  175-176
contingent protection, 163-164,      social standards, 165-166, 167,
177-178                              168, 178-179
coordination, 151-152                spillovers, 164-167
coverage of current RIAs,           trade liberalization, as
174-179                              compensation for, 168-170
credibility gains from, 168-169      transactions costs, 155-158
current RIAs, 174-180                WTO, see World Trade
defining, 149, t50-151                   Organization
depth of integration, 1 79-1 80    Portugal
economic effects of RIAs,            democracy, fostering, 199
154-155                          EU, admittance to, 19, 41
enforcement of free trade            growth and industrial location,
commitments, 168, 169-170,           18-19, 123
172-173                          migration controls, 198
environmental standards, 159,        reform, implementation of, 19,
165-166, 168, 178-179                111-112
free trade not dependent on, 167     RIAs, selected list of, 26
future prospects, 181-183            services trade restrictions,
globalization, 170-171                   161-162
harmonization, 152, 153, 154, 171  predation, 164, 183n7
human capital, 165-166, 167,       preferential trade agreements
168, 178-179                         (PTAs), 6
interest groups, 168               preferential trade liberalization, 140
internalization of externalities,  procompetitive effects
164                              policy integration, 158-164,
international norms, adoption of,        164-165, 176-177
170-174                          RIAs, 14, 31-32, 262
membership of RIA, choosuig,       procurement, public, 162-163, 171
170-174                        product differentiation, 46-47, 71
MFN status, 148, 171-173,          product standards, policy
181-182                              integration on1, 155-157,
MRAs, see mutual recognition             174-175
agreements                     protectionism, 5, 93
multilateral, 170-174                contingent protection, 163-164,
mutual recognition agreements,           177-178
149, 153,154                     emergency protection, 163
national treatment as baseline for,  external tariffs, 84
149-150                          external trade policies of customs
nondiscrimination principle, 158,        unions, 84-87
171-174                          regionalism as increasinig potential
political gains from, 168-169            for, 226-229
procompetitive effects, 158-164,    restaurant bill problem, 87
164-165, 176-177                 rules of origin, 80
product standards, 155-157,          universalism, 87
174-175                        PTAs, see preferential trade
public procurement, 162-163              agreements
recognition and acceptance of      public planning prolects, failure of,
foreign regulatory regimes,          103-1 04, 121nl
152-1 53,154                   public procurement, 162-1 63, 171



INDEX     317
Q                                     rules of thumb for, 25-26,
Qatar, 29                                261-266
selected list of, 26-29
R                                     time required (to become
R&D, see research & development          operational), 59n2
rationalization of industry, 103-104  Republic of China, see China,
reciprocity, diffuse, 224                Republic of
recognition and acceptance of       Republic of Congo, see Congo,
foreign regulatory regimes,         Republic of
152-153, 154                  research & development (R&D),
recycling requirements, economic          129-130, 134-135
effects of, 159               restaurant bill problem, 87
reform                             rest of the world (ROW), see
RIAs leading to, 108-112,              excluded countries; and see
227-228                             multilateralism and
WTO rules on RIAs, 251-255             globalization
regional integration agreements     RIAs, see regional integration
(RIAs), 1-4, see also more          agreements
specific topics, and specific  Ricupero, Rubens, 193
RIAs                          Romania, 26, 28, 29, 67, 75
burden of negotiating, 240        Rome, Treaty of, 249
choosing, reasons for, 15-16,     ROW, see rest of the world
63-64                         rules of origin, 79-81, 252
defined, 2                        Russian Federation, the, 27, 29,
developing countries, xi, 4             160, 203
difficult issues, handling,       Rwanda, 27, 28
239-242
disadvantages of, 23-25           S
historical background, 4-6        SAARC, see South Asian Association
numbers and statistics, i, 3            for Regional Cooperation
political objectives of, 7-8, 22-24,    and SAARC PTA
187-188                       SACU, see Southern African
bargaining powers,                    Customs Union
strengthening, 203-206        SADC, see Southern African
democracy, fostering, 188, 191,       Development Community
198-201                            (SADC)
economic and political          SADCC (Southern African
integration, connection             Development Coordination
