Assessment and Measurement of Regional Integration

Assessment and Measurement of Regional Integration How fast and how deep are current processes of regional integration...

Author: Philippe De Lombaerde

28 downloads 1317 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Assessment and Measurement of Regional Integration

How fast and how deep are current processes of regional integration? What is their impact for different countries and how can it be measured? How are they related to globalisation and the quest for global free trade? How important is the macroregional level of governance? Using an interdisciplinary approach, this book explores how regional integration can be quantified, evaluated and monitored. It investigates the methodological problems involved in designing monitoring tools for regional integration, and makes suggestions for designing and organising systems of indicators of regional integration. The volume is organised into three sections: • • •

Part I discusses general theoretical and methodological aspects of building indicator systems. Part II focuses on the measurement of regional economic integration and interdependence. Part III explores the possibilities for monitoring political integration, regional governance and conﬂict.

This book will be of interest to students and researchers of international relations and economics as well as policy makers and professionals within international and regional organisations. Philippe De Lombaerde is an economist and a Research Fellow at the United Nations University – Comparative Regional Integration Studies (UNU-CRIS) in Bruges, Belgium.

Routledge/Warwick Studies in Globalisation Edited by Richard Higgott and published in association with the Centre for the Study of Globalisation and Regionalisation, University of Warwick.

What is globalisation and does it matter? How can we measure it? What are its policy implications? The Centre for the Study of Globalisation and Regionalisation at the University of Warwick is an international site for the study of key questions such as these in the theory and practice of globalisation and regionalisation. Its agenda is avowedly interdisciplinary. The work of the Centre will be showcased in this new series. This series comprises two strands: Warwick Studies in Globalisation addresses the needs of students and teachers, and the titles will be published in hardback and paperback. Titles include: Globalisation and the Asia-Paciﬁc Contested territories Edited by Kris Olds, Peter Dicken, Philip F. Kelly, Lily Kong and Henry Wai-chung Yeung Regulating the Global Information Society Edited by Christopher T. Marsden Banking on Knowledge The genesis of the global development network Edited by Diane Stone Historical Materialism and Globalisation Edited by Hazel Smith and Mark Rupert Civil Society and Global Finance Edited by Jan Aart Scholte with Albrecht Schnabel Towards a Global Polity Edited by Morten Ougaard and Richard Higgott New Regionalisms in the Global Political Economy Theories and cases Edited by Shaun Breslin, Christopher W. Hughes, Nicola Phillips and Ben Rosamond

Routledge/Warwick Studies in Globalisation is a forum for innovative new research intended for a high-level specialist readership and the titles will be available in hardback only. Titles include: 1 Non-State Actors and Authority in the Global System Edited by Richard Higgott, Geoffrey Underhill and Andreas Bieler 2 Globalisation and Enlargement of the European Union Austrian and Swedish social forces in the struggle over membership Andreas Bieler 3 Rethinking Empowerment Gender and development in a global/local world Edited by Jane L. Parpart, Shirin M. Rai and Kathleen Staudt 4 Globalising Intellectual Property Rights The TRIPS agreement Duncan Matthews 5 Globalisation, Domestic Politics and Regionalism The ASEAN Free Trade Area Helen E. S. Nesadurai 6 Microregionalism and Governance in East Asia Katsuhiro Sasuga 7 Global Knowledge Networks and International Development Edited by Diane Stone and Simon Maxwell 8 Globalisation and Economic Security in East Asia Governance and institutions Edited by Helen E. S. Nesadurai 9 Regional Integration in East Asia and Europe Convergence or divergence? Edited by Bertrand Fort and Douglas Webber 10 The Group of Seven Finance ministries, central banks and global ﬁnancial governance Andrew Baker 11 Globalisation and Poverty Channels and policy responses Edited by Maurizio Bussolo and Jeffery I. Round 12 Democratisation, Governance and Regionalism in East and Southeast Asia A comparative study Edited by Ian Marsh 13 Assessment and Measurement of Regional Integration Edited by Philippe De Lombaerde

Assessment and Measurement of Regional Integration Edited by Philippe De Lombaerde

First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2006 Philippe De Lombaerde for selection and editorial matter; individual contributors, their contributions

This edition published in the Taylor & Francis e-Library, 2006. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Assessment and measurement of regional integration / edited by Philippe De Lombaerde. p. cm. – (Routledge/Warwick studies in globalisation ; 13) Includes bibliographical references and index. 1. Regional economics. 2. Regionalism. 3. International economic integration–Evaluation. 4. Regionalism (International organization)– Evaluation. I. Lombaerde, Philippe de. II. Series. HT388.A77 2006 337.1′072–dc22 2005022668 ISBN10: 0–415–36635–6 (hbk) ISBN10: 0–203–08789–5 (ebk) ISBN13: 978–0–415–36635–9 (hbk) ISBN13: 978–0–203–08789–3 (ebk)

To Liliana, Griet and María Paz P.D.L.

Contents

List of ﬁgures List of tables List of contributors Acknowledgements Introduction and summary

xi xiii xv xvii 1

PHILIPPE DE LOMBAERDE

PART I

General theoretical and methodological aspects

7

1 Indicators of regional integration: conceptual and methodological aspects

9

PHILIPPE DE LOMBAERDE AND LUK VAN LANGENHOVE

2 Towards a qualitative monitoring of regional integration

42

LUK VAN LANGENHOVE

3 Communicative indicators for the study of regional integration: a critical theory perspective

52

ANN-CHRISTIN RASCHDORF

PART II

Assessment of economic integration 4 Regional integration agreements and the geography of world trade: statistical indicators and empirical evidence LELIO IAPADRE

63

65

x Contents 5 Explaining regionalisation via relative trade intensities

86

GUILLAUME GAULIER, SÉBASTIEN JEAN AND DENIZ ÜNAL-KESENCI

6 Assessing the impact of regional trade agreements: an analytical framework for rulemaking in trade and investment

107

STEPHEN WOOLCOCK

7 Comparative ﬁscal integration indicators

130

PHILIPPE DE LOMBAERDE AND ANA-CRISTINA COSTEA

8 Indicators of real economic convergence

146

JOSÉ VILLAVERDE CASTRO

9 Measuring and explaining levels of regional economic integration

162

JENNIFER PÉDUSSEL WU

PART III

Assessing political integration, regional governance and peace

181

10 Regional integration and (good) regional governance: are common standards and indicators possible?

183

EDWARD BEST

11 Assessing accountability of global and regional organisations

215

MONICA BLAGESCU AND ROBERT LLOYD

12 Measuring the impact of regional organisations on peace building

232

RODRIGO TAVARES AND MICHAEL SCHULZ

Index

252

Figures

1.1 1.2 1.3 1.4

Dobson’s ‘Policy Conﬂict–Independence–Integration’ spectrum The independence–interaction spectrum Preconditions in phased integration processes Regional integration as a process: proposal for a conceptual framework 3.1 Parsons’ pattern variables revisited 4.1 Intra-regional trade shares 4.2 Intra-regional trade intensity (traditional Balassa indicator) 4.3 Intra-regional trade intensity (homogeneous indicator) 4.4 Intra-regional trade intensity (symmetric indicator) 4.5 Intra- and extra-regional trade intensity (EU-15) 4.6 Trade introversion 4.7 Relative trade openness 4.8 Relative intra-regional trade propensity 5.1 The polarisation of foreign trade with regard to the Triad, by country, 2001 5.2 Breakdown of Regional Trade Agreements by geographical area 8.1 Lorenz curve 8.2 Per capita income distribution – polarisation versus inequality 8.3 Convergence clubs 8.4 Stochastic kernel 9.1 Evolution of regional trade agreements in the world 1948–2002 9.2 Types of regional integration arrangements 9.3 Regional integration in 150 countries 1965–1995 12.1 Violence and peace are part of the dyadic relationship of a continuum

14 14 23 25 56 76 76 77 78 78 79 79 80 88 95 149 150 153 156 162 164 166 245

Tables

1.1 1.2

Anderson’s criteria for choosing ‘good’ economic indicators Overlapping memberships of integration arrangements for African countries 1.3 Proposal for classifying variables in a SIRI 1.4 Bensidoun and Chevallier’s indicators of regionalisation 5.1 Top relative trade intensities in America 5.2 Top relative trade intensities in Eurafrica 5.3 Top relative trade intensities in Asia-Oceania 5.4 Estimating a gravity model for RTIs, 1967–2001 6.1 Proposal for an analytical framework for rulemaking in trade and investment 7.1 The World Bank’s list of ﬁscal decentralisation indicators 9.1 Ordered probit regression 1960–98 9.2 Ordered probit regression 1987–98 9.3 Bivariate probit regression 1960–98 9.4 Bivariate probit regression 1987–98 12.1 Regional integration dimensions 12.2 Various forms of regional interventions

12 19 26 30 90 91 93 100 112 136 172 174 175 176 242 247

Contributors

Edward Best, Head of Unit, European Institute of Public Administration (EIPA), Maastricht . Monica Blagescu, Accountability Programme Manager, One World Trust, London , and PhD candidate at the University of Bath. Jose Villaverde Castro, Professor, Department of Economics, Universidad de Cantabria, Santander . Ana-Cristina Costea, Assistant to the Director, United Nations University – Comparative Regional Integration Studies (UNU-CRIS), Bruges . Philippe De Lombaerde, Research Fellow, United Nations University – Comparative Regional Integration Studies (UNU-CRIS), Bruges . Guillaume Gaulier, Economist, Centre d’Études Prospectives et d’Informations Internationales (CEPII), Paris . Lelio Iapadre, Associate Professor of International Economics and Economics of European Integration, University of L’Aquila, and Associate member of the Inter-department Centre of International Economics of the University of Rome ‘La Sapienza’. Robert Lloyd, Accountability Project Ofﬁcer, One World Trust, London . Rodrigo Tavares, PhD candidate, Department of Peace and Development Research (PADRIGU), Göteborg University . Ann-Christin Raschdorf, Political Affairs Officer, Office of the Special Representative for Sierra Leone, UNAMSIL. Michael Schulz, Senior Lecturer, Department of Peace and Development Research, and Director, Centre for Middle Eastern Studies, Göteborg University .

xvi Contributors Sébastien Jean, Senior Economist, OECD Economics Department ; External Fellow, Leverhulme Centre for Research on Globalisation and Economic Policy, University of Nottingham; and Managing Editor of Economie Internationale. Deniz Ünal-Kesenci, Economist, Centre d’Études Prospectives et d’Informations Internationales (CEPII), Paris and associated with the network of the Forum of Euro-Mediterranean Economic Institutes (FEMISE) and with the Groningen Growth and Development Centre (GGDC). Luk Van Langenhove, Director of the Comparative Regional Integration Studies Programme of the United Nations University (UNU-CRIS), Bruges and Professor at the Vrije Universiteit Brussel (VUB) . Stephen Woolcock, Associate Research Fellow, UNU-CRIS , and Lecturer in International Relations and Coordinator of the Trade Policy Unit, London School of Economics (LSE), London . Jennifer Pédussel Wu, Senior Fellow, Zentrum für Europeische Integrationsforschung – Center for European Integration Studies (ZEIb), Bonn .

Acknowledgements

This book arose from papers presented in the context of a Virtual Workshop on Indicators of Regional Integration that took place in 2003–2004 and was convened by UNU-CRIS, and in which academics and practitioners from all over the world participated. Monitoring and assessing regional integration processes is one of the research priorities of UNU-CRIS (see www.cris.unu.edu). We wish to thank all those that participated in the workshop, many of them being co-authors of this book, and those that joined the book project later. We are grateful to Luk Van Langenhove, Director of UNU-CRIS for his support, and to Kim Reynvoet and Eveline Snauwaert for their outstanding job in formatting the manuscript. We are also indebted to Heidi Bagtazo, Grace McInnes, Harriet Brinton and staff of Routledge, Kelly Hallett, Andrea Platts, Emma Wood, Jackie Mills and last, but not least, series editor Professor Richard Higgott from Warwick University. Philippe De Lombaerde Bruges

Introduction and summary Philippe De Lombaerde

Without any doubt, regional integration will continue to be one of the driving forces that will shape the world polity and economy in the years to come. Regional integration is thereby understood as covering a variety of forms and modalities. From a governance point of view, scholars are now focusing on the present and future role that the (macro-)regional level might play in a complex multilevel governance context. Both theoretical approaches and real world cases show that the regional level is likely to maintain and probably increase its importance in the global governance architecture. There is not necessarily a consensus, however, about certain causalities or tendencies. It is not clear, for example, whether regional economic integration contributes or not to multilateral economic governance, whether the regional level is an optimal level for the provision of certain public goods, whether regionalism contributes or not to regional and global political stability, whether it contributes to the alleviation of underdevelopment and poverty, whether it contributes to ‘good’ and democratic governance, whether interregionalism will become more and more important in international relations, etc. In order to ﬁnd answers to these types of questions, scholars need to have good data at their disposal and a good understanding of the dynamics and mechanics of regional integration processes. All too often facts and trends are taken for granted and not scientiﬁcally veriﬁed. New theories or theoretical ideas do not always have a solid empirical basis. On the other hand, policy-makers, regional leadership and civil society do not always have the necessary monitoring tools at their disposal to follow up these processes, deﬁne their political positions and propose and design policies and measures. This book aims to contribute to the advancement of assessment and measurement tools for monitoring regional integration processes in the world. The starting points (working hypotheses) of this joint effort include the following: • •

regional integration processes are complex and multidimensional phenomena and therefore require a multidisciplinary approach, regional integration processes are heterogeneous and cannot be modelled in a unique way,

2 Philippe De Lombaerde • •

monitoring of regional integration requires a combination of quantitative indicators and qualitative assessments, monitoring can increase democratic participation and therefore improve the quality of policy-making in the context of the regional integration processes.

The book explores the methodological problems involved in designing monitoring tools for regional integration in a systematic way and makes a number of concrete suggestions for designing and organising systems of indicators of regional integration. It also offers a critical overview of available indicators and tools and presents some new monitoring instruments. The book is organised in three parts. In the first part, general theoretical and methodological aspects of building indicator systems are discussed. The second part focuses on the measurement of regional economic integration and interdependence. And, ﬁnally, the third part explores the possibilities for monitoring political integration, regional governance and peace.

PART I Chapter 1 explores the panorama of conceptual, methodological and organisational issues that arise in the process of constructing a multidimensional system of indicators for regional integration processes. De Lombaerde and Van Langenhove also present a proposal for the classiﬁcation of regional integration variables and indicators for reference, comparison and design purposes. The authors end with a check-list of issues that should be addressed when building indicator systems. Van Langenhove, in Chapter 2, argues that next to quantitative data, studying regional integration also needs empirically based qualitative data and advocates that a qualitative monitoring and impact assessment method can be constructed upon the so-called ‘fourth generation evaluation’ model and that such an approach allows to develop a participatory monitoring and assessment scheme able to demonstrate the ‘added value’ of regional integration. The argument in favour of qualitative assessments is taken further by Ann-Christin Raschdorf in Chapter 3, stressing that communicative indicators should be included and operationalised. Special attention is given to procedural indicators for communicative participation and empowerment.

PART II Iapadre, in Chapter 4, presents a critical survey of the available statistical indicators for the measurement of intra-regional trade intensity, as well as some proposals in order to overcome their limitations (such as the symmetrical trade-introversion index). He further discusses some indicators that can be obtained by combining trade and GDP data, aimed at measuring the trade-creating effects of regional

Introduction and summary 3 integration. The indicators that are presented are then applied to the analysis of intra-regional trade within four regional integration agreements (ASEAN, EU, Mercosur and NAFTA). From this application, it becomes clear that the empirical assessment of intra-regional trade is strongly influenced by the choice of the statistical instrument used to measure its importance. When assessing and interpreting the dynamics of trade integration, this should be borne in mind. In order to evaluate whether trade within a given region is more intense than elsewhere, Gaulier, Jean and Ünal-Kesenci (Chapter 5) ﬁrst look at the ‘normal’ level of trade between partners (and, in particular, neighbours), and its ‘normal’ evolution across time. The authors argue in favour of controlling for country-pair fixed effects in the estimations when assessing the impact of regional trade arrangements (RTAs) on trade ﬂows. They ﬁrst consider different indexes to assess the regionalisation of international trade and then go on to study these observed patterns using a gravity model based on a global data set. They find that the distorting impact of RTAs on the geographical polarisation of member countries’ foreign trade appears to be far more limited than the one based on OLS estimations. The impact of regional agreements is clear as far as Western Europe or Mercosur are concerned. In many cases, however, causality is not straightforward. The aim of Woolcock’s chapter (Chapter 6) is to present an analytical framework for making qualitative assessments of the impact of RTAs. It argues that the current wave of preferential agreements necessitates augmenting the existing theoretical and analytical frameworks. The rulemaking (or deeper integration) that is involved in the new wave of preferential agreements necessitates the study of the detailed content of the agreements, which has been eschewed by many of the more model-based approaches to assessing RTAs. Furthermore, as rulemaking invariably touches on ‘governance’ issues in that trade or investment rules limit domestic regulatory autonomy, there is a need to be able to more easily identify how regional or free trade agreements impinge upon domestic regulatory policy autonomy. The chapter provides a taxonomy of the key elements of trade and investment agreements that is applicable to most issues covered by RTAs and all aspects of rulemaking, in order to facilitate more effective comparisons of RTAs, both across regions and with the multilateral level. De Lombaerde and Costea (Chapter 7), present an overview of the conceptual framework offered by ﬁscal federalism theory and the technical tools which have been designed to analyse (ﬁscal) decentralisation processes. They then go on to explore how these tools could be adapted in order to be used for the monitoring and comparative analysis of regional integration processes and, more speciﬁcally, how fiscal integration indicators could be designed. The authors present a preliminary discussion of the specific methodological and practical issues and problems that can be envisaged in doing so. Villaverde, in Chapter 8, deals with the measurement of economic convergence or divergence, a much and growingly debated issue at the regional level in Europe and elsewhere. The focus of the chapter is with real convergence; that is, with the long-term process of reducing inequalities across economies. The chapter not only explores the σ-convergence and β-convergence approaches, but also examines

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

4 Philippe De Lombaerde other less standard concepts like convergence as reduction of inequality, stochastic (or time-series) convergence, dynamics distribution convergence and spatial analysis convergence, and how to measure these. Wu’s chapter (Chapter 9), provides a more complete empirical test to identify factors contributing to different levels of participation in regional integration agreements (RIAs). Using the level of regional integration as the dependent variable, a probit model is developed and tested. The author argues that data on regional integration have not been previously gathered in a systematic manner. Therefore, a database of regional integration is developed which ranks the levels of integration across countries. The main conclusions of the chapter are, ﬁrst, that high-income, democratic countries with high populations who are landlocked and are not islands are more likely to belong to deeper levels of regional integration; and, second, that only trade uncertainty, and not other forms of economic and/or political uncertainty, appears to be of any signiﬁcance to explain the chosen level of regional integration.

PART III The aim of Best’s chapter (Chapter 10) is to explore whether regional systems can be assessed according to common ‘indicators’ of (good) regional governance. It is argued that regional systems do need to be evaluated in terms of their democratic (i.e. ‘good-governance’) quality and that such an exercise is plausible. The question which is addressed is whether one can evaluate the quality of regional institutional arrangements, in those cases where an integration process produces a regional system of signiﬁcant multilevel governance. It is argued that – beyond the basic operating standards, and assuming that the regional organisation does not constitute a negative impact on good governance within countries – the quality of a particular set of regional arrangements can be evaluated in terms of stability and legitimacy. A set of basic principles is suggested by which regional bodies and systems can be assessed: openness, accountability, participation, effectiveness and appropriateness. In the next chapter (Chapter 11), Blagescu and Lloyd further develop the concept of accountability within a multilevel governance setting. They challenge the traditional approaches to accountability and note that, in thinking about regional accountability, the focus should not lie with inter-state institutions alone; non-state actors such as transnational corporations (TNCs) and international non-governmental organisations (NGOs) have considerable impact on regional and global organisations as well. In this regard, it is argued that the lack of accountability among regional institutions cannot be amended solely through state-based accountability systems but that new approaches to accountability need to be developed. Any indicators will need to try to capture the multifaceted and multilevel understanding of accountability. In the final chapter (Chapter 12), Tavares and Schulz concentrate on the phenomenon of regionalism and its contribution to intra- and extra-regional peace

Introduction and summary 5 building by suggesting some (quantitative and qualitative) measures of regional integration, regional peace and regional intervention. After a conceptual discussion on regionalism and regionalisation and on their linkages with peace and conflict, they offer some considerations about what to measure and how to operationalise the impacts of regional organisations on the variable peace. They propose indicators that can cover how regions cope with open conﬂicts; these relate to regional integration, regional peace, as well as to interventions. Tavares and Schulz argue that the regional peace dividend cannot merely be quantitatively measured but must also include some qualitative evaluations concerning the way regional interventions are related to time, scope and goal.

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Part I

General theoretical and methodological aspects

1

Indicators of regional integration: conceptual and methodological aspects Philippe De Lombaerde and Luk Van Langenhove1

INTRODUCTION Regional integration (RI) refers to processes of complex social transformations characterised by the intensification of relations between independent sovereign states. It creates new forms of organisation, co-existing with traditional forms of state-led governance at the national level. The processes of regional integration that emerged after the Second World War, were originally mostly about trade and economics, but it has become clear that, especially since the 1980s, with the so-called ‘new regionalism’ wave, regional integration can be seen as a multidimensional process that implies, next to economic cooperation, also dimensions of politics, diplomacy, security, culture, etc.2 Trade and economic integration remain a central aspect of ongoing integration schemes. And in today’s globalising world, there seems to be a growing tension between the quest for global free trade (as managed through the WTO) and the fact that trade liberalisation is more often regionally than multilaterally organised. Indeed, the number of regional integration initiatives is steadily increasing. About 250 RTAs are currently in force (Gavin and Van Langenhove, 2003). This trend raises the question of what the impact of regional integration is, both for the countries involved and for those excluded from a given regional integration scheme. Also, given the fact that the ‘new regionalism’ is said to have at least the potential to protect countries from possible negative aspects of globalisation (Van Langenhove, 2003), one can wonder how such claims can be assessed. However, these issues are linked to broader questions about the future of the international institutional and governance architecture, and whether tendencies may be expected in the direction of multilateralism, (multi-)regionalism or a (new) combination of both (Fratianni and Pattison, 2001; Hettne, 2005). Looking for answers to these questions implies that one needs to have tools to monitor regional integration and assess its impacts. This could be realised through a system of indicators of regional integration (SIRI). There is evidence of a growing interest of policy-makers for such a system. The explicit objective to monitor RI processes in ACP countries in the framework of the Cotonou Agreement (European Commission 2002a, 2002b) and the proposal by the

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

10 Philippe De Lombaerde and Luk Van Langenhove European Central Bank (ECB) to measure institutional and economic integration (Dorrucci et al., 2002), illustrate this point. The Inter-American Development Bank is also considering as one of its priority actions within its Strategy on Regional Integration, ‘gathering, evaluation and dissemination of compatible data to measure progress in the regional integration effort’ (IADB, 2002), whereas ALADI planned the preparation of yearly reports on the evolution of integration in Latin America. Proposals for worldwide systems of indicators on other speciﬁc aspects of governance, like the database under construction for the Fiscal Decentralisation Indicators of the World Bank, also suggest the feasibility of such a system.3 And ﬁnally, the African economic integration indicators project at UNECA needs to be mentioned.4 Notwithstanding the above-mentioned interesting efforts, a SIRI able to monitor RI worldwide and based upon sound methodological premises does not exist yet. Hence, this chapter explores organisational, conceptual and methodological issues that arise in the process of constructing a SIRI. Also, a proposal regarding the dimensions and variables to be included in a SIRI will be advanced.

USERS AND PRODUCERS OF A SIRI A ﬁrst issue to tackle is ‘who’ should and could be involved in constructing and operating a SIRI as well as ‘for whom’ a SIRI has to be developed. Ideally, the construction of a SIRI should not be seen exclusively as an academic project, but also as an exercise in building up a relevant policy instrument. The recent announcement by the European Commission (EC) of its intention to monitor the progress of regional integration schemes which involve ACP countries and to condition future resource allocation in terms of demonstrated progress in these processes, is a clear illustration of the scope and importance of such a system and the possibilities of combining an academic interest with political relevance. The announcement of the EC is a logical consequence of the inclusion of regional integration as a goal and means of development and development cooperation in the framework of the Cotonou Agreement.5 So, regional organisations are likely candidates for being involved in the construction of a SIRI. Not only because, as in the case of the EC, they want to monitor RI elsewhere but also because regional organisations want to engage in monitoring their own policies. One can also imagine that individual countries could be interested in a SIRI as a tool to monitor the impact of their involvement in regional trade agreements or other regional cooperation schemes on their domestic performances. There are also academic interests in developing a SIRI. For instance, to address the question to what extent regional integration initiatives can help in reducing the possible negative consequences of globalisation. Or: to what extent have regional integration initiatives positive effects on raising the quality of life in a developing region? The above implies that many different actors can, for different reasons, be interested in a SIRI.

Indicators of regional integration 11 For cost-efﬁciency reasons, but also for political reasons, it would be useful to pool resources for the construction of a SIRI. Three types of institutions would be natural candidates for participating in such a project: multilateral institutions (UN, WB, WTO), regional institutions and the academic sector. In addition, certain interest groups that are operating internationally (NGOs, foundations) could also be interested in participating. Designing and implementing a SIRI will only be successful if conceived as a joint product between ‘users’ and ‘producers’, taking into account input from all relevant stakeholders.

GENERAL CRITERIA FOR THE IMPLEMENTATION OF A SIRI Independently from the choice of overall objectives, dimensions and contents of the system, a whole range of problems are likely to occur at the moment of implementation. These problems range from the difﬁculty of measuring a speciﬁc variable and the quality of a particular data source to the problems related to managing and funding the system. Other problems relate to the availability, generation and frequency of the data needed to feed the system, and the management of a workable and sustainable system of indicators. Sustainability is likely to involve some degree of co-responsibility from the side of the data-generating institutions. The quality of the data is needed to guarantee comparability. A crucial issue will obviously be the translation of the chosen variables into indicators. Although each variable brings with it its specificities, some general criteria should be identiﬁed for selecting appropriate indicators. Especially when systems are multidimensional (multidisciplinary) it is useful to have some generally applicable criteria. Anderson (1991) has offered an example of an attempt to identify such criteria (see Table 1.1), but it should be noted that, although thought-provoking, these criteria were developed in the context of the national economy. If a SIRI is used for comparative research, as would be in the case of an indicator system for monitoring different regional integration processes in the world, a choice is possible between traditional (‘comparative’) indicators (permitting a direct comparison between regions on their score on a particular variable) and ‘relative’ (‘reﬂexive’) indicators (comparing ﬁrst the performance of each region with its own objectives). The World Bank (2002), for example, favours relative comparisons. A concrete example of a case where both types of indicators are combined is the system of Indices of Economic Integration Effort in Africa (UNECA, 2001: 2). In that system two yardsticks are used: 1 2

the self-defined pre-determined targets for target-driven indicators (if they exist for particular integration groupings), or an average of the n best performers.

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

12 Philippe De Lombaerde and Luk Van Langenhove Table 1.1 Anderson’s criteria for choosing ‘good’ economic indicators Criteria that SHOULD NOT be used

Criteria that SHOULD be used

1 An indicator should have to carry with it an automatic evaluation.

1 An indicator, or the information it is calculated from, should be already available, or else able to be made available easily and cheaply. 2 An indicator should be relatively easy to understand.

2 An indicator should have a policy instrument which corresponds to it. 3 An indicator has to be new. 4 Proposals for sets of indicators should be based on particular theories of economic, social and human development.

3 An indicator, to work at all, must be about something measurable. 4 An indicator should measure something believed to be important or signiﬁcant in its own right, or should reﬂect or represent something important beyond what the indicator is itself a measurement of (for example, life expectancy ﬁgures might be used to indicate the general state of health of the population). This is really what makes something an indicator, rather than just a statistic. 5 There should preferably only be a short time-lag between the state of affairs referred to and the indicator becoming available. 6 An indicator should be based on information which can be used to compare different geographical areas, social groups, etc., so that a picture of distribution – and not just totals and averages – can be built up. 7 International comparability is desirable.

Source: Anderson (1991: 48–51).

Furthermore, designing an indicator system can be either conceived as an ordered presentation of the values of the selected relevant variables, permitting – for each variable – cross-country or cross-region comparisons and time series analysis, but without establishing explicit weights for the variables and their categories. Or, one can design the system so that it is based on the calculation of aggregate indicators per country, per region and/or per sector. This is when one is confronted with the so-called ‘index problem’. The weighting and aggregation procedures that are designed ‘pre-process’ the data so that the reading by the users is simpliﬁed, not necessarily its interpretation. The weighting procedure can be based on statistical criteria (based on the statistical contribution of the variables in the explanation of a goal variable), expert opinion or practical considerations (data availability, lack of knowledge or valid criteria, etc.). In any case, weighting procedures will always

Indicators of regional integration 13 be arbitrary to some extent. The World Bank (2002), for example, pointed to the problem of combining indicators applying to different topics or different regional arrangements, and suggests accompanying the quantitative data with qualitative assessments. Although it is attractive to combine both types of indicators, it would be helpful to have more clarity on the speciﬁc potential contribution of both and on the borderline between quantitative measurements and qualitative assessments. It should also be borne in mind that qualitative assessments are more difﬁcult to implement (although not necessarily to interpret) in international and intercultural contexts than quantitative ones.6 Another choice that has to be made is whether the system should conﬁne itself to descriptive measurements of observable variables or whether a combination is needed of such measurements with analytical information and estimations. The second option makes a SIRI richer for analysis but one is then confronted with the problems of non-standardised methods of analysis, difﬁculties with the data collection and the complexity involved in the interpretation of the information contained in the system.

CONCEPTS OF INTEGRATION A core issue in the development of a SIRI is the underlying deﬁnition of (regional) integration that will be used and how it will be operationalised into dimensions and variables. The concept of integration refers to a process in which units move from a condition of total or partial isolation towards a complete or partial unification. Applied to the interaction between independent sovereign states, integration refers to a process of large-scale territorial differentiation characterised by the progressive lowering of internal boundaries and the possible rising of new external boundaries. Such complex social transformation may or may not imply some kind of permanent institutional structure. Although integration at the level of states can refer to many different aspects of cooperation, it is mostly used in a context of economics and international trade. Integration then becomes economic integration and can be defined as ‘the voluntary linking in the economic domain of two or more formerly independent states to the extent that authority over key areas of domestic regulation and policy is shifted to the supranational level’ (Mattli, 1999: 41). The assessment of ‘levels of integration’ has led to the use of typologies, like in the case of Balassa’s stages approach. Although these typologies are attractive and allow the classiﬁcation of countries and regions, it should be clear that in empirical research the first problem is to position the countries or regions in a continuous multidimensional ‘space’ which can then (ex post) be ‘compartmentalised’, more or less successfully. Obviously, no unique deﬁnition of integration is available in the literature, so that here too important choices need to be made. The designers of a SIRI (from the academic world or from policy institutions), will have to make these choices as a necessary precondition for the construction of the system. The definition of

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

14 Philippe De Lombaerde and Luk Van Langenhove integration will almost necessarily imply that other related concepts will have to be deﬁned also. These include, for example, cooperation and coordination.7 It should be stressed that the delimitation of the concept of integration and the scope of the information system is not an exclusively academic exercise. Ultimately, the builders and users of the system will have to make the key decisions on the basis also of political and practical considerations. Therefore the discussion below is limited only to giving some orientations for the deﬁnition of the concept and will not present a particular and exclusive deﬁnition yet. Often, regional integration is presented as a location in a spectrum of forms of interaction between states. In Dobson’s spectrum, for example (Figure 1.1), integration is presented as the most intense form of inter-state interaction, involving common policies among states (Dobson, 1991). This way of presenting the phenomenon of integration has the advantage of showing how it is positioned visà-vis other forms of interaction, but does not necessarily clarify when cooperation becomes integration.8 In this framework, integration is opposed to conﬂict, and independence occupies a central place. However, one might argue that both conﬂict and integration imply interaction, so that interaction (voluntary or non-voluntary merger, in its extreme appearance) should rather be opposed to independence; conﬂict and integration being (negative and positive) expressions of interaction (Figure 1.2).9 For Greif (1997), integration has to be seen as the opposite of segregation. COOPERATION

CONFLICT

INDEPENDENCE

COORDINATION

INTEGRATION

Figure 1.1 Dobson’s ‘Policy Conﬂict–Independence–Integration’ spectrum Source: Dobson (1991: 3).

INTERACTION

INDEPENDENCE

POS.

NEG.

Figure 1.2 The independence–interaction spectrum

MERGER

Indicators of regional integration 15 The problem of delimiting the phenomenon exactly also has to do with the fact that integration itself is usually presented as a process with changing characteristics. It has long been established that integration can refer to both ‘states’ and ‘processes’ (Balassa, 1961: 1), but all the implications of this for the construction of a SIRI should be analysed. Because of the nature of the phenomenon, a SIRI should be able to monitor the process. It should not be forgotten, however, that within the integration processes there are significant qualitative steps, breakpoints, accelerations or crises that deserve to be addressed. Around these crucial ‘moments’ within the process, evaluations of the type ‘before/after’, as suggested by the static view of integration, are possible and useful. A SIRI should also be sufficiently flexible to consider processes that are characterised by a tendency towards integration, but allowing for phases of stagnation or even temporary disintegration. Another issue consists of the duality between ‘formal’ and ‘informal’, or ‘real’, integration. In Ondarts’ typology (Ondarts, 1992: 6), ‘real’ is opposed to ‘formal’ or ‘institutionalised’, meaning ‘based on formal agreements and discourses’; it is not opposed to ‘monetary’ or ‘ﬁnancial’ like in international economics.10 Formal and real integration can be thought of as two parallel processes which are relatively autonomous, the ﬁrst does not necessarily imply the second, nor vice versa. Formal integration will be usually state-led, while the so-called real integration can be occurring without much interference from states. We are not inclined to use the term ‘real integration’. In a way, it suggests that formal, institutionalised integration is necessarily an ‘empty box’, which in reality is not or only partly the case. There is at least an effect on the political debate, vocabulary and leadership, and on the collective consciousness and imagination; but usually it goes much further than that. The concept of interdependence is related to what is meant by ‘real’ integration. According to Cooper (1985), interdependence should be understood as a special case of ‘openness’. Openness becomes interdependence when the entity (country) under consideration suffers the feedback effects of the consequences of its own policies for the rest of the world or region. It is more than ‘mutual’ dependence, a situation whereby two entities (countries) have two-way economic relations and whereby an economic cost would occur if these relations should cease to exist. Whereas mutual dependence depends normally on the transaction volumes, interdependence would rather refer to marginal dependence from (two-way) economic relations (Cooper, 1985: 1198).11 Interdependence is likely to become a more relevant concept for small countries in a regional rather than in a global context. One should be aware, however, that formal/institutionalised integration can also have a direct impact on interdependence, through the creation of supranational policy institutions. In the research programme on the so-called ‘new regionalism’ (Hettne, 1999) and the literature on growth circles (Thant et al., 1994), real integration is also emphasised. It reminds us also that integration should not necessarily be seen as a form of inter-state interaction, but that the geo-political actors involved might well

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

16 Philippe De Lombaerde and Luk Van Langenhove be sub-national (regional) actors.12 A SIRI that is able to monitor such different modalities of integration would be superior to one that is not. However, in order to avoid a SIRI becoming too abstract and irrelevant, the system will have to be restricted to regions that have been institutionalised, albeit in a minimal way, or at least explicitly recognised as such by relevant actors (public authorities, political movements, business communities, international community, etc.). In this regard, one can refer to the different levels of ‘regionness’ as identiﬁed by Hettne and Söderbaum (2000) and determine a minimum level of regionness necessary to qualify for the application of a SIRI, provided that this concept can be satisfactorily operationalised. A mechanical identiﬁcation of ‘real regions’, based on indicators of interdependence and homogeneity, is probably not a sound basis for the selection of the regions in the system, as the system would be less policy-relevant and some of these regions might lack sense, politically speaking. However, one should be aware of the possibility that in top-down approaches to micro-regions used by certain international organisations, institutional recognition might follow a more or less mechanical identiﬁcation of the region. For example, in the case of growth triangles, we have the impression that this is not uncommon. In these cases, an ex post justification in terms of historical relations (and what does history not prove?), cultural proximity, etc. is usually not too difﬁcult either.13 Indicators in a SIRI should also (be restricted to) reﬂect the speciﬁc characteristics of the integration process. Variables that describe characteristics of the countries involved should not necessarily be included although they could be part of a background information section in the same system. There is, however, a large grey zone consisting of variables that are, as such, purely national indicators but that can easily be transformed into indicators of convergence/divergence.14 Another group of variables that are in a grey zone, are those belonging to political economy approaches to integration, such as underlying motivations of integration processes,15 role of interest groups, permeability and degree of corruption of regional institutions, etc. Without understanding the underlying motivations of a regional integration effort, it is difﬁcult to evaluate. For example, it would obviously be incorrect to judge the success or failure of pre-AFTA ASEAN on the basis of purely economic criteria. Finally, a SIRI should be capable of reﬂecting expressions of both ‘positive’ and ‘negative’ integration (Tinbergen, 1954). ‘Negative’ integration refers to the elimination of barriers; ‘positive’ integration refers to the formulation of common policies. It is possible to frame this distinction in a time dimension. Low levels of integrative ambition (Best, 1997) usually imply negative integration, whereas high levels of integrative ambition usually imply positive integration. Best (1997: 56) rightly points out, however, that it is difﬁcult to conceive negative integration without a minimum amount of positive measures. Integration should be seen as a varying mixture of both types of measures.

Indicators of regional integration 17 THE ROLE OF THEORY AND THE NORMATIVE AND EURO-CENTRED TRADITION IN INTEGRATION STUDIES Accepting a workable definition of regional integration allows the scope of a SIRI to be delineated but it needs to be complemented by choices regarding the theoretical framework used. Decisions on whether or not to include specific variables into a SIRI will also be dependent on the theoretical afﬁnities of the SIRI producers. Although it would be tempting to regard the availability of data as the first criteria, this should be avoided and theory should obviously be the major guiding principle for including variables. A theoretical framework is also needed to decide on how to structure the variables and analyse them. An example of the link between theory and the selection of indicators of regional integration is the ECB proposal to measure economic integration (Dorrucci et al., 2002). In this case, the authors base the choice of indicators of economic integration, and not only monetary integration, on optimum currency area theory (see p. 22). Integration is a multidimensional social phenomenon and its study has generally been undertaken along disciplinary lines. That partly explains why various (sometimes complementary, sometimes competing) theories exist. Our recommendation to give the SIRI a theoretical basis, does not mean that the system should be limited to the variables that constitute a particular theoretical model of integration but rather that the incorporation of each of the variables should be based on a structured and scientiﬁc argument. The existence of a variety of theories should not be an obstacle for the construction of a SIRI; eclecticism is inherent to scientiﬁc activity. In addition, we agree with Anderson that it might be a sufficient condition for selecting an indicator to be perceived as important by the community of users of the system, without necessarily having a particular theory that corroborates this. In this context it should also be noted that theories about regional integration are not developed in abstracto but are linked to what is happening in the world. As the research on the Single Market, transition issues, the EMU and EU enlargement very well illustrate, the development of integration studies has gone hand in hand with political developments, especially in post-Second World War Europe, and is very much a response to the signals emitted by the political centres and actors. That explains the normative tradition in integration studies (Bekemans et al., 2000: 55–7). The problem with normative theory is that the evaluation of the facts (actions, decisions, effects) is reduced to a mechanical application of the model labelling them as positive or negative, progress or decline, functional or dysfunctional, etc. The stages approach to economic integration, attributed to Balassa (1961), for example, has been extremely inﬂuential in academia and lends itself very well to a measurement of ‘progress’ of a particular integration scheme. Those who are familiar with regional integration issues know, however, that one has to be very prudent in applying this kind of theoretical scheme mechanically; all too often they are confused with general laws governing integration processes. In the real world, simultaneity, inversion and endogeneity is not uncommon. The ECB (Dorrucci et al., 2002) has shown, however, that Balassa’s scheme is perhaps ﬂexible enough

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

18 Philippe De Lombaerde and Luk Van Langenhove to be used in the presence of these phenomena. In other words, using an open and flexible theoretical framework is one of the preconditions of a SIRI with wide applicability. It should be further stressed that qualifying variables as theoretically relevant does not imply that the direction of any causal linkages with other variables can be easily established. In many cases there will be uncertainty about the causality and its direction, in some cases there might be a ‘double’ causality and/or feedback effects. The variables under consideration are part of a complex social system (with simultaneities and chaotic processes) in which causality often becomes difficult to trace or even irrelevant as it will be the social meaning that will be essential. Best (1997) has rightly pointed to some of these problematic relationships between certain variables in terms of the causal links between them and their direction. For example, although Olson’s Logic of Collective Action would suggest a positive relation between the number of countries involved and the difficulties encountered in the decision-making process (Olson, 1965), the real effects of the number of countries also depends on the characteristics of the individual countries and the design of the decision-making process. Majority voting might well have the opposite effect. Best also mentions some other theoretically problematic relationships, including the following: the relation between the scope of coverage of an integration agreement and its stability, the relation between the time perspective of an arrangement and its stability, and the relation between the degree of real interdependence and the degree of formal integration. Another (theoretically) problematic relationship is the one between, on the one hand, the degree of structural asymmetries between the members of an integration arrangement, and the stability and impact on economic development of the agreement, on the other.16 Again, using Olson’s logic, integration processes built on asymmetric groups where a leading (hegemonic) country perceives enough beneﬁts to justify the provision of the collective good (the integration agreement) would be expected to be the more dynamic and effective ones. 17 This idea can be seen as a regional application and implication of hegemonic stability theory (Keohane, 1980; Kindleberger, 1981; Gilpin, 1987). In a context of developing countries, the World Bank also defends asymmetric integration agreements (Collier et al., 2000). However, Stakhovitch (1991), for example, questioned this thesis on empirical grounds.

METHODOLOGICAL AND ORGANISATIONAL OPTIONS RELATED TO THE CONSTRUCTION OF A SIRI Next to preciseness in the concepts used and explicitness in the underlying theoretical adherences, one needs to make a number of methodological and organisational decisions in order to further design a SIRI. This will include choosing between specific or general SIRIs and between the use of area versus country indicators.

Indicators of regional integration 19 Systems can be more or less speciﬁc, meaning that they can cover one, a few or many aspects of integration, or that they intend to monitor specific sectors (transport, commerce, agriculture, etc.). Proposals have been made, for example, to design indicator systems for the evaluation of the public-management capacities for regional integration among groups of countries (Best, 1997), or for the measurement of the economic integration effort in Africa (UNECA 2001, 2002). On the infra-national level, the World Bank designed a system of fiscal decentralisation indicators (World Bank, sd). The builders of a SIRI will also have to decide upon the relevant level(s) of analysis. A system can be built to monitor the dynamics of a group of (integrating) countries or regions, or it can be built to monitor the participation of individual countries/regions in the integration schemes (individualised effects, policy implementation, etc.). An interesting attempt to build a multilevel SIRI is the indicator project of UNECA (2001) where indices are being estimated on four levels: 1 2 3 4

the country level, the regional level, the sectoral level, and the continental level.

A particular and related problem is that many countries, all over the world, are simultaneously a member of more than one integration arrangement. The African case illustrates this very well (Table 1.2). Currently 45 countries out of 53 are simultaneously members of (at least) three regional integration arrangements. One country (DR Congo) is even a member of four arrangements but, paradoxically, it is not known as a particularly active actor in African regional integration. In these cases of overlapping memberships, it might be difﬁcult to separate the effects of different integration agreements. In addition, it is often the case that the different agreements have similar orientations and that RIAs that involve WTO members, are generally WTO compatible. Table 1.2 Overlapping memberships of integration arrangements for African countries Category

Number of memberships of integration arrangements per country

Number of countries per category

1 2 3 4 5 Total

0 1 2 3 4

0 7 27 18 1a 53

Source: UNECA (2002). Notes a DR Congo.

20 Philippe De Lombaerde and Luk Van Langenhove Considering indicators at the country or the group level has political implications. As observed by the World Bank (2002), the evaluation of a regional arrangement, especially when it involves ‘rewards’ or ‘sanctions’ from the international community (as in the case of the Regional Cooperation Review Process (RCRP)), should be able to handle asymmetries within the groupings, such as passive or obstructive behaviour by one or a minority of members. This, in turn, can be caused by different factors, such as the occurrence of a conﬂict in one or more member countries or diverging policy preferences between the coalitions in power in different member countries. In the case of the Cotonou Agreement and the RCRP, the identification of the region as a relevant policy level is imposed from outside and is embodied in the regional strategy papers. Finally it should be noted that the evaluation of integration policies and their implementation is an essential part of any SIRI. This requires the incorporation of indicators of policy implementation and effects of regional integration. Again, this brings with it several problems. •

•

•

The useful distinction between positive and negative integration has been mentioned before, but one should be aware that positive integration might suggest more ‘policy effort’ and be captured as such by many indicators, although nothing assures ex ante that these measures have more important effects than negative integration. It is a well-known characteristic of empirical research that the observers tend to show a higher sensitivity for observations that tend to confirm the hypothesis or the existence of what is sought. As far as indicators of integration policies are concerned, the registration of policies and measures might offer a biased view in the sense that simultaneous policies (possibly in other policy areas) in the other direction (revealing disintegration, protectionism, nationalism, etc.) might or might not yield a net progress. This is even more so because integration policies tend to be common policies, whereas disintegration policies (protectionist reactions) tend to be national. Theoretically, ideal indicators would be net indicators, showing whether a given set of policies and measures taken during a period of time do or do not contribute to integration. However, these kinds of indicators are difﬁcult to construct. In its discussion of the UNECA methodology, the COMESA Secretariat expressed strong reservations over the methodology used, precisely for the reason that the UNECA indicators do not necessarily reflect the effects of programmes being undertaken by regional organisations (COMESA, 2002: 6). As an example, COMESA criticised the ranking of SADC and ECOWAS as the most successful regional organisations. According to COMESA, these rankings simply reflect the presence of a member with a large economy in each case (South Africa and Nigeria, respectively). One should therefore carefully distinguish between structural characteristics of countries and regional groupings, on the one hand, and integration policies, on the other hand.

Indicators of regional integration 21 DIMENSIONS AND CATEGORIES OF VARIABLES Once the coverage and limits of a SIRI are established, its variables should be organised systematically. As the multidimensional character of the phenomenon is obvious, the variables could be organised according to disciplinary fields (political, social, cultural, economic, etc.), and/or policy areas (trade, investment, migration, competition, agriculture, industry, infrastructure, legal cooperation, etc.). The latter could be called the traditional sectoral approach to integration. A third way of classifying the variables consists of a classiﬁcation on a functional basis, as in the input–output approach. Integration is then implicitly seen as a process where some variables act as inputs, some as outputs, while others characterise the process. The advantage of this approach is the emphasis on the output (effects) of integration; from a welfare and development point of view, that is what it is all about. The assessment of the developmental impact of regional integration processes and policies could be carried out by incorporating regional development and (social) spending indicators in the SIRI.18 However, one should be cautious in classifying variables along these lines. Double causalities and systemic relationships are more often the rule than the exception. Whereas, the World Bank (2002) stresses the importance of incorporating in a SIRI indicators that assess the long-term economic impact of regional integration or cooperation, such as indicators about cross-border movements of people, capital, information and goods and services, the European Commission (2002b) favours indicators of inputs and efforts. This brings us to the fact that there exist a number of proposals on how to classify the variables. These include the proposals respectively made by DG Development of the European Commission, UNECA and ECB. DG Development proposed the classiﬁcation of the indicators of the foreseen indicator system for monitoring economic integration in the ACP countries in the following broad categories: • • • •

regional economic cooperation, functional regional cooperation, governance, ﬁnancial issues and functioning of institutions, implementation of EDF projects and programmes.

A more detailed description of this proposal is shown in Appendix 1.1. UNECA (2001, 2002) considers eight ‘clusters of activity’ to classify the variables and indicators. These are: 1 2 3 4 5 6

trade and market integration, monetary, ﬁscal and ﬁnancial integration, transport, communications, industry, energy,

22 Philippe De Lombaerde and Luk Van Langenhove 7 8

food and agriculture, and human development and labour markets.

ECB distinguishes between institutional and economic integration (Dorrucci et al., 2002). The former is evaluated on the basis of the implementation of decisions in four dimensions, based on Balassa’s stages approach to integration: 1 2 3 4

free trade area/customs union, common market, economic union, total economic integration.19 Within the latter category, seven subcategories (and 11 variables) are considered:

i ii iii iv v vi vii

synchronisation of the business cycle, convergence of inﬂation rates, exchange rate variability, trade openness and integration, ﬁnancial market integration, convergence of interest rates, income convergence.

As a response to DG Development’s proposal, the COMESA Secretariat launched a proposal for a system of indicators with an alternative design (Appendix 1.2). The philosophy of that proposal is different in the sense that inter-regional comparisons are not the main focus, but rather the monitoring of their own integration process. COMESA considers 12 categories of variables: 1 2 3 4 5 6 7 8 9 10 11 12

trade liberalisation, trade facilitation, trade in services, transit facilitation, monetary convergence, domestic payments and settlement systems, ﬁscal environment, government intervention in the economy, capital ﬂows and foreign investment, governance issues, regulatory environment, licensing requirements.

Related to the question on how to organise the variables in a SIRI, is the question on whether and how to integrate the preconditions for regional integration, or ‘forward-looking’ variables, in the system. In the literature, several attempts can be found to measure the potentiality (feasibility and effects) of future integration

Indicators of regional integration 23 agreements. In the economists’ jargon they are referred to as ex ante studies. Of special interest for us are those attempts that permit comparison and those that are also relevant for monitoring purposes. The conception of integration as a phased process, as suggested above, allows of course to consider many ‘befores and afters’ within the same process, so that ex ante studies and methods are not necessarily without relevance for our purpose (Figure 1.3). The ex ante/ex post approach can be linked easily to the input–output typology of variables, mentioned above. The following can be considered as ‘inputs’: structural characteristics of the integrating area (number of countries, shared borders, etc.), asymmetries, capacities to integrate, commitments, governance structure, overlapping memberships, etc. The following could be considered as ‘outputs’: policy implementation, effects on ﬂows, effects on growth, etc. A special category of inputs could be called preconditions for integration. Although originally intended to assess (ex ante) the possibilities and potential of (future) integration agreements, the variables involved can also be used in a dynamic manner to evaluate the compatibility of the formal integration process with the preconditions. In addition, these preconditions are not static, they are often endogenous because of feedback effects of the integration process. Well known are the ex ante trade analyses that estimate the foreseeable effect of new regional trade agreements on trade ﬂows and welfare. These include the estimates of trade creation and trade diversion, 20 the identification of ‘natural markets’ (Krugman, 1991; Kreinin and Plummer, 1994), the tests of regionalisation of international trade and the capacity to trade of individual countries (Freudenberg et al., 1998a, 1998b), among others. However, ex ante studies have not been limited to trade issues. Best (1997), for example, analysed the public-management capacities for regional integration. Best’s starting-point is the difﬁculty (for his purpose) that integration is not a homogeneous phenomenon, that its contents and orientations change over time, and that its (theoretical) finality is not pre-established since it is impossible to identify the ‘optimal’ policy levels. In order to be practical, Best considers ‘levels of integrative ambition’; he implicitly refers hereby to formally expressed models of integration that are aimed at by a group of countries (Best, 1997: 54–5). He proposes to use Pelkmans’ detailed taxonomy to classify the levels of integrative ambition (Pelkmans, 1993), although he is aware that actual integration processes M1

M2

M3

PRECONDITIONS

TIME PRECONDITIONS PRECONDITIONS

Mi: critical moments in integration process

Figure 1.3 Preconditions in phased integration processes

24 Philippe De Lombaerde and Luk Van Langenhove are not always exact copies of the (text book) models and that the suggested sequence of these models is not always respected. In Pelkmans’ taxonomy, we ﬁnd the traditional Balassa-type (text book) stages of economic integration, but each level is broken down in more speciﬁc arrangements. Preferentialism is differentiated into partial sectorial preferentialism, sectorial integration, non-reciprocal preferential systems, and reciprocal partial preferentialism; free trade areas are differentiated according to whether they cover all industry or all sectors; customs unions are broken down into various stages of product market integration; and common markets and economic and monetary unions are also broken down in stages. Best stresses that a combination is needed of the identification of governing needs, on the one hand, and the identiﬁcation of governing capacities, on the other (Best, 1997: 56–9). He proposes as a ﬁrst step ‘to evaluate the degree of difﬁculty posed for a particular group of countries, and for each participating administration, to deal effectively with a particular set of integration objectives at a given time’ (Best, 1997: 60). The author identiﬁed nine key variables that shape the complexity of the implementation of the integration objectives (Appendix 1.3).21 The variables are: 1 2 3 4 5 6 7 8 9

number of member states, relative sizes of the participating countries, different levels of development, scope of coverage, type of impact, time perspectives, degree of real interdependence, political framework, perceptions, values and norms.

Best does not present, however, a method or a scheme that links the complexity variables to a speciﬁc institutional arrangement or speciﬁc institutional changes. Rather, he considers the variables as input in a complex negotiation process. The outcome of this process is institutional readjustment. He stresses that almost every integration agreement has implications for public management, but that there is no linear relationship between integration and the centralisation of public administration. The discussion on the subsidiarity principle illustrates this. Horizontal cooperation and partnerships are often important components of the integration architecture.

DIMENSIONS AND CATEGORIES OF VARIABLES: A PROPOSAL Taking stock of the many issues and problems presented in the previous sections, a proposal for a conceptual framework that could serve as a basis for the

Indicators of regional integration 25 construction of a SIRI will now be presented. Or at least, it can be thought of as a check-list for selecting the variables that should be incorporated in the system. The framework explicitly recognises the multidimensional and dynamic character of regional integration and distinguishes between six categories of variables, which can easily be converted into six modules of a SIRI. The logical relationships between the categories of variables are shown in Figure 1.4 and explained brieﬂy below. We consider the distinction between real and formal integration as not really appropriate; we prefer to consider parallel (but interconnected) processes of regional integration: institutional (more or less capturing what is usually called ‘formal’), political, economic, cultural, etc. The effects of integration policies and the evolution of regional interdependence will obviously have feedback effects for the institutionalisation process, thus conceptually restoring its endogenous character. We think that this is a workable way to classify the relevant variables into broad categories. Particular SIRIs will have to opt for some combination of variables of these categories. In Table 1.3, a proposition is made regarding which variables and/or sub-categories correspond to the different categories.

CAT. III: INSTITUTIONALISATION

CAT. I: ACTORS

CAT. II: STRUCTURAL FACTORS

CAT. IV: IMPLEMENTATION

FEEDBACK

CAT. V: EFFECTS

CAT. VI: INTERDEPENDENCE

ECONOMIC

POLITICAL

SECURITY

CULTURAL ETC.

TIME

Figure 1.4 Regional integration as a process: proposal for a conceptual framework

26 Philippe De Lombaerde and Luk Van Langenhove Table 1.3 Proposal for classifying variables in a SIRI Category

Sub-categories

I

Actors

Number of integration units involved (countries, regions, organisations, etc.) Number and quality of actors in the decision-making process Level of activity of the actors Actors’ opinions and perceptions (survey results) Overlapping memberships

II

Structural factors

Proximity of the actors (geographical, cultural, etc.) Structural complementarities Structural asymmetries Historical patterns of cooperation, integration and conﬂict

III Institutionalisation

Number of treaties and agreements Contents of treaties and agreements Time frames of treaties and agreements Institution building Arrangements on common policies and policy coordination Gradualism, exemptions and differential treatments

IV Implementation

Status of implementation of general treaties Status of implementation of speciﬁc agreements Degree of accomplishment of convergence criteria

V

Human development Economic growth Trade Migration Capital ﬂows

Effects

VI Interdependence

Mobility of persons Political interdependence (existence of common policy variables, de facto coordination of policies, occurrence of conﬂicts, tensions, etc.) Economic interdependence (trade, capital ﬂows, correlation of activity levels, symmetry of shocks, etc.) Information and knowledge ﬂows

Actors and structural factors The categories of ‘Actors’ (cat. I) and their structural characteristics (‘Structural factors’) (cat. II), contain information about the basic building blocks of the integration effort. The category of ‘Actors’ refers to the number and type of actors involved and their behaviour.22 The number of countries or regions involved has a direct influence on the dynamics of the decision-making process, although that relationship is not straightforward and should be seen in combination with the issue of asymmetries (see above). From an administrative and political point of view, the number and character of the policy-making and implementing levels is also important.

Indicators of regional integration 27 The new forms of integration, sometimes called ‘new regionalism’, 23 have stressed the importance of considering the participation of different types of actors involved in the (real) integration processes. When the non-governmental actors are formally organised and linked to the formal integration process, then it is rather easy to identify and evaluate their participation, and incorporate this information in an indicators system. If it is not the case, then it will be more difﬁcult to do so. Examples of formalised participation of non-governmental actors range from the Economic and Social Committee in the EU to the Private Sector Forum in the ZMM-GT (Slocum et al., 2003). It might be relevant to distinguish between primary (active?) and secondary (passive?) actors. The primary actors being persons, groups of persons and organisations that have an influence on the dynamics of the integration process; the secondary actors being all those that are affected by the process. Such secondary actors do not necessarily belong to the integrating region. The supranational actors might be seen as a third category, although they can be both a primary stakeholder or a secondary actor. In addition to their numbers, within each category of actors a list of attributes could be established to reﬂect their character and importance. The intensity of their involvement and their importance in the decision-making process could be evaluated through quantitative methods (number of meetings attended, ﬁnancial contribution, etc.) or qualitative assessments (expert opinion). The category of ‘Structural factors’ (cat. II), includes all those variables that refer to structural characteristics of the integration grouping and of its members. They should logically be restricted to variables that are directly or indirectly related to the integration process. These variables might relate to the scale of the arrangement, the structure of the grouping and of each component, the nature of the components, etc. Proximity of the actors is obviously a relevant variable to evaluate the potential and sustainability of an integration grouping. Gravity type models of economic interaction have shown signiﬁcant (negative) relationships between the intensity of economic relations between countries and their distance. 24 It has also been shown that proximity/distance is a typical multidimensional variable; physical, economic, political, cultural, linguistic and historical proximity are all relevant variables. As we mentioned before, structural asymmetries play an important role in integration processes although the direction of causalities is not clear. Its measurement can be based on variables of population, the economy, external relations, and so on.

Indicators of institutionalisation and policy implementation The actors involved in integration processes take steps (measures) that are supposed to contribute towards regional integration and the ‘Institutionalisation’ of the region and its integration effort (cat. III). These political decisions are implemented to some extent (‘Implementation’, cat. IV) and either do or do not

28 Philippe De Lombaerde and Luk Van Langenhove have certain impacts (‘Effects’, cat. V) in different areas (social, economic, cultural, etc.). Also relevant is the institutional basis on which the whole integration process rests (for example, constitution-based versus treaty-based integration processes); institutionalisation should thus be analysed on different levels. Obviously, institutional activity has quantitative (for example, number of treaties or ministerial meetings) and qualitative aspects (content of the treaties or decisions). Productivity measures might be applied to the institutional activity, thus linking policy outputs to their resource cost. The World Bank (2002) pointed to the problem that there might be discrepancies between (formal) objectives and the institutional set-up to achieve them. As a consequence, indicators that reﬂect institution building do not necessarily show progress towards reaching the integration goals. Although relevant, inconsistencies between expressions of formal integration might be difﬁcult to assess. In our opinion, it is further useful to distinguish between policy implementation (at the regional or national level) and the assessment of the impact of implemented policies.

Effects Seen from the perspective of citizens and policy-makers, this should be the most important category of variables. But from the point of view of a designer of a SIRI, it is also the most problematic category to include in the system. This is related to the fact that it is difﬁcult to isolate effects of integration from those of other phenomena. Integration is a complex and dynamic process not necessarily adequate for causal explanation. On top of that, for many aspects of integration, there are no comparable data sets or standardised research methodologies available.25 This is certainly true for the analysis of static effects of integration (directly linked to the reallocation of resources among sectors and countries), but even more so for the analysis of the dynamic effects of integration. Although researchers often concentrate on the short-term (static) effects of integration measures rather than on the dynamic ones, it should be stressed that the sign (direction) of the effects does not necessarily coincide, so that the former cannot necessarily be used as a proxy for the latter. In practice, it is often explicitly understood and accepted that short-term costs (transition costs) are the price to pay for reaping long-term beneﬁts. Other aspects of monitoring the impact of regional integration policies are outlined below: First, as Devlin and Ffrench-Davis (1998) pointed out, the methodological problems related to counterfactual analysis are particularly important: • • • •

the more complex the studied process is, the longer the time period considered, the bigger the changes involved, and the more the analysis turns on exact magnitudes. Most of these items apply in the case of integration studies.

Indicators of regional integration 29 Second, the right balance has to be struck between quantitative indicators and qualitative assessments.26 Third, a distinction should be made between the strength and the type of impacts (Best 1997). Fourth, the existence of overlapping integration agreements (see above), often with convergent objectives, makes the identiﬁcation of effects even more difﬁcult. Devlin and Ffrench-Davis (1998) add that this phenomenon might also have a cost of its own in terms of reduced transparency and higher transaction costs. Fifth, the comparison of effects between countries does not allow for analysing the distribution of effects within member countries. Integration might well favour the agglomeration of activities within the countries, might preferentially favour border areas, might have an impact on the income distribution, etc. Finally, in developing countries, integration policies are often an integral part of more general structural adjustment and liberalisation policies, so that their speciﬁc effects cannot easily be established.

Indicators of interdependence ‘Effects’ of integration, together with structural conditions and exogenous inﬂuences, can explain the degree and evolution of ‘Interdependence’ between the regional actors (cat. VI). Effects are thus attributable to specific integration policies, whereas the degree of interdependence is autonomously measured and reﬂects the evolution of interdependence on different dimensions. Interdependence is used here as a substitute for what is often called ‘real’ or ‘de facto’ integration. Interdependence tries to capture the degree of ‘regionness’ of the region, or at least some aspects of it. Regionness is also a central concept in the new regionalism approach (Hettne, 1999; Hettne and Söderbaum, 2000). Interdependence can be assessed on different dimensions, such as economical, political, cultural, security and infrastructural. These dimensions coincide broadly with those considered in the proposal for a system of indicators of interconnectedness, made by Held and others in the framework of the Global Transformations project (Held et al. 1999).27 The following dimensions are being considered in that project: • • • • • • •

political-legal indicators, military indicators, economic indicators, migration indicators, culture indicators, environment indicators, global stratiﬁcation.

Many of the indicators proposed could be transformed into indicators of regional interconnectedness.

30 Philippe De Lombaerde and Luk Van Langenhove The measurement of the degree of interdependence can be approximated via the measurement of the ﬂows (of people, goods, capital, information, etc.) that are interconnecting the actors or via direct measurements of correlations of variables (for example, symmetries in business cycles, interest rate spreads, etc.). For the forms of interdependence that are more difﬁcult to measure, like political interdependence, indirect measurements should be considered. The patterns of voting behaviour in multilateral organisations might, for example, be a possible indicator of regional policy convergence/divergence. Interdependence through trade flows is probably the most studied kind of interdependence. Its study is usually based on simple indicators such as the relative importance of intra-regional trade and its growth, which can easily be calculated. More sophisticated indicators are available, which correct the former for size effects in order to allow for methodologically sound inter-regional comparisons (Iapadre, 2000, 2001). The indicators of the degree of integration can be complemented with indicators of the direction and nature of commercial integration. This is particularly relevant from an analytical point of view. The composition of the flows induced by the integration process are good indicators of the underlying socio-economic changes that take place in the member countries. In the case of trade ﬂows, for example, indicators of intra-industry trade and of the technological content of intra-regional trade28 can easily be calculated. A proposal for a combined set of indicators on (trade) regionalisation has been made by Bensidoun and Chevallier (1998) (Table 1.4). The fact that studies on regional trade flows are relatively abundant can be explained by trade data being relatively readily available, and of sufﬁcient detail and good quality. This is much less frequently the case for data on capital ﬂows, migration and ﬂows of information. In the case of intra-regional migratory ﬂows and stocks of immigrants, for example, although it is essential information, in many countries the quality of the data is insufﬁcient or the data are lacking.29 The ex post analysis of the ﬂows of (public) funds between national governments and the supranational institutions within a group of countries (a region) also permits an evaluation of the degree of their integration, provided that these ﬂows reﬂect the actual level of organised solidarity, the importance of the supranational institutions, etc. The category of variables related to the budget in DG Development’s framework, for example, could easily be broadened to include indicators on the Table 1.4 Bensidoun and Chevallier’s indicators of regionalisation Indicators 1 2 3 4 5

Relative trade intensity index North’s share in South trade (or South’s share in North trade) Degree of currency anchor (relative volatility of the exchange rate) Intra-industry index (Aquino coefﬁcient) Similarity of South–North trade with South–Rest of the World trade

Source: Bensidoun and Chevallier (1998).

Indicators of regional integration 31 composition of revenues and expenditures and on vertical (im)balances between the national and supra-national levels.30 The conceptual framework and indicator systems developed to monitor decentralisation processes from a ﬁscal perspective might be very helpful to structure a ﬁscal module of a SIRI. Appendix 1.4 shows the list of indicators incorporated in the system of ﬁscal decentralisation indicators of the World Bank (sd). Three categories are considered: 1 2 3

main indicators, composition of sub-national revenues and grants, composition of sub-national expenditures. Vertical imbalances could also be calculated on a sectoral basis.

CONCLUSIONS Adequate monitoring tools for regional integration processes would allow better (regional) policy design and implementation, better scrutiny and participation by all stakeholders and affected groups and individuals, as well as more in-depth academic analysis of these complex social transformation processes. The aim of this chapter was to explore the conceptual and methodological aspects related to building a system of indicators of regional integration (SIRI). We have tried to demonstrate that there are multiple and interconnected issues to be considered, and that choices need to be made, related to: • • • • • •

the kind of users and producers involved in its design, the underlying concepts and theories, the scope of the system and level(s) of analysis, the adequate selection and organisation of the variables, the balance between quantitative indicators and qualitative assessments, the ‘correct’ observation and construction of indicators.

Future attempts to implement monitoring tools for regional integration processes, launched at the regional, inter-regional or supranational levels, will require due consideration of these issues raised and answers to the questions posed. The attempts will only have a chance to succeed if conceptual, technical, political, institutional and organisational aspects are simultaneously addressed.

32 Philippe De Lombaerde and Luk Van Langenhove APPENDIX 1.1 DG Development’s proposal for a system of indicators to measure regional integration and cooperation performance Categories

Subcategories

Variables

Economic integration

Trade liberalisation policy

WTO compatibility of rules of customs valuation Quality of classiﬁcation of goods Application of rules of origin Exemptions Phasing out of temporary measures Liberalisation of trade in services Importance of intra-regional trade

Other integration policies

Facilitation of investment Movement of persons Right of establishment Competition policy Creation and implementation of cohesion policy Improvement of comparable statistics Macroeconomic surveillance Trade facilitation measures

Transport

Progress towards a common transport policy Expenditure for maintenance of regional network Application of harmonised transit regulations

Maritime resources

Human and physical input for a common surveillance Human and physical input for a common evaluation of natural resources Enforcing of common quality and sanitary standards

Functional regional cooperation

... Governance, ﬁnancial issues and functioning of institutions

Institutions

Number of meetings Qualitative assessment of meetings Performance of speciﬁc institutions

Budgets

Fulﬁlment of requirements of budgetary contributions Transparency of procedures Implementation of budgets

Indicators of regional integration 33 HR Implementation of EDF projects and programmes

Progress on appraisal Decisions Disbursements

Recruitment policy Staff training

Contracts concluded Contracts implemented

Source: European Commission (2002a).

APPENDIX 1.2 COMESA’s proposal for the development of a set of regional integration indicators Categories

Variables

Trade liberalisation

Number of non-zero tariffs Highest MFN tariff Highest regional tariff Weighted average MFN tariff

Trade facilitation

Level of conformity to the WTO TBT Agreement Capacity of member states to implement mutually recognised certiﬁcation marking schemes Notiﬁcation of National Enquiry Points Ability to regulate and monitor sanitary and phytosanitary standards Use of ASYCUDA (or similar) Use of GATT valuation system Use of COMESA customs document Use of HS1996 (or later) customs classiﬁcation system

Trade in services

Establishment and publication of Contact and Enquiry Point Performance with regard to commitments Reductions in exemptions over time

Transit facilitation

Implementation of COMESA harmonised road transit charges Use of the COMESA carriers license Use of the COMESA customs bond guarantee Implementation of harmonised axle load and vehicle dimensions regulations Implementation of the COMESA third party vehicle licensing system

Monetary convergence

Inﬂation Size of the budget deﬁcit Size of the external debt Exchange rate movements continued . . .

34 Philippe De Lombaerde and Luk Van Langenhove Categories

Variables

Domestic payments and settlement systems, banking and exchange rates

Restrictions on the current account Restrictions on the capital account Level of government ownership of banks Restrictions on foreign banks to open branches and subsidiaries Government inﬂuence over the allocation of credit Restrictions on private sector companies to offer all types of ﬁnancial services, securities and insurance policies Use of domestic electronic clearing systems Restrictions on foreign ﬁnancial institutions

Fiscal environment

Weighted average income tax rate Income tax as percentage of GDP Weighted average of direct taxes on business Direct business taxes as percentage of GDP VAT rate VAT as percentage of GDP Government expenditure as percentage of GDP

Government intervention in the economy

Government consumption as percentage of the economy Government ownership of businesses and industries Share of government revenues from state-owned enterprises and from government-owned property Economic output produced by government

Capital ﬂows and foreign investment

Existence of foreign investment code providing national treatment Degree to which there are any restrictions on foreign ownership of businesses Level of restrictions on foreign ownership of land Level of restrictions on repatriation of earnings

Governance issues

Freedom of the judicial system from government inﬂuence Level to which contracts are respected and whether there is an independent arbitration of contract disputes Level of transparency and accountability of the judiciary Legally granted and protected private property rights

Regulatory environment Existence of an independent competition authority tasked with implementing a set of legally recognised rules and regulations on competition Existence of an independent telecommunications authority tasked with implementing a set of legally recognised rules and regulations on telecommunications Existence of an independent standards authority tasked with implementing a set of legally recognised rules and regulations on standards Existence of deﬁned regulations dealing with public procurement in member states Licensing requirements

Source: COMESA (2002).

Level of licensing requirements to operate a business Time taken to obtain appropriate licenses to start business operations Transparency of the licensing system

Indicators of regional integration 35 APPENDIX 1.3 Best’s key variables of complexity in regional integration arrangements Key variables

Possible breakdown

Number of member states Relative sizes of the participating countries Different levels of development Scope of coverage

Type of impact

Time perspectives Degree of real interdependence

Number of sectors covered Number of sensitive sectors excluded Length of exclusion lists Strength of impact: impact on macro-economic variables, impact on ﬂows, population affected, etc. Type of impact: distribution of costs and beneﬁts, degree of uncertainty of impact, time-lags, etc. Degree of complementarity Intra-group trade Degree of transnational interaction Degree of transgovernmental interaction Importance of political ties Importance of military alliances

Political framework

Existence of common threat Awareness of individual international impotence Existence of historical perceptions of the relationships between the participating countries

Perceptions, values and norms

(Dis)similarity of perceptions of the nature, objectives and importance of the integration process Differences in constitutional systems Differences in political systems Attitude towards institutionalisation Perception of a common (regional) identity (Dis)similarities in national value systems (Dis)similarities in social structures

Source: Best (1997: 60–66).

36 Philippe De Lombaerde and Luk Van Langenhove APPENDIX 1.4 The World Bank’s list of ﬁscal decentralisation indicators Categories

Indicators

Main indicators

Sub-national expenditures (% of total expenditures) Sub-national revenues (% of total revenues) Vertical imbalance (%) Sub-national expenditures (% of GDP) Sub-national revenues (% of GDP)

Composition of sub-national revenues and grants

Tax revenue (% of total sub-national revenues and grants) Transfers from other levels of government (% of total sub-national revenues and grants)

Composition of sub-national expenditures

General public services (% of total sub-national expenditures) Defence (% of total sub-national expenditures) Public order and safety (% of total sub-national expenditures) Education (% of total sub-national expenditures) Health (% of total sub-national expenditures) Social security and welfare (% of total sub-national expenditures) Housing and community amenities (% of total sub-national expenditures) Recreational, cultural and religious affairs and services (% of total sub-national expenditures) Fuel and energy (% of total sub-national expenditures) Agriculture, forestry, ﬁshing and hunting (% of total sub-national expenditures) Mining and mineral resources, manufacturing and construction (% of total sub-national expenditures) Transportation and communication (% of total sub-national expenditures) Other economic affairs and services (% of total sub-national expenditures) Other expenditures (% of total sub-national expenditures)

Source: World Bank (sd).

Indicators of regional integration 37 NOTES 1 The authors wish to thank Brigid Gavin and Dirk Hansohm for comments on a previous version of the text. Only the authors are responsible for its contents. 2 See for example Hettne (1999). 3 See http: //www1.worldbank.org/publicsector/decentralization/webﬁscal.pdf 4 Based upon a speciﬁc methodological framework (UNECA, 2001), data have been gathered back to 1994, the year the Abuja Treaty came into force. First results have been published in UNECA (2002). 5 This is the case for articles 28–30 of Section 3 on Regional Cooperation and Integration, as well as for article 22 on Macroeconomic and Structural Reforms and Policies; it is also in accordance with Article 11 of Annex IV of the same Agreement. 6 Although aspects of interculturality also appear at the stage of choosing and weighting the variables in the system, reﬂecting cross-cultural differences in preferences. 7 Recently, for example, the problem of conceptualisation has been illustrated very well in the context of the discussions about the RCRP proposal of the European Commission. The World Bank (2002), proposed to distinguish between ‘integration’ and ‘cooperation’ on the basis of the degree of sovereignty that countries agree to transfer to supranational institutions, but recognised that the borderline is not clear-cut. The Commission itself proposed a category of inter-state interaction called ‘functional regional cooperation’. 8 The most intense form of cooperation, namely policy coordination, refers to the coordination of common objectives, of the choice, amplitude and moment of the policy measures, or the exchange of policy information, expectations and economic structures (Van Velden, 1988: 1). 9 ‘Positive’ and ‘negative’ is used here in its usual sense, not in the sense suggested by Tinbergen. 10 Ondarts (1992) disaggregates the category of formal integration further into integration ‘by treaty’ and formal integration in sensu stricto. The latter category implies the creation of supranational entities that start to coordinate the integration process with certain autonomy. 11 The author makes a further distinction between four forms of interdependence: (1) structural interdependence, (2) interdependence between (economic) policy objectives, (3) interdependence between exogenous variables, and (4) political interdependence (Cooper, 1985: 1199–1200). As far as the measurement of the degree of interdependence is concerned, he adopts the distinction between vulnerability and sensitivity from Keohane and Nye (1977). 12 See, for example, Hettne (1999). 13 The ZMM-GT might be an example of this (Slocum et al., 2003). 14 As in the case of the COMESA proposal (Table 1.4; COMESA, 2002), this can simply be only a matter of presentation (i.e. not involving calculations of convergence indicators). 15 See, for example, Page (1998) for an exploration of the ‘real’ motivations behind 14 country groupings. 16 Structural asymmetries can be distinguished from: (1) asymmetries arising from differences in commitments by the member countries, (2) asymmetries in the decisionmaking process and in the organisational design, (3) asymmetric effects of integration, and (4) asymmetries embedded in common policies (De Lombaerde, 2002). 17 For a more recent presentation of club theory and its application to international organisations, see Fratianni and Pattison (2001). 18 See, for example, The World Bank (2001), Silva Lopes (2003). 19 It is correct to consider the four categories as dimensions rather than as stages within a single process, because the regional integration agreements are evaluated by the ECB

38 Philippe De Lombaerde and Luk Van Langenhove

20 21 22 23 24

25

26

27 28 29

30

independently in terms of their progress in each category, thus taking into account that integration in the real world does not always respect Balassa’s model and that parallel processes and inverted sequences are not exceptional. For a good overview of methods and results since Verdoorn’s pioneering studies (Verdoorn and Meyer zu Schlochtern, 1964), see El-Agraa (1999). Several of these ‘key variables’ are in reality a set of variables themselves. For an overview of actors involved in integration projects in the real world, see, for example, the mapping exercises of the WTO (2000) and UNU/CRIS (www.cris.unu. edu). See Hettne (1999). The classical references on gravity models for trade analysis are Tinbergen (1962), Pöyhönen (1963) and Linnemann (1966). For an overview of the theoretical discussion on the validity of the gravity equation (and an application to trade ﬂows in the Baltic Sea area) see Paas (2002). Paas identiﬁes three classes of theoretical foundations in the literature: (1) regional science and new economic geography, (2) microeconomic foundations, (3) trade theories. Hinojosa et al. (1996), for example, provide an overview of the debate on the (national and regional) labour market impacts of NAFTA for the United States and show the sensitivity of the results towards the type and quality of the data, and the methodology used. For illustrative purposes, we might mention some examples of combined indicator systems that have been presented as monitoring tools for integration processes. The Transport and Environment Reporting Mechanism (TERM) model of the European Environment Agency, for example, is a relevant and interesting model for monitoring a speciﬁc policy area on a regional level. The TERM is a ﬂexible instrument containing quantitative as well as qualitative data on progress made in these policy ﬁelds. The indicators-based model shows the quality of policies, the degree of their implementation and the need for more policy coordination in speciﬁc areas (EEA, 2001). Another recent example is Sustainable Calgary’s system of indicators of sustainability and the quality of life (Keough, 2002). Although conceived as a tool for the local community, it is thought to be a system that can be transformed to monitor the effects of international trade agreements. See http: //www.polity.co.uk/global/ See, for example, Devlin and Ffrench-Davis (1998). In addition, from a social policy perspective, it is necessary to distinguish between the magnitude of the ﬂows of migration and the degree to which the immigrants are actually ‘integrated’ in the host societies. Figures on temporary migration and return migration could give some indication about the structural character of migration but the ‘integration’ issue of immigrants is much more complex and its monitoring would require speciﬁc indicators. According to the typology proposed by Werth et al. (1997), three different types of indicators could be considered: (1) indicators of accessibility, dealing with the legal framework in which migrants live, (2) indicators of the living situation of migrants, referring to employment, education, welfare, housing, etc., and (3) indicators on attitudes, expectations and political participation of migrant and host population, using opinion surveys. The typology is based on the work presented in the framework of the project ‘The Integration of Immigrants: Towards Equal Opportunities’ of the European Committee on Migration (CDMG) of the Council of Europe. The complete series of papers, showing the different national perspectives, has been published as Council of Europe (1997). ‘Vertical imbalances’ refer to the degree to which national governments rely on revenues from supranational institutions to support their expenditures.

Indicators of regional integration 39 REFERENCES Anderson, V. (1991) Alternative Economic Indicators, London: Routledge. Balassa, B. (1961) The Theory of Economic Integration, Homewood, Ill: Richard Irwin. Bekemans, L., Fiorentino, R. and Van Langenhove, L. (2000) ‘Reshaping Integration Studies in Social Sciences: Challenges and Perspectives for Research’, Collegium, 20: 43–72. Bensidoun, I. and Chevallier, A. (1998) ‘Indicators of Regionalisation in North America, Asia, Central and Eastern Europe and the Mediterranean Basin’, in OECD, Globalisation, Migration and Development, 15–33, Paris: OECD. Best, E. (1997) ‘Capacities for Regional Integration: Conceptual Framework for Comparative Analysis’, in M.O. Hosli and A. Saether (eds) Free Trade Agreements and Customs Unions: Experiences, Challenges and Constraints, 51–75, Maastricht: TACIS/European Commission – EIPA. Collier, P., Schiff, M., Venables, A. and Winters, L.A. (2000) Trade Blocs, The World Bank, Washington–New York: Oxford University Press. COMESA (2002) Discussion Paper on Surveillance as a Means to Measure the Degree to Which Integration is Taking Place in the COMESA Region, COMESA Secretariat. Cooper, R. (1985) ‘Economic Interdependence and Coordination of Economic Policies’, in R.W. Jones and P.B. Kenen (eds) Handbook of International Economics, 2: 1195–1234, North-Holland, Amsterdam. Council of Europe (1997) Measurement and Indicators of Integration, Strasbourg: Council of Europe. De Lombaerde, P. (2002) ‘Integración asimétrica’, in P. De Lombaerde (ed.) Integración asimétrica y convergencia económica en las Américas, 15–32, Bogotá: Universidad Nacional-Antropos. Devlin, R. and Ffrench-Davis, R. (1998) ‘Towards an Evaluation of Regional Integration in Latin America in the 1990s’, Working Paper, INTAL, 2. Dobson, W. (1991) ‘Economic Policy Coordination: Requiem or Prologue?’, Policy Analyses in International Economics, 30. Dorrucci, E., Firpo, S., Fratzscher, M. and Mongelli, F.P. (2002) ‘European Integration: What Lessons for Other Regions? The Case of Latin America’, European Central Bank Working Paper Series, 185. EEA (2001) ‘Term 2001. Indicators Tracking Transport and Environment Integration in the European Union’, Environmental Issues Series, European Environment Agency, 23. El-Agraa, Ali M. (1999) Regional Integration. Experience, Theory and Measurement, Basingstoke and London: Macmillan Press. European Commission (2002a) ‘Regional Cooperation Review Process’, Discussion Paper, Brussels: DG Development, European Commission, (DEV/B/1/WKD(2002)). European Commission (2002b) ‘Informal Workshop to Discuss the Regional Cooperation Review Process’, Meeting Report, Brussels: DG Development, European Commission (DEV/B/1/WK D(2002)). Fratianni, M. and Pattison, J. (2001) ‘International Organisations in a World of Regional Trade Agreements: Lessons from Club Theory’, The World Economy, 24(3): 333–58. Freudenberg, M., Gaulier, G. and Ünal-Kesenci, D. (1998a) ‘La régionalisation du commerce international: Une evaluation par les intensités relatives bilatérales’, Document de Travail, Paris: CEPII (1998–05). Freudenberg, M., Gaulier, G. and Ünal-Kesenci, D. (1998b) ‘La régionalisation du commerce international’, Économie Internationale, 74: 15–41. Gavin, B. and Van Langenhove, L. (2003) ‘Trade in a World of Regions’, in G. Sampson

40 Philippe De Lombaerde and Luk Van Langenhove and S. Woolcock (eds), Regionalism, Multilateralism and Regional Integration, Tokyo: UNU Press, 277–313. Gilpin, R. (1987) The Political Economy of International Relations, New Jersey: Princeton. Greif, A. (1997) ‘Cultural Beliefs as a Common Resource in an Integrating World’, in P. Dasgupta, K.G. Malter and A. Vercelli (eds) The Economics of Transnational Commons, 238–96, Oxford: Clarendon Press. Held, D., McGrew, A., Goldblatt, D. and Perraton, J. (1999) Global Transformations: Politics, Economics and Culture, Cambridge: Polity. Hettne, B. (1999) ‘Globalisation and the New Regionalism: The Second Great Transformation’, in B. Hettne, A. Inotai and O. Sunkel (eds) Globalism and the New Regionalism, 1–24, Macmillan Press. Hettne, B. (2005) ‘Regionalism and World Order’, in M. Farrell, B. Hettne and L. Van Langenhove (eds) Global Politics of Regionalism, London: Pluto Books. Hettne, B. and Söderbaum, F. (2000) ‘Theorising the Rise of Regionness’, New Political Economy, 5(3): 457–73. Hinojosa Ojeda, R., Dowds, C., McCleery, R., Robinson, S., Runsten, D., Wolff, C. and Wolff, G. (1996) North American Integration Three Years After NAFTA. A Framework for Tracking, Modeling and Internet Accessing the National and Regional Labor Market Impacts, Los Angeles: North American Integration and Developing Center, UCLA. IADB (2002) Draft Profile Strategy. Regional Integration, Washington: IADB. Online. Available HTTP: . Iapadre, L. (2000) ‘Accordi di integrazione regionale e geograﬁa del commercio mondiale: problemi di misurazione ed evidenza empirica’, Quaderni di ricerca, Istituto nazionale per il Commercio Estero, 13. Iapadre, L. (2001) Intra-Regional Trade Intensity: A Survey of the Indicators, Powerpoint presentation Aquila: University of l’Aquila – ICE. Keohane, R. (1980) ‘The Theory of Hegemonic Stability and Changes in International Economic regimes, 1967–77’, in O.R. Holsti et al., Change in the International System, Boulder, CO.: Westview Press, Chapter 6. Keohane, R.O. and Nye, J.S. (1977) Power and Interdependence: World Politics in Transition, Boston: Little Brown. Keough, N. (2002) ‘Using Indicators to Engage the Community in Sustainability Debates’, in J.J. Kirton and V.W. MacLaren (eds) Linking Trade, Environment, and Social Cohesion. NAFTA Experiences, Global Challenges, 277–95, Aldershot: Ashgate. Kindleberger, C.P. (1981) ‘Dominance and Leadership in the International Economy: Exploitation, Public Goods, and Free Rides’, International Studies Quarterly, 25: 242–54. Kreinin, M.E. and Plummer, M.G. (1994) ‘“Natural” Economic Blocks: An Alternative Formulation’, The International Trade Journal, 8(2): 193–205. Krugman, P. (1991) ‘Is Bilateralism Bad?’, in E. Helpman and A. Razim (eds) International Trade and Policy, Cambridge: MIT Press. Linnemann, H. (1966) An Econometric Study of International Trade Flows, Amsterdam: North-Holland. Mattli, W. (1999) The Logic of Regional Integration. Europe and Beyond, Cambridge: Cambridge University Press. Olson, M. (1965) The Logic of Collective Action, Cambridge, Mass.: Harvard University Press. Ondarts, G. (1992) ‘La nueva integración’, Integración Latinoamericana, 92: 3–12.

Indicators of regional integration 41 Paas, T. (2002) ‘Gravity Approach for Exploring Baltic Sea Regional Integration in the Field of International Trade’, HWWA Discussion Paper, 180. Page, S. (1998) ‘Survey of regions’, Politeia, 17(3): 22–43. ‘Partnership Agreement Between the Members of the African, Caribbean and Paciﬁc Group of States of the One Part, and the European Community and Its Member States, of the Other Part. Signed in Cotonou on 23 June 2000’, Special Issue Cotonou Agreement, Brussels: DG Development, European Commission (September). Pelkmans, J. (1993) ‘Comparando las integraciones económicas: prerrequisitos, opciones e implementaciones’, Integración Latinoamericana, 191: 3–17. Pöyhönen, P. (1963) ‘A Tentative Model for Volume in Trade Between Countries’, Weltwirtschaftliches Archiv, 90: 91–113. Silva Lopes, P. (2003) ‘The Relative Usefulness of Social Spending Indicators in Determining Social Outcomes in Sub-Saharan Africa’, paper presented at the IIAS Conference on Shared Governance: Combating Poverty and Exclusion, Yaoundé, July. Slocum, N., Mordonu, A. and De Lombaerde, P. (2003) ‘The Zambia-Malawi-Mozambique Growth Triangle: A Case Study in New Regionalism’, in H. van Ginkel, J. Court and L. Van Langenhove (eds) Integrating Africa: Perspectives on Regional Integration and Development, 89–121, Tokyo: UNU Press. Stakhovitch, A. (1991) ‘Unas y otras integraciones’, in V.L. Urquidi and G. Vega Canovas (eds) Unas y otras integraciones. Seminario sobre integraciones regionales y subregionales, 69–90, México: El Colegio de México, Fondo de Cultura Económica. Thant, M., Tang, M. and Kakazu, H. (eds) (1994) Growth Triangles in Asia. A New Approach to Regional Economic Co-operation, Hong Kong: Oxford University Press. Tinbergen, J. (1954) International Economic Integration, Amsterdam: Elsevier. Tinbergen, J. (1962) Shaping the World Economy, New York: Twentieth Century Fund. UNECA (2001) Annual Report on Integration in Africa. Methodology for Calculating Indices of Economic Integration Effort in Africa, Addis Ababa: UN Economic Commission for Africa. UNECA (2002) Annual Report on Integration in Africa 2002, Addis Ababa: UN Economic Commission for Africa. Van Langenhove, L. (2003) ‘Regional integration and global governance’, UNU Nexions, 1–5. Van Velden, L. (1988) ‘A Survey into the Literature on Game Theory and Policy Coordination in International Economics’, Research Memorandum, University of Amsterdam. Verdoorn, P.J. and Meyer zu Schlochtern, F.J.M. (1964) ‘Trade Creation and Trade Diversion in the Common Market’, in H. Brugmans (ed.), Intégration Européenne et réalité économique, Bruges: de Tempel. Werth, M., Delfs, S. and Stevens, W. (1997) ‘Using Indicators of Integration for the Implementation of Integration Policies’, in Council of Europe, Measurement and Indicators of Integration, 187–93, Strasbourg: Council of Europe. World Bank (sd) Fiscal Decentralisation Indicators, Washington: The World Bank. World Bank (2001) World Development Indicators, Washington: The World Bank. World Bank (2002) WB Comments to the EC Draft Discussion Paper on a Regional Cooperation Review Process, Washington: The World Bank. WTO (2000) ‘Mapping of Regional Trade Agreements’, Note by the Secretariat, Committee on Regional Trade Agreements, Geneva: World Trade Organisation (WT/REG/W/41).

2

Towards a qualitative monitoring of regional integration Luk Van Langenhove

INTRODUCTION From a social scientist’s point of view, processes of regional integration can be studied in a variety of ways. In addition to theoretical approaches that try to understand the nature of regionalisation/regionalism, an empirical approach is also possible. In the latter case, one can focus on monitoring interactions and on assessing the impact of regional integration on both those included and those excluded from a given integration scheme. Both the monitoring and impact assessment processes need indicators. In line with mainstream economic thinking such indicators are usually thought of as quantitative data. In this chapter it will be argued that in addition to quantitative data, the study of regional integration also needs empirically based qualitative data. Also, it will be advocated that a qualitative monitoring and impact assessment method can be constructed upon the so-called ‘fourth generation evaluation’ model (Guba and Lincoln, 1989) and that such an approach allows a participatory monitoring and assessment scheme to develop that can demonstrate the ‘added value’ of regional integration.

THE QUALITATIVE–QUANTITATIVE DICHOTOMY IN THE SOCIAL SCIENCES Two major methodological paradigms occur within social science: there are the ‘general’ and ‘particular’ methodologies, also known as the nomothetic and idiographic approaches. The word ‘nomothetic’ is taken to characterise the search for general laws (applicable in all circumstances at any time). The word ‘idiography’ indicates a concern for what is particular to the individual case. The nomothetic approach is in line with the positivist paradigm: the seeking of general laws by employing only those procedures admitted by the natural sciences. The ‘behaviour’ of a single apple falling down from a tree, for example, can be easily explained by applying the general law of gravitation. So, explaining why apples fall with a certain speed and acceleration to the earth can be discovered by observations and

Qualitative monitoring 43 measurements, the result being a theory that applies to all apples (and for that matter to all other objects). The natural sciences do not seek to explain why one particular apple falls down from a tree at that particular time . . . Applied to the social sciences, this implies that researchers should study aggregates of events in order to detect regularities across individual particulars. The nomothetic approach has thus pushed the social sciences in a ‘positivist’ quest for reliable and replicable data (Van Langenhove, 1996). The difference between quantitative and qualitative methodologies runs, to a great extent, parallel with the above distinction. Quantitative research is based on statistical inference and, to a lesser extent, on mathematical models. For the quantitative researcher, only ﬁgures make up valuable ‘data’. Anything that cannot be expressed in ﬁgures cannot be analysed scientiﬁcally. This view is challenged by the advocates of qualitative methodologies who argue that one should first describe in detail the phenomena one studies before trying to impose quantitative measurements. Some qualitative researchers will even argue that quantiﬁcations are always ‘bad’ science. The qualitative approach that opposes the mainstream quantitative approach is not monolithic but consists of different ‘schools’ and has different names such as ethnomethodology, naturalistic inquiry, etc. A central element, however, is the focus on case studies and on descriptive assessments. Today most scholars will state that both methods have their value and that a qualitative approach does not exclude using quantitative data.

THE QUALITATIVE SIDE OF STUDYING REGIONAL INTEGRATION Regional integration is a complex process that needs to be studied from a wide range of perspectives. Obviously, economical and political-legal quantitative indicators will play an important role in such studies (see other chapters in this volume). But there is, in my view, also a need to develop a qualitative approach to the study of regional integration (Levy, 2002). Such a qualitative approach is related to three aspects of the research process: • • •

the data-gathering perspective, the units of analysis, the purpose of the analysis.

Data gathering Qualitative data are words rather than numbers. Those words can come from different sources such as interviews or content-analysis of documents. Applied to regional integration, this implies that sources of data can include treaties, ofﬁcial documents, non-papers, press articles, interviews with stakeholders, etc. If, for instance, one is interested in analysing how regional responses to risk are organised, data could include documents referring to the development of regional early

44 Luk Van Langenhove warning systems to reduce impacts of natural disasters, as well as assessments by participants of the practical functioning of such systems. The main problem with using such data is to make sure that the data-analysis is done in a sound, reliable way. One of the most advanced methodologies for handling qualitative data has been developed by Miles and Huberman (1994). More recently, this methodology has been coupled to software that can assist the researcher in organising the data as well as in producing conclusions. There is no reason to say that qualitative research is less ‘scientiﬁc’ than quantitative research, but it requires more thought and takes more time than most off-the-shelf quantitative analysis. Some authors will argue that, all in all, qualitative research is more likely to lead to useful and defensible ﬁndings.

Units of analysis Qualitative research works first and foremost with single cases. Quite often studying one single case is the objective of a project. But single cases can also be elements of a comparative analysis and there exist multiple schemes of conducting cross-case analyses. In studying regional integration, this implies that attention can be directed to single integration processes (e.g. the European integration), to single integration organisations (e.g. Mercosur) or to single aspects of a regional integration scheme such as trade or environment. Also, such cases can be used in a broader comparative project.

Purpose Perhaps the important aspect of the qualitative approach is that both its way of data gathering and its focus on single cases are related to the specific purpose of gathering intelligence about one individual process or product. As such, qualitative research has become the prime method in evaluation research where the purpose is to assess single events. In the next paragraph, we will explore in some detail how evaluation processes have evolved towards embracing qualitative research methodology.

FOUR GENERATIONS OF EVALUATION Guba and Lincoln (1989) have presented a four-stage history of evaluation practices that nicely shows how the paradigm moved from a positivist and purely quantitative approach to a qualitative approach. In a ﬁrst generation, evaluation was focused on measurement of facts. Applied to schools, this resulted in individual students being the object of evaluation. For these studies tools such as psychometric tests were available and the role of the evaluator was purely technical. The second generation shifted attention to descriptions of strengths and weaknesses of school curricula based upon measurements of the extent that curricula

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Qualitative monitoring 45 reached their stated objectives. The focus was no longer on what an individual student had learnt but on what the school had intended them to learn. In a third generation, the main focus was not the achievement of the objectives but the question whether the objectives were worthwhile. The evaluator now became a judge. These three types of evaluation occur in many different contexts and one can easily see their applicability to regional integration. In the ‘measurement’ approach, the focus is on gathering data such as foreign direct investment (FDI) or trade and these data are analysed as such. In the second approach these data are compared with stated regional integration objectives (benchmarks). In the third approach, the objectives themselves are scrutinised, often on the basis of quantitative data. Guba and Lincoln consider the three generations of evaluation they identiﬁed to be ﬂawed to the extent of having to be replaced. Three major categories of ﬂaw (cf. Guba and Lincoln, 1989) were described as detailed below.

A tendency towards managerialism The manager who hires the evaluator tends to stand outside the evaluation – his or her managerial qualities are not called into question. The manager/evaluator relationship is disempowering for those being evaluated as they decide which questions should be asked, how answers will be collected and interpreted and who will see the results. Other stakeholders are not represented and may be disenfranchised.

A failure to accommodate value pluralism When evaluation is seen to be about valuing there is the question of whose values. The claim of value-freedom within the scientiﬁc mode of inquiry is not tenable and, that being the case, the value-pluralism within societies and between cultures is a crucial matter to be attended to in an evaluation. None of the evaluation approaches of the ﬁrst three generations accommodates value differences in the slightest.

An overcommitment to the positivist paradigm of inquiry Virtually all of the ﬁrst three evaluation models use the positivist paradigm to guide their methodological work. But this extreme dependence on the methods of natural sciences has had unfortunate results: • • •

context stripping – leading to generalisation in questions and results and fatally reducing the contextual relevance and usability of the ﬁndings, overdependence on quantitative measurement – leading to the modus operandi that what cannot be measured cannot be real, the ‘coerciveness of truth’ – science claims to tell us about ‘the way things really are’ and, given managerialism and commitment to the scientific

46 Luk Van Langenhove

• •

paradigm, this locks thinking into the positivist mode and lends illegitimate support to the status quo, scientiﬁc truth is non-negotiable – if science discloses the truth about things then any other alternative explanations must be in error, the evaluator bears no moral responsibility for his or her conclusions if they are ‘scientiﬁc truth’.

Although the evaluation practices referred to by Guba and Lincoln are quite distinct from what concerns us, their critiques apply to a large extent to how regional integration is monitored. If one looks at the completed studies available, one can easily detect ﬂaws similar to those of the three generations of evaluation: • • •

a focus on quantitative measurement, a researcher perspective that does not include stakeholders’ perspectives, a ‘value-free’ approach that does not include appraisals of regional integration.

For Guba and Lincoln (1989), the answer to the shortcomings of the first three generations of evaluation practices is the adoption of a social constructionist approach to evaluation which they termed ‘fourth generation evaluation’. In my view this approach can be valuable to organise monitoring of regional integration as well. Fourth generation evaluation is organised by the claims, concerns and issues of stakeholding audiences, and it utilises the methodology of the constructionist paradigm. It includes the following processes but not necessarily in the order indicated: • •

•

Identify the full array of stakeholders who are at risk in the projected evaluation. Elicit from each stakeholder group their constructions about the issues at hand and the range of claims, concerns and issues they wish to raise in relation to it. Provide a context and a methodology through which different constructions, and different claims, concerns and issues, can be understood, critiqued and taken into account: •

•

Carry out this methodology within each stakeholder group, so that the group construction (or several, if there are intra-group conflicts) can emerge and decisions can be reached about which claims, concerns and issues should be pursued. Cross-fertilise each group with the constructions, claims, concerns and issues arising from other groups so that those items must be confronted and dealt with. This cross-fertilisation may also include constructions drawn from the literature, from other sites or from the experience of the evaluator. Any construction, claim, concern or issue may properly be introduced so long as it is open to critique and criticism.

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Qualitative monitoring 47 • • • • •

•

Generate consensus with respect to as many constructions, and their related claims, concerns and issues as possible. Prepare an agenda for negotiation on items about which there is no, or incomplete, consensus. Collect and provide the information called for in the agenda for negotiation. Establish and mediate a forum of stakeholder representatives in which negotiation can take place. Develop a report, probably several reports, that communicate to each stakeholder group any consensus on constructions and any resolutions regarding the claims, concerns and issues that they have raised (as well as regarding those raised by other groups that appear relevant to that group). Recycle the evaluation once again to take up still unresolved constructions and their attendant claims, concerns and issues.

Guba and Lincoln (1989) give ﬁve reasons for insisting upon the use of stakeholder claims, concerns and issues as the basis for deciding what information is needed in an evaluation. These are as follows: • •

• • •

Stakeholders are placed at risk by an evaluation; thus, in the interests of fairness, they deserve to have input into the process. Evaluation exposes stakeholders to exploitation, disempowerment and disenfranchisement; thus, in the interests of self-defence, they are entitled to some control over the process. Stakeholders represent a virtually untapped market for the use of evaluation ﬁndings that are responsive to self-deﬁned needs and interests. The inclusion of stakeholder inputs greatly broadens the scope and meaningfulness of an inquiry. All parties can be mutually educated to more informed and sophisticated personal constructions as well as an enhanced appreciation of the constructions of others.

QUALITATIVE MONITORING OF REGIONAL INTEGRATION There is a case to be made for developing a qualitative strategy to monitoring regional integration along the lines of the constructionist approach discussed above. Such an approach would allow the inclusion of more than just economic parameters and focus upon the interconnectedness of states in regions. This would be in line with the so-called new regionalism approach. Also, such an approach would allow the monitoring exercises to be performed with the full cooperation of the different stakeholders involved. As a ﬁrst step in developing such an approach, the following three issues can be put forward.

48 Luk Van Langenhove

Identiﬁcation of a broad set of dimensions to assess the extent and depth of interconnectedness. Inspired by Held et al. (1999), one should consider including the following dimensions in any regional integration monitoring or assessment exercise: 1

Political-legal dimension: • • • • •

2

Economic dimension: • •

3

regional identity regional civil society.

Security dimension: • •

6

transport migration refugees.

Cultural dimension: • •

5

international trade foreign direct investment.

Mobility dimension: • • •

4

diplomatic connectedness of states growth of Foreign Ministries (and/or ministries following up regional integration) development of customary international law recognised by states membership of formal international treaties international transactions of domestic political agencies.

peace natural disaster.

Cooperation dimension: • •

environment criminality.

The above list is far from exhaustive and needs to be completed. In line with what was discussed before, it should also be noted that the dimensions include much more than quantitative indicators. As such, this list complements the proposal for classifying variables in a SIRI as presented in Chapter 1: the dimensions to be considered can involve different variables of both quantitative and qualitative indicators.

Identiﬁcation of consequences of integration processes Interconnectedness is not only something that can be described, it also needs to be assessed along a number of dimensions. In my view, what one needs to do is gather

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Qualitative monitoring 49 intelligence regarding the extent to which a certain regional integration initiative has an ‘added value’. As a starting point, one can look to the following possible consequences in such an assessment. 1 2 3 4 5 6 7 8

Contribution to the creation of an appropriate enabling environment for private sector development. Contribution to the strengthening of trade integration in the region. Contribution to developing strong public sector institutions and good governance. Contribution to reducing social exclusion and fostering of an inclusive civil society. Contribution to the development of infrastructure programmes in support of economic growth and regional integration. Contribution to building environmental programmes at the regional level. Contribution to peace and security in the region. Contribution to strengthening the region’s interaction with other regions in the world.

Identiﬁcation of stakeholders involved in the regional integration process The third step in developing a qualitative approach to the monitoring of regional integration is the identiﬁcation of different stakeholders involved in the regional integration process. Regional integration stands in a double relation to people: on the one hand, people can gain or lose by regional integration (or they can remain unaffected), and on the other hand some people can have a discretionary power to make decisions about regional integration. From a formal point of view, the following types of stakeholders can be distinguished: • • •

agents beneﬁciaries victims.

‘Agents’ are those people involved in designing, promoting and implementing regional integration. People who somehow proﬁt from the development or implementation of a regional integration scheme can be called ‘beneficiaries’. Such beneﬁts can be direct or indirect (e.g. more export possibilities for one’s ﬁrm versus a better economic climate that contributes to creating someone’s job). Some beneﬁts can be directly related to the regionalisation process itself (e.g. career prospects for international civil servants). Those people who are negatively affected by the regional integration quality as ‘victims’. This can include people who suffer opportunity costs or negative side effects. A distinction should also be made between actual and potential beneﬁciaries and victims. And one also has to take into account that the same person can be agent and beneﬁciary or victim at the same time.

50 Luk Van Langenhove Stakeholders can be identiﬁed as groups that hold their own claims, concerns and issues regarding regionalisation. A ‘claim’ is any assertion that a stakeholder may introduce that is favourable for regional integration. A ‘concern’ is any assertion that is unfavourable and an ‘issue’ is any state of affairs about which reasonable persons may disagree. Applying this to regional integration means that one has to identify different groups in society that are somehow affected by the success or failure of a regionalisation process. This might prove to be a very difficult exercise as not all stakeholders are fully aware of their role. As a starting point, one could focus upon: • • • • • •

governmental officials (both those involved in external affairs and those involved in internal affairs) ofﬁcials of regional integration institutions political parties industrialists (regional) non-governmental organisations global organisations.

Together, these three preparatory steps should enable a monitoring exercise to be conducted that follows the constructionist and participative approach as advocated by Guba and Lincoln (1989). The next step would then be to draw up a ‘contract’ between the parties involved in the monitoring process. Such a contract should determine the ambition, scope and limitations of the exercise. It will be essential that the setting up of such a contract is in itself a dialogical process related to the overall question of ‘what do we want to achieve by this exercise?’. Finally, one can then start the process of drawing up lists of claims, concerns and issues.

CONCLUSIONS Regional integration can be studied as a process or as a state of the art. The latter focuses on describing what is actually going on (including the consequences of regional integration). The former deals with how a certain state of regional integration is being achieved. The qualitative and constructionist approach to monitoring regional integration advocated in this chapter puts the focus on the understanding of how and why the processes take place, as well as on studying the broad consequences of those processes. As such it opens avenues for monitoring the progress that countries make in regional integration. Rather than being satisﬁed with a description of the level of integration, such an approach allows us to study how and at what pace and costs countries achieve a progressive elimination of the imperfections of their regional integration.

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Qualitative monitoring 51 REFERENCES Guba, E.G. and Lincoln, Y. (1989) Fourth Generation Evaluation, London: Sage. Held, D., McGrew, A., Goldblatt, D. and Perraton, J. (1999) Global Transformations: Politics, Economics and Culture, Cambridge: Polity Press. Levy, J.S. (2002) ‘Qualitative Methods in International Studies’, in F.P. Harvey and M. Brecher (eds) Evaluating Methodology in International Studies, Ann Arbor: University of Michigan Press. Miles, M.B. and Huberman, M. (1994) Qualitative Data Analysis: an Updated Sourcebook, London: Sage. Van Langenhove, L. (1996) ‘The Theoretical Foundations of Experimental Psychology’, in J. Smith, R. Harré and L. Van Langenhove (eds) Rethinking Psychology, London: Sage.

3

Communicative indicators for the study of regional integration A critical theory perspective Ann-Christin Raschdorf

This chapter seeks to contribute to an ongoing discussion on what sort of qualitative indicators are required to understand and measure changing degrees of social and cultural integration in non-Western contexts. Following Jürgen Habermas’ discourse theory, it develops a set of procedural indicators for communicative participation and empowerment in regional integration processes. It argues for an application of a formal dialogical matrix that allows for a deﬁnition of varying degrees of local ownership through a measurement of processes of discursive inclusion and exclusion (Habermas, 1976a). The intent is to show how critical theory can contribute to an alternative, not state-centric understanding of regional integration from the ground up. This chapter shares many of the methodological concerns that were ﬁrst voiced by a so-called ‘fourth generation evaluation model’ that aimed ‘to organise evaluation by the claims, concerns, and issues of stakeholding audiences’ (Guba and Lincoln, 1989). However, a communicative approach differs in many ways from a constructionist approach to the monitoring of regional integration. Following Habermas’ critique of Schütz (Schütz and Luckmann, 1979), it argues that it is not sufﬁcient to ensure and monitor formal participation as such. It points at the necessity of scrutinising the contextual basis of the interaction of the involved stakeholders. A prerequisite for such a contextual understanding is a reﬂexive discursive analysis of both expert and local discourses. In other words, while a constructionist approach signiﬁes an important ﬁrst step towards a communicative turn in regional studies, this chapter argues that a wider reaching discursive approach is needed to understand all possible sources for misunderstandings and competing interpretation. The intent of this chapter is to demarcate these conceptual spaces that favour or hinder higher levels of communicative cooperation on local, national and international levels. In this context, it aims to include alternative local discourses on regional integration and their validity claims. It will further point at present conceptual shortcomings within the regional integration discourse. In this context, it will be argued that most current approaches fail to explain the full scope of regional integration insofar as they focus either on the issue of systemic complexity or on strategic cooperation and neglect the perceptual dimension of regional integration. This bias leads to a systemic exclusion of issues that are associated with concerns at lower levels (such as human security or local

Communicative indicators 53 ownership of the process). This chapter will argue for a parallel analysis of (nonstrategic/communicative) transformation processes at international, national and, most importantly, at local community level. In recent times, identity-based models have sought to remedy past theoretical shortcomings by adding a perceptual dimension to the field and studying newly emerging identification structures at international, regional and national level. However, their shortcoming lies in the fact that they do not (at least not explicitly) discuss the extent to which alternative local identities are represented, included or excluded. These questions of inclusion and exclusion can only be answered if one chooses to scrutinise further the enabling conditions of identity structures in terms of communication structures and processes. A definition of integration as a structure/outcome and process necessarily implies that we have to ask not only which common dynamics are currently used to reproduce or modify the status quo within given parameters but also which other factors rightly challenge the latter and lie beyond present conceptual categories. In this context, it is important to ensure that the term integration reflects also a critical potential. Following Habermas and Linklater (Linklater, 1998), this chapter suggests an inclusion/exclusion perspective to integration that conceptualises integration as an afﬁrming life-worldly inclusion process and as a negating systemic exclusion process at the same time. It argues that current integration processes reinforce frictions on lower levels that have to be made explicit and have to be addressed by communicative means. As a consequence, this chapter makes a case for a conscious transition of the normative and interactive barriers that characterise the local community, national and international/regional divide. It stresses the importance of participation and conceptual complementarity and coherence on all levels of cooperation. While it is beyond the scope of this chapter to develop this argument fully in a comparative regional context, it will scrutinise current indicators for integration in the United Nations framework. It will question standard institutional deﬁnitions and provide conceptual alternatives. This chapter follows a methodology that was ﬁrst outlined by Jürgen Habermas in his classic Reconstruction of Historical Materialism (Habermas, 1976b). This suggests the necessity of a de- and reconstruction of theoretical and perceptual discourses (e.g. on regional integration) and their normative foundations. It stresses that any analysis should begin with an epistemological critique of the dominant discourse of a discipline. As will be shown, structural-functional and neo-liberal assumptions inﬂuence the way we think and talk about regional integration and arguably also inﬂuence the ways in which we seek to compare it. The method of conceptual de- and reconstruction is quite straightforward and can be applied to any ﬁeld of human knowledge. It is comprised of two steps as follows: 1

The ﬁrst step lies in a de-construction of conceptual/textual certainty through an inclusion of competing normative agendas and a consequent identiﬁcation of weaknesses within the functional/neo-liberal paradigm. This textual strategy will inform the ﬁrst part of this chapter.

54 Ann-Christin Raschdorf 2

The second step is a reconstruction of the normative foundation of analysis and a development of a new set of (communicative) indicators (e.g. for regional integration).

According to Talcott Parsons (Parsons, 1949), the founding father of structural functionalism, integration specifies one of the key functions that every social system (this includes any kind of organisation, groups of individuals, communities or states and hence also regions) has to fulﬁl to ensure its survival. According to his AGIL scheme, social systems reproduce themselves through the fulﬁlment of four core tasks: •

• •

•

Adaptation: this task is derived from the necessity of a system to acquire/ reproduce adequate material resources. This function is fulfilled by the economic subsystem of a social system. Goal attainment: this task corresponds to the necessity of goal setting and implementation. This function is fulﬁlled by the political subsystem. Integration: this task highlights the necessity to maintain solidarity and coordination among the subunits of the system. This function is fulﬁlled in all (economic, political, cultural) subsystems and (cultural, personal, behavioural) action systems by the social and legal system. In other words, integration is necessarily a normative function that is safeguarded by normative agents. Latency: this task stresses the necessity to create, uphold and transmit the system’s particular cultural norms and values. This educative function is fulﬁlled by the cultural subsystem.

The above-mentioned functions are accomplished by different subsystemic actors: • • • •

in the economic domain, systemic needs for adaptation are met by fiscal bodies/large cooperations, in the political domain, governments fulﬁl goal-attainment functions, in the social domain, courts/unions/political parties ensure the systemic survival through the meeting of integration tasks, in the cultural domain the same needs are met by schools and churches.

Each subsystem operates either in terms of directive or communicative media that are: 1 2 3 4

money in the economic domain, votes in the political domain, normative inﬂuence/legitimacy in the social domain, value commitment in the cultural domain.

The latter are natural indicators for the success and stability of a system and its constitutive (economic, political, social and cultural) components. These indicators have been explored by regime theorists like Haas (Haas, 1958) and system theorists

Communicative indicators 55 like Luhmann (Luhmann, 1975). Clearly, structural-functionalism has undergone many modifications since Parsons’ days. However, one generic problem of his theory has remained for his predecessors and has been a subject of criticism ever since. While his model is one of the most adequate approaches to explain the functioning of a systemic status quo (e.g. in organisational studies), it cannot explain normative change insofar as it is lacking a critical perceptual dimension. It is this very aspect that is often forgotten in the study of international regimes and processes of regional integration. However, most interestingly, this theoretical shortcoming also has practical ramifications, insofar as Parsons’ structuralfunctional core assumptions are not only reﬂected in expert discourses but also in current peace-building practices. As will be shown below, most, if not all, UN programmes for action (seeking to rebuild failed, rough or post-conflict states, ensuring their integration in the world community) can be explained through Parsons’ blue-print. As a consequence, they also share the perceptual shortcomings associated with his static model. This is a real issue of concern insofar as peacebuilding processes are commonly understood as transformation processes seeking to promote normative change within a society/region and are aiming to rebuild social cohesion. However, it is difﬁcult to envisage how this could be done without any in-built perceptual dimensions at regional as well as at national and local level. Any critical review of the past requires an ideology-critical (life-world) perspective, which the present integration and peace-building discourse is lacking. The next section will show how a critical based approach could contribute to a better understanding of additional challenges for peace-building and regional integration. Figure 3.1 shows how the four main areas of the current UN sustainable peacebuilding strategy – sustainable development, good governance, rule of law and education – reiterate Parsons’ catalogue of functional necessities of an integrated system. The optimum systemic condition is in this instance a peaceful international order that reproduces itself through free markets and economic globalisation, democracy, universal norms like the adherence to human rights and access to Western education. These ideals have corresponding indicators. Poverty is quantified as an income under one dollar a day, economic success is measured by a state’s share in the world market, good governance means that leadership is also being made quantiﬁable – through elections/the number of votes. Even social and cultural progress is operationalised in numeric terms like number of people with internet access/literacy rates. Obviously, these indicators are focused on quantiﬁable outcomes that can legitimise ‘result-orientated’ processes. They are modelled in analogy to quantiﬁable media of the economic and political domain (money and votes) since variables like normative inﬂuence/legitimacy and value commitment are difﬁcult to operationalise, quantify and measure in a cross-cultural context. This equation is clearly problematic and has sparked many methodological debates. It has also exempliﬁed the urgent need for alternative indicators. But this is just one side of the story. While it is difﬁcult to explain a normative phenomenon (like integration) solely in quantitative terms, there is an even more worrying aspect that has remained largely unnoticed. The structural-functional parallels

56 Ann-Christin Raschdorf DEVELOPMENT

RULE OF LAW GOOD GOVERNANCE EDUCATION/ CULTURAL IDENTITY

Economic dimension Social dimension Political dimension Cultural dimension

SYSTEMIC DIMENSION INT. COMMUNITY/REGIONS/STATES

Figure 3.1 Parsons’ pattern variables revisited

between UN programmes for action and Parsons’ model show that in most current practices peace-building is equated with ﬁnding ways to guarantee the functioning of a system and not with a normative recovery of a society. In other words, current peace-building processes neglect core perceptual transformation processes in the field of social and cultural integration. They seek to establish a status quo, a particular predetermined functional outcome. They do not allow the simultaneous conceptualisation of integration as a process and outcome. Why is that? The dilemma described above echoes a theoretical debate on legitimatisation that took place in sociology roughly 20 years ago. It was sparked by some observations made by Jürgen Habermas that have remained largely unnoticed in other disciplines but that have nevertheless not lost their validity. These thoughts provide an interesting new angle on how to define or redefine processes of integration. Habermas argues that while Parsons’ model can explain the importance of integration for the order and maintenance of a society, it does not account for the fact that present communicative integration processes are disrupted and changed by systemic forces. He speaks in this context of a ‘legitimisation crisis’ that characterises all modern societies. According to Habermas, integration

Communicative indicators 57 processes are the product of cultural life-worlds (Parsons’ social/cultural domain). Habermas argues that these processes are disrupted by the system (Parsons’ economic and political domain) (Habermas, 1981). He argues that integration processes are impoverished through the systemic interference of quantiﬁable media (money/votes) which replace communicative media that are based on a shared normative understanding. In other words, the life-world is organised according to quantiﬁable systemic priorities (money/power). Habermas refers to this process as ‘colonialisation of the life-world’. Systemic priorities (money/power) alienate actors from their cultural and normative values and turn them into agents of the system. Actors become either clients of the welfare state, and global consumers with misconceived needs, or disillusioned anarchists who might turn against the system but without any normative reference point (think of anarchic opposition or violent protests at World Summits or Attac). In the long run, the growing normative anomy also threatens the survival of the system insofar as the latter loses its normative foundations and hence its integrative powers. The system has to compensate for this through an increased systemic and administrative effort in which it tries to appease its citizens through further concessions of a welfare state. This requires a renewed intervention into the life-world (and on the periphery if one thinks for example of the ongoing changes in immigration and asylum policies). As a result the public sphere – that is for Habermas the mechanism through which agents of the life-world can engage in discursive will formation, that is to say can give expression to their real needs and inﬂuence policies – is again reduced and fragmented. In this way, faith in the normative foundations of the system and hence its legitimacy is eroded. Culturally shared feelings of a common purpose and shared visions that have safeguarded processes of social and cultural integration are lost. This process results ultimately either in moral anomy and blind consumerism or extreme traditionalism and violence. Current phenomena like terrorism and weak states can be (among other things) explained through an inadequate integration of (local/national) social systems in the current world order. It can be argued that these counter-movements that challenge the merits of globalisation seek to assert and open lost public spheres that could legitimate and control the process. Habermas’ argument suggests that a stable international/ regional order (based on successful integration processes) requires participation processes in all dimensions of the public sphere, that is to say on local, national and international level. Hence, any international, regional, national or local system that is not based on complementary normative understandings ultimately undermines itself and becomes a subject of disintegration. On the indicator front this means that key areas for research have to be those communicative media that are operative in the life-world and that reflect and reproduce shared understandings. But how can these variables be operationalised if they cannot be quantiﬁed? The key to a deﬁnition of qualitative indicators for (regional) integration is, of course, communication. To establish which kind of communicative indicators can be used, we have to look at the prerequisites of discourse. Obvious choices are, for example, the degree of participation in a society, an intact public sphere and a reﬂexive state of cultural/collective (self)knowledge

58 Ann-Christin Raschdorf (e.g. willingness to redeﬁne a group identity in the light of contesting values). The latter point is particularly important for a successful development of truly inclusive identity formations. To achieve and develop fully reﬂexive and inclusive ways of thinking within a (regional/national/local) community, it would be necessary to undertake a critical but also affirmative re-reading and recovery of traditional forms of identiﬁcation and cultural knowledge. The above-mentioned aspects could be positively inﬂuenced through grass-root consultations and empowerment of the poor and women through participatory education. Of course, much of what has been said so far is not essentially new. There is a growing number of practitioners who acknowledge that every institutional and interactive practice promoting integration exists through and is limited by its terminology. This insight ﬁnds its expression, for example, in a recent trend for cultural-political correctness in the language of the United Nations. For example, the term ‘local ownership’, which was first introduced by the United Nations Development Programme (UNDP) and then later reluctantly adopted by other international actors like the World Bank (WB), is one prominent example for a new awareness and acceptance of the cultural Other, the concession of limited modern wisdom and contextual constraints. This term stresses the importance of alternative normative agendas and the necessity of local participation and empowerment. The growing acceptance of sustainable peace as a necessary precondition for sustainable development (best expressed in the Brahmini Report, the latest UNDP strategy papers and the Special Reports of the Secretary General on Conflict Prevention and Conflict Resolution in Africa) further demonstrates an increasing understanding of procedural interdependence and a change in settings of objectives. Besides standard reforms of the security sector, demilitarisation and supervision of ceasefires (through classic peace-keeping operations) and macro-economic initiatives (WB/UNDP), other social components like the promotion of the rule of law, electoral assistance, re-education and resettlement of former combatants and the support of national reconciliation commissions have been added to the catalogue of activities. The latter are coordinated through the UNDP resident representative system, which also includes the UNDAF framework, that explicitly seeks to foster and facilitate communication processes between UN agencies in the ﬁeld, local NGOs, INGOs and governments. Through the procedural inclusion of actors at all levels, processes of political participation and civil society building have been strengthened. However, despite all of these efforts, it cannot be denied that this kind of sustainable peace-building still follows the unreﬂected principles of a dominant political Western discourse and consequently has not much in common with a culturally sensitive approach that fosters international, regional or national integration from the ground up. Current buzz words in the UN discourse like ‘good’ governance still suggest that the ultimate task of any integrating practice lies in the facilitation of a structural assimilation of periphery states to Western democracies.

Communicative indicators 59 In this comparative context, any alternative set of beliefs (reﬂecting different life-worldly/cultural knowledge) must appear suspect and/or inferior. True is what is dictated by science and economic interest. Just is what is to be considered as Western law. To give a simple example: not many practitioners would equate the application of the Sharia with rule of law. This is surprising insofar as the Sharia has been part of the constitution of secular democratic countries like India for the past 50 years. Although there are some penal provisions that give clear cause for concern, much of Islamic family law is a consistent alternative legal universe that has made provisions for divorced women long before females even had property rights or could get a divorce. On the whole, there is an ongoing trend to mistrust cultural/life-worldly solutions that are pivotal for integration. Adherence to traditional values is internationally misunderstood as an indicator for less developed, in structural terms, weaker states. And nearly every form of ideology is perceived as a potential terrorist threat. This is clearly as much a result of a realist discourse (‘norms are contra-productive to order’) as it is a product of neo-liberal thinking (‘democracy equals free trade and proﬁt’) and a cultural bias (‘Western democratic states are structurally better developed (and more peaceful) than nonWestern states’). These narratives are structural equivalents to Parsons’ economic and political domains and Habermas’ notion of the system. They are major agents of colonialisation. They hide some of the most predominant causes for disintegration on the periphery. They conceal, for example, the fact that conflicts do not necessarily result from the intra-systemic inability to cope with internal problems through the application of traditional, normative solutions (e.g. weak states) but from systematic misuse or a lost/limited understanding of the latter, or external pressures of a neo-liberal world order and its (national/local) agents. There are many ways in which cultural/ life-worldly knowledge can be distorted and people can be prevented from being themselves or are hindered from engaging in reinterpretation. Put differently, although normative systemic agents like UNDP acknowledge the importance of local ownership they do not give sufﬁcient attention to the structural function that this notion has for integration processes like peace-building and sustainable development. As a result, UN agents perceive local ownership as an additional goal and not as an enabling condition. As a consequence, the current terminology still lays more stress on the necessity for a cultural transformation of the non-Western rather than the need for local reinterpretation and cultural adaptation of functional transformation processes. This conceptual shortcoming is a consequence of a missing awareness and research into the normative/life-worldly components of integration processes. This weakness is best expressed in terms of another systemic UN buzz word ‘resultorientated process’. The latter clearly specifies a functional/systemic outcome (in Habermas’ terminology ‘strategic action’). It does not describe a consensual/ communicative based process. The current peace-building and development discourse simply ignores the fact that there is not only a necessity for a – economic, political and legal – sustainable order but also for a social and cultural sustainable peace. The present regional integration discourse commits the very same error

60 Ann-Christin Raschdorf through a conceptual negligence of communicative indicators for (socially and culturally sustainable) integration. Current functionalist assumptions stress the ideal of systemic stability at the cost of cultural diversity. Neo-liberal values of globalisation and economic expansion subjugate communicative and moral needs further under strategic prerogatives of systemic survival and reinforce the wellknown pathologies of modernity/colonialisation of the life-world that affect all societies but in particular the weaker states on the periphery. Despite all lip-service, most peace-building, development and integration efforts are still part of these narratives. This chapter has tried to show that an analysis of integration must lie beyond standard answers. It can not solely equal the meeting of functional needs. A successful integration process starts with a joint communicative/critical assessment of functional and normative needs (of all parties involved) and ends with a concrete assessment of the process within the framework of a society’s own categories and values. Given the normative nature of the process, the success of integration processes could be measured by the degree of dialogue between concerned actors on the ground and those actors and agencies in power (this applies on a national, regional and international plane). Integration starts from the ground up. It is this form of participation that should ultimately serve as a cross-cultural indicator for good indigenous governance (and not, for example, the frequency of elections) and a functioning civil/international society. So far, peace-building, development and other integration strategies have not been based on a critical and creative application of cultural knowledge which has always led – in one way or the other – to a duplication of existing (Western) knowledge. This explains why practitioners have found it extremely difﬁcult to include ideals of local ownership and community empowerment in their peace-building and development agendas. Ultimately, practitioners are faced with an unsolvable dilemma. On the one hand, their functional framework for action provides clear guidelines that encourage a replication of Western political and economic structures. Their action is focused on systemic adaptation and functional/neo-liberal goal attainment. On the other hand, they are confronted by competing normative directives seeking to promote integration and latent pattern maintenance. These objectives have, strictly speaking, no reference point in their framework for action. In this context, a critical theory based approach to regional integration can provide a unique normative/life-worldly perspective on interrelated questions of regional peace and development (compare Chapter 10). Its explanatory value results from its conceptual ability to bridge the gap between the domestic and the regional/international and its capacity to overcome the system/life-world divide. Integration, empowerment, local ownership and participatory practice are all communicative terms that depend on the application of communicative reason. Following Habermas, a critical approach suggests the introduction of a life-worldly perspective to integration that respects cultural knowledge and diversity. It argues for structural reform through communicative and moral learning instead of simple adaptation. It attempts to highlight those unjust and unequal power structures that the term ‘globalisation’ seeks to conceal and offers an ideology critique.