between, 188, 201-206               Conference), see Southern
efficient RIAs, need for, 264         African Development
migration controls, 188,              Community
196-198                       Sahel, Permanent Interstate
peace and security, see peace         Committee on Drought
and security                        Control in the (CILSS), 67
policy integration, political   SAPTA, see South Asian Association
gains from, 168-169                for Regional Cooperation
procompetitive effects, 14, 31-32,      (SAARC) and SAARC PTA
262                           Saudi Arabia, 29
qualitative changes in, xi, 1-3   scale, economies of, 50-54
research prolect, xii, 3-4, 10-13  Scandinavia, 150, 204, 231
rise of, reasons for, 6-10        Schuman, Robert, 188



318     INDEX
security, see peace and security     services liberalization, 175
self-generated (endogenous) growth   tax harmonization and
theory, 125, 126, 132-133           competition, 99
Senegal, 28, 52-53, 72, 211         Southern African Development
services liberalization, 159-162,         Community (SADC), 6
175-176, 250-251                agglomeration and location of
settlement of disputes, 105, 107,         industries in Africa, lack of,
255, 258n19     ,                   140
Seychelles, 27, 28                    bargaining power, RIAs
shallow integration, 20, 151, 180        increasing, 205-206
Sierra Leone, 28                      domino regionalism, 235
Singapore, 27, 29, 231                member countries, 28
Slovak Republic, 26, 28, 29, 67, 75   natural trading partners fallacy,
Slovenia, 26, 28, 67, 75                  66, 67
social standards and policy           overlapping blocs, 76, 78
integration, 165-166, 167,      political integration and trade
168,178-179                        integration, 203
Somalia, 28                           services liberalization, 175
South Africa                          tax revenues, 94
EU agreement with, 71, 97         Southern African Development
external trade policies of customs      Coordination Conference
unions, 85                          (SADCC), see Southern
multiple memberships, 78                African Development
negotiating partners, RIAs as,          Community
239                           SouLth Korea, see Korea, Republic of
political integration and trade   South-South agreements, 262-263
integration, 203                agglomeration and location of
RIAs, selected list of, 28             industries, 139-141,
services trade restrictions, 160        142-143
tax revenues, 94                    growth and, see growth (North-
South Asian Association for               South vs. South-South
Regional Cooperation                agreements)
(SAARC) and SAARC PTA           investment, 17, 102, 116-117
(SAPTA), 29, 66, 67, 234        membership of RIA, choosing,
Southern African Customs Union            73-74
(SACU)                          natuLral trading partners fallacy,
British Colonial Office support         66, 67
for, 108                        preference for North-South
disproportionate shares of tariff-      agreements over, 15
free trade, fiscal              trade effects of RIAs between
compensation for, 97                developing countries, 59
domino regionalism, 234           Soviet Union, former, (Russian
excluded countries, trade               Federation), 29
discriminationi against, 211  Spain
external trade policies, 85         democracy, fostering, 199
member countries, 28                EU, admittance to, 41
negotiating partners, RIAs as, 239  migration controls, 198
neighborhood RIAs, 72               political integration and trade
North-South nature of, 108             integration, 203
overlapping blocs, 76, 78           reformi, implementation of,
regionalism and protectionism,          111-112
227                             RIAs, selected list of, 26



INDEX     319
specialization within region,       trade deflection, 79-81
103-104                       trade discrimination, see
spillovers                                nondiscrimination principle
growth, 125, 127-135              trade diversion, 13-14, 33-36
knowledge spillovers, see           defined, 31
knowledge capital and           discrimination causing, 210-211
spillovers                      evidence of, 212-213
policy integration, 164-167         excluded countries, 23, 212-221
Sri Lanka, 29                         export losses of excluded
St. Kitts and Nevis, 27                   countries, 213-215, 252-253
St. Lucia, 27                         external trade policies, 82
St. Vincent and the Grenadines, 27    investment switching, 220-221
"substantially all trade," 247, 248,  lobbying, 88-90
250-251, 253, 254               MFN status and reduction of, 251
Sudan, 28                             rules of origin, 79
Swaziland, 27, 28, 78                 simple analytics, 54-59
Sweden, 26, 150                       terms of trade of nonmembers,
Switzerland, 152                          RIAs affecting, 215-221
trade liberalization
T                                     agglomeration and location of
Taiwan (China), 27, 120, 160              industries, 138-141
Tanzania                              economic effects of RIAs
agglomeration and location of           compared to, 40-46, 63, 73
industries, 142                 policy integration as
collapse of East African                compensation for, 168-170
Community, 121nl, 195           preferential, 140
multiple memberships, 78            trade creation, liberalizing
public investment planning, failure     momentum of RIAs
of, 121nl                           generating, 92
RIAs, selected list of, 27, 28      unilateral, 73, 140
tariff levels and RIAs, 226-229       WTO encouragement of RIAs to
tariff preferences, economics of, see     increase, 251-252
economic effects of RIAs      trade negotiation process, effect of
tax competition, 98-99                    RIAs on, see negotiation
tax harmonization, 98-99            trade policy, regionalism as, see
tax on exports, 214                       economic effects of RIAs
tax revenues, 94-99, 265            traditional (neoclassical) growth
TFP, see total factor productivity        theory, 125
Thailand, 27, 29, 211               transactions costs, 114, 155-158, 265
time required (for RIA to become    Trans-Tasman MRA, 174
operational), 59n2            Trans-Tasman Travel Arrangement,
Togo, 28, 104                             150
total factor productivity (TFP)     Treaty of Rome, 249
growth, 127, 128-131          Trinidad and Tobago, 27
trade blocs, see regional integration  Tunisia, 26, 67
agreements trade creation,    Turkey
14, 33-36                       EU and, 81-82, 178, 234, 253
defined, 31                         migration controls and, 198
liberalizing momentum of RiAs       North-South agreements with EU,
generating, 92                      3
mixed blessing of, 89               reform, RIAs leading to, 227
simple analytics, 54-59             services trade restrictions, 160



320    INDEX
U                                     Article XXIV of GATT, 259n24
UDEAC, see Union douaniere et         BITs for FDI, 105
economique de l'AfriqLie        Brazilian imports from, 217-219
Centrale                        CARICOM market access
UEMOA, see West African                  agreements, 205
Economic and Monetary           change in attitude toward RIAs,
Union                               10
Uganda                                civil war triggered by integration,
agglomeration and location of           194-195
industries, 142                 CUSFTA, see Canada-United
collapse of East African                States Free Trade Agreement
Community, 121nl, 195              (CUSFTA)
comparative advantage, 69-71        democracy, fostering, 198
RIAs, selected list of, 27, 28      domino regionalism, 231
U.K., see United Kingdonm             economies of scale, 52
unilateralism, concerted, 243         FDI, 120
unilateral trade liberalization, 73,  FTAA negotiations, 251
140                             historical background to RIAs, 4
Union douaniere et economique de     insurance motive, 236, 237, 238
I'Afrique Centrale (UDEAC)      Israel, agreement with, 71
CEMAC, 28, 76, 140, 235             Mexico, ties to, 108
excluded countries, trade           migration controls, 197, 198
discrimination against, 211     MRAs, 149, 156-157, 173-174
increased trade between members,    multilateral negotiations and
32                                 regionalism, 230
investment and growth, 116-117      NAFTA, see North American Free
natural trading partners fallacy,       Trade Agreement (NAFTA)
66, 67                          Ouchy Convention, 259n24
neighborhood RIAs, 72               peace and security, 194-195,
openness associated with, increase      196-197
in, 38                          policy integration, 20
tax harmonization and tax             antitrust law, 165
competition, 99                   contingent protection, 163, 178
trade effects of RIAs between         harmonization, 152
developing countries, 59         investment, 150
Union economique et mon6taire           MRAs, 149, 156-157, 173-174
ouest-Africaine, see West         product standards, 156-157,
African Economic and                174
Monetary Union                    regulatory norms, adapting, 148
United Arab Emirates, 29               services trade restrictions, 159,
United Kingdom                            160, 162
agglomeration and location of         social and environmental
industries, 141                     standards, 159, 166
FDI, I 1 9                          recycling requirements, economic
insurance motive, 236                   effects of, 159
negotiating partners, RIAs as, 239  RIAs, selected list of, 26, 27
Ouchy Convention, 259n24            SADC understanding with, 206
peace and security, 189             services trade restrictions, 159,
RIAs, selected list of, 26              160, 162
United States                         Singapore seeking agreement with,
agglomeration and location of           231
industries, 141                 trade deflection, 81