Communicative indicators 61 Current static models of integration leave little or no room for normative adjustments. They can explain neither transformative change nor ideological distortions. Some theorists have suggested remedying this condition by understanding functional outcomes in their procedural context. However, this does not solve the initial problem insofar as systemic processes assign structural functions to different actors that reproduce outcomes. This is why the study of the phenomenon of integration requires a life-worldly perspective. One aspect that will be crucial for the success of regional integration concepts on the periphery is the avoidance of structural-functional analogies that suggest the possibility of copying a European model of integration. The term ‘regional integration’ must not become yet another conceptual addition to a list of strategic do and don’ts for states on the periphery in their quest to survive the rough waters of the international system. It should become an expression of real and legitimate communicative cooperation of states/regions that share a common colonial history or that feel they have other shared local reference points/interests. This chapter has argued for an inclusion of communicative indicators in the monitoring of regional integration processes. It has suggested operationalising, in addition to numeric/systemic indicators of money and power, alternative communicative variables measuring cultural legitimacy and value commitment through an inclusion of a life-worldly perspective to integration. In this context, special attention was given to procedural indicators for communicative participation and empowerment. These two variables specify cornerstones of a formal dialogical matrix that seeks to put different normative assumptions on different levels under a constant review. The latter would allow for a deﬁnition of varying degrees of local ownership through a measurement of processes of discursive inclusion and exclusion. Positive communicative indicators for local ownership would include a promotion of local discursive participation through grass-root consultations, communicative gender/minority mainstreaming, communicative empowerment of the poor and minorities, a fostering of openness and diversity of public spheres through the application of participatory methodologies and an adoption of a reﬂexive learning position of peace-building and development practitioners. Of course, this list is not comprehensive but it indicates how a communicative approach can help local actors by opening spaces for authentic participation and through reviewing the assumptions that are held by expert practitioners in the ﬁeld.

REFERENCES Guba, E.G. and Lincoln, Y. (1989) Fourth Generation Evaluation, London: Sage. Haas, E. (1958) The Uniting Europe: Political, Social and Economic Forces, 1950–1957, Stanford: Standford University Press. Habermas, J. (1976a) Vorstudien zur Theorie des kommunikativen Handelns, Frankfurt: Suhrkamp Verlag. Habermas, J. (1976b) Zur Rekonstruktion des Historischen Materialismus, Frankfurt: Suhrkamp Verlag.

62 Ann-Christin Raschdorf Habermas, J. (1981) Theorie des kommunikativen Handelns, Frankfurt: Suhrkamp Verlag. Linklater, A. (1998) The Transformation of the Political Community, Cambridge: Polity Press. Luhmann, N. (1975) Soziologische Aufklärung Bd 2, Opladen: Westdeutscher Verlag. Parsons, T. (1949) The Structure of Social Action, New York: Free Press. Schütz, A. and Luckmann, T. (1979) Strukturen der Lebenswelt, Frankfurt: Free Press.

Part II

Assessment of economic integration

4

Regional integration agreements and the geography of world trade Statistical indicators and empirical evidence Lelio Iapadre1

INTRODUCTION A process of international economic integration, propelled by strong technological and economic forces, as well as by trade liberalisation policies, characterised the second half of the twentieth century, and translated into a growth of world trade much more rapid than that of production. This phenomenon has been gradually assuming a global scope, but manifested itself with particular intensity within groups of countries tied by geographic proximity, or by historical and political factors, such as the conclusion of preferential trade agreements. Regionalism is at the centre of a growing body of theoretical and empirical research, and its implications for the multilateral trading system are still a source of controversy. At the same time, policy circles tend to adopt a pragmatic approach to the issue, aimed at recognising the strong political and institutional motivations for regional integration policies, as well as at reducing their possible economic costs (World Bank, 2000). One of the ﬁrst steps to be taken in order to assess the trade effects of regional integration agreements is to measure the actual intensity of trade among their member countries. The problem is less trivial than it could appear at ﬁrst sight, and this chapter aims to offer a contribution to its solution, by proposing and experimenting with new measurement methods. The following section contains a critical survey of the available statistical indicators for the measurement of intra-regional trade intensity, as well as some proposals which aim to overcome their limitations. There follows a discussion of some indicators that can be obtained by combining trade and GDP data, aimed at measuring the trade-creating effects of regional integration. Next, all the indicators are applied to the analysis of intra-regional trade within the four main regional integration agreements (ASEAN, European Union, Mercosur and NAFTA) in the 1990–2003 period. Some brief remarks conclude the chapter.

66 Lelio Iapadre MEASURING THE INTENSITY OF INTRA-REGIONAL TRADE In the literature on international trade, countries are often grouped in areas – or regions – that are deﬁned according to different physical, political or economic geography criteria, among which the membership of preferential trade agreements is particularly important. Trade ﬂows are deﬁned as intra-regional if both partners belong to the same region, and extra-regional if they belong to different regions. The importance of intra-regional trade (exports plus imports) is often measured by the intra-regional trade share (Si):

Si = tii /ti .

(4.1)

where: tii = region i’s intra-regional trade, ti. = region i’s total trade. At ﬁrst sight this indicator seems the most obvious choice to detect the trade effects of regional integration and is indeed widely used in empirical studies.2 However, its usefulness for both cross-region and time-series analysis is limited by some problems, which have been outlined by Anderson and Norheim (1993: 80–1), as well as by Frankel (1997: 21–5). When comparing different regions, for example, the intra-regional trade share may give misleading information, because its value is biased by the number of countries in each region and by their dimensions. Given the size of a region, as measured by its total trade, the higher the number of countries in that region, the larger its intra-regional trade share will be. In other words, splitting a region into an increasing number of countries (as happened in Central and Eastern Europe in the 1990s) raises its intra-regional trade share by transforming domestic exchange into international (intra-regional) trade. Moreover, other things being equal, a region with a high number of member countries would show a larger intra-regional trade share than a region of the same total trade size, but with a smaller number of members. The second problem is more important and subtle. In order to understand it, the concept of geographic neutrality of trade ﬂows must be introduced, deﬁned as the absence of preferential directions: the geographic distribution of a region’s trade is said to be neutral if the weight of every partner in the region’s trade is equal to its weight in world trade. If a partner is more important than what would be implied by the neutrality criterion, this reveals the presence of factors, such as common borders or regional integration agreements, that generate a preferential orientation in trade ﬂows. Using the neutrality criterion, it is easy to verify that, other things being equal, and in particular for any given number of member countries, the intraregional trade share is positively inﬂuenced by the size of the region, as measured by its total trade. In other words, even assuming that every member country’s trade is geographically neutral, larger regions would show a higher intra-regional trade share only because of their greater size, that is independently of the actual intensity of intra-regional trade.

Indicators of intra-regional trade 67 Similar arguments may be used in the case of a single region to explain why an increase in its intra-regional trade share does not necessarily imply a higher inward orientation of trade ﬂows, but may simply reﬂect a growth in the region’s relative size in world trade. Stated differently, other things being equal, the intra-regional trade share is biased by a pro-cyclical distortion. In order to eradicate these problems, one may use the trade intensity index, proposed by Brown (1949) and elaborated by Kojima (1964). In its simplest form, the intra-regional trade intensity index of the region i (Ii) is equal to the ratio between the intra-regional trade share (Si) and the region’s share in world trade (Wi):3

I i = Si /Wi = (tii / ti . )/(ti . /T )

(4.2)

where: T = world trade. This index is equal to one if the region’s weight in its own trade is equal to its weight in world trade (geographic neutrality). On the contrary, if intra-regional trade is relatively more important than trade ﬂows with the rest of the world, as is usually the case, the intra-regional trade intensity index is higher than one. It can be considered as a variant of the well-known index of revealed comparative advantages, which was proposed by Balassa (1965) in order to study trade specialisation patterns. If a region’s intra-regional trade intensity is higher than one, it can be said that the region’s trade is ‘specialised’, i.e. relatively more oriented, towards its member countries rather than towards the rest of the world. An increase of the index, revealing that the region’s importance for its own trade rises more (or falls less) than its weight in world trade, can be considered as an ex post indication of an increase in trade integration, that is a reduction of trade resistances among the region’s countries (Drysdale and Garnaut, 1982). The meaning of the intra-regional trade intensity index can be further clariﬁed using the interpretation proposed by Kunimoto (1977) for a wide class of similar indicators.4 The intra-regional trade intensity index can be seen as the ratio of the actual value of intra-regional trade ﬂows to their expected value, E(tii), under the assumption of neutrality in the regional direction of trade:

I i = tii / E(tii )

(4.3)

where: E(tii) = ti.2/T In other words, assuming that the matrix representing the geographic distribution of world trade ﬂows shows no statistical connexion between their regions of origin and destination, this would reveal the absence of preferential directions (neutrality) and would translate into values of intra- and extra-regional trade ﬂows that would be proportional to the importance of each region in world trade. These hypothetical values represent the benchmark against which the intensity of actual trade ﬂows can be evaluated. In a region, if factors of any kind foster trade relations among

68 Lelio Iapadre member countries, the value of intra-regional trade will be higher than its expected value under the neutrality assumption, which translates into an intra-regional trade intensity index higher than one. Although immune to the speciﬁc problems afﬂicting intra-regional trade shares, due to their sensitivity to the number and size of member countries, the traditional Balassa intensity index is also characterised by at least three uncomfortable features, which limit its interpretability and usefulness: • • •

range variability: the maximum value of the index is inversely related to the region’s total trade size; range asymmetry with respect to the threshold value of one; dynamic ambiguity: possible sign concordance between the changes of complementary indicators.

The following subsections will be devoted to the analysis of these problems, and point to a possible solution for each of them.

Range variability Looking at equation (4.2), it is easy to see that the actual range of values assumed by the intra-regional trade intensity index is inﬂuenced by the region’s size. In fact, whilst in the extreme case of no intra-regional ﬂows the intensity index is equal to zero for any region, its maximum value, which is reached in the opposite case when all trade is intra-regional, is inversely proportional to the region’s relative size in terms of total trade:

Max ( I i ) = T /ti .

(4.4)

In other words, the intra-regional trade intensity index ranges from zero (no intraregional trade) to a maximum value (no extra-regional trade), which is the higher, the smaller the region’s total trade. This range variability problem implies that indices computed for different regions and/or periods are not perfectly comparable among each other. At ﬁrst sight, the solution for this problem could appear to divide the intraregional trade intensity index by its maximum value. However, the result of this normalisation would simply be equal to the intra-regional trade share, arousing again the problems mentioned before. An alternative solution consists in changing the denominator of the intra-regional trade intensity index, by substituting the region’s weight in the trade of the rest of the world (Vi), which is equal to zero in the limiting case of no extra-regional trade, for the region’s weight in world trade. The result could be called homogeneous index of intra-regional trade intensity (HIi):

HI i = Si /Vi = (tii /ti . )/(t ri /t r .) where: tri = region’s i extra-regional trade, tr. = total trade of the rest of the world.

(4.5)

Indicators of intra-regional trade 69 The threshold value of this index, in the case of geographic neutrality, is equal to one, no different from its traditional Balassa formulation, but its range goes now from zero (no intra-regional trade) to infinity (no extra-regional trade), independently of the region’s trade size.

Range asymmetry The second problem of the intra-regional trade intensity index, in both the formulations discussed so far, is that its range is not symmetric around the neutrality threshold. More precisely, if intra-regional trade is lower than its expected value under the assumption of geographic neutrality, the intensity index ranges only from zero to one, whilst it goes from one to inﬁnity in the homogeneous formulation, and from one to a number which is always much higher than two in the traditional Balassa formulation, if the region’s trade reveals a preferential inward orientation. This problem may give rise to biased assessments of the index changes, depending on whether they occur above or below the neutrality threshold. In addition, it may create problems in econometric estimates involving the index. One possible solution for the asymmetry problem consists in applying to the homogeneous index the transformation proposed by Dalum et al. (1998) for the Balassa revealed comparative advantage index, which yields the following symmetric index of intra-regional trade intensity (SIi):

SI i = ( HI i − 1)/( HI i + 1) = (Si − Vi )/(Si + Vi )

(4.6)

This index ranges from minus one (no intra-regional trade) to one (no extraregional trade), and is equal to zero in the case of neutrality. 5 It is therefore a standardised transformation of the intensity index, which allows proper crossregional comparisons.

Dynamic ambiguity Each of the above-mentioned indicators for intra-regional trade can be compared with a complementary indicator, measuring the intensity of extra-regional trade. In the case of the intra-regional trade share, this complement is given by the share of extra-regional flows in the region’s total trade. It is obvious that an increase in the former implies a decrease in the latter, and vice versa. The complementary indicators of the three intensity indices discussed in the previous subsections are given by the following extra-regional trade intensity indices:

Ei = (1 − Si )/(1 − Wi )

(4.7)

HEi = (1 − Si )/(1 − Vi )

(4.8)

SEi = ( HEi − 1)/( HEi + 1)

(4.9)

70 Lelio Iapadre The value of these indicators will be the higher, the larger is the share of extraregional trade in a region’s trade, relative to the other regions’ weight in world trade, in equation (4.7), or to the intra-regional trade share of the rest of the world, taken as a single region, in equations (4.8) and (4.9). The geographic neutrality threshold is equal to one for the ﬁrst two indicators, and to zero for the third one. Unfortunately, unlike the intra- and extra-regional trade shares, all the three couples of intensity indices are afﬂicted by a common problem: the change of the intra-regional index, although having usually opposite sign than that of its complementary extra-regional index, does sometimes take the same sign, which makes it difﬁcult to interpret the dynamics of the indicators. More precisely, it can be shown that, if Ii ⫽ Ei, i.e. if the regional direction of trade is not neutral, and if the ratio between the changes of Si and Wi lies in the interval between Ii and Ei, the two complementary indices change in the same direction (condition of sign concordance). Since almost all regions are intraregionally oriented, Ii is usually higher than Ei. In this case, if:

Ei < ∆Si / ∆Wi < I i

(4.10)

then: ∆Ei ⋅ ∆I i > 0 and, more precisely, if the sign concordance condition holds, and ∆Wi > 0, then ∆Ii < 0 and ∆Ei < 0; on the contrary, if under the same condition, ∆Wi < 0, then ∆Ii > 0 and ∆Ei > 0. In other words, under condition (4.10), if a region’s trade is relatively dynamic, in the sense that its weight in world trade increases, both the intra- and the extraregional trade intensity indices decrease. On the contrary, if the region is relatively slow, the two complementary indices increase. In practice, with reference to the latter case, the increase of the intra-regional trade intensity index is due to the fact that the intra-regional trade share falls at a lower rate than the region’s weight in world trade, whilst the increase of the extra-regional trade intensity index results from a rise of the corresponding trade share which is relatively larger than the increase of the other regions’ weight in world trade. These results are reversed in the few cases where Ii < Ei, and hold also for the homogeneous and symmetric versions of the index. It should be noted that the range of values of ∆Si/∆Wi for which the complementary indices change in the same direction is equal to the difference between their levels, which means that, other things being equal, the probability of obtaining results that are dynamically ambiguous is higher when intra-regional trade intensity is either very high or very low. It is in any case difﬁcult to interpret the data, when the indices show a simultaneous increase (or fall) of both intra- and extra-regional trade intensity. Any reference to the effects of regional integration policies would be problematic, because the ﬁrst index seems to contradict the second, and vice versa. In order to solve this problem, one could make use of the ratio between the complementary indicators, which shows synthetically if the intensity of

Indicators of intra-regional trade 71 intra-regional trade is growing more or less rapidly than that of extra-regional trade. The resulting indicators of relative intra-regional trade intensity, which could be called trade introversion indices, are the following:

Ji = I i /Ei

(4.11)

HJi = HI i /HEi

(4.12)

SJi = ( HJi − 1)/( HJi + 1)

(4.13)

It is easy to see that, also in this ‘relative’ version, the homogeneous index HJi ranges from zero (no intra-regional trade) to infinity (no extra-regional trade), independently of the size of the region, and is equal to one in the threshold case of geographic neutrality. Correspondingly, its symmetric formulation SJi ranges from minus one to one, passing through the neutrality threshold of zero, and should be preferred to the others for the reasons already explained in the previous subsections.6 Moreover, a symmetric trade extroversion index (SFi) can be deﬁned as:

SFi = ( HEi /HI i − 1)/( HEi /HI i + 1)

(4.14)

and it is clear that SJi = – SFi. The dynamic ambiguity problem is absent in this case, since any change in the trade introversion index necessarily implies an opposite and equal change in the complementary extroversion index. An interesting property of the homogeneous trade introversion index is that it simultaneously measures the intensity of intra-regional trade in the target region i and in the rest of the world, taken as a single ‘complementary region’. Stated differently, if the world is divided into two regions, since, by definition, S2 = (1 – V1); V2 = (1 – S1), and vice versa, it is easy to show that:

HJ2 = [(1 − V1 )/(1 − S1 )] /(V1/S1 ) = HJ1

(4.15)

which obviously implies that SJ2 = SJ1, independently of the regions’ size. Intuitively, it is reasonable that, if the world is divided into only two regions, any level of trade introversion in one of them implies the same result in the other, with the limiting case where both regions are completely isolated from each other. The symmetric trade introversion index can therefore be proposed as the most satisfactory measure of intra-regional trade intensity. It must be stressed, however, that SJi can give only an ex post evaluation of the level of trade integration within the region, which is not enough to draw any conclusion about the strength of the underlying causal links. In other words, trade introversion indices simply measure revealed trade preferences among the region’s members, in analogy to what can be said about the Balassa specialisation index, which measures revealed comparative advantages, without pretending to give any ex ante assessment of their intensity.7

72 Lelio Iapadre THE TRADE-CREATING EFFECTS OF REGIONAL INTEGRATION: INDICATORS OF RELATIVE OPENNESS The indicators discussed in the previous section are all aimed at measuring the relative intensity of intra-regional trade with respect to the region’s trade with the rest of the world. Trade introversion can be the result of a variety of factors, including the trade diversion effects of regional integration agreements. However, an increase in trade introversion indices can also reflect an intra-regional trade creation process, without diverting effects on trade with the rest of the world. Trade introversion simply means that the region’s weight in its own trade is higher than its weight for the rest of the world, or in other words that actual intra-regional trade flows are larger than their expected level under the hypothesis of geographic neutrality. The use of the geographic neutrality criterion requires an exogenously given value of the region’s total trade. However, international integration factors affect not only the intensity of intra-regional ﬂows, as measured by trade introversion indices, but also the total value of the region’s trade. More precisely, a region’s weight in world trade (Wi) is equal to the product between the region’s share of world output (Gi) and its relative degree of openness (Oi) with respect to the world average degree of openness8 (both measured by the trade-to-GDP ratio9):

Wi = Gi ⋅ Oi = yi /Y ⋅ (ti ./yi )/(T /Y )

(4.16)

where: yi = region i’s gross domestic product (GDP), Y = world GDP. The two elements of equation (4.16) are inﬂuenced by a complex set of cyclical and structural forces.10 Speciﬁcally Gi depends on the various supply and demand factors letting the region’s gross domestic product grow faster or slower than world output. As to the relative degree of openness (Oi), its variations may be seen as reﬂecting mainly changes in resistance factors which dampen the region’s international trade as a whole (including its intra-regional component), such as those changes induced by the growing integration of national markets and by multilateral trade agreements. In comparing different regions, it should anyway be remembered that the level of trade openness can be the result of many factors, such as country size, economic structure, distance from the main markets and trade policies (in the region and abroad). The speed of integration processes may be different across the regions of the world, resulting in variations in the relative degree of openness of each region. However, changes in Oi may also be inﬂuenced by the internal and external trade creation effects of regional integration.11 Some of the problems already described for trade intensity indices also affect the relative openness indicator Oi. In order to solve them, a symmetric indicator of relative openness (SOi ) can be computed:

SOi = {(ti ./yi )/[(T − ti . )/(Y − yi )] − 1}/{(ti ./yi )/[(T − ti . )/(Y − yi )] + 1} (4.17)

Indicators of intra-regional trade 73 SOi is based on the comparison between the region’s trade-to-GDP ratio and that of the rest of the world. This index is higher than zero if the region’s degree of trade openness is higher than the world average, revealing that forces leading to market integration are stronger than in other regions, possibly also as an effect of preferential trade agreements. It can be useful to combine the measurement of intra-regional trade intensity and relative openness, in order to generate a comprehensive measure of the trade effects of regional integration. This can be done through an index of propensity to intra-regional trade (PIi) deﬁned as the intra-regional trade share (Si) divided by the region’s weight in world GDP (Gi):12

PI i = (tii /ti . )/( yi /Y )

(4.18)

It is easy to see that PIi is also equal to the product between the intra-regional trade intensity index (Ii) and the region’s relative degree of openness (Oi):

PI i = I i ⋅ Oi

(4.19)

If PIi is higher than one, the intra-regional trade share is larger than the region’s share of world output, which can be the effect of a high relative degree of openness (global trade creation) and/or of a high intra-regional trade intensity (revealed trade preferences). Adapting to this indicator the approach used by Kunimoto (1977) to interpret trade intensity indices (see equation (4.3)), it can be shown that the propensity to intra-regional trade is equal to the ratio between the actual value of intra-regional trade and its expected value, which is here proportional to the region’s weight in world GDP:

PI i = tii /E′(tii )

(4.20)

where: E′(tii ) = Gi ⋅ ti . = [( yi )2 /Y ] ⋅ (ti . / yi ) There is a clear analogy between the underlying structure of this descriptive indicator and the logic of gravity models of international trade, where the intensity of bilateral flows depends positively on the GDP of partner countries, and negatively on their ‘distance’.13 A PIi higher than one can be interpreted as a sign of a situation where, after controlling for the region’s GDP and its trade-toGDP ratio, intra-regional trade is higher than expected due to the operation of ‘proximity’ factors, such as common borders and preferential trade agreements. The two elements determining the expected value of intra-regional trade capture the effects of the region’s size and of its overall degree of trade openness, which can be related, among other things, to geographic distance and other barriers limiting the region’s international trade. Given such inﬂuences, the propensity to intra-regional trade can be directly explained by the linkages among the region’s

74 Lelio Iapadre members, although the relative importance of the various proximity factors can only be estimated with a gravity model.14 The index of intra-regional trade propensity is affected by all the problems already discussed with reference to the traditional specification of the trade intensity index. The maximum value of PIi is inversely related to the size of the region; its range is not symmetric around its neutrality threshold of one; and its changes may have the same sign as those of the complementary indicator of propensity to extra-regional trade (PEi), given by the ratio between the extraregional trade share and the rest of the world’s weight in world GDP:

PEi = (1 − Si )/(1 − Gi )

(4.21)

These problems can be solved in a way similar, but not identical, to that followed for the intra-regional trade intensity index. The difference is due to the denominator of PIi, which cannot be adjusted in the same way as that of Ii, because GDP ﬂows, unlike trade ﬂows, are not associated to a couple of partner regions, but refer only to the region where the product is produced. In order to bypass this hurdle, it is necessary to reverse the sequence followed in the construction of SIi. First, the asymmetry problem can be reduced, although not suppressed, by building a symmetric index of intra-regional trade propensity (SPIi), through a formula similar to equation (4.6):

SPI i = (PI i − 1)/(PI i + 1)

(4.22)

This index is equal to zero under the assumption of neutrality in the regional direction of trade and ranges from minus one (no intra-regional trade) to a maximum value (no extra-regional trade) which is the closer to one, the smaller the region’s weight in world GDP, biasing comparisons among different regions or periods. The second step could be to divide SPIi by its maximum value, getting a normalised index of intra-regional trade propensity (NPIi):

NPI i = SPI i /[(Y /yi − 1)/(Y /yi + 1)]

(4.23)

This index is also equal to zero under the assumption of neutrality, but its maximum value (no extra-regional trade) is equal to one independently of the region’s size. On the other hand, the comparability problem is now shifted to the minimum level of the index (no intra-regional trade), which is a negative number the closer to minus one, the larger the region’s weight in world GDP. In this case the problem is less worrisome, because the intra-regional trade share is almost always higher than the region’s weight in world GDP, so that the section of the NPIi’s range where the comparability problem matters is that of positive numbers. The last remaining problem, that is the possible sign concordance of changes in complementary indicators, has the same features as for the intensity indices. Its solution, although giving rise to a relatively complex indicator, has the advantage

Indicators of intra-regional trade 75 of solving completely the asymmetry and non-homogeneity problems which still affect both SPIi and NPIi. The first step is to compute an index of relative propensity to intra-regional trade (PJi ), given by the ratio between the two complementary indicators of intra- and extra-regional trade propensity:

PJi = PI i /PEi = (Si /Gi )/[(1 − Si )/(1 − Gi )]

(4.24)

Similarly to what is seen for the introversion indices, PJi is less ambiguous than its underlying complementary indicators because, even if their changes happen to have the same sign, it increases if a region’s propensity to intra-regional trade rises more rapidly than that to extra-regional trade, and vice versa. It is easy to see, furthermore, that the range of PJi is homogeneous: independently of the region’s size, it goes from zero (no intra-regional trade) to inﬁnity (no extra-regional trade), crossing at one the neutrality threshold. Finally, a symmetric index of relative intra-regional trade propensity (SPJi) can be obtained in the usual way:

SPJi = (PJi − 1)/(PJi + 1)

(4.25)

SPJi is equal to zero if the regional distribution of trade is neutral and ranges between minus one and one, independently of the region’s size.

RECENT TRENDS IN INTRA-REGIONAL TRADE: A COMPARISON AMONG FOUR PREFERENTIAL INTEGRATION AREAS (ASEAN, EU, MERCOSUR, NAFTA) The indicators discussed in the previous sections will now be applied to the analysis of intra-regional trade intensity in the 1990–2003 period,15 with reference to the four regional integration agreements that have the highest share of world trade: the European Union (EU), the North American Free Trade Agreement (NAFTA), the Association of South East Asia Nations (ASEAN) and the Southern Common Market (Mercosur).16 Time series of trade data have been drawn from the WTO website,17 and are measured in US dollars at current prices. Throughout the 1990–2003 period, the country composition of each region has been kept constant as it was in 2003, independently of the actual accession date of each member, in order to avoid distortions due to changes in the number of member countries.18 Figure 4.1 shows the simple intra-regional trade shares of the four regions, computed according to equation (4.1). In the case of the EU, the trend is slightly decreasing, from around 64 per cent in the early 1990s to less than 62 per cent in the past four years.19 The other three areas show a more or less pronounced upward trend in the 1990s, which in the case of Mercosur was sharply reversed during the ﬁnancial crises in Brazil and Argentina.

76 Lelio Iapadre 70% 60% 50% 40% 30% 20% 10% 0%

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

EU (15)

NAFTA (3)

ASEAN (10)

Mercosur (4)

Figure 4.1 Intra-regional trade shares 16 14 12 10 8 6 4 2 0

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

EU (15)

NAFTA (3)

ASEAN (10)

Mercosur (4)

Figure 4.2 Intra-regional trade intensity (traditional Balassa indicator)

The regions’ ranking is clearly inﬂuenced by their different trade size, as well as by the number of member countries. A completely different ranking emerges from Figure 4.2, which shows the intraregional trade intensity index, computed according to the traditional formulation of equation (4.2), resembling the Balassa index of revealed comparative advantages. The ﬁgure is dominated by Mercosur, whose intra-regional trade share was, on average, 13 times larger than the region’s weight in world trade, with an upward

Indicators of intra-regional trade 77 20

16

12

8

4

0

1990 1991 1992 1993 1994 1995 1996 1997

EU (15)

NAFTA (3)

1998 1999 2000 2001 2002 2003

ASEAN (10)

Mercosur (4)

Figure 4.3 Intra-regional trade intensity (homogeneous indicator)

trend of the intensity index, particularly strong until 1993. The other regions show much lower and more stable intensity indices. A similar picture is shown by Figure 4.3, where the intensity indices have been computed according to the homogeneous formulation of equation (4.5). The level of all the indices is obviously higher than in Figure 4.2, but the relative distances between the regions are lower, because the correction operated by equation (4.5) with respect to equation (4.2) is more pronounced for the larger regions, which lie in the bottom of the ﬁgure. Neither of the latter two figures gives a clear view of the time pattern of the indices for the three larger regions, because they are squashed by the high level of the Mercosur index. This problem is solved by the symmetric formulation of the intensity indices of equation (4.6), which has been used in Figure 4.4. Normalising the range of the index, from minus one to one, generates a much more readable ﬁgure, where all regions show signiﬁcant changes across time. As explained above, however, all the intensity indices, whatever their formulation, are exposed to a problem of dynamic ambiguity. Figure 4.5 shows this problem in the case of the European Union. In general, one expects that an increase in intra-regional trade intensity will be associated with a fall in the corresponding index of extra-regional trade, and this is what actually happened in most of the years shown in Figure 4.5. However, there are three exceptions (1993, 1997 and 2000), when both the complementary indices rose, because the condition of sign concordance (equation (4.10)) held, and the EU’s total trade grew at a lower rate than the world average. Figure 4.6 shows the indicator proposed in this chapter in order to solve the above problem, i.e. the trade introversion index, computed in its symmetric speciﬁcation (equation (4.13)). The regions’ ranking is now partly different from that

78 Lelio Iapadre 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

EU (15)

NAFTA (3)

ASEAN (10)

Mercosur (4)

Figure 4.4 Intra-regional trade intensity (symmetric indicator) –0.33

0.48 0.47 0.46

–0.34 0.45 0.44 –0.35

0.43 0.42 0.41

–0.36 0.40 0.39 0.38

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

INTRA-TRADE (left)

–0.37

EXTRA-TRADE (right)

Figure 4.5 Intra- and extra-regional trade intensity (EU-15) (symmetric indicators)

shown in the previous ﬁgures. Mercosur is conﬁrmed as the area with the highest level of intra-regional trade intensity, with an upward trend which was particularly strong in the ﬁrst three years, but was interrupted by the Argentine crisis. The other regions show similar levels of trade introversion. NAFTA is the area where trade introversion increased more, starting from relatively low levels in 1990. In the ASEAN region a fall of the index until 1995 was followed by an upsurge, particularly in the aftermath of the Asian crisis. Opposite trends appear in the European

Indicators of intra-regional trade 79 1.0

0.9

0.8

0.7

0.6 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

EU (15)

NAFTA (3)

ASEAN (10)

Mercosur (4)

Figure 4.6 Trade introversion (symmetric indicator) 0.6 0.5 0.4 0.3 0.2 0.1 0.0 –0.1 –0.2 –0.3 –0.4 –0.5 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

EU (15)

NAFTA (3)

ASEAN (10)

Mercosur (4)

Figure 4.7 Relative trade openness (symmetric indicator)

Union, where trade introversion has been slowly declining since 1995, after rising markedly in the previous years. As shown in the preceding section, trade introversion indices can be complemented with indicators of relative trade openness, aimed at capturing the total trade-creating effects of regional trade agreements and other integration factors. Figure 4.7 refers to the symmetric indicator of relative openness (equation (4.17)) and shows huge differences among the four areas, with the two American regions much less open than the rest of the world. It is a well-known property

80 Lelio Iapadre of the trade-to-GDP ratio to be inversely related to the size of the country, because the importance of the foreign markets relative to the domestic one decreases with the country’s GDP. Therefore, it is not particularly surprising that the relative trade openness of NAFTA is so low, given the GDP size of the USA. More striking is the position of Mercosur, which is probably due also to the relatively inwardlooking orientation of its member countries’ trade policies, as well as to the market access barriers faced by their exports, particularly in the agricultural sector. It is also interesting to note that no curve in Figure 4.7 shows a clear upward trend in the 1990s, which seems to reveal that the trade-creating effects of regional integration were not particularly strong, or at least not stronger than the other technological and economic factors leading to a global increase of the trade-toGDP ratio. The upsurges of the index for ASEAN in 1998 and for Mercosur in the past ﬁve years are due mainly to the sharp depreciations of the local currencies hit by the ﬁnancial crises, which translated into an increase of trade prices much higher than that of production prices. Combining the indicators of trade openness with the trade introversion indices shown in Figure 4.6, one gets the index of relative intra-regional trade propensity (equation (4.25)), which appears in Figure 4.8. This figure ranks the four regions according to the ratio between their intraregional trade share and their weight in world production. ASEAN and Mercosur show the highest levels of SPJi, but the underlying factors are completely different. In the ASEAN countries the large importance of intra-regional trade is mainly the result of their higher degree of relative openness, as shown in Figure 4.7, whilst in the case of Mercosur it stems from a particularly high level of trade introversion (Figure 4.6). 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

EU (15)

NAFTA (3)

ASEAN (10)

Mercosur (4)

Figure 4.8 Relative intra-regional trade propensity (symmetric indicator)

Indicators of intra-regional trade 81 CONCLUSIONS The ﬁgures presented in this chapter show clearly that the empirical assessment of intra-regional trade is strongly inﬂuenced by the choice of the statistical instrument used to measure its importance. The simplest indicators, such as the intra-regional trade share and the trade intensity index, suffer from several limitations, impairing their ability to evaluate the dynamics of the process in a single region, as well as to compare different regions. In this chapter, some alternative indicators have been proposed and examined. One of them in particular, the symmetric trade introversion index, seems able to solve the main problems of traditional indicators, and offers a completely different picture of the regional geography of trade patterns. The highest degree of trade introversion is shown by a region, such as Mercosur, whose process of preferential integration began relatively late with respect to other regions, and was not particularly deep. This feature is common to other developing regions, not covered by this chapter, and points to one of the possible interpretations of the index: a high trade introversion could be the outcome of reciprocal protectionism between the region and the rest of the world. It is well-known that trade policies in developing countries are more inward-looking than in developed countries, and that, at the same time, their trade opportunities are severely limited by protectionist policies in the rest of the world. This could help to explain why the trade introversion index of Mercosur appears so high in relation to other regions. However, all the four regions considered in this chapter are characterised by a moderate upward trend in their trade introversion indices, albeit with different timing. In the EU, Mercosur and NAFTA the increase was relatively stronger in the ﬁrst half of the 1990s. On the contrary, in the ASEAN region the upward trend began in 1996, after a phase of decline in the previous years. It would be difﬁcult to ascribe this common trend to an increase in protectionism, given the prevailing open orientation of trade policies in the 1990s. A possible explanation could be found in the trade effects of the preferences among members of regional integration agreements. A useful complement to the trade introversion index can be found in an indicator of relative trade openness, which is based on the comparison between a region’s trade-to-GDP ratio and the other regions’ average. This index shows if integration forces are stronger at the regional level than on a global scale, and can therefore be interpreted as a measure of revealed internal and external trade creation. ASEAN was shown to be the most open region, while Mercosur and NAFTA were largely below the world average. However, in the 1990s no region exhibited a clear increasing trend in its relative openness, which could indicate that regional tradecreating forces were not stronger than the global drift towards an increase in the trade-to-GDP ratio. The index of intra-regional trade propensity combines the measurement of trade introversion and relative openness. It can be seen as the ratio between the actual and the expected value of intra-regional trade, the latter being based on the size of

82 Lelio Iapadre the region’s GDP and on its trade openness. There is here a clear analogy with the logic of gravity models of international trade. For opposite reasons, ASEAN and Mercosur were the regions with the highest intra-regional trade propensity in the 1990s. However, the simple descriptive indicators proposed in this chapter should be interpreted with caution. For example, an increase in relative trade openness, as measured by equation (4.17), does not necessarily mean that trade policies have become more open in the region than in the rest of the world. It only shows that the trade-to-GDP ratio has grown more rapidly than the world average, which can be the result of many factors other than trade policies, such as changes in economic structure, the international fragmentation of production and the complex linkages between migrations, capital ﬂows and trade. Further research is needed in order to assess the causal relationships underlying the trade effects of regional integration. Moreover, indicators similar to those described in this chapter should be applied to the study of the geographic direction of foreign direct investment, in order to explore its links with trade integration. However, a deeper understanding of the relation between trade and investment in the context of regional integration requires the availability of comparable and detailed data on firms’ international activities, based on a new approach to the production of statistics on globalisation, centred on the enterprises more than on their transactions, which is emerging in several countries. Although these new developments are potentially very interesting, the current state of the available sources is not yet adequate to address the needs.

NOTES 1 The author would like to thank Michel Fouquin and Maurice Kugler for their very useful comments on a previous version of this chapter. The usual disclaimer applies. 2 See, for example, WTO (1995: 38–41). 3 Since no country can trade with itself, the denominator of the index should be corrected by subtracting from the region’s total trade, as well as from world trade, one n-th of the region’s total trade (where n is the number of countries in the region), as shown by Anderson and Norheim (1993: 82, Note 8). This correction ensures that the index is approximately equal to unity, if the geographic orientation of the region’s trade is not inward biased. The more similar are the trade values of member countries, the lower is the approximation error. Frankel (1997: 25–29) calls the non-corrected intensity indices ‘simple concentration ratios’. We will not use the word ‘concentration’ in this context, because it could be confused with its usual deﬁnition, which refers to the entire distribution of a region’s trade, rather than to the importance of a single element, such as intra-regional trade. The Anderson and Norheim correction is important for comparing trade intensity levels of different regions, but may be neglected if the interest is focused on the time path of intra-regional trade intensity in a single region. On the other hand, this correction does not solve other problems raised by the fact that no country can trade with itself, which were highlighted by Savage and Deutsch (1960). A more rigorous correction procedure can be found in Freudenberg et al. (1998).

Indicators of intra-regional trade 83 4 Bowen (1983) criticised trade intensity indices, arguing that in Kunimoto’s interpretation they imply a hypothetical world where every good is exported and imported by every country, which would be inconsistent with standard trade theories. Bowen’s critique applies particularly to revealed comparative advantage indices, such as that proposed by Balassa (1965), and is questionable for several reasons, including those mentioned by Vollrath (1991). Anyway, it could not be applied to the intra-regional trade intensity index, because this index does not refer to the commodity distribution of trade ﬂows. A world without geographic preferences, in which every country, although not necessarily exporting every good, trades with each partner in proportion to its importance in world trade, appears to be a reasonable yardstick for actual trade ﬂows. 5 Similar properties are shown by the hyperbolic tangent of the natural logarithm of the intensity index, proposed by Jungmittag et al. (1998) as a substitute for the Balassa formulation. 6 It is also easy to verify that Ji happens to be equal to HIi. 7 The problems discussed so far for intra-regional trade intensity indices affect also the measurement of revealed comparative advantages given by the Balassa index. Our trade introversion formula can easily be adapted to this indicator, in order to get a better measure of the intensity of trade specialisation. 8 See Anderson and Norheim (1993: 81), who use the same decomposition in the denominator of the traditional trade intensity index (equation (4.2)). In this case, corrections similar to those described in Note 3 above should be applied to the GDP data to get an unbiased measure of trade intensity. 9 The trade-to-GDP ratio is an imperfect measure of trade openness for a variety of reasons. First of all, it should be borne in mind that trade ﬂows are measured in terms of gross output (including the value of intermediate goods), whilst GDP is expressed in terms of value added. This problem cannot be easily solved because data on the value of gross domestic production is not readily available. Second, since GDP includes the services sector, trade in goods and services should be used in the numerator. However, for the purposes of this chapter, only trade in goods can be taken into account, because data on the geographic distribution of trade in services is difﬁcult to ﬁnd. Third, independently of these statistical problems, the trade-to-GDP ratio should be used very cautiously in cross-country comparisons, because it is inversely related with country size (see Anderson and Norheim 1993: 80, Note 3). Other things being equal, larger countries tend to show lower trade-to-GDP ratios, only because they face a smaller ratio between foreign and domestic markets. 10 It must be stressed that all the variables are inevitably measured at current prices, given the non-availability of regional trade data at constant prices. As a consequence of this, all the indicators are also affected by inﬂation differentials and exchange rate ﬂuctuations. The latter tend to inﬂuence trade prices more than domestic prices, which can create problems in the interpretation of the indices, particularly after currency crises, as we will see in the section on Recent trends in intra-regional trade, p. 80. 11 It must also be remembered that Gi and Oi are not completely independent of each other. For example, other things being equal, if the region’s GDP is growing faster than world output, capacity constraints could, in principle, have the effect that the pressure of domestic demand in the region makes it easier to purchase goods from external sources, thus raising the region’s relative degree of openness. 12 Similar indicators are discussed by Anderson and Norheim (1993: 84) as well as by Frankel (1997: 27). 13 The relationship between trade intensity indices and gravity models of international trade is analysed in Leamer and Stern (1970), Drysdale and Garnaut (1982) and Frankel (1997). Gaulier (2003) shows how trade intensity indices can be related to the gravity model proposed by Deardorff (1998). 14 A recent survey of the empirical evidence on the trade and investment effects of

84 Lelio Iapadre

15 16 17 18

19

regional integration, based on gravity models, can be found in Adams et al. (2003). Traditional trade intensity indices have been recently used in the context of a gravity model by Gaulier, Jean and Ünal-Kesenci (2004) and by Zhang and van Witteloostuijn (2004). Trade intensity indices for previous periods can be found, for example, in Frankel (1997: 26–29). Many other preferential trade agreements exist, but the sum of their intra-regional trade ﬂows did not exceed 1 per cent of world trade (WTO, 1999: 20). See http: //www.wto.org/english/res_e/statis_e/its2004_e/appendix_e/a03.xls and previous editions. The EU is therefore taken with 15 members (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden and the United Kingdom). NAFTA includes Canada, Mexico and the United States. ASEAN includes Brunei Darussalam, Cambodia, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand and Vietnam. Members of Mercosur are Argentina, Brazil, Paraguay and Uruguay. The fall in 1993, which seems paradoxical in the completion year of the Single Market programme, is mainly a statistical artefact, due to the introduction of a new collection system for intra-Community trade data (Intrastat), which led to a considerable underevaluation of intra-regional ﬂows (Iapadre, 1996).

REFERENCES Adams, R., Dee, P., Gali, J. and McGuire, G. (2003) ‘The Trade and Investment Effects of Preferential Trading Arrangements – Old and New Evidence’, Productivity Commission Staff Working Papers, Canberra. Anderson, K. and Norheim, H. (1993) ‘From Imperial to Regional Trade Preferences: Its Effect on Europe’s Intra- and Extra-regional Trade’, Weltwirtschaftliches Archiv, 129(1): 78–102. Balassa, B. (1965) ‘Trade Liberalisation and “Revealed” Comparative Advantage’, The Manchester School of Economic and Social Studies, 33(2): 99–123. Bowen, H.P. (1983) ‘On the Theoretical Interpretation of Indices of Trade Intensity and Revealed Comparative Advantage’, Weltwirtschaftliches Archiv, 119(3): 464–72. Brown, A.J. (1949) Applied Economics: Aspects of the World Economy in War and Peace, London: George Allen and Unwin. Dalum, B., Laursen, K. and Villumsen, G. (1998) ‘Structural Change in OECD Export Specialisation Patterns: De-specialisation and “Stickiness”’, International Review of Applied Economics, 12(3): 423–43. Deardorff, A. (1998) ‘Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World?’, in J. A. Frankel (ed.) The Regionalisation of the World Economy, Chicago: The University of Chicago Press. Drysdale, P.D. and Garnaut, R. (1982) ‘Trade Intensities and the Analysis of Bilateral Trade Flows in a Many-country World: A Survey’, Hitotsubashi Journal of Economics, 22(2): 62–84. Frankel, J.A. (1997) Regional Trading Blocs in the World Economic System, Washington: Institute for International Economics. Freudenberg, M., Gaulier, G. and Ünal-Kesenci, D. (1998) ‘La régionalistion du commerce international: une évaluation par les intensités bilatérales’, CEPII Working Paper, 98–05, Paris: Centre d’Etudes Prospectives et d’Informations Internationales. Online. Available HTTP: .

Indicators of intra-regional trade 85 Gaulier, G. (2003) ‘Trade and Convergence: Revisiting Ben-David’, CEPII Working Paper, 2003–06, Paris: Centre d’Etudes Prospectives et d’Informations Internationales. Online. Available HTTP: . Gaulier, G., Jean, S. and Ünal-Kesenci, D. (2004) ‘Regionalism and the Regionalisation of International Trade’, CEPII Working Paper, 2004–16, Paris: Centre d’Etudes Prospectives et d’Informations Internationales. Online. Available HTTP: . Iapadre, L. (1996) ‘The Intrastat System from a User’s Perspective’, in Eurostat, European Seminar on Intra-Community Trade Statistics, Proceedings of the Intrastat seminar held in Luxembourg on 13 and 14 March 1996, Luxembourg: European Commission, 305–15. Jungmittag, A., Grupp, A. and Hullmann, A. (1998) ‘Changing Patterns of Specialisation in Global High Technology Markets: An Empirical Investigation of Advanced Countries’, Quarterly Journal of Economic Research, 2: 86–98. Kojima, K. (1964) ‘The Pattern of International Trade Among Advanced Countries’, Hitotsubashi Journal of Economics, 5(1): 16–36. Kunimoto, K. (1977) ‘Typology of Trade Intensity Indices’, Hitotsubashi Journal of Economics, 17: 15–32. Leamer, E.E. and Stern, R.M. (1970) Quantitative International Economics, Chicago: Aldine. Savage, I.R. and Deutsch, K.W. (1960) ‘A Statistical Model of the Gross Analysis of Transaction Flows’, Econometrica, 28: 551–72. Vollrath, T.L. (1991) ‘A Theoretical Evaluation of Alternative Trade Intensity Measures of Revealed Comparative Advantage’, Weltwirtschaftliches Archiv, 127: 265–80. World Bank (2000) Trade blocs, A World Bank Policy Research Report, Oxford: Oxford University Press. WTO (1995) Regionalism and the world trading system, Geneva: World Trade Organisation. WTO (1999) Annual Report 1999, Geneva: World Trade Organisation. Zhang, J. and van Witteloostuijn, A. (2004) ‘Economic Openness and Trade Linkages of China: An Empirical Study of the Determinants of Chinese Trade Intensities from 1993 to 1999’, Review of World Economics, 140(2): 254–81.

5

Explaining regionalisation via relative trade intensities Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci1

Trading with a close neighbour is easier than trading with a distant partner, for a wide variety of reasons, ranging from transport costs to information asymmetries. As a consequence, it is no surprise that the bulk of world trade occurs between countries belonging to the same region (whatever the precise meaning given to this notion). Still, the surge in the number of regional trade agreements (208 in force notiﬁed to the WTO up to May 2004) has been spectacular during the past 15 years. And the link between this rising regionalism (i.e. institutional agreements across neighbouring countries) and the intensity of regionalisation in international trade (i.e. regional polarisation of trade ﬂows) raises many questions. What is the ‘normal’ level of trade between partners (and in particular neighbours), and what is its ‘normal’ evolution across time? To what extent is trade within a given region more intense than elsewhere? And if it is, does this result from trade agreements or not? Addressing this issue is a complex matter, given the intricacy of foreign trade relationships, and the complexity of their determinants. This chapter primarily updates and extends the work of Freudenberg et al. (1998), based upon the CHELEM-CEPII database, covering international trade worldwide, on a consistent and harmonised basis, for each year from 1967 to 2001. In order to characterise more accurately the extent of regionalisation, this work makes extensive use of an indicator of the geographical orientation of foreign trade, namely the CEPII’s index of bilateral, relative intensity of trade flows. These relative trade intensities (RTIs) are not only used for descriptive purposes, they are also employed for econometric estimates. The most salient particularity of these estimates is to control for country-pair ﬁxed effects when assessing the impact of regional trade arrangements (RTAs) on trade ﬂows. In other words, the estimates only focus on the relative intensity of the trade relationship between any country-pair, compared to its mean level across the period. Working exclusively in comparison to the mean level of trade intensity, for any country-pair, makes it possible to control for any pairwise-speciﬁc country links constant across the period considered, such as geographical devices (distance, natural obstacles to transportation, contiguity, etc.), language, former cultural and historical ties (colonial, for example), etc. This is a significant improvement in methodological terms, as recently pointed out by Cheng and Wall (2003), even compared to one-way ﬁxed effects, which control for country-speciﬁc variables.

Relative trade intensities 87 However, it has never been carried out before on a worldwide basis for such a long time span, to the best of our knowledge, except by Freudenberg et al. (1998) and Fontagné et al. (1999) and Cheng and Wall (2003), with trade level as explained variable in the latter two cases. Introducing these country-pair fixed effects is crucial when assessing the impact of RTAs, which tend to be signed by countries sharing some kind of proximity link (due to geography, politics or culture). These questions will be tackled in three stages. A first goal is to size up the issues. Different indexes are considered to take an overview of international trade regionalisation. The institutional context is examined, and used to further qualify empirically the regionalisation of international trade. In a second stage, these observed patterns of regionalisation are studied using a gravity model. Trade relative intensities are explained through proximity or gravity factors like physical distance, regional agreements, differences in specialisation and income. The ﬁnal section concludes.

AN OVERVIEW OF REGIONALISATION OF INTERNATIONAL TRADE The fact that international trade flows are more intense between neighbours naturally leads to a general pattern where a large part of trade ﬂows takes place within geographical regions. In this section, the pattern of regionalisation of international trade ﬂows is ﬁrst glanced at empirically with regard to the Triad regions, defined here as America, Asia-Oceania and Eurafrica.2 This general pattern is then compared with an overview of the institutional context, namely the rapid development of RTAs during the past decade.

Regionalisation at ﬁrst glance Since distance is an obstacle to trade, it does not come as a surprise that countries trade more intensively with their neighbours. As a consequence, the regionalisation of foreign trade is a natural pattern, in the sense that countries tend to trade in large part with other countries belonging to the same ‘region’. Of course, there is no straightforward, unique definition of the notion of region. However, following Freudenberg et al. (1998), this regional polarisation of trade is nicely illustrated in a triangle representing the polarisation of each country’s foreign trade across the Triad regions (Figure 5.1). This diagram summarises within a triangle the share of each country’s foreign trade (exports + imports) carried out with a partner belonging to each of the three main geographical regions, namely America, AsiaOceania and Eurafrica. Only 10 out of 80 countries fall within the equilateral inscribed triangle, meaning that 70 out of 80 countries have more than half their foreign trade concentrated within a single Triad region. The 10 remaining countries are mainly large countries, such as the US, Japan, China, India, Brazil and the EU15 (excluding intra-EU trade). It is noteworthy that world trade appears very evenly balanced across the Triad regions when intra-EU trade is not accounted for.

88 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci Eurafrica

50%

America

World

50%

50%

Asia-Oceania

Figure 5.1 The polarisation of foreign trade with regard to the Triad, by country, 2001 Source: Authors’ calculations, based on CHELEM-CEPII database. Note: Countries are labelled by their ISO code, see Appendix 5.1. In this diagram, for each country, the distance from each of the three triangle summits is inversely proportionate to the share of the corresponding Triad region in the country’s foreign trade (exports + imports). In other words, the point representing a given country is the barycentre of the system formed by the points representing each of three Triad regions, weighted by the share of this region in the country’s foreign trade. ‘W × UE’ refers to world trade, excluding intra-EU trade. ‘EU-15’ refers, to the foreign trade of the EU-15, excluding intra-EU trade.

Regional polarisation is especially strong in Eurafrica. The region accounts for more than 75 per cent of foreign trade in the majority of Western European countries, as well as in CIS countries; in Central and Eastern European countries (CEECs) and Baltic states, trade outside the region generally does not exceed 10 per cent of the total. In comparison, the regional orientation of foreign trade appears rather limited in America. This is notably due to the rather intense trade relationships maintained with European countries (and with Asia as far as Chile and Peru are concerned). Still, except for Brazil, Latin American countries trade in majority within the American region, with a particularly strong polarisation towards the US neighbours, Mexico and Canada. Asia-Oceania appears as the region exhibiting the weakest polarisation. However, except for the three largest economies in the region (Japan, China, South Korea), intra-regional trade accounts for around 55 to 60 per cent of total trade.

Relative trade intensities 89

The CEPII’s index of relative trade intensity Due to widespread differences in size and openness, the extent of trade flows’ regionalisation cannot be established by simply comparing trade shares. Instead, this section makes use of CEPII’s index of bilateral, relative intensity of trade ﬂows. This index is intended to measure ‘trade proximity’ of partner countries, by comparing observed bilateral flows (as a percentage of world trade) with theoretical ﬂows (the product of total relative weights in world trade) that reﬂect the overall capacity of the partners to trade. The index thus eliminates size effects, in order to reveal what is speciﬁcally linked to the bilateral relationship. In practical terms, relative trade intensities (RTI) are the ratios of the observed trade ﬂows to ‘natural’ ﬂows (Deutsch and Savage, 1960; Drysdale and Garnaut, 1982). The latter are determined by the geographical distribution of world trade according to the relative importance of exporters and importers respectively:

Vi , j V ⋅V Vw RTI = = i ,j w Vi ⋅ Vj Vi ⋅ Vj Vw

(5.1)

2

where Vi,j stands for trade between country i and country j, Vi for total trade of country i, Vw for total world trade. Additional corrections are made to this indicator, in order to make sure that it is neutral with respect to the country size (for a detailed description of these changes, see Freudenberg et al. (1998)). RTI has two important properties. First, it is independent from the size and openness of the two partners. Second, it is purely relative, since it focuses on the comparison of trade ﬂows of each country across its different partners, without taking into account the absolute level of ﬂows. For illustrative purposes, the pairwise combinations exhibiting the highest RTI (and at least a value of three) are displayed in Tables 5.1 to 5.3. Table 5.1 conﬁrms that trade regionalisation has reached a fairly high level in Latin America. This is particularly true within the Andean Pact and within the Mercosur. While geographical factors are obviously the main explanation, this might also result from the institutional efforts witnessed during the last two decades (see below). Even though the absolute magnitude of trade is far higher, RTIs are found to be rather limited in North America, slightly below ﬁve for the relationships of the USA with its two neighbours. These tables conﬁrm the higher level of regional integration of trade in Europe. The relative intensities are especially high for former communist countries, reﬂecting both a survival of the former COMECON trade dependencies, and the importance of the EU to them as an export market. This also results from the former integration of these new countries in a national entity (USSR, Yugoslavia) which has split up. This reﬂects the proximity between small economies, which

Ecuador Peru Venezuela Brazil Chile Peru Peru Ecuador Other America Venezuela Venezuela Peru Kazakstan Chile Colombia Other America

48.3 38.4 29.2 27.5 27.1 22.4 19.9 12.3 11.4 10.8 10.1 8.3 8.0 7.6 7.0 5.8

2001 15.8 28.5 –3.6 2.0 10.6 9.2 –6.3 1.1 2.8 2.4 –3.4 1.6 7.6 –0.5 1.9 1.7

Chg. 95–01

Relative trade intensity

Argentina Brazil Ecuador Argentina Brazil Mexico Brazil Canada Chile Brazil Peru Brazil Brazil Peru Brazil Ecuador

Partners

Other America Nigeria Other America Ecuador Peru USA Venezuela USA Venezuela Other America Other America Algeria Colombia Nigeria Latvia Serbia Monte

5.6 5.4 5.3 5.0 4.8 4.7 4.6 4.4 4.1 4.0 3.8 3.7 3.6 3.2 3.1 3.1

2001 –0.2 2.6 1.4 0.3 –0.8 –0.7 –0.4 –0.5 0.6 –2.5 –0.3 2.0 1.1 3.2 2.7 –3.8

Chg. 95–01

Relative trade intensity

Note: Partner pairs are ranked by decreasing order of relative trade intensity, among pairs including at least one country belonging to America. Column ‘Chg. 95–01’ refers to the absolute change of the RTI between 1995 and 2001 (value in 2001 minus value in 1995).

Source: Authors’ calculations, based on CHELEM-CEPII database.

Colombia Ecuador Colombia Argentina Argentina Chile Colombia Chile Venezuela Ecuador Peru Argentina Other America Brazil Chile Columbia

Partners

Table 5.1 Top relative trade intensities in America

Macedonia Bosnia-Herzego. Bosnia-Herzego. Latvia Kyrgyzstan Albania Belarus Greece Ukraine Bosnia-Herzego. Estonia Croatia Kazakstan Belarus Belarus Caucasus Romania Albania Bulgaria Estonia Estonia Kazakstan Bulgaria Czech Republic Belarus Bulgaria Bosnia-Herzego. Slovenia

Partners

Serbia Mont. Croatia Serbia Mont. Lithuania Other CIS Greece Lithuania Macedonia Other CIS Slovenia Latvia Slovenia Kyrgyzstan Ukraine Latvia Other CIS Brunei Macedonia Macedonia Lithuania Finland Other CIS Greece Slovakia Other CIS Serbia Mont. Macedonia Macedonia

286.5 247.2 151.9 104.5 100.6 99.7 78.6 70.5 68.1 67.0 64.5 61.5 58.0 57.0 56.2 55.9 50.4 44.7 43.3 35.4 35.4 34.5 32.4 31.5 31.0 29.6 29.5 29.3

2001 –134.1 17.1 –1176.1 25.3 –69.8 16.2 16.3 65.2 13.1 17.5 –9.1 –11.0 –117.3 21.0 35.4 19.9 –2.0 –32.1 –41.0 –7.7 –1.7 –20.2 7.2 –40.2 11.6 –54.5 14.7 –8.3

Chg. 95–01

Relative trade intensity

Table 5.2 Top relative trade intensities in Eurafrica

Norway Belarus Portugal Ukraine Romania Estonia Denmark Iceland Gulf Greece Bulgaria Romania Albania Lithuania Albania Ukraine Finland Russian Fed. Nigeria South Africa Gabon Belarus Austria Algeria Estonia Hungary Kazakstan Denmark

Partners

Sweden Kazakstan Spain Caucasus Other CIS Sweden Sweden Norway Pakistan Romania Turkey Serbia Mont. Belarus Ukraine Croatia Macedonia Sweden Caucasus Other Africa Other Africa Other Africa Caucasus Slovenia Turkey Russian Fed. Slovakia Other Europe Norway

11.6 11.4 11.1 11.0 10.5 10.5 10.5 10.0 10.0 10.0 9.9 9.6 9.5 9.4 9.2 9.0 8.9 8.9 8.7 8.6 8.5 8.4 8.4 8.2 8.1 8.0 8.0 7.9

2001

continued . . .

1.7 –4.6 1.8 –8.1 –1.0 1.7 2.2 0.1 2.9 4.0 1.3 –17.3 9.5 –9.2 1.0 2.4 0.5 –1.0 3.7 –1.9 4.3 4.8 1.4 2.0 –2.0 –1.6 7.6 0.3

Chg. 95–01

Relative trade intensity

Serbia Mont. Ukraine Taiwan Russian Fed. Estonia Caucasus Other Europe Kyrgyzstan Iceland Other CIS Macedonia Ukraine Serbia Mont. Caucasus Serbia Mont. Caucasus Ukraine Poland Romania Serbia Mont. Russian Fed. Italy Bulgaria Poland

28.6 28.2 27.9 26.6 21.7 20.8 20.7 20.5 19.3 19.2 19.0 19.0 18.5 16.4 16.4 16.3 15.2 14.4 13.7 13.1 12.6 12.3 12.0 11.7

2001 26.2 –6.2 27.9 –18.7 9.5 7.1 7.4 12.1 1.5 –3.3 –7.3 9.8 15.7 –11.0 6.3 7.5 4.4 6.1 6.2 –21.9 –3.2 2.6 –5.7 3.8

Chg. 95–01

Relative trade intensity

Turkey Kyrgyzstan Hungary Bulgaria Iceland Poland Czech Republic Poland Latvia Nigeria Estonia Albania Austria Turkey Croatia Finland Bosnia-Herzego. Latvia Egypt Ireland Austria Austria Finland Gulf

Partners

Non-OPEC MEast Russian Fed. Romania Russian Fed. Poland Slovakia Poland Ukraine Sweden India Ukraine Turkey Hungary Other CIS Non-OPEC MEast Russian Fed. Belarus Russian Fed. Non-OPEC MEast United Kingdom Croatia Slovakia Latvia Non-OPEC MEast

7.6 7.6 7.5 7.5 7.4 7.3 7.2 7.2 7.0 6.9 6.9 6.6 6.5 6.5 6.5 6.5 6.5 6.4 6.4 6.3 6.2 6.2 6.1 6.1

2001 –0.2 –9.2 –1.1 –3.4 7.0 0.7 –0.3 0.1 –2.0 3.6 –0.5 –1.4 –1.9 –0.2 4.8 2.0 4.5 –6.2 1.0 –0.1 1.5 1.6 –1.3 –1.1

Chg. 95–01

Relative trade intensity

Note: Partner-pairs are ranked by decreasing order of relative trade intensity, among pairs including at least one country belonging to Eurafrica. In order to save space, only the country-pairs exhibiting an RTI above 6 in 2001 are displayed. Column ‘Chg. 95–01’ refers to the absolute change of the RTI between 1995 and 2001 (value in 2001 minus value in 1995).

Source: Authors’ calculations, based on CHELEM-CEPII database.

Croatia Russian Fed. Kyrgyzstan Kazakstan Belarus Turkey Greece Belarus Denmark Russian Fed. Croatia Bulgaria Greece Bulgaria Slovenia Kazakstan Kazakstan Belarus Bulgaria Albania Lithuania Albania Albania Lithuania

Partners

Table 5.2 continued

Romania Kyrgyzstan New Zealand Camb-Lao-Vietn. Singapore Gulf Singapore Thailand Singapore Nigeria Thailand Asia-Oceania nes Malaysia Asia-Oceania nes Hong Kong Camb-Lao-Vietn. Other Gulf Other Asia-Oceania Other Africa Other Asia-Oceania Singapore Taiwan Camb-Lao-Vietn. Japan Other Africa Philippines

50.4 27.9 21.0 16.3 12.3 10.0 7.7 7.4 7.4 6.9 6.7 6.7 6.2 5.9 5.8 5.3 5.3 5.1 5.1 4.8 4.8 4.6 4.4 4.3 4.3 4.2

2001 –2.0 27.9 –0.5 –10.8 1.2 2.9 –1.0 1.6 3.2 3.6 –0.5 0.0 1.9 1.3 –0.5 –0.3 –4.4 0.1 0.6 –0.7 0.7 0.8 1.9 0.6 0.1 1.9

Chg. 95–01

Relative trade intensity

Thailand Malaysia Thailand Singapore Taiwan Australia Australia Korea Indonesia Indonesia Philippines Pakistan India Camb-Lao-Vietn. Japan Philippines China China Malaysia Japan Singapore Brunei Korea India Brunei

Partners

Other Asia-Oceania Thailand Non-OPEC MEast Thailand Camb-Lao-Vietn. Indonesia Brunei Gulf Korea Japan Singapore Non-OPEC MEast Gulf Belarus Philippines Thailand Korea Japan Philippines Thailand Other Europe New Zealand Camb-Lao-Vietn. Other Africa Korea

4.2 4.1 4.1 4.0 4.0 3.9 3.8 3.7 3.7 3.5 3.5 3.3 3.3 3.3 3.2 3.2 3.2 3.1 3.1 3.1 3.1 3.1 3.0 3.0 3.0

2001 1.6 1.1 2.7 –0.3 0.4 0.7 2.7 1.3 0.8 –0.1 1.1 0.2 –1.1 2.7 0.3 1.0 0.3 0.0 1.4 –0.1 1.0 2.3 –0.3 0.0 0.3

Chg. 95–01

Relative trade intensity

Source: Authors’ calculations, based on CHELEM-CEPII database. Note: Partner-pairs are ranked by decreasing order of relative trade intensity, among pairs including at least one country belonging to Asia-Oceania. Column ‘Chg. 95–01’ refers to the absolute change of the RTI between 1995 and 2001 (value in 2001 minus value in 1995).