INDEX     321
universalism, 87                      CRTA, 249, 254-255
Uruguay, 27, 52, 84n 16               dangers of reliance on, 265-266
U.S., see United States               democracy requirements, 201
developing countries, means of
V                                         assisting, 246-247
value-added tax (VAT), 94-99          difficult issues, handling, 241
Venezuela, 27, 75, 78, 121nl, 211     dispute settlement procedures,
Vietnam, 27, 29                           255, 258n19
Enabling Clause, 26, 249-250
W                                     enforcement issues, 249, 253
war, see peace and security          external trade policies for customs
West African Economic and                unions, 84
Monetary Union (Union           GATT, succession of WTO to, see
economique et monetaire             General Agreement on Tariffs
ouest-Africaine-UEMOA)              and Trade
agglomeration and location of       handling of RIAs by, 24-25
industries in Africa, lack of,  investment credibility, and reform,
140                                 111
CEAO, previously, see Economic      multilateralism and globalization,
Community of West Africa            223
domino regionalism, 235             multilateral negotiations and
economies of scale, 52-53               regionalism, 230
excluded countries, trade           negotiating partners, RIAs as, 239
discrimination against, 2111    number of members, 258n1 9
lobbying and external trade         policy integration
policy, 88                        APEC, 181-182
member countries, 28                  depth of integration, 179-1 80
overlapping blocs, 76                 Doha Development Agenda,
regionalism and protectionism, 227      150
tax harmonization and                 FTAA, 182
competition, 99                   global standards, 172, 173
tax revenues, 94                      product standards, 174-175
World Bank                             role in, 21, 148
International Centre for             services liberalization, 175
Settlement of Investment        policy on regionalism, 244-245
Disputes (ICSID), 105           public procurement, 163, 171
research project on regionalism    reform of rules on RIAs, 251-255
and development, xii, 3-4,
10-13                         Y
World Trade Organization (WTO)      Yugoslavia, Federal Republic of, 29
Article XXIV of GATT as rules
for RIAs, 247-251, 256        Z
bargaining powers, RIAs           Zambia, 27, 28, 94
strengthening, 204            Zimbabwe, 27, 28, 94
contingent protection, 163        Zollverein, 4



.relations. Regionnil lItegi-arioii anid Development examines regionatisim fronm the
viewpoint of dleveloping countries and provides a comprehensive account of existing
economic theory and empirical results. In a thorough analysis of the politics and
dynamics of regionalism, the book considers the relationship between regionalism and
;multilateralism  and explores the economic advantages of nondiscriminatory trade
liberalization, which the authors argue should be exploited to the maximum extent.
Designed as a practical guidc for policymakers whose countries are considering
-  membership in a trade bloc, this book will also be of interest to scholars and students of
international trade.
Giveni the remarkable increase in regional integration agreements [RIAs] in the past decade,
understanding of their economic rationale and consequences is essential. Schiff and Winters
succeed admirably in covering the conceptual, empirical, and policy issues arisingfrom WIAS in an
infonned, comprehenisible, and accessible mannier. Their book should be required reading for
policy analysts, academics, and students concerned with evaluation of RIA initiatives.
Regional Integration and Development analyzes all aspects of an issue too often debated from
simplistic viewpoints: regional integration agreements. Its authors provide the right baLance of
economic anatlysis, facts, and political perspectives, of subtleness and decisiveness, that such a
delicate topic requires. This easy-to-read book is applied economic analysis at its best.
Regional Integration and Development is essential readingfor developing countries' economic
policymakers and those who advise them. Cuttinig through the rhetoric, it assesses the effects on
development of regional integrarion agreements in a remarkably clear and readable style. The
book's concluding list of practical lessons is particularly valuable.
i ~~...... .. . :,,,, 1.. 
-4E\l .\j                ' 'Al>.  '' * .. ) ,1. AZ A  (..Mm
*  U \;'A     %! * 1>1 \..  '  ) *  "\i::  * Ii I2 'ii  X  ''.,"l'*  X  \)-\'         15078 -\  *t_.(!
-Zi~  ~         .\l.' U(: :Z  * ,  ' t  *  ,:V N x i 7 ;..* |jUls 
E~~~~;A            1 E                           i~   u If1 fi111f *1 1:llbll
i S THE WORLD BANK *                             )             i 9 780821 350782@
--\    * N I              *              4      '' Itt  A FT  ISBN: 0-8213-5078-1
XI  \  x  *  :. -X Xt t t  .\l  * (:  sI -f  >*  t i  - I. i. .*  t. .  .  s  v  \  t  \  t v ). l1 " -