Brunei Taiwan Australia Brunei Malaysia Pakistan Brunei Brunei Indonesia India Camb-Lao-Vietn. India Brunei Pakistan China Singapore Other Asia-Oceania New Zealand Brunei Australia Hong Kong Hong Kong Australia Brunei Pakistan Taiwan

Partners

Table 5.3 Top relative trade intensities in Asia-Oceania

94 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci ‘naturally’ leads to high relative intensities of trade. Noteworthily, while steep declines in RTI are experienced for several pairs of partners among former communist countries, many of them experienced an increase between 1995 and 2001. Several neighbourhood relationships are also apparent in these most intense relative trade relationships, in particular between Scandinavian countries, between Spain and Portugal, between UK and Ireland, or between African countries. Although far larger in absolute value, RTI is weaker for relationships like Germany–France (RTI = 1.5 in 2001), Germany–the Netherlands (1.9), France– Spain (3.6) or even Germany–Austria (3.8). In any case, the values obtained for largely open countries, such as European ones, vary across a more limited range than those obtained for less open countries, such as Central Asian ones. In Asia-Oceania, outside some distinctive trade relationships concerning mainly oil trade (in particular those involving Brunei), a few pairs are remarkable for the intensity of their reciprocal trade, in particular Australia–New Zealand, Malaysia–Singapore, Indonesia–Singapore, China–Hong Kong and Hong Kong– Taiwan. Among the main pairs of countries, most exhibit a RTI superior to 3, including those between Japan, China, Korea and Indonesia (with the exception of Japan–Korea, 2.4, and China–Indonesia, 1.4). Although it has been emphasised above that Asian countries make a signiﬁcant part of their trade (and in particular their exports) outside the region, this illustrates the intensity of their trade links.

Institutional context: regional trade arrangements and their inﬂuence on trade ﬂows As mentioned above, regional polarisation is certainly, at least partly, a natural pattern. But it also has a strong institutional background. It has become usual to distinguish three waves of regionalism (see e.g. Adams et al. (2003)). The first wave dates back to the 1950s, when the economic integration of Western Europe began, followed by regional agreements in Africa and in Latin America. The second wave began in the mid-1980s, with the European Single Market (1986), the Canada–US FTA, followed by the NAFTA, and a number of subregional arrangements. The third wave followed the creation of the WTO, and was characterised by the proliferation of FTAs. Figure 5.2 spectacularly illustrates the recent surge in RTAs, in particular since 1991. Up to May 2004, 208 agreements notiﬁed to the WTO were in force. This ﬁgure also points to the overwhelming importance of the Eurafrica region in this trend. This has in particular been the case since 1991, mainly as a result of the very large number of agreements signed by former communist countries, either CEECs or Former Soviet Union countries, and more recently former Yugoslavia countries. However, it is worth noting that trade agreements signed between members formerly belonging to the same country can hardly be considered as progress toward economic integration, but rather as an accompanying measure to political disintegration. Outside Europe, a number of bilateral agreements have been signed, but they are mainly linked to the intense activity in this domain of a handful of

Relative trade intensities 95 250 World

Number of RTAs

200

Eurafrica 150

100

50

America Asia-Oceania

0 58 61 64 67 70 73 76 79 82 85 88 91 94 97 00 03 Year

Figure 5.2 Breakdown of Regional Trade Agreements by geographical areaa (in force and notiﬁed to the GATT/WTOb) Source: WTO, authors’ calculations. Notes a World ﬁgures are inferior to the sum of the RTAs by region: when a trade agreement involves two or three regions, the agreement is taken into account in each region. b As of 13 October 2003.

countries, namely Mexico, Canada and Chile in America, Singapore and Australia in Asia-Oceania.

ANALYSING REGIONALISATION BY BILATERAL TRADE INTENSITIES The fact that countries trade more intensively with their neighbours does not come as a surprise, given the well-known negative impact of distance on trade relationships. Nor should the fact that the USA’s neighbours have strong trade links with the USA be counterintuitive, given the relative sizes. Although prima facie evidence is useful to characterise the extent and intensity of trade ﬂows, this means that trade flows are bound to be more intense between neighbours, and within geographical regions. Characterising properly a possible regionalisation phenomenon thus requires going beyond the mere observation of trade intensities. Speaking of regionalisation would only make sense as soon as trade intensities within a region are higher than expected, once ‘natural’ determinants of trade, such as distance, size, cultural and geographical ties, and income per capita have been accounted for. This requires estimating gravity equations. This gravity-based, econometric study of regionalisation is carried out in this section. The analysis focuses on characterising the extent of any regionalisation phenomenon, over and above natural determinants. The general approach follows

96 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci Freudenberg et al. (1998). As in this chapter, the relative trade intensity is explained by several trade resistance variables. Compared to the 1998 paper, more preferential trade agreements are taken into account.

The gravity model Assuming perfect specialisation,3 strictly positive barriers to trade and CobbDouglas utility, as in Deardorff (1998), the f.o.b. trade ﬂow between country i and j can be written as:

Vi , j =

Yi ⋅ Yj ti , j ⋅ Yw

(5.2)

where Yi is country i GDP, Yw is world GDP and ti,j is a transaction cost factor (one plus transaction cost) of Samuelson’s ‘iceberg’ form. The transaction cost can include different aspects: distance, other transport costs, cultural ties, and market accessibility, both in terms of formal and informal obstacles, of quality of infrastructure cost or of access to information. Naturally, trade agreements ought to be among the determinants of this transaction cost, insofar as reducing it is precisely their purpose. In this rather general framework, taking the logarithm of equation (5.2) provides us with a theoretical grounding of the standard gravity equation:

ln Vi , j = ln Yi + ln Yj − ln ti , j − ln Yw

(5.3)

For the purposes of our study, it is useful to note that a general form can also be derived, as far as the relative trade intensity deﬁned in equation (5.1) is concerned. Indeed, under the set-up used by Deardorff (1998),4 it follows from equation (5.2) that:

ln(RTI i , j ) = − ln(ti , j ) − ln( MPi ) − ln( MPj )

(5.4)

where

MPi = ∑ c

yc tic

denotes the market potential of country i, and index c refers to partners. yc is country c’s share in total world GDP

( yc =

Yc ). ∑ Yn n

Relative trade intensities 97 This variable is already commonly introduced in gravity equations, and belongs to what Anderson and van Wincoop (2003) call ‘multilateral resistance factors’. Gravity equations used to be estimated for trade ﬂows. Equation (5.4) shows that estimating a gravity equation for RTIs also makes sense. As shown by Gaulier (2004), the advantage of this speciﬁcation is that, except for market potential terms, only purely bilateral transaction costs enter the equation. It is thus likely to allow for a better control of multilateral transaction costs.

RTAs as determinants of trade ﬂows Based on the theoretical foundations laid out above, a gravity model is estimated. The basic model is expressed in logarithms. It has the bilateral trade ﬂow (exports plus imports) between a couple of partners as the dependent variable, and includes the following independent variables: •

relative geographic distance between partners. Based on equations (5.3) and (5.4), it is written as ln(RelativeGeoDist) = ln(di,j) + ln(MPGeoDisti) + ln(MP GeoDistj), where di,j is the geographic distance5 between country i and j and y MPGeoDisti = ∑ c ; c dic

•

dummy variables indicating adjacency, common language, former colonial tie, former common coloniser; GDP of each partner; relative economic distance between partners:

• •

EcoDisti , j =

Max{PCI i , PCI j } Min{PCI i , PCI j }

where PCIi is country i per capita income.

MPEcoDisti = ∑ c

yc EcoDistic

ln(Relative EcoDisti,j) = ln(EcoDisti,j) + ln(MPEcoDisti) + ln(MPEcoDistj); •

•

relative similarity of trade specialisation between partners. As for geographical or economic distance, this term includes the bilateral similarity between partners, but also the multilateral similarity index of each partner (see Appendix 5.2 for the deﬁnition of the indicators); dummy variables indicating RTA membership of one partner, as well as joint membership.

This model is rather standard for most variables, already extensively used in the literature. Note, however, that, based on equation (5.4), each partner’s market potential is not considered separately, but incorporated in the measurement of the relative distance. This is more consistent with the model laid out above. It means

98 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci that market potential can be interpreted as a benchmark, against which absolute distances are compared: relative distance thus refers to distance between two partners, in comparison to (weighted) average distance from each partner to the rest of the world, as measured through market potential. The relative similarity of trade specialisation is the only unusual variable here, although it has already been used in Freudenberg et al. (1998). The bilateral index measures how similar two partners’ sectoral specialisations are, as measured by revealed comparative advantages, through the index of ‘contribution to the trade balance’ (Lafay, 1990). It is arguably difficult to export a given product to a country that has a comparative advantage in this product, since competition from local producers is tough in this case. As a result, a high similarity of trade specialisations implies that strong export sectors from each partner will be faced with tough local competition from the other partner (since this sector will, on average, also correspond to a comparative advantage of the partner). Relative similarity is thus expected to have a negative inﬂuence on the magnitude of trade ﬂows. Based on equation (5.4), the model is estimated with bilateral RTI as the independent variable. As already outlined, this speciﬁcation allows the estimation to focus on purely bilateral resistance factors, thus getting rid of any disturbance linked to other variables. It is also worth emphasising that such estimates have a different meaning from those based on trade ﬂow levels. RTIs exclusively refer to the geographic orientation of foreign trade, and no information can be drawn as to the volume of trade. This implies, in particular, that potential trade diversion effect of a given RTA cannot be studied using RTIs: the results will only refer to the difference between the trade creation effect between members of the RTA, and the trade creation effect with third countries. Based on the CHELEM-CEPII database, the estimates are carried out on an annual basis across the period 1967–2001. The sample covers the whole world, with some countries grouped together, according to the CHELEM-CEPII geographical classiﬁcation (see Appendix 5.1). As a result, 58 countries are considered. CHELEM-CEPII has the unique device of putting together exhaustive, harmonised and consistent time-series describing worldwide trade over a relatively long period. These devices (and in particular the time-consistency) make it possible to carry out the estimation using two-way ﬁxed effects (actually, the within estimator is reported). In other words, the estimates only focus on the volume of trade between any country-pair, compared to its mean level across the period. The advantage of this approach is that it makes it possible to control for trade determinants speciﬁc to any country-pair, which are constant across the time span considered. Although distance, contiguity, common language and colonial ties are taken into account in the model, this is undoubtedly far from being enough to control for the speciﬁc ties existing between a pair of countries. As pointed out by Haveman and Hummels (1998), this gives rise to an omitted variable bias. This has already been controlled for by introducing country fixed effects (see e.g. Matyas (1997), or Adams et al. (2003)), and it was found to improve substantially the adequacy of the model. As argued by Cheng and Wall (2003), this is only a special case, where strong restrictions to the nature of heterogeneity of trade ﬂows

Relative trade intensities 99 are imposed. A more general model should be used when estimating gravity equations, allowing for a more general form of heterogeneity. Since specific, bilateral ties cannot be satisfactorily controlled for, they argue that standard methods for estimating gravity models of trade suffer from estimation bias due to omitted or misspeciﬁed variables. The problem can be addressed by using the twoway ﬁxed-effects model in which country-pair and period dummies are used to reﬂect the bilateral relationship between trading partners. Results by Freudenberg et al. (1998) or Cheng and Wall (2003) suggest that standard gravity estimates of the effects of integration can differ a great deal from what is obtained when heterogeneity is not accounted for. Introducing country-pair fixed effects is thus a significant improvement. To the best of our knowledge, except Freudenberg et al. (1998), Fontagné et al. (1999) and Cheng and Wall (2003), there is no example of estimating a fixed-effects gravity model on a worldwide basis for such a long time span.6 Given the exhaustive nature of our country sample, a ﬁxed-effect model is clearly better suited than a random-effect model. With 58 countries considered over 35 years, the theoretical number of observations would be 58 × 57 × 35/2 = 57,855. However, the number of non-zero observations is 55,880. A Heckman two-stage estimation procedure was performed to account for the information provided by these zero flows, but this did not signiﬁcantly alter the results. An OLS estimate carried out by pooling yearly data and including time dummies gives results in line with the literature (Table 5.4, column a). The fit (Adjusted R2 0.36) is rather low compared to usual estimates. This is due to relative trade intensity replacing trade level as left-hand-side variable: the very good ﬁts obtained in usual gravity estimates come from GDPs and other variables measuring the size of trade partners. The distance coefﬁcient is relatively close to unity, as expected. Common language, common former coloniser, contiguity and, to a lesser extent, former colonial tie have a strong positive impact on trade ﬂows. Economic distance has a very low, yet signiﬁcant, positive impact. Relative similarity in sectoral trade specialisation has a negative and signiﬁcant impact on trade ﬂows. This conﬁrms the relevance of taking into account the degree of complementarity between sector specialisations. Countries with similar specialisations seem to be more competitors than customers, one to the other. A very strong preferential impact on trade between members is found for COMECON, Mercosur, the Andean Community, and to a lesser extent CER. Coefficients for Mercosur and CAN are huge: for the latter it would imply a multiplication by 8.7 (exp(2.16)) of trade between Andean Community countries. Such impact is unlikely to stem from the agreement itself, other factors (correlated with the agreement in the cross-section dimension) must play a role. Results for the second speciﬁcation will conﬁrm this intuition. A more limited impact is found for EU, EFTA, EU-EFTA and Euromed agreements. Except NAFTA, which may be difficult to identify due to lack of points (only three countries), all preferential trade agreements (PTA) have high positive impact on relative trade intensity.

100 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci Table 5.4 Estimating a gravity model for RTIs, 1967–2001 Dependent variable (ln) Speciﬁcation Time-independent variables Contiguity Common language Common former coloniser Former colonial tie

RTI OLS (a)

RTI Within (b)

0.63 0.45 0.24 0.83

(0.03) (0.02) (0.04) (0.04)

Time-speciﬁc variables other than RTAs ln GDP ln (relative) distance –0.88 ln (relative) similarity –0.92 ln (relative) economic distance 0.05

(0.01) (0.01) (0.01)

0.37 –0.38 –0.28

(0.04) (0.02) (0.01)

Dummy for both partners belonging to the RTA EU-EFTA 0.54 (0.04) Euromed 0.28 (0.04) Israel–US–Canada FTA 0.60 (0.29) NAFTA 0.10 (0.24) EU 0.49 (0.03) EFTA 1.24 (0.05) Comecon 1.64 (0.27) Mercosur 1.93 (0.42) ASEAN 0.63 (0.10) CER 1.67 (0.32) CAN 2.16 (0.18)

0.04 0.29 0.07 0.05 0.21 0.30 0.57 0.98 0.07 0.36 0.96

(0.05) (0.03) (0.23) (0.18) (0.04) (0.07) (0.30) (0.31) (0.07) (0.28) (0.13)

Root MSE N Adj R2 F value

1.33 55 880 0.36 1677

0.83 55 880 0.03 120

Source: Authors’ calculations, based on CHELEM-CEPII database for international trade ﬂows and GDPs. Note: In column (b) (within estimator), the dependent variable is the difference of RTI to the average throughout the period, for the corresponding country-pair. Standard errors are reported in parentheses. All coefficients are statistically significant at the 5 per cent level, except those in italics. Dummy variables denoting RTA’s membership are year-speciﬁc, starting from the year of enforcement of the RTA.

Using our preferred specification, the within model (column b), turns out to deliver substantially different results. Variables without time dimension are removed as the country-pair ﬁxed effects capture the cross-section information. Other resistance variables are strongly affected. Relative geographical distance appears with a positive coefﬁcient: it is now a measure of market potential (see formulae above), only affected by change in the ‘centrality’ of countries. If neighbouring countries experience a relative rise in GDP, it has a positive impact on relative trade intensity. The sign for economic distance is reversed to negative. It means that per capita GDP convergence favours bilateral trade. The sign for

Relative trade intensities 101 similarity remains negative. The positive impact of complementarity seems to be robust. With the ﬁxed-effects speciﬁcation, PTA impacts results are much more heterogeneous: Israel–US–Canada FTA, EU-EFTA, NAFTA and ASEAN have coefﬁcients inferior to 0.1 and non-signiﬁcant. Other coefﬁcients remain signiﬁcant but with lower value or turn non-signiﬁcant with still relatively high point estimates (COMECON, CER). The explanation for such low within estimates may lie in the imperfect ﬁt of the estimates, and in the difﬁculty of identifying the impact on RTI when a shock concerns a large part of a country’s trade, as is the case for European countries with the EU, and North American countries with the NAFTA. Consider, for instance, Austria entering the EU: this will boost its trade with Italy, but also with all other EU members. As a consequence, the numerator of the RTI will increase, but so too will the denominator. At the end of the day, the change in the index will be moderate, making the impact more difﬁcult to identify. This problem is likely to be blurring the estimation for EU-EFTA and NAFTA, as well as for the EU. The corresponding coefﬁcients should therefore be interpreted carefully. However, generally speaking, fixed effects allow us to eliminate disturbing timeless bilateral ties correlated with PTA. By using only the time dimension, they should provide more reliable estimates of the impact of PTAs. For instance, ASEAN is found to have an insigniﬁcant effect on trade between its members. This suggests that ASEAN members did not increase trade with other member countries by more than they did with third countries: ASEAN did not (signiﬁcantly) increase the polarisation of foreign trade of member countries between each other. Trade in Asia like in other regions is regionalised, but institutionalised preferences do not seem to have signiﬁcantly distorted the geographical orientation of trade so far. In contrast, agreements concerning European countries had a moderate but significant impact on European flows, and agreements in South America (Mercosur and CAN) had very high impact (a coefficient of 0.98 implies a multiplier of 2.7 for trade). The magnitude of this effect is surprising. However, a free-trade area between countries which are highly protected, as are on average most South-American countries, delivers a high preference margin to each member in the market of each other member. This might explain, at least in part, the strong effect found.

CONCLUSION Regionalisation is an important device of trade across the world, both in terms of trade patterns and in terms of institutions. The link between both aspects is not straightforward to establish, since trade flows are influenced by a bunch of ties between countries, which are not easily controlled for. Beyond the thorough description of the nature and evolution of regionalisation, this chapter proposed an empirical analysis for a large number of countries covering the whole world over the period 1967–2001, based on the CHELEM-CEPII

102 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci database. This analysis updates and extends the one carried out in Freudenberg et al. (1998). The most salient particularity of our estimations is probably to control for country-pair ﬁxed effects when assessing the impact of RTAs on trade ﬂows. By only using changes across time to identify the impacts, this method makes it possible to control for any country-pair tie constant across the period under study. In so doing, a number of problems potentially blurring the econometric analysis are avoided. The drawback is that identifying the impact of regional agreements is more difﬁcult in some cases, in particular for large agreements, such as the EU and the NAFTA. Using this method has a strong influence on the results. In particular, the distorting impact of RTAs on the geographical polarisation of member countries’ foreign trade appears to be far more limited than is found based on the OLS estimations. The impact of regional agreements is clear as far as Western Europe or Mercosur are concerned. In many cases, however, regionalism might in fact have followed or accompanied regionalisation, rather than caused it.

ALB DZA ARG AUS AUT BLR AAA BIH BRA BRN BGR CAN CHL CHN COL HRV CZE DNK ECU EGY EST FIN AAW AAF

*Albania Algeria Argentina Australia Austria *Belarus Belgium–Luxembourg *Bosnia and Herzegovina Brazil Brunei Darussalam *Bulgaria Canada Chile China Colombia *Croatia *Czech Republic Denmark Ecuador Egypt *Estonia Finland Former USSR Former Yugoslavia

France Gabon Germany Greece Hong Kong * Hungary Iceland India Indonesia Ireland Israel Italy Japan *Kazakhstan *Kyrgyzstan *Latvia *Lithuania *Macedonia Malaysia Mexico Morocco Netherlands New Zealand Nigeria

Country name FRA GAB DEU GRC HKG HUN ISL IND IDN IRL ISR ITA JPN KAZ KGZ LVA LTU MKD MYS MEX MAR NLD NZL NGA

ISO Norway Pakistan Peru Philippines *Poland Portugal *Romania *Russian Federation *Serbia & Montenegro Singapore *Slovakia *Slovenia South Korea South African Union Spain Sweden Switzerland Taiwan Thailand Tunisia Turkey *Ukraine United Kingdom United States of America Venezuela

Country name

NOR PAK PER PHL POL PRT ROM RUS YUG SGP SVK SVN KOR AAI ESP SWE CHE TWN THA TUN TUR UKR GBR USA VEN

ISO

Notes The geographical classiﬁcation refers to CEPII-CHELEM database. * Countries for which data is available only since 1993. The ‘Eastern European’ countries, former Yugoslavia and former USSR are considered as single zones in the model for the 1967–2001 period.

ISO

Country name

APPENDIX 5.1 Geographical classiﬁcation

104 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci APPENDIX 5.2

Similarity of Specialisation Patterns International specialisation of countries is measured by the ‘contribution to the trade balance’ (CTB) indicator (Lafay, 1990). Unlike other indicators of specialisation, the CTB is a symmetrical indicator in the sense that it focuses not only on exports but also on imports. CTB compares observed trade balance for a product to a theoretical trade balance corresponding to an absence of specialisation. The latter is calculated so as to spread the global trade balance on the different products according to their respective weights in the country total trade.

⎡ ⎛ 1000 ⎞ ⎢ k k k k CTB = ⎜ ⎟ ⎢ X i − Mi − ∑ X i − Mi ⎝ GDPi ⎠ ⎢ k ⎣

(

k i

)

(

⎛ k k ⎜ X i + Mi k k ⎜ ⎜ ∑ X i + Mi ⎝ k

) (

)

⎞⎤ ⎟⎥ ⎟⎥ ⎟⎥ ⎠⎦

with i the country, k the product, X are the exports and M the imports. A positive contribution is interpreted as a revealed comparative advantage. By definition, the sum over all products is zero. In CHELEM database we get 72 categories of products. The ‘contribution to the trade balance’ (CTB) indicator is used to evaluate the similarity of specialisation patterns between pairs of countries. Two steps are needed to transform the CTB indicator into a similarity index. We ~ ﬁrst compute adjusted CTB, (CT B) , in order to eliminate the size effect (degree of specialisation) included in the CTB: CTB are multiplied by a coefﬁcient so that the sum of adjusted values equals 100 for positive contributions and –100 for negative contributions. Then, for each pair of countries, we add up absolute differences of adjusted CTB. The similarity will equal 100, if the two countries have the same specialisation pattern (possibly with different intensities). If each comparative advantage for country i is matched by an equal disadvantage for country j then the similarity will be 0. The similarity of specialisation patterns between country i and j, Simij, is deﬁned as follows:

Simij = 100 −

1 ˜ − CTB ˜ ∑ CTB ik jk 4 k

An index of multilateral dissimilarity is then computed as:

MultiSimi = ∑ j

yj . Simij

Relative trade intensities 105 This index is the equivalent for similarity of what market potential is for distance. It measures how the position of the country in terms of structure of specialisation, compared to all other countries, inﬂuences its global propensity to trade.

NOTES 1

2 3

4 5 6

This chapter is based on a study that beneﬁted from ﬁnancial support from the Directorate General for Trade of the European Commission. We are grateful in particular to Edouard Bourcieu, Andrea Rossi and Vincent Aussilloux for helpful comments and suggestions. The usual disclaimer applies. A large geographical area formed by Europe, the former USSR, Africa and the MiddleEast. Deardorff (1998) refers to the Heckscher-Ohlin Model with impeded trade but he mentions that his results would hold in an H–O model with frictionless trade if factor endowments differ sufﬁciently to yield perfect specialisation, in a Ricardian model, in an Armington model or in any monopolistic-competition model where differentiation implies specialisation. As in Deardorff, it is assumed that Vw = Yw. As measured based on the methodology proposed by Head and Mayer (2002). Cheng and Wall (2003) estimate an equation with country-pair ﬁxed effects on a data set including upper-middle and high-income countries (following World Bank) covering four years. They end up with 3188 observations.

REFERENCES Adams, R., Dee, P., Gali, J. and McGuire, G. (2003) ‘The Trade and Investment Effects of Preferential Trading Arrangements – Old and New Evidence’, Staff Commission Paper, Australia: Productivity Commission. Anderson, J.E. and van Wincoop, E. (2003) ‘Gravity with Gravitas: A Solution to the Border Puzzle’, American Economic Review, 93(1): 170–92. Cheng, I.-H. and Wall, H.J. (2003) ‘Controlling for Heterogeneity in Gravity Models of Trade and Integration’, Federal Reserve Bank of St. Louis Working Paper Series, 99–010D, revised May 2003. Deardorff, A.V. (1998) ‘Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World?’, in J.A. Frankel (ed.), The Regionalization of the World Economy, Chicago: The University of Chicago Press. Deutsch, K.W. and Savage, I.R. (1960) ‘A Statistical Model of the Gross Analysis of Transaction Flows’, Econometrica, 28(3). Drysdale, P. and Garnaut, R. (1982) ‘Trade Intensities and the Analysis of Bilateral Trade Flows in a Many-Country World: a Survey’, Hitotsubashi Journal of Economics, 22(2). Fontagné, L., Freudenberg, M. and Pajot, M. (1999) ‘Le potentiel d’échanges entre l’Union européenne et les PECO: un réexamen’, Revue Economique, 50(6): 1139–68. Freudenberg, M., Gaulier, G. and Ünal-Kesenci, D. (1998) ‘La régionalistion du commerce international: une évaluation par les intensités bilatérales’, CEPII Working Paper, 98–05, Paris, Centre d’Etudes Prospectives et d’Informations Internationales. Online. Available HTTP: .

106 Guillaume Gaulier, Sébastien Jean and Deniz Ünal-Kesenci Gaulier, G. (2004) ‘Trade and Convergence: Revisiting Ben-David’, mimeo. Haveman, J. and Hummels, D. (1998) ‘Trade creation and trade diversion: new empirical results’, Journal of Transnational Management Development, 3(2): 47–72. Head, K. and Mayer, T. (2002) ‘Illusory Border Effects: Distance Mismeasurement Inﬂates Estimates of Home Bias in Trade’, CEPII Working Paper, 2002–01, Paris, Centre d’Etudes Prospectives et d’Informations Internationales. Online. Available HTTP: . Lafay, G. (1990) ‘La mesure des avantages comparatifs révélés’, Economie prospective internationale, 41; reprinted as ‘The Measurement of Revealed Comparative Advantages’, in M.G. Dagenais and P.-A. Muet (1994) International Trade Modelling, London: Chapman & Hall. Matyas, L. (1997) ‘Proper econometric specification of the gravity model’, The World Economy, 21(3): 363–68.

6

Assessing the impact of regional trade agreements An analytical framework for rulemaking in trade and investment Stephen Woolcock

INTRODUCTION The aim of this chapter is to present an analytical framework for making qualitative assessments of the impact of regional or free trade agreements. It argues that the current wave of preferential agreements necessitates augmenting the existing theoretical and analytical frameworks based on customs union theory, optimal tariff theory or rational choice type of political economy approaches. The rulemaking (or deeper integration) that is involved in the new wave of preferential agreements necessitates study of the detailed content of agreements, which has been eschewed by many of the more model-based approaches to assessing regional trade agreements (RTAs). Furthermore, as rulemaking invariably touches on ‘governance’ issues, in that trade or investment rules limit domestic regulatory autonomy, there is a need to be able to more easily identify how regional or free trade agreements impinge upon domestic regulatory policy autonomy. This chapter provides a taxonomy of the key elements of trade and investment agreements that is applicable to most issues covered by RTAs and all aspects of rulemaking. Breaking down the provisions in any agreement facilitates a more effective comparison of RTAs both with one another and with the established multilateral rules. In this instance the multilateral rules used as reference are those of the World Trade Organisation, but the analytical framework could be equally applied to other international rules. The framework also enables a more systemic assessment of whether and how preferential agreements go beyond the WTO provisions. It therefore helps to assess whether and in which areas preferential agreements are simply implementing agreed multilateral principles and where they constitute a threat to existing rules or could pose difﬁculties for future multilateral rulemaking. It shows that some aspects of rulemaking are likely to embody stronger (regional) preferences than others. For example, most rules on transparency generally result in little or no preference because the information is readily available to parties in third countries. On the other hand, speciﬁc regulatory norms are more likely to constitute a preference. In this way the framework can help in assessing the trade and investment effects of deeper integration.

108 Stephen Woolcock THE NATURE OF THE ‘NEW WAVE’ OF REGIONAL AGREEMENTS1 The current wave of regional agreements or ‘new regionalism’ is characterised by a number of features (Hettne, 1996). First, there has been a signiﬁcant increase in the number of agreements either concluded or in the process of being negotiated. The World Trade Organisation (WTO) estimates that the number of RTAs in force grew from 50 in 1990 to over 300 in 2005, with a further 60 under negotiation (World Bank, 2005). Second, participation in RTAs has now become more or less universal. Up to the mid-1990s most RTAs were limited to Europe and a few involving the United States, but by 2004 almost all countries had become involved in RTAs including countries like China, Japan and India. Whereas most Asian countries eschewed RTAs until the late 1990s, by 2004 the region was one of the most active in terms of RTA negotiations. Third, RTAs now include a growing volume of rulemaking that touches on behind the border issues. Forth, trade agreements are now taking the form of intra-regional agreements, such as between the USA and Singapore or the EU and Chile as well as agreements among countries in a contiguous region.

ASSESSING THE IMPACT There is no consensus on the impact of this new wave of RTAs. Some see it as part of a competitive process of liberalisation, compatible with multilateralism, others see it as creating a ‘spaghetti bowl’ of inconsistent rules of norms (Bhagwati et al., 1999). The well-established and very rich literature on preferential trade agreements provides a partial answer to the question of the impact. This literature is based on the Vinerian model of trade creation and diversion. Although it provides the main body of economic research on preferential agreements there are a number of limitations with this approach. First, there is the problem of determining the counterfactual, in other words what might have happened in the absence of a preferential agreement. Second, the static analysis of the trade creation and diversion effects of RTAs has been ambiguous or has shown that the effects are small, which has raised questions about the motivations behind RTAs if the benefits are so limited. Third, the standard discussion of RTAs proceeds as if tariffs were the only barrier to trade (Winters, 1996). This may still be a reasonable approximation for many of the free trade agreements (FTAs) between developing countries today, but not the newer type of RTA involving developed economies. Even RTAs between developing countries are beginning to include some elements of rulemaking or deeper integration. In terms of the economic effects of deeper integration, the broad view among trade economists is that discriminatory deep integration is unlikely to be harmful except in the sense of the opportunity cost of foregoing the greater gains from nondiscriminatory integration (ibid Winters, 1996: 7; Schott, 2004: 13). In the 1980s, a number of approaches were developed that incorporated assessments of the

Analytical framework for rulemaking 109 dynamic effects of RTAs. Baldwin, for example, showed the medium-term beneﬁts that could be gained from increased competition, resulting in increasing returns on investment, increased investment and thus a higher growth trajectory (Baldwin, 1989: 247–81; CEPR, 1996). There have been some similar assessments of the growth effects of RTAs that increase market size. Adding investment to the assessment of RTAs raises the question of what impact rules on investment, including both liberalisation/access rules and investment protection, will have on investment ﬂows. As there is some evidence that some developing countries have signed RTAs with industrialised countries in order to attract foreign investment (World Bank, 2005), it would seem important to include the effects of RTAs on investment in any analysis. But this is something that is not included in customs union theory. There is a significant body of literature on the systemic effects of RTAs. Krugman initiated a series of articles with his 1991 study on the potential systemic effects of the progressive expansion of RTAs on optimal tariffs (Krugman, 1991: 7–42; 1992). This approach is again based on tariffs and neglects other elements of RTAs. Various other contributions followed Krugman’s and developed the literature, but this literature was characterised more by the elegance of the models than what it can tell us of the impact of the current wave of RTAs.2 There have also been studies that look at the issue of bloc formation and argue, for example, that a region can increase its welfare by integrating when other regions continue to follow the principle of most favoured nation status (Frankel, Stein and Wei, 1995: 61–95). There are also ‘negotiated tariff models’ that seek to identify the ‘discount rate’ at which blocs are indifferent between defecting (from most favoured nation (MFN) and the multilateral system) and cooperation. In other words, these seek to identify at which point regional integration means the countries will cease to have an incentive to cooperate. These models also discuss whether the welfare loss from retaliation by other blocs is enough to keep large blocs in check.3 These approaches all address interesting questions but their ability to provide unambiguous or comprehensive answers to the question of the impact of the current wave of RTAs is doubtful. The models are based on a number of assumptions which may not hold in the current climate. First, they assume rational unitary actors. In other words, that national governments are inﬂuenced in their decisions solely by cost beneﬁt analysis. Second, the models are invariably tariff-based and seldom if ever include non-tariff issues, investment or rulemaking issues. Third, the models tend to discount the impact of institutions. This means that they cannot take account of the regulatory harmonisation or regulatory emulation that is arguably one of the factors driving recent RTAs. The structure of RTAs appears to be shaped by existing RTAs and by existing multilateral rules. A quick look at the RTAs concluded by Mexico with its central and southern American neighbours will show that these are more or less carbon copies of the NAFTA. The EU’s agreements with its near neighbours are shaped by the existing EU regulatory regime. Finally, all RTAs have been shaped by rules and approaches developed in the GATT, OECD or WTO (OECD 2002).

110 Stephen Woolcock Bhagwati has called for a dynamic time path analysis approach to RTAs that seeks to ask whether increasing membership of RTAs will ultimately lead to world welfare increases equivalent to what might have been achieved through multilateral liberalisation, or whether RTAs will result in lower world welfare (Bhagwati et al., 1999: 72). Bhagwati has also alerted us to the danger of regional hegemons using their negotiating leverage in bilateral negotiations to achieve relative gains at the expense of smaller trading countries (Bhagwati, 1994). Such an approach could also result in competing regional approaches, a plethora of regional and bilateral agreements and a ‘spaghetti bowl’ of rules that imposes costs for traders and investors and leads to global welfare losses. While the thesis seems to chime with some of the current strategies pursued by the major trading countries, these views on the role of RTAs in the area of rulemaking are generalisations not based on any broad analysis of the speciﬁc RTA rules. Bhagwati appears to base his spaghetti bowl analogy on the complex rules of origin that have emerged as a concomitant of preferential agreements. One can certainly argue that different preferential rules of origin represent a costly complication for world trade and investment, but this does not (yet) mean that the same is the case for other areas of rulemaking or deeper integration covering investment, service sectors, technical barriers to trade, etc. Even in the case of rules of origin, recent work suggests that the North American and PanEuro models are finding wider and wider application so that rather than many different rules of origin there will be perhaps four or ﬁve. In addition to the trade economists there has also been work on the issue of compatibility of RTAs with the letter and spirit of the multilateral order as embodied in the GATT Articles XXIV and GATS Art. V. These are the exceptions for preferential agreements provided for in the GATT and General Agreement on Trade in Services (GATS). Assessing the consistency of RTAs with the GATT rules has been an important aspect of the WTO’s work (World Trade Organisation, 1995). Indeed, a Committee on Regional Trade Agreements (CRTA) was established in 1995 for this purpose.4 Work in the CRTA has, however, been difﬁcult because of the well-known ambiguities in the GATT wording with regard to RTAs, in particular the requirements that RTAs should cover substantially all trade and that an RTA should not result in an increase in the general incidence of protection. Although there has been some progress in defining the general incidence of protection and substantially all trade for duties, the question of how to apply these GATT and GATS rules to regulatory or rulemaking aspects of RTAs remains largely unanswered. Article XXIV refers to duties (tariffs) and other regulatory restrictions to commerce, but it is not clear yet how this should be applied in the case when diverse national rules are replaced by a single rule for the whole RTA. The need to address these questions has meant that the WTO and lawyers following the debate have begun to look in some detail at the substance of the RTAs.5

Analytical framework for rulemaking 111 THE NEED FOR AN ANALYTICAL FRAMEWORK In other words, while the existing work on RTAs provides a rich source of knowledge, there has been a tendency to focus on tariff-based models or models per se rather than assess the substance of the agreements themselves. In the international organisations such as the WTO and OECD6 work has focused more on the substantive provisions of agreements, but here the vested interests of the governments shape research and output, with the result that there has been little progress in the WTO.7 Given the growing number and detail of the RTAs it is understandable that researchers should seek to avoid detailed assessment of the agreements, but the task of assessing the agreements can be facilitated by an analytical framework similar to the one proposed here. In addition, many RTAs adopt similar if not identical rules. For example, the structure and detailed content of the NAFTA is replicated in numerous FTAs in the western hemisphere and, through FTAs with the USA, also in other regions. Detailed analysis of the substance of the RTAs will enable an assessment of whether they are consistent with an open, multilateral trade and investment system or not. By breaking down the RTAs, and in particular the rulemaking aspects of RTAs, into their component parts it is possible to assess their impact. For example, if the areas in which RTAs go beyond the WTO concern sectoral coverage only then any preference is subject to erosion through future multilateral or other preferential agreements. Although the preference may result in trade or investment diversion in the short to medium term it is unlikely to constitute more of a threat to the multilateral order than tariff preferences. If an RTA goes beyond the WTO in terms of transparency, this is unlikely to constitute a challenge to the system because there is seldom much if any preference in transparency rules, which enhance the flow of information and tend to promote regulatory best practice. If, on the other hand, RTAs include speciﬁc regulatory norms or rules they may be less benign. Rules that go beyond the WTO or other existing international standards may result in divergence between regional rules, which could make agreement on future international rules harder.

THE FRAMEWORK Table 6.1 summarises the key elements in any rulemaking exercise, the likely impact of such provisions, the WTO approach to rulemaking in the relevant area and form of preference that may result from regional or bilateral agreements including such deep integration/rulemaking provisions.

Helps guard against regulatory capture by national interests

Some agreements also require (c) and (d)

Facilitates compliance with national rules

(b) Notiﬁcation of secondary/ implementing regulation

Third parties have no right to participate in consultations or decision-making process

Third parties not provided with information General transparency provisions in all areas, but not always effectively implemented

Promotes best practice in regulation

(a) Notiﬁcation of relevant legislation

Transparency of rules and procedures

No discrimination between third party suppliers

Third parties do not beneﬁt from non-discrimination WTO rules embody national treatment and MFN principles but with signiﬁcant exceptions for sectors

Precludes discrimination against foreign suppliers

(a) National treatment

Principles

(b) Most favoured nation status

Preserves regional preference when applied

No provision on preferential rules of origin

Stringent rules preserve preference

Analogous to tariff preference

(a) Limit beneﬁts to regional suppliers or goods

Generally full coverage of legislation but less of secondary regulatory instruments

Rules of origin

(d) Regulatory instruments covered

(c) Type of entity/subject covered (e.g. central, state, local, independent regulator or private entities)

Central government covered and some other levels depending on the issues area

Third parties do not beneﬁt from the greater WTO-plus coverage of sectors, entities or regulatory instruments

Use of positive and negative lists

More extensive coverage implies greater ‘liberalisation’

(a) Coverage of policy area

Coverage

(b) Sector schedules

Nature of preference

Provision in WTO

Likely impact

Typical provisions

Rule element

Table 6.1 Proposal for an analytical framework for rulemaking in trade and investment

Institutional provisions

Substantive measures

(a) Common decisionmaking machinery (b) Intergovernmental institutions to oversee agreement (c) Specialist committees (i.e. for speciﬁc policy areas) (d) Technical cooperation and capacity building

(e) Mutual recognition of regulations or test results

(d) Equivalence

(c) Approximation as a general aim

(b) Partial harmonisation

(a) Harmonisation

(e) Obligation on regulator to respond to submissions

(d) Right of access to administrative procedures

(c) Opportunity for all parties to comment

Helps less developed economies develop best practice

Helps identify regulatory barriers before these create disputes

Promotes convergence on rules and/or best practice

(d) and (e) Reduce costs whilst retaining regulatory autonomy

(a) – (c ) Eliminate or reduce ‘frictional ‘costs

Limited resources for technical assistance

continued . . .

Third parties do not beneﬁt from technical assistance provided under RTA General provisions for cooperation and technical assistance Specialist committees in key policy areas

Third parties not involved in more intensive cooperation within the RTA

Preference for partners in mutual recognition

Preference for regional norms or rules over international rules

WTO intergovernmental organisation with large membership

Encourages but does not require mutual recognition or equivalence

Selective harmonisation

(a) Tight limitations on the use of safeguard measures (b) General exceptions permitting discrimination

(a) Legal persons have standing in regional/ bilateral disputes and rules ‘direct effect’

Safeguard

Implementation

(d) Regional dispute settlement

(c) Remedies (e.g. ﬁnancial penalties)

(b) Independent reviews

Typical provisions

Rule element

Table 6.1 continued

Effective implementation promotes conﬁdence and thus trade and investment ﬂows

State-to-state dispute settlement only, focuses on national legislation (rather than secondary legislation)

Third parties have no recourse to tougher and more immediate remedies and reviews

Loose discipline enables scope to discriminate against third parties

Generally broad exceptions that offer considerable scope for regulatory discretion, but some tightening e.g. SPS agreement

Tight discipline promotes conﬁdence and thus trade and investment Loose wording provides broad scope for exceptions that has a chilling effect on trade and investment

Nature of preference

Provision in WTO

Likely impact

Analytical framework for rulemaking 115 ILLUSTRATION OF THE ELEMENTS OF THE ANALYTICAL FRAMEWORK The following sections explain and illustrate the analytical framework summarised in Table 6.1 with reference to some typical policy areas in which rulemaking is important.

Coverage Policy area coverage An obvious question to start with is whether RTAs go beyond the WTO in terms of their coverage of policy areas. There are often not straightforward answers to this question. For example, there are no general provisions covering all investment in the WTO, but the GATS covers investment in services and some GATT articles, such as Article III national treatment and Article XI, have been interpreted as covering aspects of investment policy in the Trade Related Investment Measures (TRIMs) agreement. Schedules It is convenient to compare the sectoral coverage of RTAs with that of the WTO. If the RTAs go beyond the WTO, this could be an indicator of their effectiveness in opening markets and their contribution to wider liberalisation efforts. Such assessments must be made on a case-by-case basis. The WTO coverage varies between the policy areas. In the case of services, for example, the WTO’s GATS framework agreement requires non-discrimination for all services, but then includes schedules and exceptions which effectively determine the coverage of the agreement. For example, Article II of the GATS provides for exceptions from the principle of most favoured nation (MFN) treatment (Art. II). This provision has been used by all WTO members to exclude sensitive sectors. In addition to such speciﬁc exemptions, the GATS agreement provides for a combination of positive and negative list-based schedules to determine coverage. Positive lists specify which sectors are covered by the framework agreement, but coverage of subsectors is then determined by negative lists. Thus, most WTO members include ﬁnancial services in their positive lists, but then exclude certain aspects of regulation, such as certain categories of insurance. This complex approach was devised to satisfy the often competing objectives of retaining control over liberalisation and ensuring reciprocity. Regional agreements also use positive and negative lists to determine coverage. The use of the negative list approach as employed in NAFTA is generally considered to be more liberal because all activities not speciﬁcally listed are liberalised. But detailed analysis of schedules is needed before one can draw conclusions because exclusions can still be significant with negative listing. RTAs that use negative listing would generally be seen as going beyond the WTO, which uses a combination of positive and negative listing. The type of listing of exceptions is

116 Stephen Woolcock also important in investment where there are extensive provisions in RTAs but no comprehensive regime in the WTO. Another example of the use of schedules to define coverage and satisfy the demand for reciprocity can be found in the public procurement. Here the reference point for comparison is the plurilateral 1994 Government Purchasing Agreement (GPA) negotiated under GATT/WTO auspices. As with the GATS agreement, the GPA consists of a framework agreement and schedules determining coverage. Coverage of RTAs can therefore be compared to the WTO and to each other, by comparing schedules. Greater coverage of an agreement or the use of negative lists, will imply that an agreement is more liberal. The nature of preferences is fairly clear-cut in such cases and is analogous to tariff preferences. Entities covered The case of public procurement also illustrates how an analysis of the coverage of rulemaking provisions must also take account of the type of entity subject to the disciplines embodied in the rules. For example, the 1979 GPA covered only national government agencies engaged in purchasing. The 1994 GPA and an increasing number of RTAs also include state/provincial, some local government and some even private entities that benefit from special and exclusive rights granted by government that could be used as leverage to influence purchasing decisions. The deeper the coverage the more entities included and thus the greater the potential market opening. RTAs that go deeper into national public purchasing than the plurilateral rules of the WTO could therefore be said to go beyond the WTO and offer a regional preference. The case of technical barriers to trade (TBTs) also illustrates how coverage can be deﬁned by the entity subject to rules. The WTO rules on TBTs are binding for regulations or standards issued by central government only. Sub-central government is subject to the non-binding ‘best endeavours’ rules and there is only a non-binding Code of Conduct for private standards-making bodies. An analysis of RTAs must therefore consider the types of entity covered by rules. In terms of the potential impact of greater entity coverage, the presumption would be that greater coverage means greater reach of non-discriminatory rules, greater predictability, more transparency and thus more open markets. A regional or bilateral preference would exist when a larger number of entities is bound by rules on national treatment or other provisions under an RTA than under equivalent WTO rules. For example, while sub-central government might be bound by an RTA not to discriminate against suppliers from within the region, they would be able to discriminate against suppliers from third countries. The degree of preference is therefore not very clear-cut here. In a (very) few cases there may be a formal preference, such as a price preference for national or regional suppliers in public procurement contracts. In most cases, however, it is simply the case that the entity concerned has more scope to exercise discretion in its treatment of third party suppliers. As many rules concern fairness and transparency in procedures, the question becomes whether national regulators will ﬁnd it worth their while to apply

Analytical framework for rulemaking 117 different rules to third party suppliers? The experience in various RTAs suggests that they will not and will use the same procedures for third country suppliers so there will be no clear preference. Regulatory instruments covered Coverage can also be measured in terms of the type of regulatory instrument used; for example, legislation, an implementing regulation by a government department or regulatory agency, or a code of conduct or voluntary standard. Legislation will generally set out the broad policy aims and determine, for example, the scope of liberalisation and the regulatory norms and standards that must be protected. The day-to-day implementation of these objectives is then generally delegated to a regulatory agency that then takes decisions, issues guidelines, adopts standards or promulgates directives. Such decisions may be generally applicable to the whole sector or they may relate to speciﬁc companies, but all will shape market access. For example, in the ﬁeld of technical barriers to trade, technical regulations are generally mandatory and adopted in legislation setting down norms for health and safety. Legislation may also establish the procedures for conformance assessment, although this is not always the case. In addition to such mandatory regulations there are, however, a large number of voluntary standards, generally developed and promulgated by private standards bodies. These are often used as one means of showing conformance with mandatory standards. In the field of services trade, day-to-day responsibility for the regulation of competition in many sectors rests with a regulatory agency, such as the US Federal Communication Commission (FCC) that takes decisions on who can operate in the market and on what conditions. The same applies in other service sectors such as ﬁnancial services. In some service sectors professional bodies regulate key aspects of the market. Coverage of any system of rules must therefore include an assessment of what type of regulatory instrument is covered by their discipline. Greater coverage suggests a more liberal regime, because the disciplines of the rules are applied to more instruments. For example, if rules only cover legislative acts, there will remain scope for discrimination against non-national suppliers when it comes to the use of secondary instruments. A regional preference would therefore take the form of greater coverage with respect to regulatory instruments within an RTA than for third parties. As with the entity coverage, the degree of preference will depend very much on how the scope for discrimination in the use of secondary instruments is used. If regulatory bodies establish practices or procedures based on the objective pursuit of the aims of regulatory policy as set out in legislation regardless of who is the supplier or market participant, there is no preference. If discretion available in the use of secondary instruments is used to favour suppliers from the region or the bilateral partner, then there is discrimination and a preference.

118 Stephen Woolcock

Rules of origin Rules of origin are used in all areas of trade and investment policy to preserve a preference. Therefore they ﬁnd application in the ﬁeld of regulatory measures. The restrictive use of rules of origin can constitute a protectionist instrument. In the case of non-tariff measures or rulemaking, origin rules are used to defend a preference just as they are in the case of tariffs. But they have not been used as extensively in this field and tend to be more liberal than in their application to tariffs. In the services sector origin rules are used to limit the beneﬁts of market access rules to service suppliers from the region. For example, NAFTA provides that service suppliers from outside the NAFTA region may benefit from the nondiscrimination and access rules under NAFTA provided they have substantial business activities within a party to the agreement. But here as in the GATS there have (as yet) been no closer deﬁnition of origin criteria. In government procurement RTAs also include rules of origin, for example, the EU rules on public purchasing in the utilities ﬁeld enable purchasers to discriminate against non-EU suppliers if the goods to be supplied under any public contracts are not above a set percentage of EU content. But regional preferences are not possible in all aspects of public purchasing because the 1994 GPA prohibits rules of origin for purchasing by central government.

Principles Provisions in trade agreements on rulemaking are, as with border measures, covered by the non-discrimination principles of national treatment and most favoured nation treatment. National treatment means that foreign products or suppliers are treated the same as national suppliers. This is simply stated but becomes more complex in its implementation. Most favoured nation status means that regulators cannot discriminate between foreign products or suppliers. The conclusion of a regional or bilateral agreement that offers preferential treatment to signatories is, of course, in breach of the MFN principle. Although much attention is focused on national treatment in trade agreements it cannot necessarily be equated with liberalisation. If, for example, there is a national monopoly provider of a speciﬁc service, national treatment simply means that foreign competitors will be treated the same as national competitors of this supplier. In other words they will not be allowed into the market. Agreements have provided for national treatment of technical regulations for many years without this resulting in open markets. Indeed, the introduction of new regulations and standards that conformed to national treatment obligations can still result in progressive market closure. Even within regions with deep integration such as the EU national treatment had proved to be inadequate. National treatment measures have therefore either been supported by rigorous implementation and enforcement measures, or augmented by policy approximation, harmonisation or mutual recognition.

Analytical framework for rulemaking 119 The nature of preference with regard to national treatment is very clear-cut. If an RTA or bilateral agreement only extends national treatment to preferential partners there is a clear preference. With regard to MFN, any provision that discriminates between different foreign products or suppliers represents a preference.

Transparency Transparency in national policy-making and regulation has long been recognised as an important ﬁrst step towards facilitating more open markets. Consequently, all agreements include some form of transparency rule, but they vary in their scope. It is therefore helpful to differentiate between obligations to provide information on statutory measures/legislation, secondary measures, and decision-making procedures, or ‘due-process’ as it is termed in American literature. The latter relates to opportunities for interested parties to make submissions and comments on regulatory decisions. Transparency requires, at a minimum, that information on all statutory measures be published. It is also standard practice to include a requirement on notification of national statutory measures to the WTO or to the relevant institution in a regional or bilateral trade agreement. While statutory instruments generally set out the objectives of regulatory policy, the day-to-day implementation of policy takes the form of decrees, executive orders and guidelines from regulatory agencies, administrative guidance from ministries, judicial or arbitral decisions. For effective transparency in rulemaking it is therefore necessary to include secondary instruments that can set precedents and thus shape market access. For example, the conditions for interconnection to telecommunications networks are laid down by the relevant regulatory agencies. The GATS agreement and equivalent measures in RTAs on basic telecommunications therefore require notiﬁcation of such conditions. In public procurement, statutory measures set out procedures to be followed in placing public contracts, but the key information for potential suppliers comes in the form of calls for tender issued by a large number of public bodies that set out the qualifying conditions for prospective tenders and contracts. For foreign suppliers to have any chance of competing, transparency with respect to these subsidiary instruments is essential. Likewise, information on modifications to technical regulations or health standards is essential in the fields of TBT and SPS. In competition policy, precedence set in dealing with cases inﬂuences the evolution of policy, therefore transparency may need to include information on all decisions or judgments. Without effective transparency with regard to these subsidiary instruments of regulatory policy it will be, at best, very costly for foreign suppliers to follow the detailed developments in policy. At worst, close links between domestic suppliers and regulators can result in the exclusion of foreign suppliers. Some agreements may require prior notiﬁcation of any legislation or change in norms or regulatory standards. If prior notiﬁcation is provided, foreign suppliers

120 Stephen Woolcock will have an opportunity to re-tool production or modify their product or service to comply from day one with the new rules. The absence of prior notiﬁcation or short notification periods may disadvantage some foreign suppliers. National producers will be more aware of changes in the rules as they will probably have been consulted by the regulatory agency. They will therefore have an opportunity to conform from day one. Small foreign suppliers, especially those in developing countries are unlikely to have the capacity to monitor regulatory policy and could, as a result, lose market share to national suppliers during the period they are unable to comply. This is an issue in sanitary and phytosanitary measures, for example. Prior notiﬁcation facilitates consultation and rules may provide for parties to an agreement to comment on any decision or regulation. This ability to comment is generally extended to the state authorities. With deeper integration the right to comment may be extended to all those potentially affected by regulation, including legal persons (i.e. companies, consumers or other civil society NGOs) in other countries. Transparency in decision-making can be taken one step further if regulatory agencies are required to give consideration to the views of market participants or provide a reasoned response to the comments made. NAFTA, for example, requires the responsible agency in each country to respond to submissions from interested parties. This form of procedural transparency can promote best practice by ensuring that the experience of all market participants is included in any decision. The input of foreign suppliers may also help to correct the usual information bias in favour of domestic suppliers and hence help guard against regulatory capture by dominant national suppliers. Finally, some agreements, such as NAFTA, provide for open decision-making or what is equivalent to national treatment in regulatory procedures. For example, under US administrative procedures, private entities generally have extensive rights to make submissions and have regulators respond to their comments. This model has been used in NAFTA, which extends the same rights to private entities regardless of their origin. This kind of provision promotes conﬁdence that national regulatory policies are genuinely carried out in a non-discriminatory fashion. When combined with recourse to legal remedies in cases of non-compliance, they can help to ensure that national treatment or non-discrimination is effectively implemented in regulatory policy. Transparency rules can therefore facilitate improved regulatory practice, reduce uncertainty and thus enhance market access and give investors conﬁdence to invest in the market concerned. Transparency rules do, however, involve costs for the national regulators. Notifying the many thousands of implementing decisions in regulatory policy can be very costly. This is one reason for the resistance to rulemaking in some policy areas such as competition and public procurement. Increased use of information technology may reduce these costs, but the costs may still be high relative to the gains, especially for small countries. WTO provisions on transparency typically include obligations to publish and notify legislative changes. In some policy areas there are also transparency rules covering subsidiary regulatory instruments, but here coverage is not as great. The

Analytical framework for rulemaking 121 WTO Agreement on Technical Barriers to Trade and the Agreement on Sanitary and Phytosanitary Measures also require prior notiﬁcation of statutory measures, but there is no requirement for prior notiﬁcation under the GATS. A number of WTO agreements require regulators to take note of comments from national governments, but not from private parties. In general, WTO provisions are not very strong and have not always been implemented effectively. The degree of preference embodied in transparency rules is generally not signiﬁcant, but it can vary between different types of transparency. If transparency takes the form of publication of statutory and subsidiary regulatory measures, then there is no preference, since this information will be available to all. However, regional or bilateral agreements often include channels for cooperation between regulatory bodies. These could well result in more effective mutual exchange of information than is possible at the multilateral level. Such regulatory cooperation is likely to be especially important for developing country parties concluding an RTA with a developed economy, because the technical expertise and resources available in the developed partner can help ensure effective dissemination of information. There is more clearly a regional preference when participation in regulatory decision-making is limited to preferential partners. If this preference is used to facilitate collusion between the regional partners to keep out third country suppliers then the degree of preference could be important. On the other hand if the involvement of parties from within the region helps guard against regulatory capture and helps to promote best practice in regulation, this will beneﬁt all comers and there will be no effective preference.

Substantive measures The ineffectiveness of national treatment in dealing with regulatory policy issues has led to the inclusion of other substantive measures in trade agreements. As set out in Table 6.1 these can take a number of forms such as harmonisation, approximation, equivalence or mutual recognition of test results or of national regulation per se. The perceived failure of national treatment to facilitate – let alone guarantee – effective market access has led to efforts to harmonise rules such as in the EU, but also to a lesser degree in other RTAs and in the WTO. Harmonisation comes at a cost in terms of lost policy autonomy, which is important if countries have signiﬁcantly different policy preferences, such as agreements that include developed as well as developing countries. Economists also argue that welfare losses can result from harmonisation and that less rigid approaches, such as institutional or policy competition are preferable in principle. Agreements have therefore opted for partial harmonisation focused on the most important sectors or areas in which producers feel there is most to gain from deep integration. Trade agreements may also have approximation as a general aim, without any detailed commitments on implementation. Alternatively, agreements may provide for harmonisation as a long-term aim, such as in the case of the

122 Stephen Woolcock Euro-Med agreements, but include little by way of concrete means of achieving such an objective. Another alternative to harmonisation is equivalence, which is preferred to harmonisation or mutual recognition in NAFTA. This approach encourages the parties to accept each other’s regulatory norms or standards as equivalent. The attraction of this approach is that it dispenses with most of the detailed substantive rules and institutions involved with harmonisation or mutual recognition and thus limits compliance costs and entails much loss of regulatory autonomy. But there is as yet little evidence that such equivalence approaches have been effective in opening markets. National regulators retain considerable discretion that can frustrate market access. Another difﬁculty is that with no policy cooperation in setting norms or standards, equivalence may, in practice, simply mean that the standards or regulatory norms of the dominant market are adopted. While this provides enhanced access for the smaller parties, they also have to carry all the costs of adjustment. The complement to harmonisation preferred in the EU has been mutual recognition. Outside the EU mutual recognition refers to mutual recognition of compliance assessment or test results, not full mutual recognition of regulatory policy. Mutual recognition of conformance assessment means that the regulators in the importing country recognise the results of conformance assessment carried out in the exporting country. It is important to bear in mind that this generally still means that the exporter must conform to the norms or standards in the importing country. Mutual recognition of conformance assessment therefore simply means that the costs of compliance are reduced. It does not necessarily overcome potential barriers to market access resulting from different national regulations or standards. Even this limited application of mutual recognition is not without its drawbacks. Some agreements leave the regulator in the importing country a safeguard that permits rejection of products or services even when they have been certiﬁed in the exporting country as complying with the importing countries standards. This is the case for TBT provisions in NAFTA although NAFTA does require parties rejecting test results elsewhere to justify such a refusal. Another difﬁculty with mutual recognition is that importing countries expect assurances that conformance assessment systems in the partner countries are up to standard before they are ready to recognise test results from the country concerned. This can mean the inclusion of obligations in mutual recognition agreements that are tantamount to accepting the conformance assessment norms or accreditation procedures of the dominant party. The EU, for example, requires the adoption of conformance assessment equivalent to EU conformance assessment in all its bilateral mutual recognition agreements. With full mutual recognition of regulatory policy the importing country accepts that if a product or service can be safely sold in a foreign market it can be sold in the home market, thus effectively removing any barrier to market access. In practice, this generally requires de facto approximation or convergence of regulation and has only been applied in regional agreements between countries at broadly the

Analytical framework for rulemaking 123 same level of development, such as in the EU and the ANZCERTA (Australia and New Zealand). A number of agreements envisage such mutual recognition but only after what in practice is likely to be a long transition period of close regulatory cooperation in order to establish such conﬁdence. The WTO includes a good deal of hortatory language on the use of common international standards, such as in the TBT and SPS agreements. The SPS agreement is built on the foundations of harmonised standards for food safety in the shape of the CODEX Alimentarius. The Trade Related Intellectual Property rights (TRIPs) agreement of the WTO also requires the adoption of harmonised standards of protection for intellectual property. There is also partial harmonisation of regulatory norms in services, such as in the sector agreements on telecommunications and ﬁnancial services adopted in 1997 and the WTO explicitly encourages mutual recognition with respect to technical barriers to trade and services. What constitutes a preference in this area of harmonisation and mutual recognition? A regional preference could be said to exist if regional standards or regulatory norms are developed that diverge from existing international standards. If no international standards exist, regional standards could be seen as a preference, but one that would be eroded once an agreed international standard is adopted. The impact of regional standards on third parties may, however, be positive even if such a ‘preference’ exists. The adoption of one regional norm or standard may mean lower (friction) costs in trading with the region, because exporters need only comply with one standard rather than a range of different national standards. The systemic impact of such regional norms may, however, be less benign, since the evolution of conflicting regional norms may prevent or slow the adoption of genuine international norms or standards. Whether this is indeed happening is one of the central research questions with regard to the new wave of regionalism. Partial or selective harmonisation may represent a form of preference in the sense that harmonisation may be focused on those sectors in which the regional producers feel they stand to gain most from harmonisation or approximation. A regional preference may arise in the case of equivalence or mutual recognition if this approach means, in effect, the adoption of the dominant norm or standard in the region and this differs from agreed international norms. Finally, any mutual recognition agreement could be said to constitute a preference. It has been argued that this is the case and that the spread of mutual recognition agreements initiated by the EU will result in a ‘two tier’ trading system (Baldwin, 1989). The preference embodied in a mutual recognition agreement could be subject to ‘erosion’ if all parties that can satisfy the requirement and obligations of a mutual recognition agreement are allowed to negotiate accession. In practice, this will depend on how stringent the obligations are with regard to harmonisation of essential common standards, conformance assessment and accreditation.

124 Stephen Woolcock

Institutional provisions Trade agreements generally include institutional provisions ranging from common decision-making machinery at one end of the spectrum through to technical assistance at the other end. Common decision-making machinery is generally only found where there is a high degree of policy integration, such as in the EU. In the EU statutory provisions are adopted at the EU level and in some cases there are common, supranational regulatory bodies, i.e. the European Commission, that implement the common rules. This is, for example, the case in competition policy. In most policy areas, however, national legislation implements the aims of EU policy that are set out in directives and competent authorities in the EU member states deal with day-to-day regulatory issues. The rule for trade agreements and most customs unions is intergovernmental institutions. These may have general responsibility for development and implementation of the agreement, such as in the case of the WTO General Council, or the committees or councils that are created in most trade agreements. The EU RTAs with third countries create, for example, Association Councils. NAFTA has a Free Trade Commission. These are high-level bodies that meet relatively infrequently, usually once or twice a year, and have responsibility for the overall functioning of the trade agreement. Trade agreements also often set up specialist committees or commissions to deal with specific policy areas such as technical barriers to trade, government procurement or services. Such committees monitor implementation of the trade agreements in the relevant policy area, act as a forum for extending provisions, exchange information on developments and may provide a forum for initial consultations on disputes. These kinds of committees have been established in multilateral, regional or bilateral agreements, but in the WTO there are, of course, far more participants and regional level committees are likely to provide a more intimate setting facilitating more effective cooperation. Finally, trade agreements may provide for technical cooperation and capacity building. As with the other institutional provisions discussed above, this type of provision can be found on all levels. But ﬁnancial and other resources available for technical cooperation are much more thinly spread in the WTO than at the regional levels. Linking technical cooperation and capacity building with regional trade objectives may also unlock more resources. Hub or core trading entities, such as the EU and US, are more likely to provide funds if they have substantial control over the use of the funds and they serve policy objectives in a regional or bilateral setting. Closer links between partners in regional trade agreements are also likely to lead to greater availability of technical expertise or twinning arrangements between regulatory agencies. Technical cooperation and capacity building are particularly important for developing countries and can help establish the institutional infrastructure (e.g. competent and perhaps independent regulatory agencies) needed if regulatory reform and thus regulatory barriers are to be tackled effectively. Technical assistance and funds for capacity building may help to create the institutional capacity

Analytical framework for rulemaking 125 needed to implement provisions in such policy areas as TBT, SPS, competition or services. What is the impact of such institutional provisions on preferences? Preference dilution occurs when the more liberal states succeed in pressing for common policies that are more liberal than those of other partners in the agreement. Preference concentration occurs when the parties to a regional arrangement opt for trade diversion by liberalising among themselves and maintaining or perhaps even raising barriers vis-à-vis third countries. The question of preference dilution or concentration has been most pertinent for the EU because of the scope of its common decision-making machinery. But the same question can also be posed in each preferential agreement. More generally, cooperation can also help to promote regulatory best practice and thus contribute to improved economic performance in the countries concerned and to the evolution of an open trading system. On the other hand, cooperation may be used as a means of promoting the detailed speciﬁc standards of a dominant partner and can thus result in a regional preference and be detrimental to wider multilateral trade and investment. Institutional provisions can therefore create regional preferences if they result in preference concentration. An assessment of this can only be carried out after detailed study of speciﬁc cases. More generally, third parties are clearly excluded from the cooperation procedures between parties to an agreement, so this can be seen as a form of preference. Finally, third parties are unlikely to beneﬁt from the enhanced technical assistance or ﬁnancial provisions included in regional or bilateral agreements.

Safeguards All trade agreements contain safeguards. Governments are reluctant to make openended commitments to liberalisation. Thus, the GATT has always had and retains scope for safeguard actions under Article XIX. Every regional or bilateral trade agreement also has safeguards for border measures such as tariffs and quotas. It must come as no surprise then that governments wish to ensure they have scope to take safeguard measures for liberalisation of behind the border or regulatory policies. Governments have therefore included safeguard provisions that offer scope for the use of discretion in regulatory policy in regional and bilateral trade agreements. For example, the Treaty of Rome, as amended, requires the removal of all measures that have equivalent effect to quotas in restricting trade within the EU. Over the years this has been interpreted as including all regulatory policies. But Article 30 provides an exception that enables national governments to discriminate in regulatory policies or to pursue other legitimate policy objectives. Such safeguards are hedged by wording that requires such discretionary measures to be ‘proportional’ and not a disguised restriction on trade, and to use the ‘least trade restrictive’ or ‘less trade restrictive’ measures to pursue such policy objectives.

126 Stephen Woolcock The issue with such safeguard provisions for regulatory policy in regional trade agreements is therefore how much discretion exists. Tight wording will mean limited discretion for regulators to pursue a range of social, environmental or other policy objectives. Loose wording will provide discretion for regulators, which could be abused to keep out competition. Within the GATT/WTO system this has taken the form of a general exception to national treatment and MFN under Article XX of the GATT that enables governments to discriminate against foreign suppliers. The scope of this GATT safeguard for regulatory policies is drawn very wide as the original intention of the GATT was not to cover deep integration or regulatory policies. In recent years, however, concern about the abuse of the discretion offered by Article XX has led to efforts to draft tighter rules. This has found expression in the efforts over the years to tighten up the provisions of the TBT agreement and in particular the conclusion of the SPS agreement in the Uruguay Round. The latter signiﬁcantly reduces the discretion available to national regulators to control or prohibit imports of food and agricultural products on the grounds of protecting human, animal and plant health. A preference in this context would then exist if the RTA provisions were WTO plus and regulators within the region were able to discriminate vis-à-vis suppliers from third countries with regard to such safeguard measures. This would result in greater uncertainty for third country suppliers and is likely to have a chilling effect on trade and investment. Scope to discriminate against third country suppliers might also be abused to protect the national or regional markets against competition from third countries.

Implementation How rules are applied is crucial. Provisions on dispute settlement or conciliation are therefore critical because they shape who interprets the rules and how. With regard to interpretation of rules, RTAs could be said to go beyond the WTO if they interpret general exceptions to non-discrimination, such as those in GATT Article XX or equivalent provisions in regional/bilateral agreements more restrictively than the multilateral rules. Regional agreements also contribute to the evolution of trade rules by codifying the criteria for exemptions in the texts of agreements, or interpreting such criteria in judicial or quasi-judicial procedures, such as the European Court of Justice in such cases as Dassonville and Cassis de Dijon. Regional agreements sometimes include their own dispute settlement provisions. Some of these are highly developed. The EU, for example, has its own European Court to interpret European law, which prevails over national law on issues relating to the European single market. NAFTA also offers a comprehensive tripartite dispute settlement mechanism including conciliation as well as quasi-judicial panels analogous to that of the WTO. Other regional agreements may rely more on conciliation and consultation in the regional or bilateral councils that oversee agreements. Speciﬁc regional dispute settlement is likely to facilitate more rapid processing of disputes than in the WTO and thus contribute to better implementation of trade commitments at the regional level than the multilateral level.

Analytical framework for rulemaking 127 Agreements may also offer reviews and remedies in cases where regulatory decisions are thought to be discriminatory. For example, the GATS agreement on basic telecommunications provides that suppliers, both domestic and foreign, should have recourse to a review of decisions concerning interconnection to the network, a key area affecting market access. In government procurement the plurilateral GPA provides aggrieved parties that believe they have been unfairly treated in the granting of a public contract with the option of mounting a ‘bid challenge’. If reviews are independent then these can help to ensure that regulatory policy is non-discriminatory. If remedies are readily available they can ensure swift enforcement and prevent delay and prevarication when it comes to compliance. In measuring the RTAs against the WTO standard of provision, it will therefore be necessary to assess whether an RTA offers more effective or more immediate remedies than the WTO. There is also the question of who has the right to bring a case in any dispute. In the WTO only states can bring cases, but in deeper integration agreements there may be rights for legal persons (i.e. companies, trade associations, NGOs, etc.) to bring cases. This is the case within the EU, where actions can be initiated against national implementation of European regulation under ‘direct effect’. It is also the case in NAFTA and NAFTA-type FTAs that provide for investor–state dispute settlement. If actively used such rights of action for legal persons, combined with regional level dispute settlement, can represent as great a challenge to national regulatory autonomy and state sovereignty as harmonisation. Dispute settlement provisions at the regional level may result in more effective implementation. When the regional agreements mirror multilateral provisions, as is frequently the case, this can mean that regional actions result in more effective implementation of the multilateral principles. But if regional interpretations go beyond what is generally accepted in the WTO, this raises the question of how to reconcile regional and multilateral interpretations in the future. The WTO provides a fairly strong dispute settlement system. Following the Uruguay Round the establishment and adoption of WTO panels is now more or less automatic. There is also an Appellate Body to rule on questions of WTO law. The WTO also offers scope for consultation and conciliation. Given the strength of the WTO system, countries participating in regional trade agreements are likely to wish to retain their right to bring cases in the WTO. What the WTO does not offer, however, is rapid review and remedies or investor–state dispute settlement. Although WTO dispute settlement cases have to be processed in line with a tight timetable, cases can still take 18 months to complete. If one adds to this the time taken by governments to comply with decisions, it is likely to be years before actors can get any redress for discriminatory application of regulatory policies under the WTO. Preference in the area of dispute settlement therefore takes the form of more immediate and rapid remedies for states and companies from participating states compared to third parties as well as the potential for divergent regional interpretations of key issues, such as proportionality or least trade restrictive provisions.

128 Stephen Woolcock CONCLUSIONS This chapter has set out to fulfil the modest objective of identifying the core elements in any trade agreement and, in particular, in agreements that include elements of rulemaking. Through illustrative examples, it has shown that the framework can be applied to all policy areas or cases, although some will tend to ﬁt the framework better than others. In one chapter it has not been possible to apply the framework fully to a speciﬁc case, but this has been done elsewhere (Woolcock, 2004). The approach provides a framework for a qualitative but systematic assessment of the impact of the current wave of regional and bilateral trade and investment agreements. It facilitates assessments of where such agreements go beyond the existing WTO rules and thus their likely impact on both market opening and domestic regulatory autonomy. The framework also facilitates an initial assessment of the likely degree of preference in such regional level rules. For example, if the regional rules go beyond WTO rules with respect to procedural measures such as transparency, institution building or cooperation, they are less likely to constitute a signiﬁcant degree of preference than when they go beyond international agreed norms or standards with regard to substantive measures. Within these two broad categories the framework also offers a means of differentiating between those elements of RTAs that are compatible and those that are likely to create more difﬁculties for multilateral trade and investment. NOTES 1

2 3 4

5

6

7

There is an important issue of terminology. In this chapter we shall use regional trade agreements to include both agreements between countries within the same region and free trade agreements between countries in different regions that may in some cases be bilateral agreements. An alternative and more general term would be to simply use preferential agreements. See Winter (1996) for a summary of this work. See Bond and Syropoulos (1996). The tasks of the Committee on Regional Trade Agreements are in fact: (1) to provide legal analysis of the RIAs (and their compatibility with GATT rules); (2) to make horizontal comparisons (inventories of RIA provisions covering no-tariff barriers have been produced); and (3) to debate the context and economic aspects of RTAs. See, for example, Joel Trachtman ‘Towards open regionalism? Standardisation and Regional Integration Under Article XXIV of the GATT’ paper presented to the WTO Seminar on ‘The Changing Architecture of the Global Trading System; Regionalism and the WTO’, April 2002, WTO, Geneva and Aaditya Mattoo and Carsten Fink ‘Regional Agreement and Trade in Services: Policy Issues’, paper for the WTO Seminar, April 2002, WTO. For an example of the OECD work see Ken Heydon ‘RTA Market Access and Regulatory Provisions: Regulatory Provisions in Regional Trade Agreements: “Singapore Issues”’ paper for the WTO Seminar on Regionalism and the WTO, 26 April 2002, WTO. After ten years there are signs of progress in the WTO Committee on Regional Trade Agreements that point to improvements in the transparency with regard to RTAs, but there are still difﬁculties reaching an agreed interpretation of the key GATT provisions that would facilitate assessments of whether RTAs are GATT-compatible.

Analytical framework for rulemaking 129 REFERENCES Baldwin, R. (1989) ‘The Growth Effects of 1992’, Economic Policy, 9: 247–81. Bhagwati, J. (1994) ‘Threats to the World Trading System, Income Distribution and the Selﬁsh Hegemon’, Journal of International Affairs, 48. Bhagwati, J., Panagariya, A. and Krishna, P. (1999) (eds) Trading Blocs, Cambridge: MIT Press. Bond, E.W. and Syropoulos, C. (1996) ‘The Size of Trading Blocs, Market Power and World Welfare Effects’, Journal of International Economics, 40: 411–37. CEPR (1996) ‘Growth and European Integration: Towards an Empirical Assessment’, CEPR Paper Series, 1393. Frankel, J.A., Stein, E. and Wei, S.-J. (1995) ‘Trading Blocs and the Americas: The Natural, the Unnatural and the Supernatural’, Journal of Development Economics, 47: 61–96. Hettne, B. (1996) ‘Globalisation, the New Regionalism in East Asia’, in T. Tanaka and T. Inoguchi Globalisation and Regionalism, selected papers of the UNU Global Seminar 96 Shonan Session, 2–6 September 1996. Krugman, P. (1991) ‘The Moves Towards Free Trade Zones’, in Policy Implications of Trade and Currency Zones, Federal Reserve Bank of Kansas City, 7–42. Krugman, P. (1992) ‘Is Bilateralism Bad?’, in E. Helpman and A. Razin (eds) International Trade Policy, Cambridge: MIT Press. OECD (2002) Regional Trade Agreements and the Multilateral Trading System, November 2002. Schott, J. (2004) (ed.) Free Trade Agreements. US Strategies and Priorities, Washington: Institute of International Economics, 13. Winters, A. (1996) ‘Assessing Regional Integration Arrangements’, World Bank Research Working Papers. Woolcock, S. (2004) ‘The Interaction Between Levels of Rulemaking in Public Procurement’, paper presented on the LSE/CRIS Workshop on Interaction Between Levels of Rulemaking in Trade and Investment, Brussels, London School of Economics, International Trade Policy Unit, 16 December 2004. Online. Available HTTP: . World Bank (2005) Global Economic Prospects 2005: Trade Regionalism and Development. World Trade Organisation (1995) Regionalism and the World Trading System, Geneva: WTO.

7

Comparative ﬁscal integration indicators Philippe De Lombaerde and Ana-Cristina Costea

INTRODUCTION The literature on decentralisation and ﬁscal federalism presents us with a relatively sophisticated theoretical framework and a set of policy-relevant analytical tools to study the relations between different policy levels in a national environment. The aim of this chapter is to explore to what extent these conceptual and theoretical tools can be translated into a toolbox which is adapted to a multilevel government setting with supra-national (macro-regional) government levels. We will ﬁrst present an overview of the conceptual framework offered by ﬁscal federalism theory and the technical tools which have been designed to analyse (ﬁscal) decentralisation processes. We will then explore how these tools could be adapted in order to be used for the monitoring and comparative analysis of regional integration processes and discuss the specific methodological and practical issues and problems that we envisage.

DECENTRALISATION AND INTEGRATION IN FISCAL FEDERALISM THEORY Since the late 1970s and 1980s, administrative and/or political decentralisation have attracted a lot of attention, not only in academia but also in actual public administration practice and politics, both in the North and in the South. Different reasons explain this phenomenon: 1 2 3 4

an almost general (global) tendency towards more democratic forms of government in the post-war period, the apparent relative failure of centralised development models and the need for structural adjustment, the supporting role of multilateral organisations, especially in developing countries, the new surge of neo-liberal ideology and its view on the role of the (central) state,

Comparative ﬁscal integration indicators 131 5

the development of fiscal federalism theory and its ‘demonstration’ of the (causal) linkages between decentralisation, on the one hand, and efﬁciency and democracy, on the other.

Although ﬁscal federalism is obviously concerned with a broader area than strict budgetary matters, we will primarily focus on these. As Oates put it: ‘As a subﬁeld of public ﬁnance, ﬁscal federalism addresses the vertical structure of the public sector. It explores, both in normative and positive terms, the roles of the different levels of government and the ways in which they relate to one another through such instruments as intergovernmental grants’ (Oates, 1999: 1120). The theories of fiscal federalism are thus concerned with the functioning of multilevel government structures. This includes the allocation of taxation powers and control over public expenditure among the different levels of government, but also the relations between the different levels in the form of grants and credits. Although the theories of fiscal federalism have been developed in the context of public ﬁnance in national economies in federalist states, it is justiﬁed to consider them as more general theories of vertical government structures (Oates, 1999). They are therefore important building blocks of a theoretical framework for the analysis of the macro-regional policy level and its linkages with other (national, sub-national) government levels.1

Basic ﬁscal federalism theory and decentralisation The theories of ﬁscal federalism draw on concepts and theories from public ﬁnance, economic geography and public choice (Samuelson, 1954, 1955; Musgrave, 1959; Oates, 1972; Tiebout, 1956, 1961; Brennan and Buchanan, 1977, 1978, 1980; Buchanan, 1980). The so-called basic or traditional theory of fiscal federalism consists of a general normative framework for the assignment of powers and functions and the corresponding ﬁscal instruments to different levels of government (Musgrave, 1959; Oates, 1972). This assignment depends on the optimal level of provision of public goods. On the basis of the principle of allocative efﬁciency, a multilevel government system is normally superior to a centralised government system. This explains why the theory, in a context of (relatively) centralised states, has mainly been used to justify decentralisation policies. In addition, ﬁscal federalism theories have been associated with neo-liberal ideology; one of the theses of ﬁscal federalism being that decentralisation leads (and should lead) to less government intervention and a smaller state. According to the basic theory, the efﬁcient level of output of a ‘local’ public good is the level where the sum of residents’ marginal beneﬁts equals marginal cost. This level will normally vary across jurisdictions for two reasons: 1 2

cost differentials, determining differences in marginal costs, differences in preferences, determining differences in marginal beneﬁts.

The normative proposition which follows from this is known as the ‘decentralisation theorem’ and can be formulated as follows:

132 Philippe De Lombaerde and Ana-Cristina Costea ‘in the absence of cost-savings from the centralised provision of a [local public] good and of interjurisdictional externalities, the level of welfare will always be at least as high (and typically higher) if Pareto-efﬁcient levels of consumption are provided in each jurisdiction than if any single, uniform level of consumption is maintained across all jurisdictions’. (Oates, 1972: 54) It is further understood that information asymmetries exist between central and decentralised levels of government, only the latter having access to perfect information on preferences and cost conditions, and/or that political constraints exist that prevent a diversiﬁed supply of local public goods (Oates, 1999). Although mobility of the subjects of the jurisdictions is not a necessary condition for the precepts of ﬁscal federalism in favour of multilevel (decentralised) government to hold, it does make the case stronger. In the Tiebout model (1956), mobile households choose to reside in the jurisdiction which offers the mix of public goods and taxes which is closest to their preferences. Local governments ‘compete’ with each other and produce a ﬁrst-best solution.2 Brennan and Buchanan (1980), build further on the Tiebout model and combine it with insights from monopoly theory and public choice theory. In their view, internal mechanisms make that governments (‘Leviathans’) seek their own growth through the extraction of more and more resources from the economy. Decentralisation is an efﬁcient alternative to explicit ﬁscal rules (constraints) on the taxing power of governments because it promotes intergovernmental competition. It is believed that at the local level a market solution exists which determines the optimal level of expenditure on public goods, which is a departure from the views of Samuelson and Musgrave. Unlike Tiebout, Brennan and Buchanan do not present this as a ﬁrst-best solution and they are conscious of the political economy aspects of intergovernmental relations. For Brennan and Buchanan decentralisation is assimilated with democracy. A logical implication of Brennan and Buchanan’s views is that decentralisation (multilevel government) minimises the overall size of the public sector. Hence, both macro-economic and micro-economic efﬁciency are reached with decentralisation.3 Fiscal federalism also addresses the optimal choice of the type of fiscal instrument (tax) for each government level (the ‘tax assignment problem’). It is shown that this choice should basically depend on the mobility of the economic units. A system of multilevel government is usually characterised by additional ﬁnancial ﬂows in the form of grants (transfers) or credits. Grants can be conditional, placing conditions on their use by the recipient government, or unconditional. Fiscal federalism theory prescribes that conditional matching grants should be in place to the extent that inter-jurisdictional external effects occur. A system of unconditional grants should exist for redistributive (ﬁscal equalisation) purposes, which is a form of cohesion policy, or for revenue-sharing purposes when local taxation is not efﬁcient.

Comparative ﬁscal integration indicators 133 As Oates (1999) rightly points out, the precepts of ﬁscal federalism should be seen more as general ‘guidelines’ than ﬁrm ‘principles’. The theories do not permit to unambiguously determine the optimal level of any specific collective good.4 This point of view can be extended to the tax assignment issue and the design of a system of transfers. The basic theory of fiscal federalism does not provide us either with clear principles on the optimal geographical extention of jurisdictions. It only suggests that there will be a trade-off between the need to internalise positive and negative spillovers (supporting larger jurisdictions) and the need for local differentiation (supporting smaller jurisdictions). Basic ﬁscal federalism theory is a static theory, providing arguments in favour of a decentralised multilevel government structure. There is no theoretical model available that establishes a long-run relationship between decentralisation and economic growth or development. Empirical evidence is not straightforward. Although decentralisation may contribute to development in a particular stage of development, decentralisation should also be seen as a consequence of complex economic and political development processes. As pointed out by Prud’homme (1995) and others, in a developing country context, the hypothesis of efﬁcient local governments is not necessarily valid. The potential contribution of decentralisation to economic development depends on the historical, institutional and political contexts and the design of local ﬁscal instruments, intergovernmental transfers and intergovernmental debt ﬁnancing instruments. More recent advances include the incorporation of political goals in the objective function (Inman and Rubinfeld, 1997a, 1997b) and the reinterpretation of the optimality of the decentralised state from a neo-institutional perspective (Weingast, 1995; McKinnon and Nechyba, 1997). It has been shown that in the presence of economies of scope in the provision of public goods, a centralised government can be optimal (Alesina and Spoalore, 2003: 26–29).

Fiscal federalism and regional integration Fiscal federalism theories are perfectly capable of including supra-national (regional) government levels in the analysis. This is particularly relevant when there are relatively high (or growing) levels of intra-regional mobility of capital and labour (consumers, taxpayers), which widen the scope for externalities. Nevertheless, their application to regional integration problems has been relatively limited, although their inﬂuence on the design of the European ﬁscal arrangements should not be underestimated.5 Tinbergen (1954) and Cooper (1977, 1995) extended the general ideas of optimum policy levels and functional federalism to an international macro-regional context. According to Tinbergen (1954), the optimal (efficient) policy level is based on two criteria: 1

the level should be ‘low enough’ in order to minimise transaction costs and maximise citizens’ participation, and

134 Philippe De Lombaerde and Ana-Cristina Costea 2

the level should be ‘high enough’ so that all those affected by the policy measures are covered, or, in other words, so that external effects are minimised.

Cooper (1977, 1995) states that for the provision of certain public goods (and only for public goods), regional integration can be Pareto-optimal in view of reaching the social goals. He stresses that a variety of factors inﬂuence the determination of the optimal level of government intervention; some of them point in the direction of high levels (large jurisdictions, although not global), other in the direction of low levels (small jurisdictions). The ﬁrst category of factors includes: technical economies of scale (depending on the concerned good), positive or negative external effects (which usually also constitute a case of economies of scale, for example, when the marginal cost of additional consumption of public services is almost zero), and the mobility of the subjects of restrictive (regulatory or redistributive) policies. The second category of factors includes individual preference diversity in relation to collective goods. These goods can be ‘systemic’ (policy regimes, value systems, etc.) or ‘speciﬁc’ (science parks, infrastructural projects, etc.). This logic leads to a system of functional federalism in which the optimal level of provision of the collective goods by the public agent is determined for each good individually. Such a system is complex and is characterised by overlapping jurisdictions and partial sovereignties. Although there are several examples of supra-national (macro-regional) cooperation in order to provide speciﬁc public goods (regional central banks, regional development banks, specific regional organisations) and although functional federalism may be optimal from a static point of view, this is not necessarily the case from a dynamic perspective. As acknowledged by Cooper, as technology and preferences vary with time, functional federalism would imply frequent changes in the integration schemes, causing transition costs and possibly generating conﬂicts with the bureaucratic interests of the existing schemes. In addition, with ﬁscal federalism, the possibility exists to negotiate and to seek political deals on hierarchically higher levels in relation to the provision of collective goods on lower levels, independently from local interests. Fiscal federalist thinking logically implies that if there is a long-term tendency towards the mobilisation of goods, services, production factors and information, and towards deeper interdependence, more regional integration initiatives might be expected.6 The relationship between country size and economic integration is a two-way relationship. While the incentive for economic integration is greater for smaller countries, facing smaller domestic markets, because of the liberalising aspects of RIAs, regional integration reduces the size of viable countries (economies) (Alesina and Spoalore, 2003: 81). As the authors show, this becomes clearer when a distinction is made between the political size of a country and the size of its market. Only in a world with complete autarky, do these two concepts coincide. Casella presents a dynamic version of fiscal federalism with spatial and institutional dimensions. Her thesis is that market integration requires centralised regional institutions in order to minimise transaction costs and in order to provide

Comparative ﬁscal integration indicators 135 an adequate level of policy coordination. With deepening integration, the roles of the different regional components of the macro-region are being redefined and competition among them increases. This may lead to a more heterogeneous demand for public goods and a revision of the allocation of powers and their provision within the macro-region (Casella, 1994). The logic of the development of market-based integration leads also to a system of functional federalism (Casella and Frey, 1992). As a final remark, fiscal federalism and related theories explain the optimal policy level in a multilevel government structure. To the extent that actual policy levels differ from the optimal levels, the normative implication of the theories points in the direction of decentralisation or integration. Theoretically speaking, subsidiarity is thus not a necessary implication of ﬁscal federalism. In addition, since optimal policy levels change in time (because the ‘needs’ of the markets and the citizens change in time), integration (and, therefore, decentralisation) should logically be seen as a reversible process. This is opposed to the political debates where integration (and decentralisation) is usually presented as a unidirectional cumulative process.

FISCAL DECENTRALISATION INDICATORS In this chapter we are not overly concerned with the validity of ﬁscal federalism as a normative theory. For our purpose, the importance of ﬁscal federalism lies in the fact that, in order to test the theory, a toolbox with concepts, techniques and indicators has been designed to analyse the vertical aspects of public ﬁnance. The measurement tools that have been developed to analyse multilevelled ﬁscal relations range from simple descriptive measures to more analytical measures. Obvious indicators include those that allow the monitoring of the relative importance of each governance level, measured on the income or expenses side. Usually, complex and country-specific combinations of transfer and borrowing mechanisms are in place within federations in order to preserve the political stability of federal systems and to allow all the components of the government system to perform their assigned tasks adequately. These arrangements are the consequence of the existence of so-called vertical and horizontal imbalances. Vertical imbalances follow from an unequal distribution of revenues and expenditures (ﬁscal capacity) among various levels of government within a federation. Horizontal imbalances reﬂect the differences between ﬁscal capacities of a given governance level (Ahmad and Craig, 1997: 76). These are equally relevant from a ﬁscal federalism perspective, because it is often both technically and politically easier to address these imbalances when a superior governance level (and thus vertical payments) is involved than through a – theoretically possible – horizontal system of equalisation payments. Cross-country comparisons are made difficult because of differences in the institutional and legal frameworks, data quality and availability, public accounting principles and practices.7 The work done by the World Bank in constructing

136 Philippe De Lombaerde and Ana-Cristina Costea Table 7.1 The World Bank’s list of ﬁscal decentralisation indicators Categories

Indicators

Main indicators

Sub-national expenditures (% of total expenditures) Sub-national revenues (% of total revenues) Vertical imbalance (%) Sub-national expenditures (% of GDP) Sub-national revenues (% of GDP)

Composition of sub-national revenues and grants

Tax revenue (% of total sub-national revenues and grants) Transfers from other levels of government (% of total sub-national revenues and grants)

Composition of sub-national expenditures

General public services (% of total sub-national expenditures) Defence (% of total sub-national expenditures) Public order and safety (% of total sub-national expenditures) Education (% of total sub-national expenditures) Health (% of total sub-national expenditures) Social security and welfare (% of total sub-national expenditures) Housing and community amenities (% of total sub-national expenditures) Recreational, cultural and religious affairs and services (% of total sub-national expenditures) Fuel and energy (% of total sub-national expenditures) Agriculture, forestry, ﬁshing and hunting (% of total sub-national expenditures) Mining and mineral resources, manufacturing and construction (% of total sub-national expenditures) Transportation and communication (% of total sub-national expenditures) Other economic affairs and services (% of total sub-national expenditures) Other expenditures (% of total sub-national expenditures)

Source: World Bank (sd).

indicators and setting up a database has been very important in this respect (World Bank, sd). Table 7.1 presents an overview of the indicators included in the Bank’s database. They are organised in three categories: 1 2 3

main indicators, indicators showing the composition of sub-national revenues and grants, and indicators showing the composition of sub-national expenditures.

FISCAL INTEGRATION INDICATORS: A PROPOSAL In this section we present a list of possible ﬁscal integration indicators. They have been inspired by the decentralisation indicators of the World Bank, in turn derived from IMF’s Government Finance Statistics.

Comparative ﬁscal integration indicators 137 Supra-national expenditures (as per cent of total expenditures)

=

TOTEXPr − CURTRAr − CAPTRAr *100 ∑ TOTEXPj − CURTRAj − CAPTRAj

j =l , p ,c ,r

where: TOTEXP = total expenditures CURTRAj = current transfers from level j to all other governance levels CAPTRAj = capital transfers from level j to all other governance levels l = local level p = provincial (or departmental, regional) level, i.e. refers to the intermediate sub-national levels of governance c = central (national) level r = regional (supra-national, macro-regional) level. Correspondingly, national expenditures (as percentage of total expenditures)

=

∑

TOTEXPi − CURTRAi − CAPTRAi

∑

TOTEXPj − CURTRAj − CAPTRAj

i =l , p ,c

*100

j =l , p ,c ,r

⎞ ⎛ TOTEXPr − CURTRAr − CAPTRAr ⎜ = 100 − * 100⎟ ⎜⎜ ∑ TOTEPj − CURTRAj − CAPTRAj ⎟⎟ ⎝ j=l,p,c,r ⎠ Regional revenues (as percentage of total revenues)

=

TOTREVr * 100 ∑ TOTREVj

j =l , p ,c ,r

where: TOTREV = total revenue. Correspondingly, national revenues (as percentage of total revenues)

∑

⎛ ⎞ TOTREVr ⎜ = * 100 = 100 − * 100⎟ ⎜⎜ ∑ TOTREVj ⎟⎟ ∑ TOTREVj ⎝ j =l , p ,c ,r ⎠ j =l , p ,c ,r i =l , p ,c

TOTREVi

138 Philippe De Lombaerde and Ana-Cristina Costea Vertical imbalance (per cent)

∑

TRAi, r

i =l , p ,c

=

∑

TOTEXPi

* 100

i =l , p ,c

where: TRAi,r = transfers from levels i to the regional governance level. Supra-national expenditures (as percentage of regional GDP)

TOTEXPr * 100 ∑ GDPc

=

c=1..C

where: C = number of countries in the regional integration grouping. Supra-national revenue (as percentage of regional GDP)

=

TOTREVr * 100 ∑ GDPc c =1..C

For each level of government, the composition of expenditures (per cent)

EXPs * 100 TOTEXP

= where:

EXPs = speciﬁc expenditures category. As far as the composition of revenues is concerned, the following indicators are relevant: Tax revenue (as percentage of total revenues) at the regional level

=

TAXREVr * 100 TOTREVr

Transfers from other levels of government (as percentage of total revenues)

Comparative ﬁscal integration indicators 139

=

∑

TRAi, r

i =l , p ,c

TOTREVr

* 100

Horizontal imbalances (per cent), measured in terms of transfers

=

∑ ∑ TRAi, j

i =1..C j =1..C

∑ TOTEXPi

* 100

i =1..C

Alternatively, horizontal imbalances (as percentage of expenditures) in terms of ﬁscal balances

=

∑

i =1..C

TOTREVi − TOTEXPi

∑ TOTEXPi

* 100

i =1..C

Similar indicators can be constructed, substituting TOTEXP by TOTREV or GDP in the denominator. FISCAL INTEGRATION INDICATORS: ISSUES The construction of a comprehensive database with comparable ﬁscal data at the regional level and the calculation of the proposed fiscal integration indicators is not an easy task. From a first exploration of available data we arrived at the following set of considerations on the possible scope of such an exercise and the problems that are likely to present themselves.8

Regional integration and public ﬁnance: initial remarks Looking at regional integration through a budgetary lens allows us to have an idea of the relative political importance of the regional governance level, the degree of actorness of the regional organisation and the degree of intra-regional cohesion and solidarity (horizontal imbalances), and is therefore an interesting complement to existing comparative indicators of regional integration processes.9 It makes the regional integration processes also more transparent which, in turn, might increase participation in the processes and hence their sustainability. The ﬁscal lens allows us to get an idea of the functions that regional ﬁnances are fulﬁlling; for example, following Musgrave’s categorisation: allocative, redistributive or stabilising function. The ﬁscal lens also reﬂects the institutional architecture at the regional level. The existence and development of a supra-national institutional framework, the establishment of a customs union, the creation of a common agricultural policy, etc. all have a crucial incidence on ﬁscal integration.

140 Philippe De Lombaerde and Ana-Cristina Costea The degree of fiscal autonomy of the regional institutions (or, importance of ‘own resources’) is also a crucial factor. In the real world, the supra-national level draws most or all of its revenue from transfers and not from regional taxes. It is likely, however, that with the further development of the regional government level the ‘tax-assignment problem’ (McLure, 1983) in the fiscal federalism literature will receive also more attention in regional integration studies. Finally, compared to the national and sub-national governance levels, regional governance levels tend to be less stable. In addition, there is a trend towards more instability and a more complex distribution of shared competences (Laffan, 1997: 19). This makes a systematic monitoring and the comparison of figures crosscountry and over time more difﬁcult.

Regional budgets complexity and the type of regional organisation Regional budgets have also evolved according to the ‘wave’ or ‘generation’ of regionalism to which the organisation belongs. In the ‘ﬁrst generation’ of regionalism – based upon the idea of a linear process of economic integration involving the combination of separate (national) economies into larger economic regions – the regional budget was mostly limited to the expenses related to the functioning of a small Secretariat and a Council of Ministers, with the possibility of creating supplementary budgets approved by the Council to meet extraordinary expenditures of the Secretariat.10 In the ‘second generation’ regional integration or ‘new regionalism’, integration can also imply other matters than pure trade creation such as social policies, political integration, justice, security, culture. The creation of new political institutions with autonomous status and often a smaller membership than the main bodies of the regional organisation – such as regional Parliaments and regional Courts of Justice – has consequences for ﬁnances at the regional level involving the creation of additional budgets, which often operate through different rules than the core budget of the community/regional organisation. As an example, while the core Southern African Development Community (SADC) budget is financed through member states’ contributions based on a formula combining GDP and the population of one country, the SADC Parliamentary Forum – an autonomous regional institution of the Southern African Development Community – is financed through annual and mandatory equal contributions of US30,000 by each of the 12 member Parliaments of the Forum. In some cases, following a monitoring of the evolution of the new institutions, measures are taken in order to allow these new budgets to be integrated into the main budget of the regional organisation in question. Also, with the move towards increased integration, many regional organisations have followed the European model with the creation of funds for cohesion, infrastructure development, and project ﬁnancing next to the core budget of the regional organisation, with a different governance structure and membership. Examples include: the Fund for Cooperation, Compensation and Development of

Comparative ﬁscal integration indicators 141 the Common Market for Eastern and Southern Africa (comprising an Infrastructure Fund and an Adjustment Facility); the Fund for Cooperation, Compensation and Development of the Economic Community of West African States; the South Asian Development Fund, the ASEAN Fund (consisting of a Seed Fund Account invested to realise income for the ASEAN Fund and a Projects Fund Account). Not much analysis has been performed yet on the budgetary effects of these funds for the different member countries of a regional organisation given their relatively recent implementation in most cases as well as the difﬁcult task of tracking country shares in multinational projects.

Vertical and horizontal imbalances Theoretically, vertical imbalances are not necessarily accompanied by horizontal imbalances, although in practice this will usually be the case. Horizontal imbalances refer to differential fiscal capacities of states (within a regional scheme), which can be accompanied by equalisation transfers between states or from (via) the macro-regional level. In a national context, horizontal imbalances are only measured in very few cases (e.g. Australia, Canada, Germany) (Ahmad and Craig, 1997: 76). Thus, in the fiscal federalism literature, the concept of horizontal imbalances does not refer to the redistributive role of the centre (national state) in general; it is a more speciﬁc concept. The discussion on net contributions in the EU, for example, is therefore broader in scope. It should also be noted that establishing the redistributive or equalising role of regional finances, supposes the choice of an evaluation criterion. Different criteria are available (per capita GDP, GDP, population, ﬁscal base, etc.) and can be picked according to political inclinations and agendas. Horizontal ‘equilibrium’ or ‘neutrality’ are therefore not unique concepts. The measurement of horizontal imbalances is difﬁcult and depends on whether one focuses on ﬁscal capacities or on equalising payments, and whether one only includes direct (visible) ﬂows or also indirect (invisible) ﬂows, and whether one includes only horizontal ﬂows or also ﬂows that pass through the regional level. In existing regional integration schemes, horizontal payments are (still) practically non-existent. As is the case in certain federal countries like Australia and Canada, when an integrated system of equalisation grants exists and simultaneously addresses vertical and horizontal imbalances (Ahmad and Craig, 1997: 77), it might be very difﬁcult to disentangle the two types of imbalances.

‘Net’ contributions This is a particularly sensitive issue and an issue for domestic electoral use. It has certainly been a sensitive one in Europe. In the European case, the issue has caused many discussions, especially originating in ‘net contributing countries’ like the UK, Germany and the Netherlands (Laffan, 1997: 18, 47–60). It is technically difﬁcult to calculate the value added of a regional integration scheme and the ‘net contribution’ or ‘net beneﬁt’ for individual member states.

142 Philippe De Lombaerde and Ana-Cristina Costea Many economists will share the view that a calculus based on ﬁscal data does not make much sense, provided that a large part (if not the largest part) of the benefits of regional integration is simply not visible when looking at the fiscal accounts. The beneﬁts of free trade and capital movements, of more competition, of larger domestic markets, of more political stability, etc. cannot be assessed in fiscal terms. And even if one looks at the fiscal picture, indirect effects, which should not be underestimated, are not captured.11 If a calculation of net contributions is undertaken anyway, several problems appear. On the income side, the calculus is particularly difficult when own (regional) resources exist. On the expenses side, although part of the expenses at the regional level is assignable to member states, part is necessarily general in nature.

In-kind contributions For a number of regional organisations, in-kind contributions from, for example, the host country of the regional organisation’s headquarters or from the regional hegemon, although of limited importance in absolute terms, can be important in relative terms, especially when the regional organisations have limited budgets. These in-kind contributions can cover secondment of (national or sub-national) government officials, provision of facilities and support for functioning of the headquarters and other regional institutions, tax benefits or exemptions for the regional organisation, etc. Failing to take these contributions into account distorts the ﬁscal picture.

Extra-regional funding External funding for regional organisations can be important, especially for developing countries. The assistance received from donors towards the ﬁnancing of various programmes can signiﬁcantly inﬂuence the performance of the regional organisation. As an example, ECOWAS received in 2002 external funding from the Japanese government, the Canadian government (child protection), USAID (gas pipeline project), the ADB, the European Union and IFAD (ECOWAS/ UEMOA project). In other cases, funds with single country contribution from inside or outside the organisation are sometimes set up under the name of the regional organisation in order to answer different needs: e.g. the ASEAN Cultural Fund of ﬁve billion yen ﬁnanced by the Japanese government and managed by two banks established in ASEAN countries.

Transparency, public accounting standards and data availability The comparison of fiscal integration indicators is complicated by differences in public accounting standards and differences in the categorisation of revenue

Comparative ﬁscal integration indicators 143 and expenses, both cross-country and from a vertical perspective. When different budgetary rules are followed from one regional organisation to another (for example, rules for carrying over commitment and/or payment appropriations), comparison is made more difﬁcult. If there is no rule of unity, the coexistence of several regional organisations will complicate the compilation of data on total budgets. This consolidation is usually not undertaken by regional organisations. On the other hand, when budgetary rules exist (and are followed), they can be used to check the quality of the data.12 Except for the EU, few other organisations have tried to increase transparency by publishing their budgets systematically on websites or in the Annual Reports issued by the Secretariats. In the cases where such reports exist, often there is no detailed breakdown which might allow for a ﬁscal federalism analysis based on country contributions and receipts from the regional budget. In a few cases, the annual budgets of the organisations are presented online by the audit firms in charge of assessing the regional organisation in question. Reports by the international ﬁrms of external auditors, such as PriceWaterhouseCoopers13 for SICA and Ernst&Young for the Association of Caribbean States, are useful tools for research. General materials such as the Yearbook of International Organisations, published by the Union of International Associations14 give only the total budget of regional organisations It should be noted, ﬁnally, that the role of regional parliaments in controlling the regional organisations’ budgets, is in most regions still non-existent or – at best – in its early stages.

CONCLUSIONS The toolbox of concepts and indicators that has been produced as a by-product of the fiscal federalism approach, in order to empirically assess vertical fiscal relations at the country level, can be adapted to cover also the supra-national (macro-regional) level without major problems. These types of ﬁscal integration indicators would be a welcome complement to existing indicators of economic integration and interdependence. However, a ﬁrst exploration of the possibilities of constructing such a system of indicators encounters a series of problems. These are related to the existence of different regional institutional architectures and ﬁnancing arrangements, accounting standards, different data qualities, different degrees of extra-regional funding, and so on. The introduction of a uniform reporting format should be encouraged.

NOTES 1 For an exploratory application of ﬁscal federalism to the global level see, for example, Boadway (2003). 2 See also Buchanan and Goetz (1992).

144 Philippe De Lombaerde and Ana-Cristina Costea 3 For critical positions on this point see, for example, Anderson and Tollison (1988), Flowers (1988) and González (1994). 4 On this issue in a developing country context see also Shah (1994). 5 There was, for example, a clear inﬂuence of ﬁscal federalism on the MacDougall Report (1977) and the Sound Money: Stable Finances Report of 1993 (Laffan, 1997: 29). 6 Bureau and Champsaur (1992) raised the issue of redistribution policies in a macroregion (the EU) characterised by increasing taxpayer mobility. 7 See also Ter-Minassian (1997). 8 We analysed ﬁscal data for the EU, COMESA, SADC, ACS, SICA, APEC, ASEAN and Paciﬁc Islands Forum. 9 On the relevance of a ﬁscal approach to regional integration in the Europan context, see Laffan (1997: Chapter 2). 10 See, for instance, COMESA Treaty, Chapter 30, Financial Provisions, Art. 166 §7. ECOWAS Treaty, Art. 71. 11 For example, structural funds ﬂowing from the regional level to Spain might well be translated into beneﬁts for Germany in the form of exports of capital goods, a more (politically) stable export market, etc. 12 See Laffan (1997: 36–40) on the European case. 13 PriceWaterhouseCoopers, Estados ﬁnancieros de la Secretaría General del Sistema de la Integración Centroamericana, 2002, San Salvador, 28 March 2003. 14 The Union of International Associations, Yearbook of International Organisations, Vol. 1, 1998–1999, München: KG Saur.

REFERENCES Ahmad, E. and Craig, J. (1997) ‘Intergovernmental Transfers’, in T. Ter-Minassian (ed.), Fiscal Federalism in Theory and Practice, IMF, Washington DC, 73–107. Alesina, A. and Spoalore, E. (2003) The Size of Nations, Cambridge (Mass.): MIT Press. Anderson G. and Tollison, R. (1988) ‘Legislative Monopoly and the Size of Government’, Southern Economic Journal, 54(3): 529–45. Boadway, R. (2003) ‘National Taxation, Fiscal Federalism and Global Taxation’, WIDER Discussion Paper (WDP 2003/87). Brennan, G. and Buchanan, J. (1977) ‘Towards a Tax Constitution for Leviathan’, Journal of Public Economics, 8: 225–73. Brennan, G. and Buchanan, J. (1978) ‘Tax Instruments as Constraints on the Disposition of Public Revenues’, Journal of Public Economics, 9: 301–18. Brennan, G. and Buchanan, J. (1980) The Power to Tax: Analytical Foundations of a Fiscal Constitution, Cambridge: Cambridge University Press. Buchanan, J. (1980) ‘Federalism and Fiscal Equity’, in Grewal et al., The Economics of Federalism, Australian National University Press. Buchanan, J. and Goetz, C. (1992) ‘Efﬁciency Limits of Fiscal Mobility: An Assessment of the Tiebout Model’, Journal of Public Economics, 23: 25–43. Bureau, D. and Champsaur, P. (1992) ‘Fiscal Federalism and European Economic Uniﬁcation’, American Economic Review Papers and Proceedings, 82: 88–92. Casella, A. (1994) ‘Trade as an Engine of Political Change: A Parable’, Economica, 61: 267–84. Casella, A. and Frey, B. (1992) ‘Federalism and Clubs: Towards an Economic Theory of Overlapping Political Jurisdictions’, European Economic Review, 36: 639–49. Cooper, R. (1977) ‘Worldwide versus Regional Integration. The Optimum Size of the

Comparative ﬁscal integration indicators 145 Integrated Area’, in F. Machlup (ed.), Economic Integration Worldwide, Regional Sectoral, International Economic Association – Macmillan, London, 41–53. Cooper, R. (1995) ‘Worldwide Regional Integration: Is There an Optimal Size of the Integrated Area?’, in R. Garnaut. and P. Drysdale (eds), Asia Pacific Regionalism. Readings in International Economic Relations, Sydney: Harper Educational, 11–19. Flowers, M. (1988) ‘Shared Tax Source in a Leviathan Model of Federalism’, Public Finance Quarterly, 16(1): 67–77. González, J.I. (1994) ‘Un ordenamiento territorial de corte ﬁscalista’, in N.A. Cifuentes et al., Diez años de descentralización. Resultados y perspectivas, Bogotá: FESCOL, 99–126. Inman, R.P. and Rubinfeld, D.L. (1997a) ‘The Political Economy of Federalism’, in: D. Mueller (ed.), Perspectives on Public Choice: A Handbook, Cambridge: Cambridge University Press, 73–105. Inman, R.P. and Rubinfeld, D.L. (1997b) ‘Rethinking Federalism’, Journal of Economic Perspectives, 11(4): 43–64. Laffan, B. (1997) The Finances of the European Union, Basingstoke: Macmillan Press. McKinnon, R.I. and Nechyba, T. (1997) ‘Competition in Federal Systems: The Role of Political and Financial Constraints’, in J. Ferejohn and B. Weingast (eds), The New Federalism: Can the States be Trusted?, Stanford: Hoover Institution Press, 3–61. McLure, C.E. Jr. (ed.) (1983) Tax Assignment in Federal Countries, Canberra: Australian National University Press. Musgrave, R.M. (1959) The Theory of Public Finance, New York: McGraw-Hill. Oates, W.E. (1972) Fiscal Federalism, New York: Harcourt Brace Jovanovich. Oates, W.E. (1999) ‘An Essay on Fiscal Federalism’, Journal of Economic Literature, XXXVII: 1120–49. Prud’homme, R. (1995) ‘On the Dangers of Decentralisation’, The World Bank Research Observer, 201–10. Samuelson, P. (1954) ‘The Pure Theory of Public Expenditures’, Review of Economics and Statistics, XXXVI(4): 387–89. Samuelson, P. (1955) ‘Diagrammatic Exposition of a Pure Theory of Public Expenditures’, Review of Economics and Statistics, XXXVII(4): 350–56. Shah, A. (1994) The Reform of Intergovernmental Fiscal Relations in Developing and Emerging Market Economies, World Bank, Washington DC. Ter-Minassian, T. (1997) ‘Intergovernmental Fiscal Relations in a Macroeconomic Perspective: An Overview’, in T. Ter-Minassian (ed.), Fiscal Federalism in Theory and Practice, IMF, Washington DC, 3–24. Tiebout, C. (1956) ‘A Pure Theory of Local Expenditures’, Journal of Political Economy, 64: 416–24. Tiebout, C.M. (1961) An Economic Theory of Fiscal Decentralization in Public Finance: Needs, Sources and Utilization, Princeton NJ: Princeton University Press. Tinbergen, J. (1954) (2nd edn 1964) International Economic Integration, Amsterdam: Elsevier. Weingast, B.R. (1995) ‘The Economic Role of Political Institutions: Market-Preserving Federalism and Economic Development’, Journal of Law, Economics and Organisation, 11: 1–31. World Bank (sd) Fiscal Decentralisation Indicators, World Bank, Washington DC.

8

Indicators of real economic convergence José Villaverde Castro

INTRODUCTION Whether poor economies tend to converge towards rich ones or else to diverge over time is an issue that has attracted the attention of policy-makers and academics alike for some decades. Economic convergence or divergence is a topic of considerable interest and debate, not only for validating or otherwise the two leading and competing growth models (the neoclassical and the endogenous growth approaches) but also for its policy-oriented implications. Generally speaking, the presence of convergence is considered as a valid test in favour of the neoclassical growth model as opposed to the endogenous models that predict divergence in most cases. The convergence issue has been revived in the past two decades thanks to the seminal works by Abramovitz (1986), Baumol (1986) and Barro and Sala-i-Martin (1991, 1992), to name just a few. Convergence analysis has even gained momentum by the development of endogenous growth models that allow a major role for economic policy. Unfortunately, the available empirical evidence does not provide unambiguous support for either of the two aforementioned growth frameworks, although it does point to the existence of conditional convergence. At the regional level, the issue of economic convergence is also much debated. A look at growth performances of member countries of integration schemes in Europe, Asia and Latin America over a relatively long period of time does not show unambiguous empirical support for the convergence hypothesis either (De Lombaerde, 2002). This is in sharp contrast with the fact that, as in the case of the EU, for the majority of commentators, economists and politicians economic convergence is an expected (if not necessary) outcome of regional economic integration processes (Meeusen and Villaverde, 2002b). The empirical assessments of the effects of the elimination of barriers to ﬂows of goods, services and production factors on growth (and its spatial distribution) and the debate on cohesion policies in the EU have considerably contributed to the development of the conceptual framework and the methodological toolbox for studying economic convergence.1 According to the Oxford Dictionary, ‘convergence’ is a tendency to become similar or identical. Although this definition is somehow illuminating, when

Indicators of real economic convergence 147 economists talk about convergence they usually refer to what is known either as nominal or real convergence.2 Nominal convergence relates to the process of nominal variables approaching stability levels. It is mainly concerned with such things as, for instance, convergence in interest rates and inﬂation rates, but also stability in the exchange rate, the government deﬁcit and/or the government debt, both as a percentage of GDP. Broadly speaking, real convergence is understood to mean the approximation in the levels of economic welfare or development across economies. Thus, real convergence relates mainly to the time performance of variables such as per capita income, productivity, unemployment rate and so on.3 In its simplest form, real convergence implies a long-run tendency towards the equalisation of per capita income levels across economies (Abramovitz, 1986). Nevertheless, the relevant point is that real convergence is a multifaceted concept that, as it has been posed by Quah (1997: 27) ‘reﬂects on – among other things – polarisation, income distribution and inequality’. The focus of this chapter is on real convergence.4 That is with the long-term process of reducing inequalities across economies. One difﬁculty is that because there are different concepts of convergence, there are also different ways to measure it. The two most popular approaches, according to the terminology coined by Barro and Sala-i-Martin (1991, 1992), are the so-called σ -convergence and β-convergence. Albeit different, these two concepts are not only complementary (in that they respond to different questions), but, as Sala-i-Martin (1996) has shown, they are also closely related: some type of β-convergence is a necessary but not sufﬁcient condition for σ-convergence. This chapter will not only explore σ-convergence and β-convergence approaches, but will also examine other less standard concepts like convergence as reduction of inequality, stochastic (or time-series) convergence, dynamics distribution convergence and spatial analysis convergence.5 The chapter is organised as follows. The following section reviews convergence from the point of view of the inequality literature; this includes the σ-convergence approach. The next section portrays a simple description of the well-known β-convergence approach. Subsequent sections contain the main traits of the timeseries approach to convergence and a general view of the spatial analysis convergence. In all these four sections, the objective is how best to measure real economic convergence. The ﬁnal section gives a summary and presents the conclusions.

CONVERGENCE AS A REDUCTION OF INEQUALITY The simplest concept of economic convergence refers to the reduction of per capita income inequality across a sample of economies (countries, regions, states, provinces, etc). In order to measure it, a whole array of inequality indicators has been proposed, the three most popular being some summary measure of dispersion, the Gini index and the Theil index. In particular, it is said that there exists σ-convergence if the dispersion of the income per capita in a cross-section of economies diminishes over time. In

148 José Villaverde Castro other words, if σt+T < σt, where σ is the indicator of dispersion, there exists σconvergence. To test for σ-convergence either the variance, the standard deviation or the coefﬁcient of variation6 are conventionally used. By far the most frequently employed summary statistics for measuring dispersion are the variance and the standard deviation. However, these two indicators are unsatisfactory descriptive measures of dispersion in that their value is related to the units of measurement. Instead, the coefﬁcient of variation is independent of the units of measurement, and this is why it is used here as an indicator of σ-convergence. As is well known, the conventional coefﬁcient of variation, CV1, is deﬁned as

CV1 =

V1 Y

(8.1)

where V1 stands for the variance

⎡ 1 ⎢V1 = N ⎣

∑ (Yi − Y ) N

i =1

2

⎤ ⎥, ⎦

Yi the per capita income of the i-th economy, Y᎐ the mean per capita income and N the size of the population sample. CV1 is a non-weighted measure of inequality as it does not incorporate the population size of each economy. This index can be transformed into a new weighted one (CV2) by simply employing a weighted variance V2 given by the expression

V2 =

1 N

∑ (Yi − Y ) ( pi − p), N

2

i =1

where pi refers to the population’s share of the i-th economy. Thus

CV2 =

V2 Y

(8.2)

A very popular inequality indicator is the Gini coefﬁcient. This is derived from the Lorenz curve (Figure 8.1), a cumulative frequency curve which plots the cumulative share of population of the economies in the sample on the X-axis and the cumulative share of total income of these same economies on the Y-axis. In both cases the economies are ranked according to their per capita income from bottom to top. The Gini index (G) corresponds to twice the area between the Lorenz curve and the 45º line. Its general expression is given by

G=

1 2Y

N

N

∑ ∑ pi pj Yi − Yj i =1 J =1

(8.3)

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Cumulative % of income

Indicators of real economic convergence 149

Cumulative % of population

Figure 8.1 Lorenz curve

where all symbols have already been deﬁned. The Gini coefﬁcient ranges from 0 (complete equality) to 1 (complete inequality). The Theil inequality index, on the other hand, is derived from the notion of entropy in information theory. The T(1) version is given by the expression

T (1) =

1 N

N

⎛Y ⎞

⎛Y ⎞

∑ ⎜⎝ Yi ⎟⎠ log ⎜⎝ Yi ⎟⎠

(8.4)

1

Among other interesting properties,7 the Theil index T(1) is additively decomposable,8 meaning that – among other possibilities – it can be decomposed in a way such that K K ⎛Y ⎞ T (1) = ∑ skT (1)k + ∑ sk log ⎜ k ⎟ ⎝Y ⎠ k =1 k =1

(8.5)

where the ﬁrst term describes inequality within each of the K population groups in the sample and the second term measures inequality between these groups; sk is the income share of group k in total income. Finally, it is interesting to consider the notion of polarisation. Although this notion is closely related to that of inequality, it is different from it. According to Esteban and Ray (1994: 821) the concept of polarisation emerges because ‘the axioms of inequality measurement (. . .) fail to adequately distinguish between “convergence” to the global mean and “clustering” around local means’. Polarisation – a concept used to compare the homogeneity of a group with the

150 José Villaverde Castro

1

2

3

4

5

6

Figure 8.2 Per capita income distribution – polarisation versus inequality

overall heterogeneity of the sample population (across groups) – refers to the formation of clusters around local poles. The difference between inequality and polarisation can be easily applied to the case of per capita income distribution. For example, assume that this distribution is uniform over income levels 1 to 6 (Figure 8.2). Now, consider a transformation that causes the income of all. The economies with an income level between 1 and 3 to converge to 2 and the income of all the economies with an income level between 4 and 6 to collapse to 5. Although the Theil index will show a decline of inequality, polarisation (clustering around two poles, the poor economies at an income level of 2 and the rich economies at income level 5) will increase. A simple index of polarisation (PI) is given by: N

N

PI = ∑ ∑ pi1+α pj Yi − Yj

(8.6)

i =1 j =1

where pi (j) stands for the weight of population of economy i(j) in the sample and α is an index between 1 and 1.6 that measures the polarisation sensitivity. The smaller the sensitivity to polarisation, the closer the notion of polarisation is to that of inequality. CROSS-SECTION CONVERGENCE (β-CONVERGENCE) The concept of β-convergence implies that poor economies grow faster than rich economies. Accordingly, a statistical negative relationship between the growth rate of per capita income and its initial level is expected to be found in a cross-section

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Indicators of real economic convergence 151 analysis of these economies. Neoclassical growth theory (Solow, 1956), assuming that technology exhibits diminishing returns to physical and human capital, predicts β-convergence:9 this is because the further an economy is away from its steady-state level of capital per capita, the faster is the growth of capital and income levels. The neoclassical model not only predicts convergence of each economy to its steady-state level but also that the speed of convergence is inversely related to the distance from its steady-state position or long-run equilibrium. β-convergence may be absolute or conditional. The difference is related to the existence of key parameters in the economies that compose the sample: technology, savings rate, population growth rate and depreciation rate. If these parameters are identical for all the economies, neoclassical growth theory predicts absolute (or unconditional) β-convergence. This is often referred to as catchingup theory. If, on the contrary, the aforementioned parameters differ across economies, neoclassical growth theory predicts conditional β-convergence. In the ﬁrst case, all the economies tend to have the same per capita income in the steady-state whereas in the second, each economy tends to its own steady-state or long-run equilibrium.10 When some economies do not converge to the same steady-state but certain groups of countries converge to a particular steady-state, it is said that they form a convergence club. Within each club, economies converge to each other, but convergence does not happen across different clubs. Chatterji (1993) and Chatterji and Dewhurst (1996) have proposed a simple methodology to estimate the existence of convergence clubs (see below). As has been shown by Barro and Sala-i-Martin (1995) the transitional growth process in the neoclassical model can be represented as

(1/T )log(YiT /Yi 0 ) = α + β log(Yi 0 ) + ε iT

(8.7)

where i indexes the economy, Y is, once again, per capita income, 0 and T are the initial and final year of the sample and ε is the error term. Thus, typically, the existence of absolute β-convergence is tested by regressing the growth in per capita income on its initial level for the cross-section of economies under consideration.11 If β <0 and is statistically signiﬁcant, it is inferred that there is β-convergence;12 if, on the contrary, β >0, then it is said there is β-divergence. It can be shown that β = [(1–ebT)/T], where b represents the rate of convergence. That is, b indicates the speed at which per capita income approaches its steady-state level. The number of years needed to ﬁll in half the gap to the steady state is then given by the expression 0.5 = exp–bT. The regression approach generally used to test the existence of β-convergence has been criticised by some economists – the seminal papers are Friedman (1992) and Quah (1993a). They point out that the aforementioned technique can produce biased estimates of β-convergence because it is subject to the Galton’s fallacy. To solve the problem, Friedman (1992) proposes the use of the coefficient of variation, suggesting that it provides an unbiased estimate of β-convergence; this is the notion of convergence that has been previously referred to as σ-convergence.

152 José Villaverde Castro Quah (1993a) proposes a completely different measure of convergence (the analysis of the dynamics of evolving cross-country distributions of per capita income), which will be discussed below. Another proposal to solve the bias of the regression approach has been put forward by Boyle and McCarthy in two papers (1997, 1999). These authors propose an indicator of rank concordance (Kendall’s index of rank concordance), labelled as γ-convergence, which, in addition to σ-convergence, offers a good test for β-convergence. This new indicator also has the virtue of capturing the changes in the ordinal ranking of the economies in the per capita income distribution. The index can be computed either in a multi-annual version or in a binary version. In the second case, the index is given by the expression

γ = Variance (RPCYiT + RPCYi 0 )/Variance ( 2 * RPCYi 0 )

(8.8)

where RPCYi. refers to the ranks of per capita income of the i-th economy. This rank concordance index ranges from 0 to 1, the closer the index is to 0 the greater the mobility within the income distribution. Simple regression models of the type shown in equation (8.7) rule out the possibility that some particular groups of economies may form a convergence club. As was pointed out above, there exists a convergence club when a speciﬁc group of economies tend to converge among themselves but diverge from economies belonging to other group(s). Fortunately, as Chatterji and Dewhurst (1996) have shown, equation (8.7) can be modiﬁed in order to incorporate the possibility of the existence of convergence clubs. This implies rewriting equation (8.7) so that the key variable is the natural log of the relative per capita income.13 That is:

(1/T )[log(YiT /Yi 0 ) − log(YAT /YA 0 )] = α + β [log(Yi 0 ) − log(YA 0 )] + ε iT (8.9) If, for simplicity, we drop α and εiT in (8.7) and denote by ziT = [log(YiT/Yi0)–log (YAT/YA0)] and by zi0 = [log (Yi0)–log(YA0)], equation (8.9) can be rewritten as

(1/T )ziT = βzi 0

(8.10)

In order to allow for the existence of multiple (k) convergence clubs, equation (8.10) can be reformulated as K

(1/T )zit = ∑ βk ( zi ,t −T )k = φ( zi ,t −T )

(8.11)

k =1

where the functional form φ depends on the data. A graph of ziT against zi0, shows the gap of each economy to the average at the ﬁnal and initial years. If this line is then compared to a 45º line (Figure 8.3), four different situations become apparent. If the initial situation is between E1 and E3 the gap to the average will tend to decline and the economy will converge to the equilibrium point E0. On the other

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Indicators of real economic convergence 153 Zit

E4

E3

45˚

E0

E1

E2

Zi0

Zt = ΦZ0

Figure 8.3 Convergence clubs

hand, if the initial gap is greater than E1 (E3), then the gap will tend to increase over time and will converge to E2 (E4). This implies the existence of multiple steady states or locally stable states. In other words, this implies the existence of different convergence clubs.

TIME-SERIES CONVERGENCE (STOCHASTIC CONVERGENCE) Standard tests of β-convergence are carried out by using cross-sectional techniques. Bernard and Durlauf (1995, 1996), for example, consider the so-called unit root or stochastic convergence. According to these authors, stochastic convergence applies if per capita income disparities between two economies follow a stationary process, with a zero mean. This implies that the two economies have reached their own steady state and that shocks are not persistent but short-lived. Obviously, when per capita income differences between these two economies contain a unit root, stochastic convergence fails to hold. There are at least three interrelated concepts of stochastic (or times-series) convergence. The concept of strong convergence – or asymptotically perfect convergence according to Bernard and Durlauf (1995) – holds when the difference between two time series Yit and Yjt, where i and j are any pair of economies within

154 José Villaverde Castro the sample, contain neither a unit root nor a time trend, either deterministic or stochastic. In formal terms

limT →∞ E(YiT − YjT ξ0 ) = 0

(8.12)

where ξ0 captures all relevant information at time T. This implies that between the two series there exists a cointegration vector (1, –1). This concept of convergence has been criticised because it is rather strict: for the strong convergence to exist it is necessary that the long-run expected value (forecast) of the per capita income differences between the two economies is equal to zero. Thus, this concept of convergence is not very useful when the objective is to determine whether convergence has taken place in the past or not. An alternative, weaker concept of stochastic convergence (sometimes called asymptotically relative convergence) implies that the aforementioned disparities between any two economies within the sample do not need to converge to zero but to a finite constant. Following Bernard and Durlauf (1996: 165), this definition ‘considers the behaviour of the output differences between two economies over a ﬁxed time interval and equates convergence with the tendency of the difference to narrow’. This can be written as

E(Yi ,T − Yj ,T ξt ) < (Yi ,0 − Yj ,0 )

(8.13)

where 0 refers to the present and T to some year in the future. According to this deﬁnition, the difference between the two time series should also be stationary, but now the time trend can be deterministic. Once again, the only cointegration vector between these two series can be (1, –1). Finally, the less strict concept of stochastic convergence holds when, although the two time series have different trends, there exists a functional relationship between them in such a way that

limT →∞ E(YiT − βYjT ξ0 ) = 0; β > 0

(8.14)

In this case, both series are also cointegrated, but the cointegration vector is now (1, – β). Although the stochastic approach to convergence solves some of the problems of the standard, cross-section convergence approach, it also has some important shortcomings. Of particular relevance to our case, and as is widely known, unit root tests (augmented Dickey–Fuller14 and/or Phillips–Perron) suffer from low statistical power in ﬁnite samples, which implies that they might lead to failures in rejecting the null-hypothesis of non-stationarity. This is closely related to the existence of gaps (or structural breaks) in the series, in which case it is possible to accept the existence of unit roots when they do not exist. Under these circumstances, it is possible to wrongly admit or reject the existence of convergence. Recently, more powerful tests (panel unit root tests) have been proposed to address the issue of low statistical power of univariate unit root tests.

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Indicators of real economic convergence 155 DISTRIBUTION DYNAMICS As has been previously mentioned, both standard σ and β-convergence approaches, but mainly the latter, have been heavily criticised on the grounds that they incur in Galton’s fallacy. This implies that negative β coefﬁcients are compatible with the absence of convergence (Quah, 1993a). Although Sala-i-Martin (1994) argues that β-convergence offers interesting insights about the mobility of income within a given distribution and that σ-convergence shows how the distribution evolves over time, it is true that standard convergence analysis is, as a general rule, uninformative about the dynamics of per capita income distribution across countries. A cross-section regression cannot capture the essence of a convergence process in that it has implicitly a dynamic component; at best, the cross-section regression approach can just approximate this process. Quah’s approach is, in this regard, much more promising in that it offers additional insights into the phenomenon of convergence by emphasising not only a particular snapshot of the income distribution but the evolution of the entire income distribution over time. In particular, the distribution dynamics approach involves estimating both the density functions associated with the entire income distribution at each point in time and tracking how it evolves over time. A density function f(y) is the mathematical counterpart of a smooth curve that represents the probability distribution of a continuous random variable Y, where y refers to a speciﬁc value of Y. The density function f(y) informs about the location, shape and other external characteristics of the income distribution Y at a given point in time. The external shape of the density function may be unimodal, bimodal or multimodal. This shows whether there is convergence, polarisation or stratiﬁcation. In the typical case of bimodality, or twin-peakedness studied by Quah (1996) there is a clear polarisation between a group of rich economies and a group of poor economics, whereas the middle-income class economies tend to disappear. Generally speaking, density functions are estimated non-parametrically by using the kernel method. This implies the substitution of the boxes in a histogram by smooth ‘bumps’. The kernel function determines the shape of the ‘bumps’, a crucial element being the bandwidth that controls the smoothness of the density. Silverman (1986) has proposed a method for selecting the bandwidth that is currently the most widely used in density function estimations. Regarding intra-distribution dynamics, a standard methodology currently employed to trace movements within a distribution is based on the computation of transition probability matrices. Thus, if FO and FT refer to the initial and final distribution, the link between them can be defined as FT = MT*FO where MT represents the transition probability matrix. Operator MT is approximated by dividing the income distribution into intervals or ‘income states’. Due to the lack of sound theoretical methods to obtain an appropriate partition of the income distribution, the selection of ‘income states’ is somewhat arbitrary. After having divided the distribution into ‘income states’, it is necessary to observe how many of the economies which are in a given income state in the initial period end up in that very state or elsewhere. Thus, each entry in the transition matrix refers to the

156 José Villaverde Castro probability that an economy in a given state income transits to another different state or stays in the initial one. The interpretation of the transition matrix is, then, as follows: the elements (probabilities: row probabilities add up to 1) on the diagonal indicate persistence while all the other elements in the matrix represent mobility. If we assume that t → ∝, the transition matrix allows us to take a longrun view of the evolution of the entire distribution, if nothing structural were to change in the system being analysed: the result is the so-called ‘ergodic’ distribution15 or long-run steady state. The main drawback of this approach is the arbitrariness when choosing the number and size of the ‘income states’. The stochastic kernel approach solves this problem by replacing the discrete income states by a continuum of states: thus, a stochastic kernel can be interpreted as the counterpart of a transition probability matrix with an inﬁnite number of rows and columns. The stochastic kernels are represented as three-dimensional diagrams and contour plots (Overman and Puga, 2002), as shown in Figure 8.4. To read the three-dimensional diagram it is useful to think of period 0(T) axis as the rows (columns) of the transition matrix. Then, starting from any point on the 0-axis, any slice parallel to the T-axis traces out a probability density, describing the likelihood of transition, over T years, into different parts of the income space, conditional on beginning at a speciﬁc income (point) at time 0. On the other side, the lines on the contour plots connect points of the same height (density) on the corresponding three-dimensional diagrams. Then, on the contour plots it is easy to understand that: 1

GDP 0

50

100

150

200

250

2

if the probability mass concentrates along the positive slope diagonal, this indicates persistence; if the probability mass concentrates along the negative slope diagonal, this indicates overtaking of the economies in the ranking;

GDP 0 GDP 0

50

100

150 GDP 0

Figure 8.4 Stochastic kernel

200

250

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Indicators of real economic convergence 157 3

4

if the probability mass runs parallel to the T-axis, this indicates that the probability of being in any state at period T is independent of their position in the initial (O) period; and if the probability mass runs parallel to axis T, this indicates convergence.

The main shortcoming of this approach is that its interpretation is not as direct and clear-cut as that of the transition probability matrix because it does not offer quantitative information about the degree of mobility (or persistence); it only offers qualitative information.

SPATIAL ANALYSIS CONVERGENCE In all the aforementioned notions of convergence, the spatial dimension of the data under consideration is completely neglected. Because these spatial effects are largely ignored, problems of model misspeciﬁcation in the previous convergence analysis may arise; in particular, OLS estimates ignoring spatial effects – as in the standard β-convergence approach – will be inefficient and/or biased (Anselin, 1988). Spatial effects refer to spatial dependence (autocorrelation) and/or spatial heterogeneity, it not being an easy task to differentiate between them from a practical point of view. Spatial autocorrelation implies that the observations (economies) in cross-sectional data are not independent. Following Anselin (1988: 11), it means ‘the existence of a functional relationship between what happens at one point in space and what happens elsewhere’. Spatial dependence can originate either as a result of a true spatial interaction among the economies (substantive spatial dependence) or as a measurement error problem (nuisance spatial dependence). Spatial heterogeneity comes from the lack of homogeneity of the economies under consideration. The statistical device usually employed to test for the presence of spatial effects (spatial dependence) is the Moran’s I statistic expressed as n

n

∑ ∑ wij yit yjt ⎛ n⎞

It = ⎜ ⎟ ⎝ s0 ⎠

i =1 j =1 n n

(8.15)

∑ ∑ yit yjt i =1 j =1

where wij is an element of a weight matrix W, n is the number of economies, yit is the log of per capita income of the economy i in year t, and s0 is equal to the sum of all elements of the matrix W.16 W is a non-stochastic square matrix of which the elements represent the intensity of the interdependence between each pair of economies; in the simplest case wij = 1 if the economies i and j are neighbours and 0 otherwise.

158 José Villaverde Castro When the Moran’s coefﬁcients are statistically signiﬁcant, they provide support for the hypothesis of global spatial effects or global spatial dependence.17 This being the case, the β-convergence approach needs to be reformulated in order to consider the effects of this spatial dependence. There are three main alternative speciﬁcations of the β-convergence unconditional equation in order to consider spatial dependence:18 the spatial error model, the spatial lag model and the spatial cross-regressive model. The first alternative takes place when the error term follows a spatial autoregressive process expressed as εt = λW εt + ut, where λ is a scalar spatial error coefﬁcient and ut is normally distributed with a zero mean and a constant variance. The β-convergence equation becomes

(1/T )log(YiT /Yi 0 ) = α + β log(Yi 0 ) + ( I − λW )uiT

(8.16)

In the second speciﬁcation of spatial dependence – coming from actual interaction among economies – the regression equation is

(1/T )log(YiT /Yi 0 ) = α + β log(Yi 0 ) + ( ρW /T )log(YiT /Yi 0 ) + εT

(8.17)

where ρ is the scalar lag parameter. Finally, in the third model speciﬁcation – where the spatial variable is the independent variable – the spatial lag of the starting incomes is added to the original β-convergence unconditional equation, which becomes

(1/T )log(YiT /Yi 0 ) = α + β log(Yi 0 ) + τW log(Yi 0 ) + ε iT

(8.18)

SUMMARY AND CONCLUSIONS This chapter has addressed the topic of how to measure real economic convergence. Importantly, it has shown that there are different concepts of convergence and that each one is in need of its own type of measurement technique. However, the concepts of σ-convergence and β-convergence continue to be the most popular methods of measurement. The former method evaluates convergence through the time evolution of a summary measure of dispersion and the latter method through the estimate of cross-section regression. As has been shown, there are other ways to evaluate the convergence process that are also illuminating. Among these are the Gini and Theil inequality measures. These indices permit the disaggregation of the convergence process (or divergence) in different components. The extension by Chatterji and Dewhurst to deal with the potential existence of convergence clubs is also particularly interesting. This same topic is addressed, although from a different perspective, in the works pioneered by Quah. Density functions not only allow the estimation of the external form of the distribution but also the potential existence of convergence clubs. This same approach, via the estimate of either transition matrices or stochastic kernels, also gives insights into the degree of mobility within the distribution. Finally, and in

Indicators of real economic convergence 159 order to take into consideration the spatial dimension of the income distribution, spatial dependence can also be included in the analysis, rendering in this way a more accurate estimate of the convergence process.

NOTES 1 For recent overviews see, for example, Meeusen and Villaverde (2002a), Funck and Pizzati (2003) and the other papers in the Special Issue on EMU and Cohesion of the Journal of Common Market Studies, 41(5), 2003. 2 The relationship between nominal and real convergence has also been the subject of considerable debate and attention. For a summary, see Viñals (1994). 3 For simplicity of notation and exposition we will always refer to per capita income convergence, although it should be evident that the same principles carry over to other relevant economic variables. 4 From now on we will always refer to real convergence. 5 A straightforward application of some of these approaches can be seen in Villaverde (2003), Villaverde and Sánchez-Robles (2002) and Villaverde and Maza (2003). 6 For an interesting review of the different conclusions that may be obtained when using different indicators of dispersion see, for instance, Dalgaard and Vastrup (2001). 7 The T(1) index, as a member of the generalised entropy family, satisﬁes the axioms of symmetry (or anonymity), population replication, mean independence (homogeneity) and Dalton–Pigou principle of transfers. In order to normalise T(1) between 0 and 1, it sufﬁces to divide T(1) by log(N). 8 Although it is less commonly used, the Gini coefﬁcient can also be decomposed in three terms: a within group term, a between group term and an interaction or residual term. For a recent reference, see Dickey (2001). 9 For a critical view of β-convergence as being considered either as a useful test of the neoclassical growth theory or a robust measure of convergence (or divergence), see Cheshire and Carbonaro (1995). Endogenous growth models (Romer, 1986 and Rebelo, 1991) consider that income convergence is just one possible outcome. The main point of these models is that, although individual inputs can be subject to decreasing returns, the positive externalities associated with some of these inputs (technology, human capital, etc.) that have public good characteristics can result in constant or increasing returns at the aggregate level. 10 Although plausible and illustrative, conditional β-convergence is also a weaker and less robust concept than absolute β-convergence. 11 The model can be modiﬁed to include in the RHS the term ψ Xi0, where Xi0 denotes a set of control variables at the initial year, with associated coefﬁcients ψ , determining the individual equilibrium rates of the economies. A negative value for β is considered to imply conditional convergence. 12 Chatterji (1993) maintains that this result is representative of weak convergence, the existence of strong convergence requiring both the existence of a steady state in which per capita income is equalised and the presence of dynamic forces driving the economy, in the long run, to this steady state. For this to happen, Chatterji (1993) shows that –2<β <0. 13 Chatterji (1993) and Chatterji and Dewhurst (1996) take as the relevant variable the log of the ratio between the per capita income of the leader economy to the per capita income of any other economy. This procedure is accurate if the leader economy does not change over time; because this may not be the case, we prefer to take as the relevant variable the log of the ratio of any per capita income economy to the average (subscript A) per capita income. 14 To run a Dickey–Fuller test on a time series Yt it is necessary to estimate the equation:

160 José Villaverde Castro Yi = α + βYt–1 + εt t = 1, . . . ,T

15 16 17 18

and run a one-sided t-test on the hypothesis that β = 1. Thus H0: β = 1; HA: β <1, so if the estimated β is signiﬁcantly less than 1, then the null hypothesis of non-stationarity can be rejected; this is tantamount to saying that the time series Yt has a unit root or is integrated of order 1: I(1). To test whether β = 1 is the same as to test whether θ = 0 in equation ∆Yi = Yt – Yt–1 = α + θ Yt–1 + εt, where θ = β–1. As has been pointed out by Quah (1993b: 431) ‘nothing . . . enforces the existence or uniqueness of an ergodic distribution’. Usually, the weight matrix is row standardised so the elements of the rows sum to 1. For other possible speciﬁcations of the weight matrix see, for instance, Cliff and Ord (1981). When the values of I are larger (smaller) than the expected value E[It(k)] = –1/(n–1), this indicates the existence of a positive (negative) spatial autocorrelation. Anselin (2003) offers a new taxonomy of formal models of spatial effects in crosssectional data.

REFERENCES Abramovitz, M. (1986) ‘Catching Up, Forging Ahead and Falling Behind’, Journal of Economic History, 46: 385–406. Anselin, L. (1988) Spatial Econometrics: Methods and Models, Dordrecht: Kluwer. Anselin, L. (2003) ‘Spatial Externalities, Spatial Multipliers and Spatial Econometrics’, International Regional Science Review, 2: 153–66. Barro, R. and Sala-i-Martin, X. (1991) ‘Convergence Across States and Regions’, Brooking Papers on Economic Activity, 1: 107–58. Barro, R. and Sala-i-Martin, X. (1992) ‘Convergence’, Journal of Political Economy, 2: 223–51. Barro, R. and Sala-i-Martin, X. (1995) Economic Growth, New York: McGraw-Hill. Baumol, W. (1986) ‘Productivity Growth, Convergence and Welfare: What the Long-Run Data Show’, American Economic Review, 5: 1072–85. Bernard, A. and Durlauf, S. (1995) ‘Convergence in International Outputs’, Journal of Applied Econometrics, 10: 97–108. Bernard, A. and Durlauf, S. (1996) ‘Interpreting Tests of the Convergence Hypothesis’, Journal of Econometrics, 71: 161–73. Boyle, G.E. and McCarthy, T.G. (1997) ‘A simple measure of β-convergence’, Oxford Bulletin of Economics and Statistics, 2: 257–64. Boyle, G.E. and McCarthy, T.G. (1999) ‘Simple Measures of Convergence in Per Capita GDP: A Note on Some Further International Evidence’, Applied Economic Letters, 6: 343–47. Chatterji, M. (1993) ‘Convergence Clubs and Endogenous Growth’, Oxford Review of Economic Policy, 4: 57–69. Chatterji, M. and Dewhurst, J.H. (1996) ‘Convergence Clubs and Relative Economic Performance in Great Britain: 1977–91’, Regional Studies, 1: 31–40 Cheshire, P. and Carbonaro, G. (1995) ‘Convergence–Divergence in Regional Growth Rates: An Empty Black Box?’, in H.W. Amstrong and R.W. Vickerman (eds), Convergence and Divergence Among European Regions, London: Pion Limited. Cilff, A. and Ord, J.K. (1981) Spatial Processes: Models and Applications, London: Pion. Dalgaard, C.J. and Vastrup, J. (2001) ‘On the Measurement of σ-convergence’, Economic Letters, 70: 283–87.

Indicators of real economic convergence 161 De Lombaerde, P. (2002) ‘Integración asimétrica’, in P. De Lombaerde (ed.), Integración asimétrica y convergencia económica en las Américas, Universidad Nacional de Colombia – Antropos, 16–32. Dickey, H. (2001) ‘Regional Earnings Inequality in Great Britain: A Decomposition Analysis’, Regional Studies, 7: 605–12. Esteban, J.M. and Ray, D. (1994) ‘On the Measurement of Polarisation’, Econometrica, 4: 819–51. Friedman, M.J. (1992) ‘Do Old Fallacies Ever Die?’, Journal of Economic Literature, 30: 2129–32. Funck, B. and Pizzati, L. (eds) (2003) European Integration, Regional Policy, and Growth, Washington D.C.: The World Bank. Meeusen, W. and Villaverde, J. (eds) (2002a) Convergence Issues in the European Union, Cheltenham: Edward Elgar. Meeusen, W. and Villaverde, J. (2002b) ‘Introduction and Outline’, in W. Meeusen and J. Villaverde (eds), Convergence Issues in the European Union, 1–5, Cheltenham: Edward Elgar. Overman, H. and Puga, D. (2002) ‘Unemployment Clusters Across European Regions and Countries’, Economic Policy, 34: 115–47. Quah, D. (1993a) ‘Galton’s Fallacy and Test of the Convergence Hypothesis’, The Scandinavian Journal of Economics, 95: 427–43. Quah, D. (1993b) ‘Empirical Cross-section Dynamics in Economic Growth’, European Economic Review, 2–3: 426–34. Quah, D. (1996) ‘Twin Peaks: Growth and Convergence in Models of Distribution Dynamics’, Economic Journal, 106: 1045–55. Quah, D. (1997) ‘Empirics for Economic Growth and Distribution. Polarisation, Stratiﬁcation and Convergence Clubs’, Journal of Economic Growth, 1: 27–59. Rebelo, S. (1991) ‘Long-run Policy Analysis and Long-run Growth’, Journal of Political Economy, 3: 500–21. Romer, P. (1986) ‘Increasing Returns and Long-run Growth’, Journal of Political Economy, 5: 1002–37. Sala-i-Martin, X. (1994) ‘Cross Sectional Regressions and the Empirics of Economic Growth’, European Economic Review, 3–4: 739–47. Sala-i-Martin, X. (1996) ‘Regional Cohesion: Evidence and Theories of Regional Growth and Convergence’, European Economic Review, 6: 1325–52. Silverman, B.W. (1986) Density Estimation for Statistics and Data Analysis, New York: Chapman and Hall. Solow, R. (1956) ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal of Economics, 70: 65–94. Villaverde, J. (2003) ‘Regional Convergence, Polarisation and Mobility in the European Union, 1980–96’, Journal of European Integration, 1: 73–86. Villaverde, J. and Maza, A. (2003) ‘Desigualdades regionales y dependencia espacial en la Unión Europea’, CLM-Economía, 2: 109–28. Villaverde, J. and Sánchez-Robles, B. (2002) ‘Convergence or Twin Peaks? The Spanish Case’, in W. Meeusen and J. Villaverde (eds), Convergence Issues in the European Union, 41–60, Cheltenham: Edward Elgar. Viñals, J. (1994) ‘Es posible la convergencia en España? En busca del tiempo perdido’, Banco de España, Documento de Trabajo, 9430.

9

Measuring and explaining levels of regional economic integration Jennifer Pédussel Wu 1

INTRODUCTION The fall of the Berlin Wall precipitated a shift in the motivation of those countries considering the possibility of integrating regionally as alliances shifted dramatically from the Cold War based bi-polar regime to a new multi-polar regime. In the past several years, more regional integration arrangements (RIAs) have been notiﬁed to the WTO than in the entire history of the GATT. As seen in Figure 9.1, there have been 130 notiﬁcations since 1995 versus 124 during the period 1948–94. Schiff (2000) has examined the proliferation of free trade agreements (FTAs) and customs unions (CUs) and found that the ratio of FTAs to CUs has increased in the 1990s, particularly in the case of North–South agreements.2 Furthermore, the characteristics of RIAs have also shifted. Fernandez and Portes (1998) note that regional trade agreements (RTAs) have not only increased in number but have also deepened and widened. A search for an explanation of the proliferation of RIAs has led to the conclusion that more than the traditional gains from trade are provided by the establishment Number of RTAS 300 250 200 Establishment of the WTO 150 100 50 0 1948 50

55

60

65

70

75

80

85

90

95

Figure 9.1 Evolution of regional trade agreements in the world 1948–2002 Source: WTO Secretariat, 2003

00 2002

Levels of regional economic integration 163 of these types of organisations. Whalley (1998) depicted RTAs as insurance arrangements. Perroni and Whalley (2000) argued that RTAs are sought by small countries as protection against a global trade war. Wu (2005) has shown a potential explanation to be that the formation of RIAs may aid in reducing risks and uncertainty associated with foreign transactions, thereby leading to expanded trade and investment. As the world shifted away from the Cold War alliance regime, the resulting uncertainty encouraged the formation of RIAs. Much empirical work has been done in the context of CUs and FTAs. This work is generally divisible into two groups: computable general equilibrium models of trade and trade policy; and econometric analysis of the impact of trade ﬂows. In most cases, FTA formation is taken to be exogenous. In contrast to this literature, Baier and Bergstrand (2004) systematically look at the economic factors that inﬂuence the likelihood of pairs forming an FTA. Looking solely at these economic characteristics, the empirical model they develop is able to correctly predict 85 per cent of the number of existing FTAs in the year 1996. Mansfield, Milner, and Rosendorff (2002) empirically ﬁnd that pairs of democratic countries are roughly four times as likely to form a preferential trade agreement. The empirical literature, however, has not generalised to all forms of regional integration, nor have noneconomic and economic characteristics been systematically examined together. Recognising the fact that one explanation may not apply to recent RIA proliferation, it is essential to examine the conditions under which such arrangements come into existence. This chapter commences a discussion of whether there are certain characteristics, economic and/or political, that contribute to the probability of a country joining a particular regional integration arrangement. A correct hypothesis depends on the ability to isolate relevant characteristics of member states. Are there common traits among RIAs? If so, what characteristics inﬂuence the respective level of RIA participation? Economic and institutional development characteristics are examined with particular attention. This chapter thus has two objectives. The ﬁrst is to isolate those common traits of RIAs and define contributing characteristics. The second is to test explicitly whether economic and/or political uncertainty is a positive contributor to the chosen level of trade integration. Four general measures of uncertainty are used in this chapter: trade uncertainty, political uncertainty, business cycle uncertainty and price uncertainty. The various measures used are described in detail. In addition, this chapter provides a more complete empirical test using the different levels of RIA participation. Data on regional integration has not been previously gathered in a systematic manner. A database of regional integration was developed which ranks the levels of integration across countries (Wu, 2001). Using this database, a probit model is developed and tested where the level of regional integration is the dependent variable. The findings of this chapter are that democracy, geographic characteristics, membership in multilateral organisations, per capita income, economic freedom and trade openness, are significant contributors to the chosen level of regional integration. Per capita income is a good indicator of whether a country increases

164 Jennifer Pédussel Wu Features

Removal of internal quotas and tariffs

Common external tariff

Free mobility of factors of production

Harmonised monetary and fiscal policy

Integration of political and economic policy

Forms Sectoral cooperation (1) Preferential or free trade area (2) Customs union (3) Common market (4) Economic union (5) Supranational union (na)

Figure 9.2 Types of regional integration arrangements

its level of integration, as deﬁned in Figure 9.2. Furthermore, global trade regime uncertainty does positively contribute to the level of regional integration.

DATA AND HYPOTHESIS

A data set of regional integration levels The data set of regional integration speciﬁes a natural ordering of regional integration levels. These ranks are sequential – economic union is not established without ﬁrst passing through a customs union.3 In addition, the step between a free trade area and a customs union is not comparable to the step between a customs union and an economic union. These different levels are presented in Figure 9.2. I construct an ordered index using the categories described in Figure 9.2. These rankings are described below. •

•

•

Regional integration agreements (RIAs) cover all of these different arrangements including those involving countries where members are not all from the same geographic region. Sectoral trading agreements are very limited. They may be thought of as reﬂecting trade cooperation rather than actual integration. Sectoral cooperation limits the agreement to well-deﬁned sectors such as grain or steel. Sectoral trading is coded as level (1) integration. Preferential trade agreements or areas (PTAs) require participants to eliminate or lower tariffs on each other’s imports. Participants are allowed to retain the ability to determine the existence and the level of restrictions to be imposed on non-members. Typically, a free trade area (FTA) involves zero tariffs between member states, although usually only extended to selected goods and services. Both PTAs and FTAs are coded as level (2) of regional integration although PTAs are not admissible under WTO rules. PTAs are

Levels of regional economic integration 165

•

•

•

•

considered shelters for inefﬁcient industries in the member countries and are viewed as harming non-member countries due to the economic loss incurred when parties to the agreement divert trade to themselves and exclude outsiders. Customs unions (CUs) are similar to FTAs except that participating countries agree to adopt uniform import tariffs and common restrictions to outside countries in the form of a common external tariff (CET). The creation of the CET aids in the negation of trade deﬂection and rules of origin issues. The establishment of the CET implies some form of common decision-making and the development of common institutions to aid in the regulation of common trade interests and the redistribution of the CET revenue. The CET is the defining factor of a customs union which is coded level (3) of regional integration. Common markets (CMs) require a much greater degree of political and economic cooperation than the three previously discussed arrangements. The free movement of the factors of production characterises a common market. In order to ensure competition on a level playing ﬁeld, these agreements are often complemented with the harmonisation of policies in health and safety, social security and education. Free movement of factors is thus the key attribute of this type of RIA which is coded as level (4). An economic union (EU) reﬂects a higher and increasingly complex level of regional cooperation. The addition of monetary and fiscal harmonisation requires the formation of supranational institutions and organisations that typically have binding decision-making power over the members. An EU also presupposes common taxation and other common procedures. While there is increased political unification in an EU, the individual sovereign states maintain their separate identities. Only the European Union countries participating in the European Monetary Union have attained this level, level (5). A supranational union covers both the political and the economic realms. Member states abandon goals of preserving national sovereignty. An example of this type of uniﬁcation is the United States or the uniﬁcation of East and West Germany, i.e. a federation. The supranational union is not included in the current data set.

Each country was coded with only the highest level of RIA in which it participates. For example, Botswana, a member of the ﬂedgling Southern African Development Community, is also a member of the South African Customs Union and thus is coded with a (3) for the latter customs union membership – the highest level of RIA in which it participates. The data set therefore captures the deepest form of regional integration to which a country belongs.4 Monetary union in this framework, unfortunately, is unable to be distinguished from common market integration. This poses a problem when coding the CFA zones in West and Central Africa. Thus, the highest level of trade integration is taken and monetary union will not override the existence of the FTA in the case of West and Central Africa.

166 Jennifer Pédussel Wu 120 100 80 60 40 20 0 1965

1975

SC

FTA

1985

CU

CM

1995

Total

Figure 9.3 Regional integration in 150 countries 1965–1995 Note: The ﬁgure shows the number of included regional integration arrangements in each year. There are a total of 102 multilateral arrangements in 1995. Note that the WTO counts each bilateral agreement whereas, here, the entire regional integration arrangement is counted as one agreement. Sectoral Cooperation (SC), Free Trade Agreement (FTA), Customs Union (CU) and Common Market (CM). Participation is determined as outlined as Figure 9.2. Source: author’s calculations.

In total, 150 countries were included in the data set. Inclusion was based on available data. Figure 9.3 shows the evolution of the levels of regional integration in the sample. The former Soviet Union and Russia are not included. Various political issues make the coding of the USSR difﬁcult.5 The majority of the ten European Union Accession Countries is included from 1989 onwards. If a country’s membership in a particular type of regional integration agreement could not be verified to have attained significance using the criteria outlined above, its membership in the RIA was not included.

Factors inﬂuencing RIA formation Measures of uncertainty One objective of this chapter is to investigate how greater political and economic uncertainty has provided an incentive for countries to increase their participation in regional integration arrangements. The change in the global structure of trading alliances, brought on by the end of the Cold War, led countries to seek multilateral agreements in an attempt to alleviate augmented risks. Furthermore, the supranational institution governing dispute settlement, the WTO, emerged in 1995. Four measures of uncertainty are examined in this chapter: trade uncertainty, political uncertainty, business cycle uncertainty and price uncertainty.

Levels of regional economic integration 167 Trade uncertainty can be measured through a country’s openness in terms of exports and imports as a ratio of gross domestic product (GDP). Rodrik (1987) has argued that the more open a country is to international trade, the more uncertainty is created for participants in the economy.6 Openness to international trade is associated with uncertainty because a country with a higher proportion of international trade dependence is more likely to have its terms of trade adversely affected by the international community. Since countries may choose trade integration as a method to mitigate risks associated with trade, we expect there to be a positive relationship between trade openness and RIA formation. There are several measures of trade openness. The traditional measure of trade openness, from the Penn World Tables 9.6 (Heston et al. 2002), is the ratio of imports and exports to GDP.7 A second measure, taken from the World Bank, uses tariffs measured by import duties as a percentage of total imports. A third measure, the Trade and Openness Index (TOI), published by Economic Freedom of the World (EFW), is designed to measure the degree to which countries’ policies interfere with international exchange. In contrast to the more traditional measures, there are four basic components to this index: tariff rates, the blackmarket exchange premium, restrictions on capital movement and the actual size of the trade sector as compared to the expected size. This index was derived for 109 countries in 1998 and slightly fewer countries in previous years. It is only available for the later part of the 1990s. In 1998, the highest-ranking countries were Hong Kong, Singapore, Estonia, Belgium, Ireland, Netherlands, Germany and Luxembourg. As with the traditional measure of openness, a positive relationship between this index and levels of regional integration is expected. Political uncertainty may be measured in a number of ways. Two potential measures, one of corruption and another of economic freedom, are presented. The ﬁrst measure is the level of corruption in a country. High levels of corruption make it more likely that countries try to use regional integration agreements to end trade risks brought about by the capricious behaviour of domestic government representatives. The corruption index published by Transparency International is a constructed ranking where the purpose is to assess the corruption level of 52 countries as perceived by business people, risk analysts and the general public. The perceptions of this group may not always be a fair reﬂection of the actual state of affairs and does not necessarily reﬂect the concrete level of corruption in the surveyed countries and thus, due to its subjectivity, trends are difﬁcult to determine. High levels of corruption, and therefore greater political uncertainty, should provide a positive incentive for regional integration. The second potential measure of political uncertainty is the index of economic freedom. This index provides a higher rating for those countries with institutions and policies that are consistent with price stability, the rule of law and secure property rights, smaller governments and free trade. Freer trade is deﬁned as lower tariffs and fewer non-tariff barriers consistent with regional integration. The EFW index relies on quantitative variables to develop its rating gradations for each component.8 Institutional improvement has been shown to be a prerequisite for growth and development and thus more secure trade. Institutional development is

168 Jennifer Pédussel Wu expected to be positively related to regional integration as countries secure property rights and continue to deepen their levels of regional integration. Several other ways of measuring political stability presented themselves. None were signiﬁcant. One potential measure is to use the mean number of revolutions. Another measure of stability is based upon the integrity of central bank policies. While both were initially included in the ordered probit analysis, neither was a significant indicator of the level of regional integration and therefore are not reported in the results section.9 Business cycle uncertainty is another type of uncertainty contributing to greater levels of regional integration. There is risk associated with the general health of the economy for individual producers as they are subject to the vicissitude of their domestic environment. A more open economy may expose a country to the vagary of other countries’ business cycles; however, by joining an RIA, the country may also afford itself the opportunity to take advantage of the ‘portfolio effect’ of all these business cycles and thus reap the benefits of more stable production and income. Indeed, Mattli (1999) finds that downturns in business cycles spur the formation of cooperative agreements. Furthermore, there is growing evidence that volatility is positively correlated with the rate of growth (this includes most Western European countries). Business cycle uncertainty is generally measured using GDP growth, calculated from t to t + 1, as a rough measure of year-to-year uncertainty. GDP comes from the Penn World Tables 9.6 data. We expect a negative relationship between positive GDP growth and the level of RIA. Price uncertainty is proxied with inﬂation rates. It has been shown that average inﬂation rates are lower in more open economies.10 Inﬂation rates are often used as measures of domestic instability, soundness of economic policies, and as a reflection of the importance of international markets. Global financial markets interact on a daily basis. It serves to discipline those countries in the developing world with extremely high inflation rates and thus discourages the adoption of domestic policies that promote price uncertainty. Inﬂation is measured as changes in the personal consumption price levels of each country. All data come from the Penn World Tables 9.6 and are on an annual basis and are PPP measures adjusted for the exchange rate. Thus far we have examined the various measures of uncertainty. In order to proceed further, we still need to present other data to be used as control variables.

Other data Democracy It is the practice of the European Union to forbid membership until the applying country has fulﬁlled three conditions: it must be a functioning market economy, have a democratic political system, and accept the acquis communautaire. A country is thus required to obtain a certain level of institutional and economic development before it is accepted as a European Union member. The importance of institutional development was discussed in the section on political uncertainty.

Levels of regional economic integration 169 Whether a country is a democracy is a measurable characteristic. As previously cited, there is evidence in the political economy literature that democracies are more inclined to join an RTA (Rosendorff (2000), Mansfield et al. (2002)). Democracy is measured on a 0–10 scale and is the institutionalised democracy score given in the Polity IV data set (see Marshall et al. 2002). A second measure for institutional development, highly correlated with democracy, is the measure for polity, also from the Polity IV dataset. Polity is measured on a –10 to +10 scale where –10 is autocracy and +10 is democracy. The correlation between polity and democracy is almost 1 to 1. Country characteristics and income To control for country-speciﬁc characteristics, several geographic indicators are used: population, land area, a dummy variable for whether a country is landlocked and a dummy variable for whether a country is an island nation. A measure for remoteness is also used. Remoteness is deﬁned as the inverse of the log (distance to nearest trading partner/GDP).11 As a further defining characteristic, I create dummy variables for a country’s participation in the OECD and the IMF. GATT/WTO participation is not included because the relationship with RIA formation is almost 1 to 1. Membership in other multilateral organisations is a suitable instrument for multilateral participation. If participation in the multilateral system is important, as hypothesised, we would expect positive relationships between membership in such establishments and participation in an RIA. Per capita GDP is also included as a control variable since per capita income is a good economic indicator of a country’s development level and potential partner countries may choose not to invite less developed countries to join an existing RIA. Finally, although per capita income should theoretically have a positive relationship with the levels of regional integration, there is also the possibility of reverse causality. When countries liberalise their trade, often as a result of joining an RIA, levels of trade increase and as a result the per capita income of a country also improves. Certain levels of per capita income as well as the development of stable institutions may be required of countries seeking to join already existing RIAs. The result is that countries interested in joining RIAs or deepening their commitments will follow policies designed to enhance these aspects and causation is the reverse of the hypothesis. A way to remove the reverse causality problem is to use instrumental variable estimation techniques. Unfortunately, the ordered probit model is far from ideal for instrumental variable estimation. As estimation techniques improve, future assessment of this difﬁculty may be possible.12 GDP per capita is calculated from data provided by the Penn World Tables. The reported results use terms of trade adjusted per capita GDP. I now turn to a description of the model.

170 Jennifer Pédussel Wu MODEL SPECIFICATION The basic model As noted, nearly every country in the world is a member of, or in the process of discussing membership in, one or more regional integration arrangements.13 From Figures 9.1 and 9.3, it is apparent that increasing numbers of countries have made a multinomial choice. Thus, in any one year, a country chooses between sectoral cooperation, a free trade agreement, a customs union or a common market.14 When the outcomes are clearly ordered, we should take into account the fact that the dependent variable is both discrete and ordinal (as in this case). We cannot use linear regression to identify the causal relationship between RIA levels and explanatory characteristics since it would treat the difference between sectoral cooperation and PTA as identical to the difference between a customs union and a common market. This is obviously not the case. Under this setting, the most commonly used and most appropriate estimation technique is either an ordered probit or an ordered logit model. Here, I choose to estimate an ordered probit model.15 Regarding the level of regional integration, a country follows a speciﬁc set of steps to institutionally develop the framework necessary to support ﬁrst a free trade agreement and then a customs union.16 In addition, the data leads us to infer that the dependent, ranked levels of regional integration reﬂect an underlying qualitative variable, i.e. the chosen level of cooperation. Therefore an ordered probit estimation is more suitable than that of an ordered logit. Given the characteristics of a country: income, measures of uncertainty, its institutions and geography, we endeavour to estimate the probabilities that this country chooses a given level of regional integration. A three-way choice of regional integration, i.e. sectoral cooperation, FTA or customs union, is modelled by an ordered probit model where the latent, or unobserved, continuous variable, Yi* is a linear combination of the form:

Yi * = β ′xi + ε i

(9.1)

where β is a vector or unknown parameters to be estimated, xi is a vector of explanatory variables for country i, e.g. the demographic and economic characteristics described in the previous sections, and εi are independent normal disturbances with mean zero and unit variance. To estimate this model, we use maximum likelihood estimation. The observed discrete dependent variable, Yi, takes on its lowest value (sectoral cooperation) if Yi* ≤ 0, its medium value (FTA) if 0 < Yi* ≤ τ, and its highest value (customs union) if τ < Yi*. Under the restriction that 0 < τ, this speciﬁcation implies that:

Prob(Yi = Sectoral Cooperation) = Φ(− β ′xi ) Prob(Yi = FTA) = Φ (τ − β ′xi ) − Φ (− β ′xi ) Prob(Yi = Customs Union) = 1 − Φ (τ − β ′xi )

(9.2)

Levels of regional economic integration 171 where τ is the threshold value to be estimated and Φ is the standard normal cumulative distribution function.17 Standard practice is to set the ﬁrst threshold value to 0. Equation (9.2) above is interpreted as the probability of observing an FTA is equal to the probability that the estimated linear function, plus a random error, is within the range of the estimated threshold between 0 and τ. The constant in an ordered probit is set to zero. There are several ways to interpret the results of the ordered probit. The signs and statistical signiﬁcance of the coefﬁcients present some information about the model. The absolute magnitude of the coefﬁcient estimates has no meaning. The impact on the intermediate classiﬁcation value (FTA) of a change in the explanatory variables is indeterminate whereas on the lowest and highest, the impact is irrevocable. A positive signed coefﬁcient indicates a decrease in the probability of the lower threshold and an increase in the probability of the higher threshold, i.e. the change in the likelihood of observing the ﬁrst and last choices. The statistical signiﬁcance thus indicates a relationship between the independent variables and the dependent variable, but it does not indicate the magnitude of the effect. Two periods are examined: the overall period 1960–98 and the later period 1987–98. Note that the latter, 11-year span, marks a decade of signiﬁcant international changes: a large number of RTAs were notiﬁed to the WTO in the 1990s, the Cold War alliances dissolved and a number of new countries formed. The results of the estimation are examined in the results section below.

RESULTS

1960–98 estimates The empirical analysis of levels of regional integration suggests that economic freedom and institutional development, while not conclusive, do play a role in the choice of regional integration agreement. Examining Table 9.1, both trade uncertainty, as measured using the traditional measure of openness, and democracy are positive and signiﬁcant. This implies that increased trade uncertainty and the characteristic of democracy both contribute to the probability of a country choosing higher levels of regional integration. GDP per capita is also significant and positive. This implies that wealthier countries have a greater probability of belonging to a high level RIA. This result is brought about by the European Union. However, island nations are less likely to increase their level of RIA participation, as are large nations, while landlocked countries are positively correlated with an RIA. Population is also a positive and signiﬁcant contributing factor. This implies that nations with large populations are more inclined to increase their level of RIA participation. When a tariff measure is instead used as a measure for trade uncertainty, it is signiﬁcant at the 20 per cent level while all other variables retain their signs and their significance. It is important to note that when the Corruption Index from Transparency International is combined with the traditional measure of openness, neither measure is signiﬁcant. This implies that either the combination of political

172 Jennifer Pédussel Wu Table 9.1 Ordered probit regression 1960–98 (Y = regional integration) (1)

(2)

(3)

(4)

(5)

(6)

GDP per capita

0.346 (0.023)

0.273 (0.036)

0.483 (0.108)

0.443 (0.109)

0.344 (0.23)

0.348 (0.023)

Democracy

0.056 (0.005)

0.075 (0.007)

0.134 (0.021)

0.152 (0.021)

0.057 (0.005)

0.051 (0.005)

Population

0.726 (0.166)

0.1.032 (0.200)

–0.291 (0.261)

–0.176 (0.258)

0.577 (0.160)

0.581 (0.160)

Land area

–1.236 (0.126)

–1.593 (0.164)

–1.094 (0.192)

–0.960 (0.194)

–1.398 (0.127)

–1.40 (0.128)

Landlocked country

0.298 (0.045)

0.179 (0.667)

–0.361a (0.179)

–0.418 (0.180)

0.308 (0.045)

0.310 (0.045)

Island

–1.00 (0.059)

–0.633 (0.076)

–1.34 (0.258)

–1.395 (0.264)

–0.984 (0.060)

–0.979 (0.059)

Openness

0.142 (0.031)

0.217 (0.124) –0.008a (0.003)

Tariffs TI Corruption Index

–0.013 (0.038)

–0.267 (0.038)

GDP growth

0.003 (0.003)

Inﬂation rate

–0.001 (0.001) N = 4186

N = 1976

N = 329

N = 327

N = 4130

N = 4132

Notes Dependent variable is regional integration; ordered as described in the text. Openness is the log of exports and imports as a percentage of GDP. The Corruption Index comes from Transparency International. Tariffs are import duties as a percentage of imports from the World Bank, WDI. Maximum likelihood estimation of ordered probit, four-way choice model. Bold face indicates ρ-value signiﬁcance at 1 per cent. a indicates signiﬁcance at 10 per cent.

and economic uncertainty does not contribute to a choice of RIA level, or there is signiﬁcant collinearity. We rule out collinearity since the correlation coefﬁcient is less than 50 per cent. In addition, population and being a landlocked country are no longer signiﬁcant. Very high levels of political uncertainty may impair the ability of a government to commit to participation in an RIA. In addition, governments may not have access to the resources necessary for the pursuit of new institutional constructs, i.e. membership in an RIA. However, none of the other measures of political uncertainty were signiﬁcant. Neither business cycle uncertainty nor price uncertainty is a significant contributing factor to the probability that a country increases its level of RIA

Levels of regional economic integration 173 participation. If the percentage of government expenditure (size of government) is accounted for, trade uncertainty is also no longer signiﬁcant. This result conﬁrms the idea that countries expand social spending to help combat trade uncertainty as postulated by Rodrik (1987). However, when remoteness is included as a measure of distance between countries, it is negative and highly signiﬁcant. Thus, remote countries are unlikely to form ‘deep’ RIAs as are island nations and those with large land areas combined with large populations. This implies that denser, more contiguous countries are more likely to form ‘deeper’ RIAs. Membership in a multilateral organisation such as the IMF or the OECD is also positive and signiﬁcant. This implies that countries that are members of multilateral organisations are more likely to join deeper levels of RIAs. Thus, multilateral participation as hypothesised, has become more signiﬁcant and is important in the decision to join an RIA.

1987–98 estimates Are the characteristics of the past decade in the data signiﬁcantly different from those of the overall period? Examining Table 9.2, we observe that most signs and significance levels are the same except for the traditional measure of trade openness; now insigniﬁcant as a measure of trade uncertainty. When the Trade and Openness Index (TOI) from EFW is used in the place of the traditional trade openness measure, it is negative and insigniﬁcant. Per capita GDP is signiﬁcant and positive. However, when the TOI index and the traditional measure of trade openness are together in an order probit, neither are significant and both are negative. I interpret this as evidence that trade uncertainty alone had a modest impact on the increase in regional integration post-Cold War. Price uncertainty and business cycle uncertainty are negative but not signiﬁcant. Signiﬁcance of business cycle and price uncertainty would positively contribute to the probability of belonging to an RIA; inferred to be a Latin American effect. However, per capita income, democracy and the geographic characteristics seem to be better indicators of the probability of participation in a certain level of RIA in this later period.

Bivariate probit estimates In order to control for the possibility of bias due to presence of highly developed, industrial countries in deeper forms of RIAs, the sample is collapsed to a bivariate probit model. In doing so, the RIA indicator is collapsed to a dummy variable where participation in an FTA or any higher level of regional integration is denoted by a 1. Sectoral cooperation is not considered a formal FTA and thus is coded with non-participation as a zero. The results are reported in Table 9.3. Examining Table 9.3, it is immediately apparent that the results do not differ greatly from the ordered probit results for the same time period. All signs and significance levels remain the same except for tariffs, which are no longer

174 Jennifer Pédussel Wu Table 9.2 Ordered probit regression 1987–98 (Y = regional integration) (1)

(2)

(3)

(4)

(5)

(6)

GDP per capita

0.259 (0.032)

0.391 (0.047)

0.507 (0.117)

0.256 (0.032)

0.263 (0.032)

0.444 (0.057)

Democracy

0.093 (0.009)

0.089 (0.010)

0.139 (0.023)

0.097 (0.009)

0.096 (0.009)

0.114 (0.016)

Population

–0.338 0.220

0.188 (0.231)

–0.451a (0.268)

–0.404 (0.212)

–0.403 (0.213)

–0.345 (0.334)

Land area

–0.432 0.114

–0.749 (0.141)

–0.980 (0.168)

–0.501 (0.114)

–0.502 (0.114)

–0.291 (0.154)

Landlocked country

0.158 (0.071)

0.085 (0.095)

–0.357a (0.194)

0.182a (0.071)

0.189 (0.072)

0.291 (0.116)

Island

–0.865 (0.101)

–0.569 (0.112)

–1.52 (0.295)

–0.882 (0.101)

–0.869 (0.099)

–0.913 (0.213)

Openness

0.069 (0.057)

Tariffs

–0.002 (0.004)

TI Corruption Index

0.002 (0.043)

GDP growth

–0.004 (0.004)

Inﬂation rate

–0.002 (0.001)

Trade Policy IEF

–0.046 (0.051) N = 1459

N = 934

N = 279 N = 1428

N = 1430

N = 456

Notes Dependent variable is regional integration; ordered as described in the text. Openness is the log of exports and imports as a percent of GDP. The Corruption Index comes from Transparency International. Tariffs are import duties as a percent of imports from the World Bank, WDI. Trade Policy measure comes from EFW. Maximum likelihood estimation of ordered probit, four-way choice model. Bold face indicates ρ-value signiﬁcance at 1 per cent. a indicates signiﬁcance at 10 per cent.

signiﬁcant. Now, we are able to conclude that the probability of joining an RIA is inﬂuenced by the democracy level as well as geographic characteristics. Only trade uncertainty as defined through the traditional trade openness measure, is of contributive importance to a country’s participation in an FTA or another form of regional integration arrangement. Once again, we examine the post-1987 time period to investigate whether there are signiﬁcant differences in characteristics. From Table 9.4, it appears that in the post-Cold War period, none of the four measures of uncertainty: trade, political,

Levels of regional economic integration 175 Table 9.3 Bivariate probit regression 1960–98 (Y = regional integration) (1)

(2)

(3)

(4)

(5)

GDP per capita

0.146 (0.012)

0.173 (0.019)

0.209 (0.055)

0.145 (0.011)

0.147 (0.011)

Democracy

0.024 (0.002)

0.028 (0.004)

0.035 (0.009)

0.024 (0.002)

0.023 (0.002)

Population

0.688 (0.083)

0.718 (0.114)

0.187 (0.155)

0.630 (0.083)

0.633 (0.083)

Land area

–0.541 (0.054)

–0.549 (0.069)

–0.348 (0.106)

–0.595 (0.053)

–0.00 (0.00)

0.178 (0.22)

0.155 (0.035)

–0.098 (0.103)

0.184 (0.022)

0.184 (0.022)

Island

–0.267 (0.015)

–0.228 (0.03)

–0.589 (0.075)

–0.267 (0.016)

–0.265 (0.016)

Openness

0.056 (0.012)

Landlocked country

Tariffs

0.002 (0.002)

TI Corruption Index

–0.013 (0.018)

GDP growth

0.001 (0.001)

Inﬂation rate

0.0003 (0.046) N = 4186

N = 1976

N = 329

N = 4130

N = 4132

Notes Dependent variable is regional integration. Openness is the log of exports and imports as a percentage of GDP. The Corruption Index comes from Transparency International. Tariffs are import duties as a percent of imports from the World Bank, WDI. Maximum likelihood estimation of bivariate probit. Bold face indicates ρ-value signiﬁcance at 1 per cent.

price or business cycle uncertainty, is a good indicator of whether a country will belong to an RIA. Per capita income seems to be a good indicator when other institutional development, as shown by the level of democracy, is taken into account. Geographic characteristics are evidently much less important in this period. None of the included geographic characteristics are consistently signiﬁcant. This suggests that in this later time period, the other non-geographic factors were essential. In addition, it is clear that there is a difference between RIAs formed in the early and later periods.

176 Jennifer Pédussel Wu Table 9.4 Bivariate probit regression 1987–98 (Y = regional integration) (1)

(2)

(3)

(4)

(5)

(6)

GDP per capita

0.128 (0.018)

0.251 (0.029)

0.193 (0.054)

0.123 (0.018)

0.127 (0.018)

0.182 (0.036)

Democracy

0.029 (0.004)

0.027 (0.006)

0.030 (0.009)

0.031 (0.005)

0.030 (0.004)

0.036 (0.008)

Population

0.126 (0.121)

0.291 (0.133)

0.078 (0.139)

0.116 (0.120)

0.117 (0.120)

–0.159 (0.194)

Land area

–0.115 (0.080)

–0.076 (0.099)

–0.237 (0.097)

–0.095 (0.080)

–0.097 (0.080)

0.187 (0.129)

Landlocked country

0.087a (0.036)

0.103 (0.052)

–0.091 (0.103)

0.098 (0.036)

0.100 (0.036)

0.072 (0.066)

Island

–0.301 (0.035)

–0.218 (0.048)

–0.603 (0.084)

–0.311 (0.035)

–0.305 (0.035)

–0.317 (0.076)

Openness

–0.007 (0.027)

Tariffs

0.004 (0.002)

TI Corruption Index

–0.129 (0.019)

GDP growth

–0.001 (0.002)

Inﬂation rate

–0.001 (0.001)

Trade Policy IEF

0.028 (0.027) N = 1459

N = 934

N = 279 N = 1428

N = 1430

N = 456

Notes Dependent variable is regional integration. Openness is the log of exports and imports as a percentage of GDP. The Corruption Index comes from Transparency International. Tariffs are import duties as a percentage of imports from the World Bank, WDI. Maximum likelihood estimation of bivariate probit. Bold face indicates ρ-value signiﬁcance at 1 per cent. a indicates signiﬁcance at 10 per cent.

CONCLUSION This chapter had two objectives: ﬁrst, to isolate those common traits of RIAs and deﬁne contributing characteristics. Second, to explicitly test for whether economic and/or political uncertainty is a positive contribution factor to the chosen level of trade integration. Four measures of uncertainty were tested: trade uncertainty, political uncertainty, business cycle uncertainty and price uncertainty. As a result of the natural ordering of the regional integration data set, developed in this chapter, an ordered probit maximum likelihood estimation was used to examine

Levels of regional economic integration 177 those factors which were likely to contribute to the probability that a country belongs to a regional integration agreement. Only trade uncertainty was of any signiﬁcance. This chapter has provided a more complete empirical test to identify factors contributing to different levels of RIA participation. Using the level of regional integration as the dependent variable, a probit model was developed and tested. Data on regional integration has not previously been gathered in a systematic manner. A database of regional integration was developed which ranks the levels of integration across countries. What are the characteristics of countries joining regional integration arrangements? High-income, democratic countries with high populations which are landlocked and are not islands are more likely to belong to deeper levels of regional integration. It is expected that developing countries do not have solid institutions designed to support deep levels of RIA. This is reflected in the importance of democracy, a variable consistently signiﬁcant across time. High per capita income is a good indicator of countries belonging to higher levels of regional integration which may be referred to as a European Union effect. Per capita income is positive and statistically significant. From Table 9.4, it is apparent that when the ordered regional integration levels are collapsed into a binary system for the 1987–98 period, with the exception of democracy and per capita income, most indicators no longer signiﬁcantly contribute to enhance the probability that a country belongs to an RIA. This result signals that there has been a shift in countries’ motivations. One plausible hypothesis is that the international community experienced a shift. This shift, in my opinion, is a result of the fall of the Berlin Wall and the subsequent end of the Cold War.

NOTES 1 I have beneﬁted from discussions with Harry P. Bowen, John DiNardo, Arindam Ghosh, Bernd Hayo, Ann Largey, Kenneth Kletzer, Martin C. McGuire, Stergios Skaperdas, Jürgen von Hagen, S.Y. Wu and seminar and conference participants at the University of California, Irvine, the Centre for European Integration Studies, UNUCRIS, and the University of Nantes. I thank Dorina Berxholi for research assistance. The views expressed here and any errors are mine. 2 ‘North’ is considered industrially developed while ‘South’ is considered less developed. 3 Recent work, e.g. Bird and Rajan (2002), conﬁrms the ordering of the choice of levels of economic integration. 4 This is unfortunate since only one part (the deepening effect) of RIA formation is able to be studied. Future studies may wish to take into account the multi-RIA participation aspect. 5 Other data for the former Soviet Union are also unavailable. 6 Rodrik posits that to combat the uncertainty generated by a more open economy, social spending should increase. 7 This measure has several theoretical difﬁculties, but here the problems help generate the sought-after measure. One problem with using the imports and exports ratio is that it mitigates the importance of services in the economy since trade is usually high in

178 Jennifer Pédussel Wu

8

9 10 11

12 13 14 15

16

17

merchandise. A way around this would be to weight the openness ratio by services. However, here, we hypothesise that high service sectors are an additional measure of development and, thus, low openness ratios may also indicate that an economy is not as susceptible to uncertainty due to shifts in the international trade regime. This is true for the United States, which has a low openness ratio and a low probability of being affected by international trade instability. In contrast, the Heritage Index relies on qualitative evaluations used to assign component ratings to each country. While the Heritage Index rates slightly more countries, the EFW index provides a more objective ﬁgure and is therefore the preferred measure in this chapter. Revolutions and coups and central bank dependence have also been shown in the macro-economic literature to be strongly associated with average inﬂation rates. See, for example, Romer (1993). Data on remoteness, total land area and total population, as well as some of the other political uncertainty measures, were graciously provided by Andrew Rose and are located on his website at http: //faculty.haas.berkeley.edu/arose/RecRes.htm. Rose (2004), interestingly, found very little signiﬁcant relationship between WTO membership, trade liberalisation and RTA membership. This endogeneity issue has been partially addressed by Baier and Bergstrand (2004). . Agreements concluded among developing countries account for 15 per cent of the total number of notiﬁed agreements to the WTO. Since the data are only up to and including 1998, economic union is not a choice as it did not become effective for the European Union until 1999. The difference between the ordered logit and ordered probit models lies in the assumed distribution of error terms. In practice, there is no signiﬁcant difference between the two results. I estimate the ordered probit model because, with the current data, the implications of the exercise are more readily explained. Exceptions are the Central and West African Monetary Unions, two groups of former French colonies who have had their currencies tied to the French franc (CFA) since independence. Although the WAMU has been characterised by some trade integration, it lacks a common external tariff and its trade liberalisation programme is ad hoc with each member determining its own degree of protection. Sectoral cooperation was thus determined to be the overall level of regional trade integration. For an example of a general ordered probit model, see Brownstone and Golub (1992) or Becker and Kennedy (1992). See also Greene (1997).

REFERENCES Baier, L.S. and Bergstrand, J.H. (2004) ‘Economic Determinants of Free Trade Agreements’, Journal of International Economics, 64(1): 29–63. Becker, W.E. and Kennedy, P.E. (1992) ‘A Graphical Exposition of the Ordered Probit’, Econometric Theory, 8: 127–31. Bird, G. and Rajan, R.S. (2002) ‘The Political Economy of a Trade-First Approach to Regionalism’, CIES Discussion Paper, 0218, Adelaide University. Brownstone, D. and Golub, T. (1992) ‘The Effectiveness of Ridesharing Incentives: Discrete-choice Models of Commuting in Southern California’, Regional Science and Urban Economics, 22: 5–24. Economic Freedom of the World (2001) Annual Report. Online. Available HTTP: . Fernandez, R. and Portes, J. (1998) ‘Returns to Regionalism: An Analysis of Nontraditional

Levels of regional economic integration 179 Gains from Regional Trade Agreements’, The World Bank Economic Review, 12: 197–220. Greene, W.H. (1997) Econometric Analysis, Upper Saddle River: Prentice-Hall. The Heritage Foundation Index of Economic Freedom. Online. Available HTTP: (accessed 2001). Heston, A., Summers, R. and Aten, B. (2002) The Penn World Table Version 6.1, Centre for International Comparisons at the University of Pennsylvania (CICUP). Mansﬁeld, E.D., Milner, H.V. and Rosendorff, P.B. (2002) ‘Why Democracies Cooperate More: Electoral Control and International Trade Agreements’, International Organisation, 56(3): 477–513. Marshall, M., Jaggers, K. and Gurr, T. (2002) Polity IV Project: Political Regime Characteristics and Transitions, 1800–2002. Online. Available HTTP:. Mattli, W. (1999) The Logic of Regional Integration, Cambridge: Cambridge University Press. Perroni, C. and Whalley, J. (2000) ‘The New Regionalism: Trade Liberalisation or Insurance?’, Canadian Journal of Economics, 3: 1–24. Rodrik, D. (1987) Has Globalisation Gone Too Far?, Washington: Institute for International Economics. Romer, D. (1993) ‘Openness and Inﬂation: Theory and Evidence’, The Quarterly Journal of Economics, 108: 869–903. Rose, A. (2004) ‘Do WTO Members have More Liberal Trade Policy?’, Journal of International Economics, 63(2): 209–35. Rosendorff, P. (2000) ‘Do Democracies Trade More Freely?’, unpublished manuscript. Online. Available HTTP: . Schiff, M. (2000) ‘Multilateral Trade Liberalisation, Political Disintegration, and the Choice of Free Trade Agreements versus Customs Unions’, World Bank Working Paper. Transparency International. Online. Available HTTP: . Whalley, J. (1998) ‘Why Do Countries Seek Regional Trade Agreements?’, in J.A. Frankel, (ed.) The Regionalisation of the World Economy, Chicago: University of Chicago Press. World Bank. Online. Available HTTP: . World Development Indicators (1999) Washington: The World Bank. Wu, J. (2001) Economic Integration, Institutions, and Protection Against Risk, Diss: University of California, Irvine. Ann Arbor: UMI. Wu, J. (2005) ‘Trade Agreements as Self-Protection’, Review of International Economics, 13(3): 472–84.

Part III

Assessing political integration, regional governance and peace

10 Regional integration and (good) regional governance Are common standards and indicators possible? Edward Best

INTRODUCTION Most attempts to establish indicators of regional integration have been intended to measure how ‘regional’ a region actually is. In other words, to what degree and in what respects does the relationship between a group of countries present signiﬁcant differences from the relationship between each of those countries and the rest of the world? On the one hand, this may take the form of classifying the formal preferential agreements concluded between the countries according to a taxonomy of intentions. What have those countries’ public authorities and leading actors indicated that they wish to be in terms of regional grouping? In traditional economic integration this has meant an ascending set of general goals – free trade, customs union, common market, economic union – although more exhaustive classiﬁcations have been proposed according to coverage and ambition, ranging from partial sectorial preferentialism to economic and monetary union with a corresponding escalation of positive policy requirements (e.g. Pelkmans, 1993). In terms of political structures, taxonomies range from simple intergovernmental agreements involving no sharing of sovereignty (the case of most preferential economic agreements) through loose confederal structures to semi-federal unions of states such as the EU today and, exceptionally, to full political union in the sense of a single federal state. On the other hand, indicators are pursued as measurements of the real degree of integration which has been reached in terms of interdependence between economies, solidarity between governments and between societies, and perceptions of common identity. Economic interdependence can be indicated by the relative importance of intra-regional ﬂows of trade, investment, labour, migration; patterns of price convergence; and so on. One can assess relative degrees of political cooperation by looking at voting patterns between the component countries in international organisations compared to non-members. One can try to quantify and measure the relative intensity of transactions and communications as a further indicator of the existence of a ‘transnational community’.1 One can also carry out opinion polls on perceptions of common interests and common identity.2

184 Edward Best This kind of distinction between the ‘formal’ and the ‘real’, however, must be qualiﬁed when one looks at the actual processes by which regionalism in fact operates. The dimension of ‘institutionalisation’, in the sense of the demonstrable existence of formal agreements, organs, decision-making rules, compliance mechanisms and so on, always needs to be seen in either or both perspectives (that is, as a formal statement of intention or a manifestation of real integration) depending on the particular regional arrangement in question. As emphasised by sociological institutionalist and social constructivist approaches, actors’ interests, preferences and perceptions of identity are not ﬁxed but may be shaped by participation in broader institutional contexts and systemic processes, of which more below. Moreover, the more attention is placed on the subjective and discursive nature of ‘regionness’ (Hettne and Söderbaum, 2000) or ‘regionhood’,3 the harder it necessarily becomes to conceive of objective indicators by which to measure things.4 Even descriptive classification in any absolute terms is difficult. The aim of this chapter is to explore whether it may nevertheless be not only possible but in fact more valuable to move beyond classiﬁcation and purely empirical comparison towards some form of normative evaluation. The aim can never be to assess whether one structure is more or less similar to another. No particular regional model or structure has any inherent superiority. Yet there are three interrelated ways in which one may legitimately try to ‘rate’ regional systems. The first, focusing on real results, is whether a regional system is likely to contribute to a net increase in welfare of the populations. This requires in the ﬁrst place a methodology capable of establishing distinct causal relationships between public policies, economic trends and social benefits. However, it may still be unclear whether short-term welfare losses would continue, while the domestic distribution of costs and beneﬁts may vary. Moreover, there may be compelling political or cultural rationales for membership – often enjoying broad and deep popular support – which may obscure and even outweigh more material evaluations. An overall cost-beneﬁt analysis will always be broad and complex, but the underlying principle may be assumed to be generally accepted as legitimate. The second approach gives priority to stability of the system. The question here is whether the institutional arrangements (meaning not just the formal organs, but the whole set of regional actors, norms, structures and policies) are capable of managing the pressures generated by a particular regional process, taking into account the speciﬁc historical background and domestic structures of the countries involved. Each ‘equation’ between complexities and capacities in this sense will be a unique formula reflecting a different combination of variables such as the number of member states, relative sizes of the participating countries, different levels of development, scope of coverage, type of impact, time perspectives, degree of real interdependence, political framework, and underlying perceptions, values and norms (Best, 1997; Pelkmans, 1987). The third way, which is the subject of this chapter, tries to combine and to go beyond these two approaches by looking at regional systems holistically from the perspective of governance.

Indicators of regional governance 185 REGIONALISM AND (GOOD) GOVERNANCE There are probably as many deﬁnitions of ‘governance’ as there are of ‘regions’. However, there are by now a number of common understandings which make the concept of ‘(good) governance’ a more appropriate framework than ‘(democratic) government’ for evaluating the democratic quality of regional systems. At the national level, it has long been noted that the ways in which societies order themselves do not only depend on the existence of ‘a functionally differentiated group of “governors”’. Governance is not only a verbal noun meaning ‘the activity of governing’ but also includes the multiple mechanisms of ‘socialisation and social control’ by which human conduct is regulated (Finer, 1970: 4–6). It is now accepted practice to talk in terms of governance, if only as ‘a broader notion than “government” (whose principal elements include the constitution, the legislature, the executive and the judiciary). Governance involves the nature of interaction between such formally deﬁned institutions, and those of civil society’ (Corkery, 1999: 15). Beyond this, different usages have arisen, fed by developments in a number of disciplines and contexts. The common ground between them is the recognition that it is not only the formal organs of national states that matter when it comes to understanding how business is actually done in the world. However, there may be quite different emphases and implications. One perspective stresses that a ‘good’ state is necessary but insufficient for countries to achieve good overall development. This is the concept of ‘good governance’ as it has emerged over the 1990s, largely in the sphere of international development cooperation (although it must be seen in parallel to the debate on ‘corporate governance’) following a 1989 World Bank study on Sub-Saharan Africa. It increasingly has become accepted that sustainable development depends not only on the existence of an efﬁcient and accountable government but on the involvement of society as a whole – government and ‘civil society’ – within the rule of law and a system of rights. In Osborne’s neat summary, after shifting priorities from development projects in the 1950s through state-directed development plans and strategies in the 1960s, integrated development programmes in the 1970s, structural adjustment policies and public management reform in the 1980s, a consensus had emerged by the early 1990s that politics mattered. Deﬁnitions and priorities varied, but constant ‘good governance’ keywords were participation, transparency and accountability (Osborne, 1999). Good governance, according to the UNDP, is thus ‘a subset of governance, wherein public resources and problems are managed effectively, efﬁciently and in response to critical needs of society. Effective democratic forms of governance rely on public participation, accountability and transparency.’ (UNDP, 1997: 9) While it is emphasised that the role of citizens and civil society is essential, the existence of effective organs of state remains indispensable, in order to guarantee the rule of law (including the accountability of public officials) and respect for basic rights.

186 Edward Best Like the rule of law itself, these basic principles are held to be of universal validity, but it is up to different societies themselves how they choose to organise themselves in their light. Borrowing a recent formulation given by the OECD (Kondo, 2002: 7), the basic principles of good governance include: • • •

accountability, meaning that it is possible to identify and hold public ofﬁcials to account for their actions; transparency, meaning that reliable, relevant and timely information about the activities of government is available to the public; openness, meaning governments that listen to citizens and businesses, and take their suggestions into account when designing and implementing public policies. While challenges are similar across countries, and principles of good governance are widely accepted, there is plenty of room for different approaches, national priorities and institutional solutions to achieve transparent, accountable and open government. (Kondo, 2002: 7)

Various efforts are indeed being made to establish not only universal principles and standards but benchmarks, indicators and comparative methodologies. The World Bank thus offers an overview of ‘Indicators of Governance and Institutional Quality’ to measure, for example, the severity of corruption, the extent of civil liberties, bureaucratic efﬁciency, the rule of law, and the predictability of policymaking, as well as ‘toolkits’ which ‘set out the principles which experience suggests should underpin public sector governance arrangements and provide methods for assessing the degree to which specific country arrangements are consistent with those principles’.5 Another set of ‘governance’ approaches, however, draws attention to the fact that a state may not in fact be there, or at least not be fully in control. On the one hand, there may simply not be any accepted higher level of recognised authority, as in the international system. The concept of ‘governance without government’ was thus developed by Rosenau and others in the late 1980s as a result of the observation of the important role played in international affairs by non-state actors as well as governments, and of the fact that common norms and rules could be determined and respected internationally even in the absence of a central international authority with coercive powers. Norms and practices are shaped by transnational economic interests, ‘epistemic communities’ of experts and a variety of non-governmental organisations; and governance is ‘a system of rule that is as dependent on intersubjective meanings as on formally sanctioned constitutions and charters’ (Rosenau, 1992). On the other hand, the role – indeed the power – of central governments has changed, partly because they have been ‘hollowed out’ as a consequence of internationalisation (often accompanied by domestic decentralisation) and partly in the

Indicators of regional governance 187 context of ‘new public management’ and ‘reinventing government’. Governments swapped direct for indirect controls as they privatised and contracted out services, while creating executive agencies for independent implementation of politicallydeﬁned goals. All this led to function-based policy networks which were broader in membership than before (including independent agencies as well as both private and voluntary sectors) and in which central departments were no longer necessarily the ‘fulcrum, or focal organisation’. ‘Governance’ in this sense is thus understood to be a particular way of policy management based not on hierarchical authority but on trust and negotiation between partners, and operating through networks composed of public and private actors from different levels (Rhodes, 2000). While all this may promise a greater degree not only of flexibility but also of participation (at least for those participating), it also poses new challenges for transparency and accountability. When it comes to a transnational expert community, a multilevel policy network or a regional cooperation programme, all of which play important roles in how the world is governed in practice, ‘the assumption of institutional hierarchy which underpins so many discussions of bureaucratic accountability no longer holds. Accountability can no longer be speciﬁc to an institution but must ﬁt the substantive policy and the several institutions contributing to it’ (Rhodes, 1997: 58–9). It is suggested here that a ‘governance’ framework based on these two perspectives, and the tension between them, is a natural starting point for thinking about how to evaluate the democratic quality of regional arrangements themselves. In the first place, it may help establish minimum universal standards which are more broadly applicable and acceptable than those tied to particular governmental structures. Regional systems cannot be judged simply in terms of national practice, and there is in any case a wide variety of national practices. By the same token, this approach recognises that in almost all cases the regional level of governance – like the international system or new forms of policy management within countries – does not rest mainly, if at all, on central hierarchical authority. Nor does it rest on some overriding cultural identity which can obviate the need for democratic and/or utilitarian justiﬁcation. Consequently, the ways in which ‘good governance’ is respected must inevitably take non-traditional forms, and may imply adapting the usual litanies of principles. The issue must be split into two parts. First, how do regional arrangements affect standards of governance within countries? Second, if regional arrangements do lead to the establishment of some degree of ‘multilevel governance’, does this constitute a net increase in the overall amount of ‘good governance’ available to people?

Intergovernmental systems A number of regional organisations have been created with the primary goal of cooperating in order to consolidate democratic values and practices. These include the Council of Europe and the associated European Convention on Fundamental Rights and Freedoms, and a few other regional systems for political cooperation

188 Edward Best and the protection of human rights such as the Organisation of American States (OAS). Principles are stated in regional treaties, charters and conventions – to which a recent addition is the Inter-American Democratic Charter adopted by the General Assembly of the OAS in Peru in September 2001. These principles may be elaborated in common standards and disseminated through policy instruments and guidelines, exchange of best practices, training programmes. A good example is the Council of Europe’s integrated project on ‘Making democratic institutions work’. Signatory governments can be obliged to comply by the courts under international and often also national law. An increasing number of mutual evaluation processes also exist with the aim of securing improvements in governance through peer pressure.6 The Group of States against Corruption (GRECO), for example, produces reports, which may be made public, on national compliance with the anti-corruption instruments of the Council of Europe. The Follow-up Mechanism to the 1996 Inter-American Convention Against Corruption likewise entails questionnaires, reviews of information provided and national progress reports, as well as mechanisms for legal cooperation. There have recently been innovations in this direction elsewhere. An African Peer Review Mechanism (APRM) was established in 2001 in the context of the New Partnership for Africa’s Development (NEPAD) and the transformation of the Organisation of African Unity (OAU) into the African Union (AU). Key objectives are laid down concerning consolidation in all countries of ‘a constitutional political order in which democracy, respect for human rights, the rule of law, the separation of powers and effective, responsive public service are realised to ensure sustainable development and a peaceful and stable society’. These are then broken down into more speciﬁc objectives, accompanied by standards (mainly international treaties and declarations), indicative criteria and examples of indicators. The procedure is to be overseen, under the aegis of an APR Heads of State Forum, by a Panel of Eminent Persons, on the basis of a report from a Country Review Team.7 This is emphatically presented not as a vehicle for pressure by outsiders but as ‘an instrument voluntarily acceded to by Member States of the African Union as an African self-monitoring mechanism’. As of January 2004, 16 of the 53 AU members had signed up to the APRM.8 It remains to be seen what may happen in practice, but the principle has received widespread encouragement as a new way of structuring regional governance development. Democratic practice within countries can also be positively promoted by broader regional systems. The EC/EU has always been based on an insistence that member states should share and respect fundamental rights and democratic values. In the perspective of an enlargement bringing in many countries from Central and Eastern Europe which have not, at least since the Second World War, enjoyed governments with such democratic traditions, the EU has not only made this more explicit in the founding treaties, but also written in provisions for monitoring and imposition of sanctions on member states which committed serious breaches of those standards. The Common Market of the South (Mercosur), created in 1991 between Argentina, Brazil, Paraguay and Uruguay (with the later association of Chile,

Indicators of regional governance 189 Bolivia and Peru), may to some extent also be seen as a mutual commitment to continued democratisation – and was employed as a means of pressure to prevent a possible relapse in the case of Paraguay in 1996. On the other hand, it should not be forgotten that regional organisations can have a negative political influence. This was most notoriously the case of the regional bodies dominated by the Cold War superpowers: the Committee for Mutual Economic Assistance (Comecon) and Warsaw Pact in the case of the Soviet Union; and the mechanisms of the OAS, which were sometimes used by the US to prop up dictatorships in Latin America. Still today, however, there is open concern in the EU and elsewhere that, for example, the absolute insistence of the Association of South-East Asian Nations (ASEAN) on political non-interference and its inclusion of Myanmar in fact amounts to propping up an anti-democratic regime. Assuming the existence of an agreed check-list of universal values and standards, one can imagine a set of standards and indicators for comparing the performance of the regional systems themselves in contributing to the enhancement of good governance within each of their respective member states. For example: • • • •

Do the regional reference documents (treaties, declarations, programmes, etc.) conform to universal norms and values? Do national reference documents conform to regional ones? Do regional mechanisms (judicial, political, social) exist by which pressure can be exerted on national authorities to comply with regional norms? Do member states comply with these pressures?

At the same time, the basic principles of good governance can be applied to the regional arrangements themselves even if there is no disposition to go beyond intergovernmentalism, functional cooperation or partial preferentialism. Are the public ofﬁcials and other actors responsible for regional programmes effectively accountable for how resources are managed? Are stakeholders well informed as to the objectives and risks of speciﬁc projects, and do they have the opportunity to make their views known?

Regional integration systems Things become more complex when people have to comply with binding decisions which are adopted beyond national structures. In terms of taxonomies of regionalism, this means differentiating purely intergovernmental frameworks from ‘regional integration’ systems, deﬁned as formal unions of states which retain legal personality but which agree to constitute a shared economic and social space partially governed by distinct common regional rules or uniform national laws. If a regional system really does constitute an additional level of governance which has a tangible impact on national systems and individual rights, then it is only proper to ask whether the resulting ‘multilevel’ arrangements constitute a net

190 Edward Best increase in the overall welfare and good-governance possibilities afforded to citizens, both in their design and in their operation. If citizens have to accept common rules, even in limited areas (especially if these are not adopted by unanimity), then additional legitimacy requirements arise which go beyond the basic principles of good governance. Indeed, the very possibility of satisfying those principles (transparency, accountability and openness/ participation) can be seen to be in inverse proportion to the level of organisation. The issue is not only that integration may have clear ‘sovereignty costs’ for states (which may translate into political costs for their governments). Even if integration does produce practical benefits, individuals still face a ‘democratic dilemma’ (Dahl, 1994) by which there is a trade-off, as the scale of organisation increases, between a reduction in relative individual inﬂuence over collective decisions, on the one hand, and an increase in collective inﬂuence over global results, on the other. These legitimacy requirements may, it is suggested, be summarised as follows. • • •

•

Citizens feel that the dimensions and composition of the unit in which binding common rules are made are appropriate. Citizens recognise the rightfulness of the authorities. Citizens accept the basic goals and principles on which the system is organised and are convinced of the existence of bodies guaranteeing respect for these principles. Citizens believe that meaningful possibilities for participation in decision making are open to them.

These requirements can be satisﬁed in different ways, with different structures, and with different trade-offs between them, depending on the particular circumstances of the countries participating in the regional arrangements. However, two general propositions may be advanced. •

•

There will be a direct relationship between the perceived impact of regional arrangements on people’s choices and the importance which people attach to the satisfaction of legitimacy requirements for regional action. There will be an inverse relationship between the strength of perceptions of common identity and the importance of ‘utilitarian’ justifications for the existence of the higher level.

This is in large part a question of the balance between input legitimacy and output legitimacy ‘Output legitimacy’ means that people agree that a particular structure should exist and its rules should be respected because of the benefits it brings. Social acceptance is thus instrumental and conditional, as well as independent of any affective relation. ‘Input legitimacy’, on the other hand, means that social acceptance of the structure in question derives more from a belief in the rightfulness of the system of rulemaking. This does not consist only in that system’s respect for due process (transparency, legality, accountability) and its openness to

Indicators of regional governance 191 popular participation, but also in the existence of ‘deeper validating legal Rules of Recognition (rules about rules)’ and a sense of shared identity or ‘we-feeling’ (Weiler, 1991, 1997). Where integration processes do produce systems of multilevel governance, therefore, any comparative framework for assessment of the democratic quality of the arrangements (that is, the good governance of the whole) has to go beyond the basic principles already discussed to take into account deeper questions of social legitimacy. One of the most important recent proposals in this respect has been the European Commission’s White Paper on European Governance (CEC, 2001), in which it proposed five principles of good governance to be applied in the European Union. Three of these principles – ‘openness’, ‘participation’ and ‘accountability’ – basically refer to the same things as those propounded internationally as means to evaluate good practice in public management although, whereas the OECD formulation of ‘openness’ given above is much closer to ‘participation’, the Commission uses ‘openness’ to mean an active version of ‘transparency’. •

•

•

Openness – the institutions should work in a more open manner. Together with the member states, they should actively communicate about what the EU does and the decisions it takes. They should use language that is accessible and understandable for the general public. This is of particular importance in order to improve the conﬁdence in complex institutions. Participation – the quality, relevance and effectiveness of EU policies depend on ensuring wide participation throughout the policy chain – from conception to implementation. Improved participation is likely to create more conﬁdence in the end result and in the institutions which deliver policies. Participation crucially depends on central governments following an inclusive approach when developing and implementing EU policies. Accountability – roles in the legislative and executive processes need to be clearer. Each of the EU Institutions must explain and take responsibility for what it does in Europe. But there is also a need for greater clarity and responsibility from member states and all those involved in developing and implementing EU policy at whatever level.

However, the other two principles – ‘effectiveness’ and ‘coherence’ – are essentially utilitarian justiﬁcations for the ‘Community method’ as a means to give citizens the beneﬁts of European integration. •

•

Effectiveness – policies must be effective and timely, delivering what is needed on the basis of clear objectives, an evaluation of future impact and, where available, of past experience. Effectiveness also depends on implementing EU policies in a proportionate manner and on taking decisions at the most appropriate level. Coherence – policies and action must be coherent and easily understood. The need for coherence in the Union is increasing: the range of tasks has grown;

192 Edward Best enlargement will increase diversity; challenges such as climate and demographic change cross the boundaries of the sectorial policies on which the Union has been built; regional and local authorities are increasingly involved in EU policies. Coherence requires political leadership and a strong responsibility on the part of the Institutions to ensure a consistent approach within a complex system. There is a great deal of validity in this approach, to be sure. In the absence of some overriding supranational cultural uniﬁer, regional arrangements which imply costs in terms of national sovereignty, individual participation and resource allocation will not long be accepted if they are not seen to give adequate results. Assuming that the regional system is capable of bringing a net increase in welfare over time, then ensuring the stability of that system is in itself a necessary dimension of good regional governance. This means, crucially, that regional institutional arrangements must be appropriate – not only in that they ‘match’ the real needs of the integration process in a functional perspective, but also in that they satisfy legitimacy requirements in ways which ‘ﬁt’ the political context. First, the democratic quality of regional structures does not rest on their apparent resemblance to national structures. Here, indeed, one faces two further ‘democratic dilemmas’. On the one hand, there is a problem of recognition. One cannot transfer national structures of any type to regional level in any simple way (except if the process is genuinely a process of political uniﬁcation). Yet people’s perceptions of legitimate institutions are inevitably determined by national experience. As Schmitter has noted with regard to Europe: If . . . the non-state Euro-polity will have to come up with novel institutions in order to democratise itself, then both politicians and citizens may have considerable difficulty in recognising these novel rules and practices as ‘democratic’. The more novel (and, perhaps, functionally appropriate) these citizen rights and decision-rules are, the greater the initial obstacles to their being accorded legitimacy. (Schmitter, 1996: 16) On the other hand, there is a risk of rejection. The constitution of regional structures with the formal trappings of national democratic institutions may be resented – precisely because they are given such formal democratic status – as an unwarranted challenge to national bodies which enjoy greater social legitimacy. Simple assertions that regional institutional arrangements should mirror national ones may thus create a no-win situation in which the regional system will be damned if it does and damned if it doesn’t. Second, neither the merits nor the viability of any particular structure, instrument or policy can be deduced simply by analogy from any other region. The ‘appropriateness’ of one or another depends on the particular characteristics of the regional process. To copy structures created in other contexts – as has so often

Indicators of regional governance 193 seemed to be the case of the EU institutions – may in fact only weaken the effectiveness and the legitimacy of the regional system. It is therefore suggested that a modiﬁed version of the European Commission’s set of principles may be used to assess the quality of regional governance structures: • • • • •

openness (in the sense of transparency plus active communications) accountability participation effectiveness appropriateness.

The following section aims to give an example of the principle of ‘appropriateness’ by looking in general at how the principle of ‘participation’ may be applied in different regional contexts, and then taking the speciﬁc case of the formal structures for public participation established in sub-regional integration arrangements in Latin America.

APPROPRIATENESS: THE CASE OF REGIONAL FORA FOR PUBLIC PARTICIPATION The principle of ‘participation’ is ideally more ambitious than, for example, the definition given by the European Commission in the White Paper on Governance. As formulated by the OECD’s study on government–citizen relations, it should mean going beyond citizens’ access to information towards meaningful consultation by the authorities and eventually into ‘active participation’ by which citizens (and organisations of civil society) ‘actively engage in deﬁning process and content of policy-making’. However, as recognised by the OECD, this level of participation is still rare within OECD countries.9 One must therefore be realistic as to what one can expect on the scale of – and in the terms of – a complex regional system (especially if it is as large as the European Union). ‘Participation’ is one aspect of (good) regional governance, and regional institutional arrangements should therefore promote it. The kinds of formal structures which will be most effective in doing so are those which are most appropriate to the regional context. This means that the structures which are created should match immediate regional realities: notably the real strength of organisations of civil society, and the ability and/or willingness of national authorities and private interests to ﬁnance participatory mechanisms. It also implies that the basic assumptions behind the creation of such mechanisms should reﬂect the real dynamics of the regional process in question. At this level of integration the process in question is one of creating a ‘regional community’, in which there is a ‘mutually reinforcing relationship between the “formal” region, deﬁned by the community of states, and the “real” region, in which a transnationalised regional civil society also has a role to play’ (Hettne and Söderbaum, 2000: 466). However, this can take very different forms.

194 Edward Best In the ﬁrst place, the nature of the interaction will vary according to what had been the main starting point – or ‘integration source’ – for the whole regional process. This can be presented in the form of a continuum with, at one extreme, an initial impetus characterised by shared elite interests (state and/or business) based on a rational calculation of net beneﬁts and, at the other extreme, the growth of transnational contacts, interdependences and perceptions of common identity. In simple terms, it may be more a top-down or more a bottom-up process. Europe provides two cases which clearly reﬂect this basic contrast. Despite the existence of historical and cultural ties, the European Community has been primarily a topdown process in which governments and elites made strong mutual commitments in order to achieve ambitious long-term goals, and for this purpose they established a strong institutional system. Even though the objective beneﬁts are considerable, it has been a constant challenge to try to involve lower levels (national parliaments, regions, local authorities, civil society organisations, individuals) in order to maintain public support. For decades, programmes have been ﬁnanced with the aim of bringing about greater transnational exchanges and interaction and thus creating better understanding and stronger perceptions of common identity. By contrast, Nordic cooperation was, from an institutional perspective, literally built ‘upwards’. The starting point, dating back to the mid-nineteenth century, may be characterised as transnationalism in the sense of the growth of Nordic organisations of civil society, economic interaction and social communications. Next was inter-parliamentary cooperation, starting formally in 1907 with the Nordic Inter-Parliamentary Union, and given new form with the establishment of the Nordic Council in 1952. On top of this came ‘ad hoc intergovernmentalism’ and only at the end the Nordic Council organisations.10 Similar contrasts can be seen elsewhere. The issue of ‘participation’ has quite different connotations in these two types of context. If the regional process is based primarily on rational choice in pursuit of elite interests, and if the goals are understood by the contracting parties to be complex and long term, then participating states may create ‘commitment institutions’ in order to increase the prospects of effective compliance over time (Mattli, 1999). Intergovernmentalist theories thus afﬁrm that it is primarily this rational acceptance of the need to ensure the credibility of commitments that explains the institutional choice for a pooling of sovereignty (decision-making other than by unanimity), delegation of sovereignty (the attribution of powers to regional organs) or ‘legalisation’ (Abbott et al., 2000; Abbott and Snidal, 2000). In other words, states do not trust each other – or themselves – to comply with ambitious commitments in the face of temptation (opportunism). They set up ‘higher’ institutions to bind themselves, and these institutions later require social legitimation. In this perspective, participation is seen to have various utilitarian advantages for the effectiveness and stability of the regional system. Social participation is an important means to help ensure compliance by national authorities with the agreements reached. This can apply in the legal sphere – compared to inter-state dispute resolution, by ‘linking direct access for domestic actors to domestic legal

Indicators of regional governance 195 enforcement, transnational dispute resolution opens up an additional source of political pressure for compliance’ (Keohane et al., 2000: 477). And, as integration advances, it may be an important means of monitoring at the practical level. Participation may also be seen as an important element in helping the rule-makers to ‘get it right’ and ensure effective implementation, as seems to be the main thrust of the European Commission’s deﬁnition in the Governance White Paper (CEC, 2001). More broadly, if governments mean business and there is a significant impact on their societies, then sooner or later they will need to take steps to maintain the legitimacy of the whole process in the face of the negative short-term impact on certain economic sectors or social groups and the general ‘sovereignty costs’ for the country concerned. In this perspective, regional social and economic fora should act as ‘legitimating institutions’ for inter-state ‘commitment institutions’. Whether or not they succeed will depend in the short term on their capacity to exercise inﬂuence over decision-making, and thus maintain their own credibility. In the longer term it will depend on their ability to serve as socialisation mechanisms in the community-building process – and thereby build popular identiﬁcation with the institutions. If, on the other hand, the starting point of a regional process is more an incipient system of inter-societal transactions, then participation as such will probably be less of an issue. If anything, it is more likely that the problem will be whether a highly participatory and consensual approach can produce results. The Nordic countries achieved a Common Labour Market and a Passport Union in the 1950s, as well as profound integration in the social sphere (for example, social security, family law, education, health) and extensive cooperation in other ‘low-politics’ areas such as the environment, transport or culture. They succeeded where agreement came naturally, inasmuch as these were areas of low sensitivity (and existing consensus) as well as of objective similarities, but notably failed in economic integration while shunning political union. Most notably, negotiations for market integration failed twice (the customs union in the late 1940s and an economic union in the late 1960s) as did efforts to establish a defence pact in the late 1940s. Success would have required adoption of institutional arrangements for which the underlying conditions were simply not present, at least in the Nordic context itself. In such a regional context, the question of participation is more relevant in what Hettne and Söderbaum term the creation of ‘regional society’, the level of ‘regionness’ preceding a regional community. This perspective has much in common with Adler and Barnett’s (1998) analysis of the development of ‘security communities’, with its emphasis on structural variables and process variables (transactions, organisations and social learning) in the construction of mutual trust and collective identity. Regional organisations may help by establishing norms of behaviour; by serving as ‘sites of socialisation and learning’; and because of their ‘often underestimated capacity to ‘engineer’ the very conditions – for example, cultural homogeneity, a belief in a common fate, and norms of unilateral selfrestraint – that assist in their development’ (Adler and Barnett, 1998). Regional fora of this sort have a quite different role from that of legitimating regional

196 Edward Best ‘commitment institutions’. This is to foment and give form to transnational associations and intersocietal interactions which may or may not give rise, in the particular grouping of countries, to deeper levels of economic and political integration. If there are not in fact any ‘commitment institutions’ by which states have agreed to generate and ensure respect for common rules, then it will not be ‘appropriate’ to establish a formal body designed to participate in common rules, whether this takes the form of a ‘parliament’, a consultative sectoral organ, or anything else. Good regional governance in terms of participation cannot be assessed simply by the existence of formal bodies. It must be asked in the first place whether the structures strike an appropriate balance, in the circumstances, between the functions of consultation and legitimation of rules, on the one hand, and the role of socialisation and learning, on the other.

Socialisation mechanisms A second question concerns precisely the ways in which socialisation may take place. For present purposes it is proposed to order the multiple mechanisms involved according to the two basic ‘diffusion pathways’ suggested by Checkel (1999). Again, these can be placed against a continuum. At one end is what is supposed to happen between the individual representatives involved in regional fora – that is, ‘social learning, where agents – typically élite decision-makers – adopt prescriptions embodied in norms; they then become internalised and constitute a set of shared intersubjective understandings that make behavioural claims’. At the other end is ‘societal mobilisation’, in which ‘non-state actors and policy networks are united in their support for norms; they then mobilise and coerce decision-makers to change state policy. Norms are not necessarily internalised by the élites.’ The governance criteria by which one may try to assess a regional structure for public participation depend on the real demands of each situation. If the main need is to promote transnational associations and people-to-people contacts, then the appropriate structure is likely to be different from what is called for if the main requirements of participation are representativeness and expertise in providing inputs to common rulemaking.

The case of regional structures for participation in Europe and Latin America As a modest case study, one can take the case of the structures which have been created with the aim of ensuring participation in the various sub-regional arrangements in Latin America.

Indicators of regional governance 197 A European model? The formal institutional structure of the European Community has been particularly inﬂuential in Latin America, and indeed it has been actively encouraged in the framework of the region-to-region relationships. Most attention in other regions has tended to focus on the European Economic and Social Committee (ESC) in view of its apparent institutional status and multi-sectorial composition, with its three Groups representing Employers, Workers and Various Interests.11 Some in Central America, for example, have lamented that their governments have never found the political will to go ‘so far’ as to create anything like the ESC, which is ‘part of the decision-making process’ in the EC (de Monterrosa, 1999: 77, 78). Yet the ESC is only one of the many institutionalised channels available at European level. Citizens and interest groups can provide input to the European Commission in the formulation of legislative proposals, through direct contacts (including new internet sites), participation in different kinds of standing consultative committees (the Economic and Social Committee, the various Social Dialogue committees and other sectorial committees such as the consumer committee). A ‘Civil Dialogue’ has grown up to promote cooperation with civil society in the social field, assisted by financial support from the Commission, which supports a wide variety of non-governmental organisations in other ﬁelds. Citizens elect and can petition the members of the European Parliament, which has acquired increasing powers both of decision-making and of scrutiny; they can address themselves to the European Ombudsman in cases of alleged maladministration by the European institutions. And they can either indirectly (through their national government, the Commission or, since the entry into force of the Nice Treaty, the Parliament) or directly, subject to certain conditions and limitations, challenge acts of the institutions before the Court of Justice. Moreover, in reality the ESC has very limited inﬂuence on decision-making and is generally seen as of little signiﬁcance in the formal policy process. There are, to be sure, formal institutional constraints, but the ESC has basically been overtaken by developments. It represented a traditional corporatist vision of economic and social participation, largely inspired by the existence of tripartite consultative bodies in ﬁve of the original six members of the EEC, at a time when the European Parliament was not widely expected to be able to fulﬁl the function of sectorial representation. Since then, however, the Parliament has grown in importance as a channel for representation of interests. The ESC, on the other hand, has been bypassed as organised interests came to prefer direct contacts with the Commission and the other institutions (Van der Voort, 1998). In the 1990s, the ESC tried to ﬁnd a new role ‘as the representative body of civil society organisations’ at European level.12 The Nice Treaty, which entered into force in February 2003, has indeed changed the deﬁnition of the members to be representatives of ‘the various economic and social components of organised civil society’. Yet here again, the ESC is competing, despite its inherent formal advantages, with a variety of other actors. In this context, the Committee at the end of the 1990s adopted a logic related to the special characteristics of a ‘European identity’. At national level, it argues, identity is deﬁned by:

198 Edward Best a common nationality . . . and a common culture, language and set of values. However, when it comes to democratic policy-making at European level, additional identity criteria are required to create a European identity . . . the democratic process at European level – even more so than at national level – must provide a range of participatory structures . . . which reflect the heterogeneous nature of the European identity. The European Parliament is elected by citizens ‘exercising their democratic rights as part of their national (territorial) identity. But people’s identity is also defined by membership of interest groups in the diverse shape of civil society groups.’ The ESC represents those identity criteria and ‘this enables the Committee to promote democratisation at European level, and to show Parliament that it provides genuine added value in the democratic European decision-making process’.13 This idea is valuable, although the way it has been tied to institutional interests can give the impression of trying to develop ‘a European citizen for the Economic and Social Committee’, rather than an Economic and Social Committee for the citizen (Smismans, 1999). Despite the important contribution made by many of its Members, the ESC is institutionally challenged in terms both of the weight of its inﬂuence in formal decision-making processes, and of the credibility of its claims to representativeness. Social and economic actors in other regions have also looked with interest at the European Social Dialogue, by which the European ‘social partners’ do play a direct role in decision-making. The Commission must consult them on initiatives in the social ﬁeld, ﬁrst on the possible direction and then on proposed content. The social partners may then choose to negotiate directly between themselves and, if a framework agreement is reached, then opt either to implement it by voluntary means or to ask the Commission and Council to turn it into European law.14 As of 2003, three cross-industry agreements and two sectorial agreements had been implemented by Council Directive. Yet the future of this form of participation is unclear, while important political issues have arisen. The representativeness of the main private employers’ organisation at European cross-industry level – the Union of Industrial and Employers’ Confederations of Europe (UNICE) – was in fact legally challenged by the European Association of Small and Medium Enterprises (UEAPME). Moreover, even if particular organisations are deemed sufﬁciently representative, it is questioned in many quarters whether this procedure is acceptable at all in terms of democratic legitimacy, since the European Parliament plays no formal role in this procedure: can management and labour organisations really be a sufficient source of democratic legitimation for European law? (Britz and Schmidt, 2000). The social dialogue in the 1990s prompted ‘the launch of a new policy objective: the building over time of a strong civil dialogue at European level to take its place alongside the policy dialogue with the national authorities and the social dialogue with the Social Partners’. This civil dialogue would have two aims:

Indicators of regional governance 199 to ensure that the views and grassroots experience of the voluntary sector can be systematically taken into account by policy-makers at European level so that policies can be tailored more to real needs, and to disseminate information from the European level down to the local level, so that citizens are aware of developments, can feel part of the construction of Europe and can see the relevance of it to their own situation, thus increasing transparency and promoting citizenship. (CEC, 1997) The Commission has thus ﬁnanced the activities of the Platform of European Social Non-governmental Organisations. These have asked for clear guarantees of ‘core-funding of NGOs by the Commission . . . [which] is vital if their independence from commercial interests is not to be compromised and to ensure that they can be an effective relay between the institutions and citizens who form them.’15 In 2003 the Commission presented a Proposal for a ‘Council Decision establishing a Community action programme to promote active European citizenship (civic participation)’ which aims to place these and other funding arrangements on a proper legal basis.16 Yet it remains a difﬁcult task to establish on what grounds some NGOs should be funded and others not, while important questions arise regarding the autonomy of civil society organisations vis-à-vis the authorities in such circumstances. Again, the question of representativeness arises as soon as such organisations aim and claim to go beyond the function of presenting information and arguments to rulemakers (and, in the other direction, to their own members and the public at large). Jacques Delors’ Address to the ﬁrst Convention on Civil Society should be recalled in this respect. While stressing the importance for the construction of Europe of civil society organisations, he argued that: Some economic and social councils attempt to gain a wider public hearing by claiming they are the representatives of civil society . . . Civil society and civil society organisations should not be confused . . . Civil society organisations must not give in to the temptation of saying they represent the general interest. They may identify the general interest in their discussions. But that is quite a different thing. As for the associative interests that flourish around the European Commission, it would be dangerous for those involved to become too much part of the system: to believe they alone have the right to represent society . . . this poses serious problems of representativeness.17

Subregional structures in Latin America In the past decades, a variety of formal structures aimed at promoting participation have been put in place in Central America, the Andean Community and Mercosur.18

200 Edward Best Central America The situation of contemporary fora in Central America has to be seen against two different historical trends which largely reflect the underlying top-down and bottom-up approaches suggested above. Attempts at political (re-)union since the early nineteenth century have repeatedly failed and, in the modern period, there has been very limited success for formal regional organisations and market integration. Yet there has also been a constant, ‘mushroom-like’, growth of regional organisations in almost every area of social and economic activity. Indeed, the ﬁrst post-war regional organisations to be established (before the intergovernmental schemes) were the public university confederation in 1948 and the regional institute of nutrition in 1949.19 On the formal side of things, regional social and economic fora have been weak, or simply absent. There was no such consultative body in the Organisation of Central American States (ODECA) created in 1951, nor in the institutional arrangements of the economic integration process which took place from the late 1950s onwards – the Central American Common Market (CACM) – although the private-sector federations were key players from the start in influencing the process. The attempts to restructure the CACM in the early 1970s included new proposals to bring in social sectors. The 1976 proposal for a Framework Treaty to create a Central American Economic and Social Community (CESCA) followed the EC model quite closely, including a permanent consultative body – an Economic and Social Committee (see Mariscal, 1989). Yet the whole process came to nothing because of the lack of support by the governments – and the whole process was then rapidly overwhelmed by the political and military crises which dominated the region from the late 1970s to the early 1990s. As Central American integration began to be restructured in the 1990s, amid important processes of democratisation, the question of the participation of civil society in the integration process has been prominent in the debate. The Protocol of Tegucigalpa, signed in 1991 and entering into force in 1993, transformed ODECA into the Central American Integration System (SICA). It established as ‘Organs’ of the system: the Meeting of Presidents, Council of Ministers, Executive Committee and Secretariat-General. It recognised as ‘parts’ of the system, the Central American Parliament (PARLACEN), Central American Court of Justice (CCJ) and a Consultative Committee composed of ‘the entrepreneurial, labour and academic sectors and the other main driving forces of the Central American Isthmus which represent the economic, social and cultural sectors and are committed to the endeavour of isthmian integration’. The function of the Committee (CC-SICA) is ‘to advise the Secretariat-General on the policy of the organisation in the development of the programmes which it carries out’. The Committee is made up of 26 organisations, including private-sector federations, trade-union confederations, academic networks and regional associations of indigenous peoples, municipalities, and so on. Its task is to advise the Secretariat-General of SICA on integration policy; formulate recommendations and initiatives to the various regional bodies in order to promote integration; communicate the aims and

Indicators of regional governance 201 achievements of SICA to its members and national representatives; and so on. The Committee has been affected in various ways by the institutional complexity in Central America. There has also been some overlap between the public/ international structures and those set up by the private/civil world to promote integration. A Central American Council for Intersectorial Coordination (CACI) was set up in June 1991, and a ‘Civil Initiative for Central American Integration’ (ICIC) came into being in 1994, including many of the same organisations which form part of the CC-SICA. Various efforts have been taken to strengthen its impact, such as the creation in November 1997 of three Commissions – Agenda and Incidence, Organisation and Functioning, and Political Union – but little appears to have come of this. The confederation of public universities (CSUCA) has tried to serve as ‘facilitator’, with initiatives such as the creation of an internet ‘Integration Forum’ and the proposal in November 2000 to establish an ‘Agenda of Central American Civil Society’. Yet the CC-SICA has seemed to remain a purely ‘symbolic’ instance of civil participation (Campos, 1998; Monge, 2000: 281). Andean Community A formal body for consultation was foreseen from the very beginning of the Andean Pact in 1969: the original arrangements included an Economic and Social Advisory Committee made up of representatives of employers and workers. However, the Committee played a very limited role. In 1983, separate employers’ and workers’ councils were created as advisory organs, with ofﬁcial ﬁnance initially provided for participation in coordination activities. The Andean Employers’ Consultative Council and the Andean Labour Consultative Council – were given their present formal bases in July 1998, following the 1996 transformation of the Andean institutional arrangements into the Andean Integration System (or Andean Community). The April 1998 Presidential Summit in Guayaquil explicitly requested the Secretary-General to prepare a proposal for the ‘organised participation of civil society in the construction of the Andean Community, which should be complementary to the participation of employers and workers’. Other forms of participation have been explored, including proposals for a formal economic and social consultative committee with three sectors as in the European case; or for a council of Andean civil society in which organisations from all sectors could participate (Ramírez, 2000). The main change in practice has been the modiﬁcation of the ‘Simón Rodríguez’ Convention which has provided the formal basis for discussion of labour issues in the past. Although at ﬁrst there was talk of this being expanded into a forum also including ‘civil society’, the Protocol replacing the original Convention which was approved at the XIII Presidential Summit on 24 June 2001 makes no such reference. It is a tripartite body devoted to ‘socio-labour’ questions, with a ‘Conference’ composed of the Ministers of Labour, the Coordinators of the National Chapters of the Employers’ Council and the Coordinators of the National Chapters of the Labour Council, supported by specialised working committees.

202 Edward Best Mercosur The March 1991 Treaty of Asunción which brought together Argentina, Brazil, Paraguay and Uruguay in the Common Market of the South (Mercosur) did not provide for any institutionalised forum for the participation of non-governmental actors at regional level. However, Chapter VII of the Rules of Procedure of the Common Market Group (CMG),20 adopted in December 1991, explicitly provides for participation by the private sector in the ‘preparatory phase’ of the work of SubWorking Groups and Commissions, although not in the ‘decision-making phase’. Moreover, the CMG in late 1991 created a new Sub-Working Group No. 11 on ‘Labour Issues’. In early 1992, this was renamed ‘Labour relations, employment and social security’, with a tripartite participation which, by way of derogation from the general rule on private-sector participation, established that the representation of employers and workers from each country would be the same as that of the countries to the International Labour Conference.21 Since 1994, two main developments have taken place. First, the Economic and Social Consultative Forum (Foro Consultivo Económico-Social del Mercosur – FCES) was created in the Protocol of Ouro Preto which established the deﬁnitive institutional structure of Mercosur, coming into force in December 1995. It is composed of an equal number (nine) of representatives from each country. There is an equal representation of employers and workers. Each of the ‘national sections’ determines how to include ‘Other Interests’, whose nature has varied significantly according to the country: consumer organisations in Argentina and Brazil; cooperatives in Paraguay; and cooperatives, university circles, and NGOs in the case of Uruguay. A second major development was the institutionalisation of a tripartite regional structure concerning labour rights. A ‘Charter of Fundamental Rights’ was included in the tasks to be carried out by Sub-Group 11 in the work programme approved in Las Leñas in June 1992. Work on such a charter – which seems to have been understood by others from the beginning more as a ‘Social Charter’ – was effectively suspended after 1994, but taken up again in 1997. The result was the ‘Socio-Labour Declaration’ (Declaración Sociolaboral) adopted by the Presidential Summit in Río in December 1998. The Declaration recommended establishing a tripartite Regional Social-Labour Commission (Comisión Sociolaboral Regional). This was formally created by the CMG in March 1999 and its Rules of Procedure approved by the CMG in April 2000.22 There are to be 12 full members from each member state. The members from the employers’ and labour sectors are to be designated ‘by the organisations which are respectively most representative in each member state, in accordance with their national practices’. It mainly focuses on principles and rights in the area of labour, but also includes the encouragement of ‘social dialogue’ both at national and regional level (Art. 13) and the commitment of member states to adopt employment-promotion policies (Art. 14).

Indicators of regional governance 203

Structural problems in making regional fora work Representativeness and resources The problem of resources is even more acute in Latin America than in Europe, with the inevitable consequence of unequal access. Yet solutions along the lines of ‘core-funding’ in the interests of NGOs’ independence and balanced representation are even more improbable, given that ﬁnancial resources are not as available (except from outsiders) – and many of the problems of selection, representativeness and accountability are just as difficult to overcome. In the case of FCESMercosur, the Forum’s mandate can only be described as wildly out of line with the resources. The Committee is to evaluate the social and economic impact of integration policies; propose norms and economic and social policies; carry out research and organise events; and so on. Yet the FCES does not receive public ﬁnancing at any level and, as some of its own leaders have recognised, it simply does not have the technical capacity to provide inputs (e.g. Padrón, 1998; Ermida, 2000). In Central America, the activities of the Consultative Committee have to be ﬁnanced by the member organisations themselves, which in practice tends to mean that participation is based on projects financed by outside sources, with obvious consequences for continuity (Monge, 2000: 283). Requests have been made to the Presidential Summit for increased support, notably at the meeting in Costa Rica in December 2002, but with limited results. In an ambitious and advanced integration process, as in any political system, there may therefore be a case for public action to ensure a reasonable balance in interest representation. The natural ‘pluralist’ tendency is towards unequal access to decision-makers by those private interests which are best organised and have the most resources (large ﬁrms, industrial lobbies) at the expense of more ‘diffuse interests’ such as consumers and environmentalists. The European Commission therefore devotes signiﬁcant resources to funding European public-interest groups, many of which depend on this funding for their actions. Some such central mechanism may be the only way in theory to deal with the problem of unequal access. Yet there are obvious practical problems. On the one hand, if the central organs try to consult everyone without discrimination, then there will be no clear result at all, as UNICE, for example, strongly argued in its preliminary contribution to the European Governance debate in 2000. On the other hand, to offer ﬁnance for contrary opinions could sow the seeds of increased debate over issues about which there is in principle no reasonable disagreement (Lebessis and Paterson, 2001). Even assuming the availability of resources, how should organisations be chosen for support? The European Commission thus tries to evaluate an NGO in terms of its representativeness; the transparency and mode or operation of its organisation; its track record as regards competence to advise in a speciﬁc ﬁeld; and its capacity to work as catalyst for exchange of information and opinions between the Commission and the citizens. In any regional context, there will have to be some effort to evaluate which organisations are most ‘deserving’ on absolute grounds,

204 Edward Best such as demonstrable inﬂuence over and support by national constituents; mutual recognition between the parts, if appropriate; and transparency and accountability. Yet this will inevitably have a speciﬁc instrumental focus. Civil society organisations – and the regional fora in which they serve – must be able to bring something which is felt to be helpful to the regional policy-making process. Governments and international public bodies will have to do more in the way of transparency and of listening, but this is not a one-way street. Economic and social associations, and civil-society organisations in general, must offer decision-makers something useful if they are to carry weight. As already noted, Opinions of the European ESC are generally not considered useful in any specific way. The same seems to apply elsewhere. In the Andean Community, for example, the presence of the Labour Council is said to be little more than formal and ritual and the proposals made by the union sector lack depth, due to a lack of resources and the gap between the representatives and their organisations (Ramírez, 2000). The same story is told in the Caribbean – dispersion (even conflict) between organisations, their lack of technical capacity, problems of representativeness and so on (Jácome, 2000: 257–8). Here there may inevitably be something of a circular dynamic, inasmuch as it is those bodies which have the greatest capacity to be useful which will be supported. Yet at a time (and in most areas of the world) when regional public resources are scarce, there will be little choice but to try to come up with input which is felt to matter. Regional fora and regional socialisation The weakness of transmission mechanisms between representatives and bases is not only a cause for concern with regard to the ‘instrumental’ value of civil-society organisations in respect of decision-making. The gap also has serious implications for the socialisation process. The point is not that socialisation does not happen. It is whether regional fora can effectively inﬂuence the perceptions of grassroots organisations and citizens with regard to integration. A clear gap has been noted in Central America, for example, between the representatives in the formal consultative process and the bases of the organisations – made worse by the lack of adequate information to the public in general and the dispersion of civil society itself (Monge, 2000; de Monterrosa, 1999; Campos, 1998).23 Regional fora, except in small groupings with small populations, are likely to have a limited direct impact: representatives can undergo signiﬁcant socialisation between themselves without much impact on the perceptions of national societies. Socialisation takes place in the ﬁrst instance between ‘boundary-role occupants’ who ‘tend to be socially, politically and culturally distant from members of their sponsoring state or organisation and close to more similarly situated individuals from other countries with whom they are in continuous interaction (Mingst and Warkentin, 1996: 184). This kind of ‘summit socialisation’ may in some cases be as far as things go. It is interesting to read, for example, statements of concern not only that CARICOM has been limited to intergovernmental committees, with the danger that grassroot organisations and the private sector are not adequately

Indicators of regional governance 205 involved, but also that this essentially intergovernmental organisation needs to do more ‘to improve its interaction with member states’ (which reminds one of the old joke about the Andean institutions: ‘so supranational that no one knew they were there’). At most, socialisation may only reach into those sectors of national political, economic and administrative life which are directly involved. One may speculate that this could actually lead to the creation of a new distance between a caste of regional actors and the citizenry at large (just as globalisation is creating a new kind – and scale – of transnational class division). Moreover, where there is, as in Europe, intensive interaction between ofﬁcials, this may create a false impression within political-administrative circles as to the real depth of international socialisation – or social internalisation – which has actually taken place between societies. Something of this nature may well have been present before the Danish referendum of 1992 and the Irish referendum of 2001, when, in both cases, there was a clear, and apparently unexpected gap between the political class and the public.

Making the forum ﬁt the needs: ‘democratic deﬁcit’ or ‘integration deﬁcit’? In the European Union, the problem of integration may well be to have gone ahead of (and conceivably against) public opinion – whether it is the governments, in ditching national discretion in ever-wider areas, or the Court of Justice, in developing federal-ish legal principles. The challenge has been to ﬁnd means of formal democratisation, mainly through an extension of the powers of the European Parliament, and social legitimation, through new forms of participation of civil society. This is not a pattern which is common elsewhere – and Europe does not yet seem to have the answers to its own problems in this respect. This begs the question: in addition to the kind of structural problems indicated above, may there not be problems in the basic conceptualisation and design of the institutionalised fora for the participation in regional integration processes? Legitimation without commitment? First, what happens if ‘legitimation’ is presented as a need (usually as a ‘deﬁcit’) and regional fora are created – but the ‘commitment institutions’ are not actually there? None of the institutional arrangements established to date in the Americas have in the end entailed any such commitment. Yet, for other reasons, regional fora have emerged which seem to justify their existence, at least in part, on the grounds that they should control regional decisions. From this perspective, it may seem, for example, that the very role of the Central American Consultative Committee is indeed ‘irrelevant’, inasmuch as that it is supposed to advise an entity, the Secretariat-General of SICA, which is not a decision-making organ but an administrative-executive body (de Monterrosa, 1999). Yet it is far from clear that Central American governments are prepared to

206 Edward Best sacriﬁce national discretion in order to adopt common decisions. Even if they were, it would still be open to doubt what degree of speciﬁc input to the formulation of concrete common decisions could really be expected from such a heterogeneous grouping, given the predictable difﬁculties in reaching consensus over details, and the lack of technical capacity to deal with many issues. Even more striking is the case of the Central American Parliament, which was ﬁrst created with European support in the 1980s in the context of the regional peace process and has been maintained until the present, with strong links to the European Parliament. Its members are directly elected, a fact that has created a positive impression in some quarters as a manifestation of a commitment to deep integration. However, this superﬁcially supranational body exists in the absence of a real supranational legal or institutional system (and without the participation of Costa Rica). Its main role and challenge would seem to be not so much to acquire a formal role in the adoption of common regional decisions (far less to make regional laws), as to provide credible leadership in promoting cooperation and developing a modern Central American identity. However, the mismatch between the Central American Parliament’s pretensions and formal characteristics, on the other hand, and the needs and expectations of Central Americans, on the other, has generally contributed to a weakening of popular support for formal integration rather than the opposite. Institutionalisation without credibility? A related problem derives from perceptions that the formal regional framework is not the most attractive or credible channel for economic interests (or civil society organisations) to pursue. The case of the European ESC has already been noted above. Another example may be drawn from the Andean Community. The weakness of the Employers Council is partly a problem of representativeness but is more a reflection of the underlying realities, notably the division of the private sector into three groups: transnational companies which can implement regional strategies without any integration agreement; the large national economic interests, which prefer to use their direct political influence and do not want any institutionalised representation since informal methods give better results; and the medium and small companies which just try to follow the others (Ramírez, 2000). Integration without ﬁxed borders? A third issue arises where regional fora are set up as sources of pressure on governments to move further in the direction of formal integration, and probably to establish stronger, greater ‘commitment institutions’ (i.e. societal mobilisation applied to the states), but there are no clear borders within which that integration and institutionalisation should take place. This is particularly evident in the Americas. The FTAA – or any agreement with North America – can seem to provide an alternative structure, thus creating the potential for dilution of deeper sub-regional arrangements, at least in the economic ﬁeld. The multiplicity of levels

Indicators of regional governance 207 of membership is chronic in Central America, even between the core ﬁve countries, with additional complications deriving from the participation of Panamá, and now Belize and even the Dominican Republic, quite apart from Mexico and the rest (including the very curious case of Taiwanese participation in Central American affairs). Regional decisions or parallel actions? In circumstances of low commitment, low legalisation and ﬂuid borders, it may be asked whether it is in fact a major role at all of regional fora to try to inﬂuence regional decisions, in the sense of the adoption of common binding rules. It may be more realistic and effective to concentrate on, ﬁrst, the national dimension of law and, second, informal mechanisms to ensure compliance – or just promote convergence. This does in fact appear to be happening both in the Americas and in Europe. These are basic NAFTA principles, by which each country binds itself to apply national laws and opens itself to transnational pressure from its partners and from individuals. In Mercosur, the Regional Social-Labour Commission is mandated to review the annual reports submitted by each of the member states, and to formulate action plans and recommendations with the aim of promoting application of the social rights contained in the Declaration. This is to be done in the first place through each of the National Commissions. Some argue that a body with the legal capability to impose sanctions is still required in order to ensure effective ‘compliance’ with guidelines, such as an International Labour Tribunal with authority within Mercosur (see e.g. Ruiz-Tagle, 2000). While recognising that questions remain as to the prospects for consensus, the technical capacity of the Commission(s) and the real ability of the arrangements to inﬂuence public opinion, however, all this can equally be seen as a very important move away from the emphasis on inﬂuencing common decisions at the regional level and away from law as the principal instrument, towards a system which relies for fulﬁlment of shared commitments on domestic forces (legal, political, public-opinion) and peer pressure between the states. In this respect, there seems to be some convergence with the evolution of employment strategy in the European Union and the emergence of the Open Method of Coordination. Perhaps most interesting is the case of Central America. A parallel interparliamentary system has in fact emerged (FOPREL), which is based on the national assemblies and aims to bring about convergence of national laws around common models, very much in the spirit of Nordic ‘parallel national action’. This would, in many observers’ and actors’ opinion, be much more appropriate to the needs of Central America. Legitimation or community-building? Continuing this point, it may well be argued that regional fora in Latin America should concentrate their attention on transnational community-building.24

208 Edward Best The issue in Mercosur may be less whether the FCES’s opinions are heeded in the intergovernmental decision-making organs, and more how the Forum can contribute to the further emergence of ‘regionalisation’ in Mercosur, in the sense that ‘many social actors are starting to take the “region” and the regional processes into account in the design strategies of action: these include scientiﬁc and university communities, feminists, environmentalists, indigenous peoples, human rights groups and artistic communities’ (Jelin, 2001: 94). Central America appears to be a case in which this emphasis would be particularly appropriate. Deep ‘commitment institutions’ remain elusive, partly because of internal differences and disputes – and partly because the allure of broader cooperation and integration projects has still further reduced any pressure to join forces. There is indeed a large ‘gap’ between the regional and the national agendas, but the answer may not be to create regional fora which attempt – with all the problems outlined above – to close the gap and inﬂuence regional decisions. In this respect it is interesting to quote the 1996 ‘Position’ of the Consultative Committee of SICA with regard to the ‘evaluation’ of the Central American institutional system then beginning to be carried out. ‘In the face of a vertical, external integration, based in market competition and exclusion, which only permits the integration of the big corporations, there should be a strengthening of horizontal, participative integration, between the peoples. . . . ’25 This partly misses the point – market integration based on credible commitments can bring social as well as economic beneﬁts. Yet the emphasis on ‘horizontal’ integration is spot on. Central America has had problems in consolidating formal political and economic integration projects, but has been exceedingly (even excessively) rich in organisations of regional cooperation, ranging from bodies with signiﬁcant ﬁxed installations and staff, to virtual ‘instances’ of coordination. The ﬁrst were created between 1948 and 1955, often with UN support, in the areas of universities, nutrition, agricultural health, public administration and industrial research and technology. Others were set up during the ﬁrst period of integration, in the areas of air trafﬁc control, tourism, hydraulic resources, telecommunications, education and culture, science and technology, electrical energy, maritime transport. Yet more sprang up from the late 1980s and the early 1990s in agriculture, natural disasters, the environment, water, hydrocarbons, social security, housing, sport, ‘social integration’, anti-drugs. The challenge remains how to reconcile the pursuit of the potential beneﬁts of deeper integration in certain respects (which requires the centralisation of selected core strategic functions, although not the regional bodies themselves, certainly not as they stand) with coherent non-centralisation in functional cooperation, and the deepening of transnational ties between societies.

Is inappropriateness indicated? In many cases, these structures were set up with the best intentions of increasing openness and public participation, and some positive results can certainly be identiﬁed. However, on balance, the evidence suggests that many formal fora have not been effective, largely because they were not appropriate to the realities of the

Indicators of regional governance 209 political context. They have rested on questionable assumptions about the needs of the particular regional system and the legitimacy expectations of the societies involved, quite apart from the practical realities of regional networks and civil society organisations. As a result they have often ended up – even started up – relying on external support. It is not clear that there is a signiﬁcant impact in terms of socialisation and societal mobilisation (which is the most meaningful form of regional ‘participation’ in the long term). And in some cases these structures may even have been counter-productive, by creating an undeservedly negative image of regional institutional arrangements as a whole.

CONCLUSIONS The aim of this chapter has been to explore whether regional systems can be assessed according to common ‘indicators’ of (good) regional governance. It is argued that regional systems do need to be evaluated in terms of their democratic (i.e. ‘good-governance’) quality and that such an exercise is plausible. However, there are some important caveats at the start. In the ﬁrst place, such an evaluation cannot be based on the degree to which a system appears to conform to any particular set of institutional arrangements, or even particular governance structures. Second, one cannot expect to come up with ‘indicators’ in the sense of quantiﬁable measurements relating to any absolute scale. Nor should one presume to give a relative ranking in any absolute sense. It may be possible to establish some common principles and standards, however, by adopting a ‘governance’ approach which merges the following assumptions. On the one hand, there are some principles of societal organisation and behaviour (e.g. transparency, accountability, participation) which are nearly universally recognised as shared norms and which can be promoted without insisting on particular institutional structures. On the other hand, traditional statecentred paradigms are no longer sufﬁcient to understand how decisions are made and conduct is regulated either nationally or internationally. Regional arrangements which go beyond the nation-state must therefore be evaluated in new ways, taking into account the democratic dilemmas involved concerning both scale and the tension between publicly-recognisable images and the actual needs of a system. Regional bodies need to respect the most basic good-governance principles such as transparency and accountability, whether they are supranational institutions, intergovernmental organisations, regional projects or transnational private associations. Standards can be drawn up and monitoring structures established more or less as happens at national level. The question addressed here, however, has been more whether one can evaluate the quality of regional institutional arrangements (independently of the standards applied to the operation of the organisations), in those cases where an integration process produces a regional system of signiﬁcant multilevel governance. In this

210 Edward Best perspective, there is quite a different relationship between the standards and structures which can be applied at regional and at national level. Designers and analysts of regional systems face two temptations: on the one hand, to try simply to apply models drawn from the experience of national governments; and on the other hand, simply to apply models drawn without adaptation from the experience of other regions. Yet experience indicates that both generally fail, and can lead to the creation of regional bodies which absorb resources but are quite out of political context as well as out of proportion to real systemic needs. It is therefore argued that – beyond the basic operating standards, and assuming that the regional organisation does not constitute a negative impact on good governance within countries – the quality of a particular set of regional arrangements can be evaluated in terms of stability and legitimacy. This means not only that the institutional arrangements have to be speciﬁcally designed to respond to the particular pressures and complexities involved, but also that they must do so in ways which are appropriate to local realities. A set of basic principles is therefore suggested by which regional bodies and systems can be assessed: openness, accountability, participation, effectiveness and appropriateness. The aim is only to consider a possible set of common principles and standards, and not at this stage to propose speciﬁc indicators. Indeed, it is stressed that one kind of indicator which is commonly used – namely, the existence of formal institutional structures and instruments – is not only often assumed to indicate a degree of real integration, without further question. In practice, and most notably in the framework of inter-regional cooperation, this dimension is implicitly also a standard – that is, an indicator that the regional arrangements correspond to a particular institutional model which is assumed to represent good practice. It may be, but not necessarily for everyone, and certainly not in the same way.

NOTES 1 For a good example of an application of this approach to a particular region, see Sundelius (1978). 2 For example, this has been periodically carried out for the EU by Eurobarometer in ‘vertical’ terms of whether people see themselves as ‘European only, European and nationality, nationality and European, or nationality only’. Interestingly, a more horizontal, ‘transnational’ approach was reﬂected in the survey carried out in Central America in 2002 by CID Gallup Latinoamérica which asked whether respondents were strongly in agreement with the propositions that the Central American countries ‘are fraternal countries’ or that the respondent was indifferent to being a national of his/ her own country or of any other Central American country. . 3 In this perspective, regions are seen as ‘a system of intentional acts’; ‘a “rational” system with statehood properties’; ‘reciprocal achievements of social actors’ (in which case ‘regioniﬁcation’ is involved); and as ‘generators and communicators of meaning and identity to social and personal actors’ (Van Langenhove, 2003).

Indicators of regional governance 211 4 Or rather, non-things – some proponents of this perspective argue that it is in fact ‘inappropriate to try to quantify and measure a concept, such as “identity” or “regionness” or “integration”. To do so would be to commit the error of reiﬁcation’ (Slocum and Van Langenhove, 2003: 8). 5 ; . 6 This approach has been particularly developed by the Organisation for Economic Cooperation and Development (OECD), which is not regional in any territorial sense. 7 ‘Objectives, Standards, Criteria and Indicators for the Africa Peer Review Mechanism (“the APRM”)’, NEPAD/HSGIC-03-2003/APRM/Guideline/OSCI 9 March 2003; ‘African Peer Review Mechanism: Organisation and Processes’, NEPAD/HGSIC-32003/APRM/Guideline/O&P, 9 March 2003. 8 Algeria, Burkina Faso, Cameroon, Congo, Ethiopia, Mozambique, Nigeria, Ghana, Kenya, Rwanda, Senegal, Uganda, Gabon, Mauritius, Mali, South Africa , 19 January 2004. 9 The useful distinctions proposed by the OECD in its study on government–citizen relations are as follows: ‘Information is a one-way relation in which government produces and delivers information for use by citizens. It covers both ‘passive’ access to information upon demand by citizens and ‘active’ measures by government to disseminate information to citizens . . . Consultation is a two-way relation in which citizens provide feedback to government. It is based on the prior deﬁnition by government of the issue on which citizens’ views are being sought and requires the provision of information . . . Active participation is a relation based on partnership with government, in which citizens actively engage in deﬁning process and content of policymaking. It acknowledges equal standing for citizens in setting the agenda, proposing policy options and shaping the policy dialogue – although the responsibility for the ﬁnal decision or policy formulation rests with government’ (OECD, 2001: 23). 10 Signiﬁcantly, in the period in which the greatest advances were made, the Nordic Council was based on resolutions of the respective parliaments. It was not until 1962 that the Helsinki Treaty placed it on a more formal international basis. The Nordic Council of Ministers was only created in 1971 (see Nielsson, 1990; Sundelius and Wiklund, 1979; Sundelius, 1978). 11 The Treaty until recently spoke of ‘representatives of the various categories of economic and social activity, in particular, representatives of producers, farmers, carriers, workers, dealers, craftsmen, professional occupations and representatives of the general public’ (Article 257 TEC). 12 Opinion of the Economic and Social Committee on ‘The role and contribution of civil society organisations in the building of Europe’ (1999/C329/10) adopted on 22 September 1999. OJ C 329, 17.11.99. Par. 10.2.3 (emphasis added). 13 Ibid., 10.3. 14 See Articles 138 and 139 of the Treaty establishing the European Community. 15 ‘The Commission and Non-Governmental Organisations: Building a Stronger Partnership’, Response of the Platform of European Social NGOs, Brussels, 27 April 2000. 16 COM (2003) 276 ﬁnal of 27 May 2003. 17 First Convention on Civil Society organised at European level, Brussels, 15–16 October 1999. 18 The Caribbean Community (CARICOM) has a Joint Consultative Group comprising representatives of the Caribbean Association of Industry and Commerce and the Caribbean Congress of Labour and of NGOs. Civil society groups are represented by the Caribbean Policy Development Centre (CPDC), a network of 23 NGOs ranging from the Caribbean Conference of Churches to the Windward Islands Farmers Association. However, the CPDC has not received any formal status vis-à-vis CARICOM.

212 Edward Best 19 The Consejo Superior Universitario Centroamericano (CSUCA) and the Instituto de Nutrición de Centroamérica y Panamá (INCAP). This leading role of education in integrationist movements goes back even further, to the time of the 1907 peace conference which saw not only a Central American International Ofﬁce and a Central American Court, but also a Central American Pedagogical Institute. 20 MERCOSUR/CMC/DEC. No. 04/91 Reglamento Interno del Grupo Mercado Común. 21 MERCOSUR/GMC/RES. No. 12/92. This later became Sub-Group No. 10. 22 MERCOSUR/GMC/RES. No. 12/00 Reglamento Interno de la Comisión Sociolaboral del Mercosur. 23 An interesting image in use in Central America is reported in Edelman (2003): ‘when a leader originates at the base [and then] becomes bureaucratised and distant from the base, the people say that he’s become like a kite (se papaloteó), that he goes up and up into the sky, and then suddently the string breaks and he’s lost’. 24 It is interesting to compare the Latin American processes with recent ASEAN initiatives, such as the institutional recommendations in the ASEAN Eminent Persons Group’s Report on Vision 2020. ‘While ASEAN governments operate by way of consensus building, broad principles and procedures, rather than ‘structures and institutional frameworks’, it is argued that there may be scope for the formation of ASEANtype institutions in a variety of ﬁelds such as education, science and tehcnology, good governance and culture; and the ASEAN Secretariat should be strengthened in order to carry out its coordinating role and preparation of consultative meetings. Yet the emphasis remains on transnational and transgovernmental relations: all public and private institutions in ASEAN should be urged to play a positive role in promoting an ASEAN community spirit by expanding their cooperation with their ASEAN counterparts.’ 25 ‘Posición del Comité Consultivo del Sistema de la Integración Centroamericana ante la evaluación de la institucionalidad regional centroamericana’, Coronado, Costa Rica, 23 September 1996.

REFERENCES Abbott, K.W. and Snidal, D. (2000) ‘Hard and Soft Law in International Governance’, International Organisation, 54 (3): 421–56. Abbott, K.W., Keohane, R.O., Moravscik, A., Slaughter, A.-M. and Snidal, D. (2000) ‘The Concept of Legalisation’, International Organisation, 54(3): 401–19. Adler, E. and Barnett, M. (1998) ‘A framework for the study of security communities’, Security Communities, 29–65, Cambridge: CUP. Best, E. (1997) ‘Capacities for Regional Integration: a Conceptual Framework for Comparative Analysis’, in M.O. Hosli and A. Saether (eds) Free Trade Agreements and Customs Unions. Experiences, Challenges and Constraints, Brussels and Maastricht: European Commission TACIS and EIPA, 51–78. Britz, G. and Schmidt, M. (2000) ‘The Institutionalised Participation of Management and Labour in the Legislative Activities of the European Community: A Challenge to the Principle of Democracy under Community Law’, European Law Journal, 6(1): 45–71. Campos, W. (1998) ‘La participación social en la integración social centroamericana: la ICIC’, Participación de la sociedad civil en los procesos de integración, 255–68, Montevideo and San José: CEFIR/CLAEH/ALOP. Checkel, J. (1999) ‘Social construction and integration’, Journal of European Public Policy, 6(4): 545–60.

Indicators of regional governance 213 CEC (Commission of the European Communities) (1997) Communication on ‘Promoting the Role of Voluntary Organisations and Foundations in Europe’, COM (97) 241 ﬁnal. CEC (2001) ‘European Governance: A White Paper’, COM (2001) 428, Brussels. Corkery, J. (1999) ‘Introductory Report’, in J. Corkery (ed.) Gouvernance: Concepts et Applications/Governance: Concepts and Applications, 9–20, Bruxelles/Brussels: IISA/IIAS. Dahl, R. (1994) ‘A Democratic Dilemma: System Effectiveness versus Citizen Participation’, Political Science Quarterly, 109(1): 23–34. Edelman, M. (2003), ‘When Networks Don’t Work: The Rise and Fall and Rise of Civil Society Initiatives in Central America’, paper prepared for the 2003 meeting of the Latin American Studies Association, Dallas, Texas. Ermida, O. (2000) ‘La ciudadanía laboral en el Mercosur’ in B. Podestá et al. Ciudadanía y mundialización. La sociedad civil ante la integración regional: CIDEAL/CEFIR/ INVESP,175–209, Madrid/ Montevideo/Caracas. Finer, S.E. (1970) Comparative Government, Harmondsworth: Penguin. Hettne, B. and Söderbaum, F. (2000) ‘Theorising the Rise of Regionness’, New Political Economy, 5(3): 457–73. Jácome, F. (2000) ‘La dimensión social de los procesos de integración en el Caribe’, in Podestá et al. Ciudadanía y mundialización, 239–63. Jelin, E. (2001) ‘Cultural Movements and Social Actors in the New Regional Scenarios: The Case of Mercosur’, International Political Science Review, 22(1): 85–98. Keohane, Robert O., Moravscik, A. and Slaughter, A.-M. (2000) ‘Legalised Dispute Resolution: Interstate and Transnational’, International Organisation, 54(3): 447–88. Kondo, S. (2002) ‘Fostering Dialogue to Strengthen Good Governance’, in Public Sector Transparency and Accountability: Making It Happen, 7–12, Paris: OECD. Lebessis, N. and Paterson, J. (2001) ‘Developing New Modes of Governance’, in Governance in the European Union, 259–305, Brussels: European Commission. Mariscal, N. (1989) Integración económica y poder-político en Centroamérica, San Salvador: UCA (2nd edn). Mattli, W. (1999) The Logic of Regional Integration. Europe and Beyond, Cambridge: CUP. Mingst, K.A. and Warkentin, C.P. (1996) ‘What Difference Does Culture Make in Multilateral Negotiations?’, Global Governance, 2: 169–88. Monge, H. (2000) ‘Participación de la sociedad civil centroamericana en el proceso de integración regional’, in Podestá et al. Ciudadanía y mundialización, 277–88. de Monterrosa, C. (1999) ‘La sociedad civil frente a la institucionalidad del Sistema de la Integración Centroamericana (SICA)’, in R. Grinspun, C. Alvarenga and Y. Shamsie (eds) Hacia una Integración desde Abajo: Participación, Sociedad Civil e Integración Centroamericana, Agencia Canadiense de Cooperación, Proyecto SICA/CERLAC/ PAR, 71–83. Nielsson, G. (1990) ‘The parallel national action process’, in A.J.R. Groom and P. Taylor (eds) Frameworks for International Co-operation, 78–107, London: Pinter. OECD (Organisation for Economic Co-operation and Development) (2001) Citizens as Partners. Information, Consultation and Public Participation in Policy-Making, Paris: OECD. Osborne, D. (1999) ‘Governance, Partnership and Development’, in J. Corkery (ed.) Gouvernance: Concepts et Applications/Governance: Concepts and Applications, 37–65, Bruxelles/Brussels: IISA/IIAS. Padrón, A. (1998) ‘El Foro Consultivo Económico-Social del Mercosur’, in Participación

214 Edward Best de la sociedad civil en los procesos de integración, 245–54, Montevideo/San José: CEFIR/CLAEH/ALOP. Pelkmans, J. (1987) ‘Alternative Methods of Economic Integration’, in ASEAN at the Crossroads, 421–36, Kuala Lumpur: ISIS. Pelkmans, J. (1993) ‘Comparando las integraciones economicas: prerrequisitos, opciones e implementaciones’, Integracion Latinoamericana, l91: 3–17. Ramírez, S. (2000) ‘El aspecto social y la participación en la comunidad andina’, in Podestá et al. Ciudadanía y mundialización, 211–35. Rhodes, R.A.W. (1997) Understanding Governance. Policy Networks, Governance Reﬂexivity and Accountability, Buckingham and Philadelphia: Open University Press. Rhodes, R.A.W. (2000) ‘The Governance Narrative: Key Findings and Lessons from the ESRC’s Whitehall Programme’, Public Administration, 78(2): 345–63. Rosenau, James N. (1992) ‘Governance, Order and Change in World Politics’, in J.N. Rosenau and E.-O. Czempiel (eds) Governance Without Government: Order and Change in World Politics, 1–30, Cambridge: CUP. Ruiz-Tagle, J. (2000) ‘Las organizaciones sindicales frente a la exclusión social en el Mercosur’, Nueva Sociedad, 169. Schmitter, Philippe C. (1996) ‘Is it Really Possible to Democratise the Euro-Polity?’, Arena Working Paper, 10. Slocum, N. and Van Langenhove, L. (2003) ‘The Meaning of Regional Integration: Introducing Positioning Theory in Regional Integration Studies’, UNU-CRIS e-Working Paper, W-2003/5. Smismans, S. (1999) ‘An Economic and Social Committee for the Citizen, or a Citizen for the Economic and Social Committee?’, European Public Law, 5(4): 557–82. Sundelius, B. (1978) Managing Transnationalism in Northern Europe, Boulder: Westview. Sundelius, B. and Wiklund, C. (1979) ‘The Nordic Community: The Ugly Duckling of Regional Cooperation’, Journal of Common Market Studies, XVIII(1): 59–75. UNDP (United Nations Development Programme) (1997) ‘Reconceptualising Governance’, Discussion paper 2, Management Development and Governance Division, UNDP, New York. Van der Voort, W. (1998) ‘The Economic and Social Committee’, in M.P.C.M. Van Schendelen (ed.) EU Committees as Influential Policymakers, 250–73, Aldershot: Ashgate. Van Langenhove, L. (2003) ‘Theorising Regionhood’, UNU-CRIS e-Working Paper, W-2003/1. Weiler, J.H.H. (1991) ‘Problems of Legitimacy in Post-1992 Europe’, Aussenwirtschaft, 46(3–4): 411–37. Weiler, J.H.H. (1997) ‘Legitimacy and Democracy of Union Governance’, in G. Edwards and A. Pijpers (eds) The Politics of European Treaty Reform, London and Washington: Pinter.

11 Assessing accountability of global and regional organisations Monica Blagescu and Robert Lloyd 1

This chapter provides a detailed discussion of the nature of accountability within multilevel governance, identifies some of the reasons why organisations that operate across borders – be it at the regional or international levels – need to be accountable, and suggests how the accountability of such organisations can be assessed by drawing on the experiences of the One World Trust’s Global Accountability Project in developing global accountability indicators. In addressing the issue of the accountability of regional organisations, this chapter will not conﬁne itself to intergovernmental organisations, rather its focus will be on regional governance arrangements. Governance means more than government; it extends beyond state institutions to encompass non-state and private actors. This is in recognition that decision-making power has become dispersed. Actors such as non-governmental organisations and corporations have considerable inﬂuence over decision-making at national, regional and international levels. If accountability is conceptualised as the ability to hold power to account, then we must widen our gaze to cover all actors that wield power at the regional and global levels. Within the context of this chapter, therefore, the concept of regional organisations extends beyond that of intergovernmental organisations (IGOs), to include international non-governmental organisations (INGOs) and transnational corporations (TNCs).

ACCOUNTABILITY IN THE INTERNATIONAL CONTEXT

What is accountability? Although the concept of accountability has gained broad usage among policymakers at the national, regional and international levels, there is little agreement over what is actually meant by the term. At its simplest, accountability is about the processes by which individuals, organisations and other entities are answerable for their actions and the consequences that follow from them (GAP, 2003). Around this, there is consensus. However, where positions diverge is around the issue of who is entitled to hold these individuals and organisations to account, for what, and what mechanisms they should have at their disposal to do so.

216 Monica Blagescu and Robert Lloyd Traditional understandings of accountability According to traditional understandings, an accountability relationship exists when a principal delegates authority to an agent to act in their interests. Central to this view is that only those with formal authority over an agent – those that have delegated authority to it – have the right to demand accountability. The principle focus is on ‘internal accountability’ (Keohane, 2003), on how an agent is accountable to its internal stakeholders, usually those that live in the same polity or community, or work within the same organisation. This approach is often used to conceptualise the accountability relationship between politicians and the electorate and company directors and shareholders. Within this traditional view, holding an agent to account requires clearly deﬁned roles and responsibilities, regular reporting and monitoring of behaviour against these roles and the ability for principals to impose sanctions for breaches of responsibilities. Accountability is largely seen as an end stage activity where judgement is passed on results and actions already taken. We argue that this understanding is too narrow. Accountability needs to be more encompassing if it is to ensure that those with power are truly answerable for their actions to those they affect. The stakeholder view of accountability An alternative view is the stakeholder approach to accountability. This argues that the ability to demand accountability should not be determined by delegation of authority alone. The right to hold an organisation to account should be invested in ‘any group or individual who can affect or is affected by . . . an organisation’ (Freeman, 1984). This approach recognises that the impact of an organisation’s or individual’s actions is diffuse and thus responsibility should be also. Although an individual may not have delegated authority to an organisation to act in their interest, the activities of the latter may impact substantially on the former to warrant the establishment of an accountability relationship. This view is more wide ranging than traditional understandings. It emphasises both internal and external accountabilities and the need for organisations to balance the demands of each. The stakeholder view also recognises that accountability is more than an endstage activity. To ensure that an organisation is responsible for its actions, stakeholders need to be involved at every stage of the decision-making process. Passing judgement after a decision is made limits the extent to which an organisation/ individual can be held to account. Accountability needs to be ex ante, it needs to be an ongoing process. Understanding accountability on these terms extends the limits of the concept beyond its role as a disciplinary mechanism and towards its use as a transformative power (Young 2000). An organisation that is accountable to multiple stakeholders not only ensures that decisions are effective in meeting the needs of those interests, but also forces decisions to be made in a more equitable and fairer manner. As Young (2000: 118) posits, ‘speaking across difference in a context of public accountability often reduces mutual ignorance about one another’s situation, or

Accountability of regional organisations 217 misunderstanding of one another’s values, intentions, and perceptions, and gives everyone the enlarged thought necessary to come to more reasonable and fairer solutions to problems’. Decision making is more likely to be responsive to local needs the more it involves those directly affected by a decision. A system of multilevel governance widely characterised by accountability at all its levels also increases the probability of making decisions which benefit the poor and marginalised (Newell and Bellour, 2002). This more open and participatory approach unlocks the potential of accountability as an agent for organisational change and learning. Accountability that is pursued on an ongoing basis opens up space for those affected by an organisation’s policies to input into the decision-making process. This in turn creates positive feedback loops that enable organisations to learn from what is effective and what is not. When understood on these terms, accountability is no longer simply a mechanism for disciplining power, but also a force for social and organisational change. Yet the stakeholder approach to accountability is not without its problems. It presents significant challenges to organisations, as it requires balancing the needs of multiple, oftentimes competing, interests. Each stakeholder has different resources (particularly ﬁnancial) and capacity; each has different degrees of access to reliable information; each has different needs and expectations; each has different interests in the success or failure of a particular organisation; and different abilities to help or hinder its activities. Furthermore, organisations themselves have vested interests in each of the stakeholder groups. The result is a complex web of competing accountabilities, in which power imbalances characterise both the relationships between organisations and stakeholders, and between stakeholders themselves. In order for stakeholder accountability to produce greater organisational accountability, balance must be struck in how these different stakeholders are engaged. Traditional notions of accountability as described above, have limited use in helping us develop an understanding of accountability for organisations that operate across borders. The impact of decisions made at regional and global levels is too diffuse for an accountability relationship to be understood within the limited parameters of the delegation of authority between principal and agent. To be effective in holding power to account, accountability must be understood according to the stakeholder approach where all actors signiﬁcantly impacted by a decision have input into it. Based on this understanding, the next section will outline what accountability means at the global level, identify where the accountability gaps are and suggest ways in which they can be plugged.

Accountability at the global level Decision making at levels above that of the nation state is an unavoidable reality. The web of connections that binds us globally through communication, trade and finance mean that few problems are confined in their impact to within state boarders. Problems such as terrorism, global warming and ﬁnancial instability are

218 Monica Blagescu and Robert Lloyd global in nature. They have a global impact and require coordinated state action. With the multiplication and diversiﬁcation of these problems, however, there is growing recognition that states alone lack the human and financial capacity to manage them. To plug these capacity gaps, non-state actors such as TNCs and INGOs have emerged as important actors in the governance of global affairs. With this in mind, the institutional landscape of global governance2 should be viewed as a multilayered, multidimensional and multi-actoral complex in which diverse agencies participate in the formulation and conduct of global public policy (Held and McGrew, 2002). While this provides greater ﬂexibility and capacity for addressing global problem solving, it poses new and significant challenges for accountability. Actors that operate across state borders have been invested with the political power and inﬂuence to determine policies that have a diffuse impact across a number of levels from the local to the national, to the regional and to the global. Yet they wield this power in the absence of adequate accountability mechanisms. There are, for example, over 300 intergovernmental organisations (IGOs) making decisions in cities such as New York, Washington and Geneva that have a substantial global impact, yet they remain inadequately accountable to the communities they affect (GAP, 2003). IGOs are supposed to gain most of their accountability from the states that make up their membership; however, few within the electorates of these countries will know the name of their IGO representative, let alone what decisions are being taken on their behalf. The reality is that few governments feel it is necessary to consult with their citizens in developing positions at the international level. For voters in developing countries this accountability gap is even more pronounced as their government representatives often have limited power to inﬂuence an IGO’s decisions. In the worst cases, governments may not even be democratically elected. Other channels for representation do exist within IGOs, such as through civil society organisations (CSOs), but these remain in nascent stages of development. Although there is no doubting that CSOs have an impact on global organisations, there are questions surrounding their ability to serve as an effective and legitimate channel for citizen representation. This is largely due to their questionable accountability. Transnational corporations (TNCs) and international non-governmental organisations (INGOs) play a similarly powerful and as of yet unaccountable role in the governance of global affairs. There are currently over 60,000 TNCs and 40,000 INGOs making decisions that profoundly affect individuals and communities, particularly the poor, around the world (GAP, 2003). TNCs are said to be held accountable by state regulation, however, increasing mobility of private capital has largely undermined this mechanism, especially in the developing world. Likewise, INGOs may have a strong accountability relationship with their donors given the financial leverage they can exert, but what about with their beneficiaries? How accountable are they to those they claim to represent? In many ways, the channels for holding the INGO and TNC sectors to account are even narrower than those for IGOs.

Accountability of regional organisations 219 To complicate things further, new organisational forms now exist at the local, national, regional and global levels between IGOs, TNCs and INGOs, where responsibilities and obligations are unclear. Who is ultimately accountable for the delivery of public services through public–private partnerships, for example, the state or the private partner? In sum, the increasing complexity and density of the relationships between actors across national, regional and global levels has led to the emergence of a multitude of accountability gaps (Newell and Bellour, 2002) – ﬁssures between those that govern and those that are governed that prevent the latter from having a say in, and inﬂuence over, decisions that signiﬁcantly impact upon their lives. The future legitimacy and effectiveness of multilevel governance rests on our ability to plug these gaps. Accountability gaps at regional level The accountability of organisations that operate at the regional level has yet to be explored in the context of this chapter. In many ways, however, this debate, mirrors that of global organisations. Regional organisations are a necessary and critical middle ground between international and national levels. Dealing with regional issues, such as natural resources for example, requires regional management. Nation states cannot effectively address the issues alone, and global solutions are beyond their scope. Regional organisations also present forums through which greater coordination in global organisations can be achieved. Through the African Union, for example, the 53 member states developed a common trade position for the WTO Ministerial Meeting in Cancun. Regional organisations encourage coordination and cooperation among the multiple levels of governance. In developing and enforcing rules and common positions, however, it is imperative that regional organisations are both accountable to those they represent and those that will be impacted by their decisions. The accountability mechanisms which currently guide their business are largely ineffective. Analogous to global organisations, the formal channels for representation among regional organisations are ill-equipped to ensure meaningful accountability. For example, Benz (2001) highlights how weaknesses in the accountability chain between a country’s citizens and its national parliament and the national parliament and the European Council undermines the European Union’s democratic accountability. In principle, Council members are individually accountable to their respective parliaments, thus citizens should be able to hold them to account through their MPs. In practice though, the relations between national governments and parliaments in EU affairs are informal and citizens are rarely privy to what Council members are doing. Even though national parliaments hold a veto power over EU decisions, it is thought that if they use these powers they will undermine the effectiveness of the EU policy-making process. Instead, national parliaments opt for ex post control of the negotiating behaviour of Council members. This informal mechanism could ensure notional accountability as it at least informs citizens of

220 Monica Blagescu and Robert Lloyd their government’s EU policy. However, politicians often fear such information could be used against the government, thus in practice it is never disclosed. As a result, conflicts between government and national parliament happen largely in secret through informal consultations. Benz’s conclusion is that, as long as European policies are debated behind closed doors, citizens are in no position to inﬂuence them by their voting behaviour (Benz, 2001: 17). What this reveals is that the current state-based understandings of accountability that are said to underlie regional institutions, are largely ineffective. The way in which regional economic agreements are negotiated also contributes to the lack of accountability at the regional level. Take the Free Trade Agreement of the Americas (FTAA) for example; backed up by a rule enforcement mechanism similar to that of the WTO’s dispute settlement mechanism this is a regional economic agreement which will undoubtedly have considerable impact on the policies of national governments (Barlow and Clarke, 2003). Groups such as Public Citizen have claimed that it will lead to massive job loss, a race to the bottom in labour and environmental standards, attacks on local democracy and sovereignty, unsafe food and increasingly corporate control through privatisation and deregulation. Whatever the validity of these arguments, it is clear that the FTAA lacks legitimacy in the eyes of many citizens because of the untransparent and unaccountable way in which the agreement is being negotiated. Although corporations and corporate lobby groups have been involved in the negotiation process from its conception, the public have had to wait years to see what is on the table and this only after a successful public campaign for greater transparency. An agreement negotiated under these terms will be plagued by claims of illegitimacy and protest. This is a clear example of how and why regional organisations need to become more accountable to those they impact. Although INGOs and TNCs do not have the authority to enforce policies at the regional level, they can still have a wide-ranging impact on individuals and communities. They exercise this inﬂuence through their lobbying power. TNCs use their economic leverage to shape policies and regulations in their favour, while INGOs, lacking the economic power of the corporate sector, exploit their aptitude for moral persuasion and public mobilisation to push their demands. A report published in 2003 by the European Parliament, for example, found that more than 70 per cent of EU lobbyists worked for corporate interests, while 20 per cent represented non-governmental organisations such as trade unions, public health organisations and environmental groups. Ensuring accountability of both these sets of organisations is thus also necessary for regional governance arrangements to be accountable to those they affect. Although indicating that both sectors are inﬂuential in inﬂuencing regional structures, the disproportionate inﬂuence of the corporate sector in relation to INGOs should not be overlooked. Traditional forms of accountability are no longer appropriate for understanding accountability in the context of multilevel governance. State-based accountability is ill-equipped to provide those inﬂuenced by a global or regional decision with an adequate voice in how that decision is made. Furthermore, given states no longer have a monopoly over power at the regional level, mechanisms also need to be

Accountability of regional organisations 221 developed that hold TNCs and INGOs to account. It is clear that new understandings of what accountability means and how it can be institutionalised need to be developed. Why do we need to plug the accountability gaps in multilevel governance? First, the legitimacy of an organisation depends on its accountability. Global and regional institutions such as the World Bank, the IMF, WTO, the EU and NAFTA are coming under increasing scrutiny because they are not accountable for their actions, or more accurately they are accountable for their actions to only their most powerful stakeholders. For an organisation to be legitimate it must have the backing of all relevant stakeholders. If organisations that operate across state borders fail to recognise the need for greater accountability within their decisionmaking structures and thus the inclusion of those most affected by their decisions, protest and conﬂict will continue to accompany their activities. In the longer term, there is the danger that this may lead to states refusing to delegate necessary additional authority to these institutions (Kahler, 2005) and thus the efﬁcacy of their activities being undermined. Second, as discussed above, increased accountability at the global and regional levels is likely to produce better policy outcomes. The sense of ownership that is created through effective accountability mechanisms leads to more relevant decision making and ultimately to better implementation. The IMF and World Bank’s recent decisions to use Poverty Reductions Strategy Papers (PRSPs) as their principle framework for lending highlights this. Developed through a participatory process, PRSPs are designed to generate national ownership by getting all stakeholders to buy into the process. Genuine involvement of stakeholders in the decision-making process is likely to lead to policies that will respond to the needs of the poor. The openness and implied fairness of the process makes it more accountable to those it will affect. Decisions are more likely to be implemented and sustainable if a social consensus exists around them. Such policies are most likely to emerge from an organisation accountable to all its stakeholders. Third, accountability gaps need to the plugged because of the inherent right of citizens to input into the decisions that impact upon their lives. As the world becomes more interconnected and as a greater number of policies are decided at the regional and global levels, so we need to adjust our understanding of citizenship to reﬂect this. Held (2005) proposes the idea of ‘multilevel citizenship’ where the meaning of citizenship shifts from membership in a community which bestows, for those that qualify, particular rights and duties to an alternative principle of world order in which all persons have equivalent rights and duties in the cross-cutting spheres of decision making which affect their vital needs and interests. It posits the idea of a global political order in which people can enjoy an equality of status with respect to

222 Monica Blagescu and Robert Lloyd the fundamental processes and institutions which govern their life expectancy and life chances. (Held, 2005: 262) Last, for regional organisations in particular, increasing accountability can have important positive impacts on furthering regional integration. The higher level of stakeholder participation that is fostered through increased organisational accountability can serve to build stronger cross-border interpersonal and interorganisational relations. If regional policies are developed through engaging all stakeholders rather than just through elite pacts, issue-specific, epistemic communities can be created that push regional integration beyond merely inter-state relations and towards the forging of social networks and trust between non-state actors as well. How do we plug these gaps? Given the many competing understandings of accountability that exist at the national level, it is of little surprise that there is a lack of consensus on the meaning of the term at the transnational level. Accountability above the level of the state is an emerging concept that scholars have to date devoted limited time to exploring (Held and Koenig-Archibugi, 2005). As a result, when confronted with the problem of global accountability gaps scholars and activists tend to draw on what they know from the national level. They make direct analogies with how power is kept in check within nation states and propose elected representatives. Commentators such as Woods and Narlikar (2001) and Grant and Keohane (2004), however, warn against this suggestion. They argue that elections in and of themselves will not increase accountability because rarely do the electorate use votes to hold politicians to account. In fact, most of the time votes are used to demonstrate party allegiance and support for future policies (Woods and Narlikar, 2001). In this regard, elections should not be expected to play the principle role of accountability at the global level, as they are not even the principle form of accountability at the national level (Woods and Narlikar, 2001: 17). In addition, as it was argued before, the ex ante accountability promoted through elections has, in isolation from other mechanisms, limited potential for holding power to account anyway. What we argue is that the efﬁcacy of an accountability mechanism is largely determined by the principles and standards that underpin it. The One World Trust, through its Global Accountability Project (GAP), argues that in order for an organisation to be accountable four core dimensions must feature in all its policies, procedures and practices at all levels and stages of decision making and implementation, in relationship with both its internal and external stakeholders. These are transparency, participation, evaluation and complaints and redress. The higher the quality and the more embedded these mechanisms are in all organisational policies, processes and practices, the more accountable an organisation is.

Accountability of regional organisations 223 Without common standards and widely agreed-upon measures ‘accountability can be exercised in a discretionary and discriminatory way, where different standards are set for different actors, depending on the political goals being pursued’ (Newell and Bellour 2002: 11). The ﬁrst step in bringing about greater accountability beyond the nation state therefore must be to ensure that all actors have a common understanding of what it means and agree on a set of principles and standards applicable across all sectors and at all levels.

Assessing accountability: The Global Accountability Project In 2003, the One World Trust published the report Power Without Accountability? which, for the first time ever, tried to measure the accountability of global organisations. The report compared the accountability of 18 global organisations from the IGO, TNC and INGO sectors across two of eight dimensions identiﬁed as crucial to fostering greater organisational accountability. These dimensions fell into two categories: those that ensured internal accountability and those that ensured external accountability. Within the former were: member control, appointment of senior staff, compliance mechanisms and evaluation processes. Within the latter: external stakeholder engagement, corporate social responsibility, complaints mechanisms and access to information. As a pilot exercise, the report only scored and ranked organisations according to member control and access to online information. It did this by developing indicators that allowed the performance of organisations according to these dimensions to be quantiﬁed and measured. There were two key ﬁndings: ﬁrst, only a few members exerted real control over the organisations examined. The IGOs in the report were particularly unaccountable, with institutionalised member dominance characterising organisations such as the World Bank and BIS. Similarly, TNCs suffered from a form of minority control as a result of the rise in the number of institutional investors. INGOs on the whole were able to skirt this problem through mechanisms that prevented minority control of the executive. Second, INGOs are less transparent than IGOs and TNCs, providing less online information about their activities, and all organisations across the three sectors provide limited information about their decision-making processes. Since the publication of the report, the thinking on what accountability means within the global context has progressed. Through a process of external consultation and internal learning based on the initial pilot report, the GAP team have reassessed the project and the original eight dimensions, and collapsed them into just four. Through the involvement of a wide range of academics, policy-makers and practitioners, research is now being undertaken to identify principles and guidelines for each of these dimensions. Quantitative and qualitative indicators will also be developed for each of the dimensions, in order to offer users a more practical understanding of how accountability can be operationalised. This will also form the basis for a Global Accountability Index that will be generated by collecting information on the accountability of approximately 30 global organisations. It is planned that the

224 Monica Blagescu and Robert Lloyd Index be updated regularly in order to record progress. The Global Accountability Index will identify speciﬁc areas for improvement within the accountability of the assessed organisations; it will identify best and worst accountability practices; and will develop a global perspective on accountability trends and challenges. It is envisaged that the Global Accountability Framework will be used by stakeholder groups to assess the accountability of global organisations and by global organisations and organisations that operate at other levels to assess their own accountability. Below is a brief description of the four dimensions in the Global Accountability Framework. Transparency At its most basic level transparency refers to the degree of information provided by an organisation. A transparent organisation is one that provides a free ﬂow of information to all its stakeholders. The provision of information in and of itself, however, does not automatically lead to accountability. For transparency to be meaningful it also needs to relate more broadly to openness, speciﬁcally openness in decision-making procedures. A transparent organisation must have transparent rules of decision-making that ensure all stakeholders have sufﬁcient information on who is doing what and at what stage in the decision-making process, from policy formulation to implementation. Furthermore, for transparency to be effective, information also needs to be presented in an appropriate way, tailored to stakeholders’ needs. Only under these conditions will transparency allow for wider, more informed participation and thus greater accountability. Participation Participation encompasses both internal participation (member control) and external participation (external stakeholder engagement). The former refers to an organisation’s internal governing systems and the degree of control its members have over the decision-making process. The more equitable member representation is, the more accountable an organisation will be. The latter refers to the degree to which an organisation involves external stakeholders in its decision-making processes. There is a scale of stakeholder engagement; at the lower end stakeholders have limited decision-making power. Although stakeholders are perhaps consulted, decision making rests somewhere else. At the higher end of the scale, stakeholders have a high degree of decision-making power, they are co-decisionmakers. For an organisation to be accountable, external stakeholders must be engaged at the highest levels of this scale and the organisation should be clear and open about the processes through which it chooses to engage with one group of stakeholders over another. Evaluation This dimension refers to the existence and effectiveness of tools and procedures to evaluate organisational performance. There are two types of evaluation that help

Accountability of regional organisations 225 increase an organisation’s accountability: external evaluations, which aim to assess whether and to what extent organisational goals and objectives have been achieved; and internal evaluations, in which staff assess their own progress, either against objectives of externally funded programmes or against internal goals and missions. One important test of an organisation’s commitment to accountability is its preparedness to have its activities evaluated by a competent independent authority. External evaluations can improve an organisation’s accountability not just by assessing performance, but also by building capacity for self-evaluation and encouraging organisational learning. However, for evaluation processes to be effective an organisation must have the capacity to implement their recommendations. Tools of evaluation can be combined with processes of participation to develop external downward accountability mechanisms through systematic involvement of stakeholders in evaluating organisations (participatory evaluation). Evaluation mechanisms require policies and targets against which performance can be assessed. Communications systems are also required to provide stakeholders with information about performance. Complaints and redress This dimension refers to the mechanisms through which stakeholders – both internal and external, can register their complaints against an organisation and have them reviewed and acted upon. This needs to include both tools such as ombudsmen, and processes such as judicial review or public inquiries for complaints and redress against a decision to be effective. Complaint and redress mechanisms are critical to ensuring an organisation’s accountability. They provide the crucial link in the chain between access to information, participation and evaluation. After having been informed and consulted, stakeholders need to have the means to register their complaints and have their grievances properly addressed. If no such mechanisms exist, efforts to hold an organisation to account can only ever be partially effective. The Trust is currently completing the Global Accountability Framework. The full development of indicators will start in the second half of 2005. The next section will outline how these indicators are being developed, but, ﬁrst, will provide a brief discussion on why we should monitor/measure accountability in the ﬁrst place.

Why measure global accountability? The reasons why global accountability should be measured are five-fold. First, there is much rhetoric about organisations seeking to increase their accountability. Developing a set of indicators that allows this to be measured enables these claims to be monitored and evaluated. Second, measurement shifts the debate around global accountability beyond purely theoretical and largely ideological understandings and begins to ground it in objective empirical analyses. Third, indicators

226 Monica Blagescu and Robert Lloyd establish benchmarks, objectives and goals for organisations seeking to become more accountable. They provide organisations with a tool for identifying in what areas their accountability is lacking and where it is well developed. Fourth, indicators will help external stakeholders to lobby for greater accountability, providing them with a tool to gauge how accountable an organisation is and where its accountability is weak. Fifth, indicators will identify best and worst practices among organisations and sectors and thus hopefully stimulate inter-organisational and inter-sectoral learning.

How to construct accountability indicators? The diverse mandates and operating styles of both regional and global institutions present unique challenges to any attempts to develop common indicators that assess their accountability. Yet the ﬁrst step in developing accountability indicators is to agree on a set of principles applicable across the board. Since a one-size-ﬁtsall approach ignores the complexities of the issues at stake, we do not intend to develop a toolbox outlining specific procedures for each of the accountability dimensions; organisations will use different policies and procedures to interact with their stakeholders according to the level at which they work and the type of issue they are trying to address. The indicators developed to assess the degree to which these principles are implemented would not make any claim to capture the diversity and complexities of each individual organisation. They would only provide a starting place on which more elaborate indicators could be based. At the regional level in particular, more elaborate sets of indicators that measure in depth the accountability of speciﬁc types of organisations could be developed, both for each sector – IGOs, TNCs, INGOs – and for separate thematic groups within the sectors, i.e. in the case of IGOs, trade, security, environment, etc. In order to measure accountability, indicators need to be developed that capture the different facets of an organisation’s accountability mechanisms. The Trust is interested in, and has identiﬁed, three. First, the extent to which an accountability mechanism exists. Existence indicators provide an objective assessment of the presence or absence of an accountability mechanism within an organisation. Second, the embeddedness of an accountability mechanism. Here, indicators tease out the difference between an organisation’s accountability in principle and in practice by measuring such things as the extent to which an organisation has developed the necessary infrastructure and capacity to support a given accountability mechanism. Third, the quality of an accountability mechanism. Quality indicators measure the actual implementation of an accountability mechanism, whether it has been successful in its purpose or not; they also try to tease out the power relations that shape the effectiveness of an accountability mechanism. Indicators reflecting the presence, embeddedness and quality of a particular accountability mechanism will be scored and weighted by both the GAP team and expert groups. Indicator scores will then be aggregated to give an organisation a score within a particular dimension. Existence indicators will be scored with a simple yes or no, embeddedness and quality indicators will require a more

Accountability of regional organisations 227 subjective assessment based on the opinions of different stakeholders. As with the previous report, the scores for each indicator will be weighted and totalled to give a ﬁnal score of 100 (zero is the worst possible score, 100 the best) for each of the four dimensions of the global accountability framework. All scores will then be compiled into a composite Global Accountability Index. Methodological issues to consider when constructing global accountability indicators In constructing indicators of global and regional accountability a number of methodological issues need to be considered. For example, how do we assess accountability mechanisms in a meaningful way? Assessing only the presence or absence on paper of accountability mechanisms has proven not to be adequate; there are many ways that a board level decision to implement an accountability mechanism can lead to no change in the accountability of the organisation. There is a whole series of management and governance processes which can help or hinder the implementation of accountability mechanisms throughout the organisation which we expect to analyse for the next report: are staff offered the right incentives to implement particular policies or do other organisational policies prevent this from happening? Are members of staff trained in how to implement the relevant policies and do they not have the necessary means to do so? Another issue to be considered is that of the correlation between the accountability dimensions. One way might be to develop proxy indicators which take us closer to understanding the relation between them. For example, it might be possible to examine the impact of an information disclosure policy on the types of groups and individuals accessing particular information and so draw some conclusions about the impact of transparency on participation. How is the data being collected? The methods used for data collection will impact upon the end results (UNDP, 2004). In a desk-based study, for example, the quality of analysis is only as good as the quality of available published information. The first GAP report suffered in this regard. Based primarily on published material and what organisations made available on their websites it had problems accessing up-to-date information, and was only able to measure the formal accountability mechanisms within an organisation through its indicators. The often more important informal processes could not be captured with such a data collection method. The picture of organisational accountability that the indicators provided was thus partial. More interactive methods of data collection, on the other hand, allow more room for in-depth data collection. These types of methods vary. Participatory workshops, for example, are ideal for free thinking; however, they are time consuming and the data they produce is often not comparable. Interviews and questionnaires are less participatory but at the same time produce comparable results that allow respondents to provide subjective opinions, just within the conﬁnes of the question (UNDP, 2004). Ideally, in collecting data for accountability indicators, a combination of all these methods would be used.

228 Monica Blagescu and Robert Lloyd What are the primary sources of data going to be? Sources of data for measuring accountability will vary according to the level of accountability that is being measured. For example, to measure the existence of an accountability mechanism within an organisation, sources such as policy statements, statements of intent and formal structures are appropriate. Trying to assess the embeddedness and quality of an organisation’s accountability mechanisms, on the other hand, is more difficult. Polls of expert groups such as key external and internal stakeholders or individuals who have extensive experience with the institutions are most appropriate as they provide informed opinions comparable across organisations. However, the reliability of expert polls does depend on the ability of the groups to provide expert objective assessment of the accountability dimensions (Kaufman et al., 2003). Thus, their credentials and experience need to be sound. Moreover, it is important to get an even balance of representation among the expert group, so as not to bias the results in any way through selection. Underlying every indicator there will be a normative assumption. For example, on the most general level GAP has made the normative assumption that accountability is a good thing. In this regard, there is a need to ensure that the normative theory upon which assumptions are made is correct (UNDP, 2004). Beyond the collection of data to measure the existence of accountability mechanisms, it becomes very difﬁcult to construct an objective-based indicator of accountability. This is because objective data on such issues as the quality and embeddedness of accountability is impossible to obtain. For example, although most organisations now have some sort of mechanism in place for internal complaints and redress, the quality and effectiveness of such a mechanism will vary greatly. Moreover, subjective perceptions are often the only way to uncover the informal power relations that distort the operations of formal accountability mechanisms.

CONCLUSION This chapter has provided a detailed overview of the discussions that surround accountability at the global level and has attempted to relate them to regional organisations. It has challenged traditional approaches to accountability and stressed that an organisation is not only accountable to those that have delegated authority to it, but also to those that are signiﬁcantly affected by its decisions. It was acknowledged that although this makes accountability complex, reconciling the needs of different stakeholders lies at the crux of an accountable organisation. This chapter has also noted that, in thinking about regional accountability, the focus should not lie with inter-state institutions alone; non-state actors such as TNCs and INGOs have considerable impact on regional and global organisations as well. In this regard, it was argued that the lack of accountability among regional institutions cannot be amended solely through state-based accountability systems. New approaches to accountability need to be developed. The One World Trust has been leading the way in this regard. It has been argued that in order for an organisation

Accountability of regional organisations 229 to be accountable, transparency, participation, evaluation and complaints and redress need to feature in all its policies, procedures and practices, at all levels and stages of decision making and implementation, in its relationships with both its internal and external stakeholders. The higher the quality and level of embeddedness of these characteristics in all organisational policies, processes and practices, the more accountable the organisation is. Any indicators need to try to capture this multifaceted and multilevel understanding of accountability. It is important to note that, in developing a better understanding of accountability, the Trust has been developing principles that can apply across the three sectors of governance. These have been developed at the most general level. Their purpose has been to ensure a common understanding among actors as to what accountability means above the level of the nation state. However, when developing indicators, it is necessary to move from the general level to the specific. Indicators at regional level would need to reﬂect the nature of each sector. In terms of indicators, and the more ambitious endeavour of ranking organisations according to their accountability scores, in the case of GAP this serves a very speciﬁc purpose – it is a means to push for accountability reform of organisations that operate at the global level. At regional level, a separate, more extensive set of accountability indicators might provide insights into the degree of integration of that organisation’s stakeholders. The process through which the GAP team develops these principles is as important as its outcome. Accountability principles for each of the dimensions are being developed by engaging representatives from each of the three sectors and their respective stakeholder groups. Through this process, a meaningful crosssectoral learning is ensured, and agreement on common accountability principles is easier to reach. The link between greater regional accountability and regional integration is, as yet, an unexplored issue. The idea that stakeholder engagement around regional issues could serve to build greater cooperation and trust between non-state actors and in turn extend regional integration beyond greater inter-state cooperation deserves more attention.

GOOD PRACTICE IN ACCOUNTABILITY By highlighting differences in transparency and member control, the ﬁrst report aims to encourage all international organisations to raise their standards of accountability. The questions below describe some of the key ways organisations can do so within the two dimensions studied. As more decisions are taken at the global level, and as more actors join those already on the global stage, the type of analysis provided by this report will become increasingly necessary to enable people to assess competing claims for accountability and legitimacy.

230 Monica Blagescu and Robert Lloyd

Governance: member control – good practice • • • •

• •

Are all members fairly represented on the governing body? Do all members have the power to add items to the agenda of governing body meetings? Do all members have the power to nominate, elect and dismiss individuals on the executive? Are there mechanisms in place to ensure equitable representation of all members on the executive (where the executive body is composed of member delegates)? Are amendments to the governing articles subject to at least a two-thirds majority? Does a majority of members (75 per cent or more) hold a majority of the votes?

Access to online information – good practice • • • • • • • •

• •

Is a description of the objectives, targets and activities available? Are evaluations of main activities available? Can the public identify all key members of the organisation? Is there a public record of the number of votes each member holds? Is a meaningful description of key decision-making bodies available to the public? Are individuals on the executive body publicly identiﬁed? Are the agendas, draft papers and minutes of both governing and executive body meetings available to the public? Is there an information disclosure policy available which clearly states the types of documents the organisation does and does not disclose, stating the reasons for non-disclosure? Are annual reports publicly available and do they contain externally audited ﬁnancial information? Is the above information available in the languages of those with a stake in the organisations?

A full copy of the report is available at http://www.oneworldtrust.org/ documents/GAP2003.pdf. For more details on the One World Trust and the Global Accountability Project, see www.oneworldtrust.org

NOTES 1 2

The views represented here are those of the authors and do not necessarily reﬂect the views of their respective organisations. Intergovernmental organisations, corporations and non-governmental organisations

Accountability of regional organisations 231 that operate at regional level constitute one of the levels of overlapping authority that characterise decision making in today’s system of multilevel governance. For the purpose of this discussion, regional becomes a subset of global unless stated otherwise.

REFERENCES Barlow, M. and Clarke, T. (2003) Making the Links: A Peoples’ Guide to the WTO and the FTAA, Canada: Council of the Canadians and the Polaris Institute. Online. Available HTTP: . Benz, A. (2001) ‘Restoring Accountability in Multilevel Governance’, prepared for the ECPR joint session workshop, Workshop 5 ‘Governance and Democratic Legitimacy’. Online. Available HTTP: . Freeman, R. (1984) Strategic Management: A Stakeholder Approach, Boston: Pitman. Global Accountability Project (GAP) (2003) Power Without Accountability?, London: One World Trust. Grant, R.W. and Keohane, R.O. (2004) Accountability and Abuses of Power in World Politics. Online. Available HTTP: . Held, D. (2005) ‘Democratic Accountability and Political Effectiveness from a Cosmopolitan Perspective’, in D. Held and M. Koenig-Archibugi, Global Governance and Public Accountability, Oxford: Blackwell Publishing, 240–67. Held, D. and McGrew, A. (2002) Globalisation/Anti-Globalisation, Cambridge: Polity Press. Held, D. and Koenig-Archibugi, M. (2005) ‘Introduction’, in D. Held and M. KoenigArchibugi, Global Governance and Public Accountability, Oxford: Blackwell publishing. Kahler, M. (2005) ‘Deﬁning Accountability Up: The Global Economic Multilaterals’, in D. Held and M. Koenig-Archibugi, Global Governance and Public Accountability, Oxford: Blackwell Publishing, 8–34. Kaufmann, D., Kraay, A. and Mastruzz, M. (2003) Governance Matters III: Governance Indicators for 1996–2002, World Bank Policy Research Working Paper 3106. Online. Available HTTP: . Keohane, R.O. (2003) ‘Global Governance and Democratic Accountability’, in D. Held and M. Keonig-Archibugi (eds) Taming Globalisation. Frontiers of Governance, Cambridge: Polity Press, 130–59. Newell, P. and Bellour, S. (2002) Mapping Accountability: Origins, Contexts and Implications for Development, Brighton: IDS Working Paper. UNDP (2004) Governance Indicators: A User’s Guide, Oslo: UNDP. Young, I.M. (2000) Inclusion and Democracy, New York: Oxford University Press. Woods, N. and Narlikar, A. (2001) Governance and the Limits of Accountability: The WTO, the IMF, and the World Bank, UNESCO.

12 Measuring the impact of regional organisations on peace building Rodrigo Tavares and Michael Schulz

INTRODUCTION This chapter concentrates on the phenomenon of regionalism and its contribution to intra- and extra-regional peace building1 by suggesting some (quantitative and qualitative) measures of regional integration, regional peace and regional intervention. Regionalism2 is not a recent phenomenon. History books are ﬁlled with descriptive accounts of cross-border military alliances and economic arrangements in several areas of the world. Over the past 20 years, nonetheless, regionalism has gone through a process of transformation, in how it is both empirically constructed and theoretically explained. Unlike the narrow top-down, state-centric, Cold War inspired and security-oriented regionalism of the 1950–80 period, this new trend, the so-called ‘new-regionalism’,3 is substantially broader in three different aspects: agents, sectors and direction, as will be clariﬁed later (Schulz et al., 2001). Coupled with the recent mushrooming of regional projects that cover virtually the whole world (Farrell et al., 2005), another issue maintains its position at the forefront of the international relations agenda: the phenomenon of war. Quantitative surveys reveal that the global trend in major armed conflicts has continued to decrease in the post-Cold War period; in 2004 the general magnitude of global warfare has fallen to its lowest level since the late 1950s (Marshall and Gurr, 2005). Despite this we are still widely experiencing the dreadful consequences of violent conﬂicts: according to the data provided by Conﬂict Barometer/ Heidelberg Institute on International Conﬂict Research, in 2004, there were 230 political conﬂicts. Of these, 3 were wars and 33 were serious crises: a total of 36 conflicts carried out with a massive amount of violence; and 51 conflicts were levelled as crises, meaning that violence was used only occasionally.4 The literature on the nexus regionalism and peace, can be divided into two different categories. Within the old regionalism, classical studies include Deutsch’s considerations on sentimental relations and security communities (1957, see also Merrit and Russett, 1981), Mitrany’s functionalist theory based on the integrative role played by technical elites on low-politics (1966, 1975), Russett’s study on regional integration and conﬂict (1967), and Nye’s survey of micro- and macro-regional organisations (1971). These studies focused on the likely peace opportunities provided by speciﬁc regional actors such as security communities

Impact of ROs on peace building 233 (Deutsch), functional institutions (in the case of Mitrany), regional geographical groupings (Russett), or micro-regional economic organisations and macro-regional political organisations (Nye). The major shortcoming of this literature is that although in a way ‘revolutionary’ the four theories were nevertheless produced within an academic atmosphere dominated by political realism and designed to interpret a world in which sovereignty, anarchy and security were formative principles. Within the ‘new regionalism’, the attempts to analyse the link regionalism/ peace are of six types. We ﬁnd fresh interpretations of the security communities question (Adler and Barnett, 1998; Adler et al., 2006); an elaboration of the nexus regionalisation/peace along levels of ‘regionness’ (Tavares 2004a); a normative conceptualisation of regions as ‘zones of peace’ (Kacowicz, 1998: 9; Singer and Wildavsky, 1993: 14); the application of the democratic peace theory to particular regions (see, for example, Kivimäki, 2001, for Southeast Asia; Geeraerts and Stouthuysen, 1999, for Europe); the study of regions as regional security complexes (Buzan and Wæver, 2003) or regional security orders (Lake and Morgan, 1997); and ﬁnally, just descriptive studies, with a focus on security, of particular areas or regions of the world like Europe, South African Development Community (SADC) and the Economic Community of Western African States (ECOWAS) (Laakso, 2002); West Africa (Francis, 2001); Asia (Hettne, 2002) or Southern Africa (Maclean, 1999). Although most of these works are compelling contributions to the debate, none of them provides new insights on how to measure the impact of regional integration and regional organisations’ intervention in peace building. In the first part of this chapter, we elaborate on the contemporary context of macro-regional politics and clarify important concepts in regional theory. We will contend that the world is becoming regionalised and that regions are beginning to perform an increasingly substantial role in global social relations. This will enable us to understand more effectively the current necessity of measuring the role of regional organisations in peace building. We focus on the driving forces that encouraged the United Nations in dedicating extra attention to regional organisations as peace agents. Finally, the last part of this chapter offers some considerations about what to measure and how to operationalise the impacts of regional organisations on the variable peace.

REGIONALISATION AS A PEACE TOOL To understand the recent phenomenon of regionalism two notions ought to be deﬁned: ‘region’, and ‘old’/’new’ regionalism. The difﬁculty in ﬁnding a satisfying deﬁnition of ‘region’ that meets the demanding expectations of the different epistemologies involved in regional studies is a symptom of the high degree of comprehensiveness that the ﬁeld has attained. As Nye reminds us ‘[M]any hours were wasted at the 1945 UN Conference in San Francisco trying to define it precisely’ (1968: 6).5 Etymologically speaking, ‘region’ derives from the Latin word regio, which refers to an administrative area or broad geographical area

234 Rodrigo Tavares and Michael Schulz distinguished by similar features. Yet as we dig deeper we can also ﬁnd that the Latin word that gave birth to regio was regere meaning ‘to direct, to rule’. History tells us therefore that ‘region’ not only has a geographical but also a political connotation. Henceforth in this chapter, ‘region’ (or macro-region) will be regarded as a cognitive construction that spills over state borders, based on territoriality, with a certain degree of singularity, socially molded by a body of different actors, and motivated by different (and sometimes contradictory) principles. By deﬁning region in this way, we are moving away from the classical perspectives of regionalism that marked the study of the subject up until the 1980s, such as functionalism and neofunctionalism, neorealism, liberal institutionalism or economic integration theory. This new conceptualisation of regionalism is broader than this so-called old regionalism (or the narrow agenda of regional studies) in three related aspects. •

•

•

Agents: The actors of regionalisation can be found in the state–civil society– market triangle and vary from elite to grass-root level, from the individual to the community, and from formal to informal elements. This category encompasses an endless number of agents such as: non-governmental organisations (NGOs), diasporas, trade unions, media, national companies, multinational corporations, lobby groups, governmental powers, networks, research groups, international organisations and so on. Vectors and motivation: It is not only the maximisation of security that instigates actors to cooperate. Integration can also be sparked off in other ﬁelds, for instance at the social, environmental, political or economic levels. The pattern of behaviour is not dependent on conditions independent of actors, but on the condition of action itself. And action can be based on self-interests or unselﬁshness (or both). Direction: Unlike what rational theories of international relations postulate, regions are not given entities that obey permanent historical structures. Regions are constructed spontaneously by human action and social practices (Grugel and Hout, 1999: 9) in the wake of a permanent redeﬁnition of processes and interests. Society and people make each other during an ongoing, two-way process.

The transition from the old to the new regionalism is fuelled by a change in the global context (Palmer, 1991; Hettne et al., 1999–2001). While the old regionalism, with its classical theories of regional peace dwell, was dominated by the bipolar Cold War structure, the new one is connected to a multipolar order where the Westphalian system of sovereign states is eroding and economic, social and political interdependence between international actors is increasing. But despite the recent enthusiasm in the political, cultural and mainly economic impact of regionalism, the literature on the ‘new regionalism’, upon which this chapter is built, has not produced mechanisms to understand the potential correlation that exists between the variables regional integration, peace and intervention. In an explorative way, this is the principal objective of this chapter. We will kick-off by presenting a conceptual definition of these variables so that the

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Impact of ROs on peace building 235 operationalisation of the correlation between them becomes more discernible. By highlighting the potential relationship between these three variables it will be possible to address questions such as ‘What is the correlation between the level of regional integration and the level of intra-regional peace?’, ‘What is the real contribution for peace of an intervention carried out by a regional governmental organisation?’, ‘What is the correlation between the level of regional integration and the type, time frame, impact or locale of an intervention?’, ‘Does an intervention (carried out, for example, by a regional organisation) have any impact on the level of regional integration?’, ‘Are more peaceful (or violent) regions more (or less) prone to intervene? What type, time frame, impact of intervention?’ or ‘Is regional peace contributing to increased regional integration?’. Despite the almost endless possibilities that are generated by the conjugation of these three variables, this chapter will put its main emphasis on the intervention of regional organisations in the construction of peace.

INTERVENTION: THEORETICAL DEFINITIONS The Thirty Years War (1618–48) marked the transition from a feudal and theological framework to an international system predicated on national sovereignty, non-recognition of any sort of superior authority, demarcation of borders and, principally, on the postulate of non-interference in the internal affairs of individual states. To a large extent, this is the international configuration that dominated most of the twentieth century. It was believed that states have to follow these rules in order to enable an orderly conduct of international relations. Legally, the norm of non-interference in each other’s national affairs was codiﬁed in the Charter of the United Nations. Article 2(4), for example, prohibits member states from exercising the threat or use of force against the territorial integrity or political independence of another state. Although this is still a formative principle in international relations, it is found that this model cannot make sufﬁcient allowance for the growing interdependence in the international system. In an increasingly interdependent world a great variety of forms of interpenetration occurs, as a result of which the notion of intervention, stemming from the conceptual universe of the early nation-state system, has become too narrow to encompass the phenomena of interference which are a structural characteristic of the contemporary international system (Leurdijk, 1986). Swinging between these two paradigms it is not surprising that in the literature concerned with ‘intervention’ the attribute that more easily stands out is the lack of agreement on the deﬁnition. Rosenau who examined a series of deﬁnitions in the classical chapter ‘Intervention as a Scientiﬁc Concept’, came to the conclusion that ‘uniformity in usage does not characterise the concept of intervention’ (1971: 283). He also pointed out that the literature is ‘pervaded with discussions of military interventions, propaganda interventions, economic interventions, diplomatic interventions, and ideological interventions, not to mention customs interventions and other highly speciﬁc actions through which one state experiences the impact of another’ (1971: 344–45).

236 Rodrigo Tavares and Michael Schulz For politicians and legal experts the central challenge now is to defend, on the one hand, the norm of non-intervention as staunchly as possible, because it is one of the basic rules of the international system and, on the other hand, to ﬁnd ground on which it is possible to allow for the practice of intervention. Nonetheless, a minimum deﬁnition of ‘intervention’, to which the majority of authors relate, could be interference in the internal affairs of other states by one state, a regional organisation or by the United Nations. Article 2(7) of the UN Charter provides that the United Nations shall not intervene in matters which are essentially within the domestic jurisdiction of any state, except where enforcement measures are permitted under the charter. These measures are only justified when international peace and stability are at risk or, more recently, in the case of massive violation of human rights, genocide, repression of populations (humanitarian intervention). In this task, the intervention of the United Nations can come through sanctions, the use of peacekeeping operations or, a new development, through peace enforcement. But if this is the approach the United Nations adopts with interventions, there is still no universal acceptance of how to understand and operationalise the concept. Björn Hettne, for instance, posits that four types of interventions could be possible: the unilateral, the bilateral, the plurilateral and the multilateral (2001: 24). The unilateral intervention type would come when one state takes action visà-vis a particular conﬂict zone. One example is the 1983 US’s unilateral ‘urgent fury’ operation in Grenada. Bilateral interventions are a form where two states agree, more or less voluntary, that the other state will intervene. The Soviet Union’s intervention in Afghanistan, in 1979, could be seen as such an example. The plurilateral form can be a regional formation, a group of states, or an organisation such as NATO, that decide to intervene. The multilateral form is when the UN is involved in the intervention. More conceptually, Rosenau famously approaches intervention as ‘the behaviour of one international actor towards another . . . whenever the form of the behaviour constitutes a sharp break with then-existing forms and whenever it is directed at changing or preserving the structure of political authority in the target society’ (1971: 292). Pointing in the same direction, Young deﬁnes interventions as ‘organised and systematic activities across recognised boundaries aimed at affecting the political authority structures of the target’ (1968: 178). Building on these definitions, we define intervention as a particular action (political, economic, military) with the purpose of affecting the management of a civil or international conflict, be it physically or structurally violent, carried out by a third party, be it an individual person, an NGO, a state, a regional organisation or the United Nations.

NEXUS REGIONALISM/PEACE: THE CONTEXT OF REGIONAL ORGANISATIONS . . . Despite the theoretical frameworks on the link regionalisation/peace outlined at the outset, the current context of international relations is demanding more thorough scientiﬁc assessments. After the Second World War, several organisations such as

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Impact of ROs on peace building 237 the Organisation of American States (OAS), the Arab League or the Organisation of African Unity (OAU) were oriented to a conﬂict-control role. However, their endeavours had limited importance (Nye, 1971: 172) since they were founded on the principles of sovereignty and non-interference in regions dominated by intrastate conflicts. Notwithstanding, over the past decade some changes in the global scene have spurred renewed attention to the potential role played by regional organisations in handling violent conﬂict. First, although the sovereignty axiom is still prominent, regional actors seduced by the potentially positive effects of economic integration and suffering the effects of globalisation are showing signs of being less orthodox and more malleable about foreign interventions. Second, in the context of the Cold War, regional conﬂicts were subordinated to the particular interest of the superpowers. Within their alliances, the Soviet Union and the United States were effective in suppressing or triggering regional turbulence in order to maximise their strategic interests. After the Cold War, a renovated interest in the global level and in the possibility of a new approach in terms of international peace building was generated. However, the weak ﬁnancial conditions of the United Nations (UN), the poor record in its peacekeeping missions of the 1990s (Somalia, Bosnia, Haiti) and political pressure from Western powers for reform is obscuring the feasibility of the UN and the global level as a self-sufficient conflict management tool. As Haas already pointed out during the Cold War: ‘regional security arrangements grow in direct proportion to disappointment with the UN collective security system’ (1969: 93). These drawbacks have persuaded the UN to reconsider the role of regional organisations in managing conﬂicts, particularly in Africa. As early as in 1992, Secretary-General BoutrosGhali in his ‘Agenda for Peace’ proposed ‘regional action as a matter of decentralisations, delegations and cooperation with United Nations efforts as means of easing the burden on the Council’ (paragraph 64) (italics added). In the same year, Russia put forward a proposal aimed at improving cooperation between the UN and regional organisations (UN Doc. A/AC/1992/CRP.1, Add.5). The proposal suggested that regional organisations should set up peace-making forces to be at the disposal of the Security Council. In 1994, the General Assembly adopted a declaration highlighting, along the same lines, how the contribution of regional organisations, under the supervision of the UN, shall be maximised (A/RES/49/57, 9/12/1994). On the same track, a series of meetings has been held since the mid1990s designed to develop a strategic partnership between the global body and regional organisations. This has taken two forms. First, as a series of six high-level meetings of the UN Secretary-General and UN specialised agencies with regional organisations (held in August 1994, February 1996, July 1998, February 2001, July 2003 and July 2005. Second, as two meetings6 between the Security Council (SC) and regional organisations (held in April 2003 and July 2004).7

238 Rodrigo Tavares and Michael Schulz . . . AND THE POLITICAL TRADE-OFF In historical terms, the duel between regional and global approaches to international peace building has continued throughout the twentieth century and is still being vehemently discussed. Globalists or advocates of the global level in conﬂict management would stress, like Woodrow Wilson did, the role of global norms or global collective security as the most appropriate strategy. Regionalists, like Churchill, Monnet or Franklin Roosevelt, would recognise that macro-regions have distinctive attributes enabling them to handle the conﬂict endogenously generated. Although these approaches are not by any means mutually exclusive, some confusion still exists over the best way to go. As Lepgold points out ‘there is no single formula for allocating security responsibilities either to global or regional forums. Conflict-management incentives vary according to the issues at stake, the actors involved, and the regional context in which a problem originates’ (2003: 10). According to chapter VIII8 of the United Nations Charter, regional organisations can be, and have been, empowered to handle intra-regional and extra-regional conﬂicts. As debated in the second Security Council meeting (July 2004), regional organisations offer comparative advantages.9 First, as the members of an organisation share the same cultural background, they are likely to be more in tune with a conﬂict at hand. Second, personal relationships with the leaders have developed in the past, which leads to greater understanding of the situation and may result in fruitful dialogue based on personal trust. Third, as time is of the essence in a crisis situation, regional organisations could offer a more timely response, compared to bureaucratic global organisations such as the UN or foreign states. Fourth, as the members of a regional organisation are the ones who would suffer more directly the impacts of the conﬂict, they have a legitimate vital interest at stake in preserving regional stability. These positive scenarios possibly inspired David Lake and Patrick Morgan to observe, ‘efforts to cope with violent conflicts, as well as to achieve order and security, will primarily involve arrangements and actions devised and implemented at the regional level’ (1997: 5). Despite all these apparent qualities, some disadvantages are nonetheless striking. First and foremost, some regional organisations, like the African Union (AU), SADC or ECOWAS, suffer from financial incapacity. Second, although some organisations have the will, they nevertheless lack military and organic structure to undertake peace operations. As General Cheick Oumar Diarra, the deputy executive secretary of ECOWAS, pointed out at a conference in Addis Ababa, Ethiopia, in November 2003: I must confess that one of the weaknesses of ECOWAS has been lack of capacity in programme design and management, . . . We had problems with the European Union and some other donors, primarily because we lacked the requisite capacity to utilise and manage the resources put at our disposal even though we were determined to succeed. (see http: //www.irinnews.org)

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Impact of ROs on peace building 239 Third, regional organisations may be perceived to lack impartiality.10 In previous interventions, like Ethiopia in Somalia or Nigeria in Liberia, the impartiality or neutrality of these countries, working in a regional framework, was questioned (Ghebremeskel, 2002). Finally, the decision-making process in most regional organisations is based on consensus which, ultimately, slows down the process and hinders quick responses to emerging problems. The trade-off between comparative advantages and disadvantages is not totally clear and therefore more work has to be done on the subject. As Koﬁ Annan pointed out in 2004, so while our cooperation is being enhanced, we have to consider more thoroughly the comparative strengths of different organisations, be they global, regional or sub-regional, and move towards the creation of strategic partnerships that meet today’s and tomorrow’s challenges [. . .] United Nations cooperation with regional organisations had been broad, but often ad hoc and more institutionalised channels of cooperation would help ensure more efﬁciency and effectiveness and perhaps even economies of scale.11 The regional strategy in international conflict management was inaugurated by the Arab League intervention in Palestine in 1948.12 Over the past few years more regional organisations have assumed responsibility in conflict management. ECOWAS has been responsible for peacekeeping missions in Côte d’Ivoire, Sierra Leone, Guinea Bissau and Liberia and decided, in 2003, to set up a rapid response military force and a US$5 million Peace Fund. Also illustrative is the African Union’s endeavour to address violent conﬂicts in Africa. In 2004, it set up a Peace and Security Council and in mid-2003 it deployed its ﬁrst peacekeeping mission (AMIB – African Mission in Burundi), which includes overseeing the implementation of cease-ﬁre agreements and the supporting disarmament and demobilisation in Burundi. In 2004, the African Union has also intervened in the Darfur/Sudan crisis through mediation and the deployment of a 5000-troop peacekeeping mission (sanctioned by the UN Security Council). The European Union has also been particularly committed to crisis management, as can be illustrated by the deployment of civilian and military operations in the Former Yugoslavian Republic of Macedonia, Georgia, Democratic Republic of Congo, Bosnia and Herzegovina, Iraq, Sudan, Aceh/Indonesia, and Palestine. In Haiti, the cooperation between the United Nations mission (UN Stabilisation Mission in Haiti – MINUSTAH) and the Organisation of American States (OAS) and the Caribbean Community (CARICOM) was also of relevance in mitigating the political unrest that swept the country in early 2004.

WHAT TO MEASURE? The endeavour to measure regional conﬂict management and regional intervention is not completely new. From Joseph S. Nye’s classical work (1971) we will make use of his measures (including slight modiﬁcations) that will serve as guides when

240 Rodrigo Tavares and Michael Schulz measuring regional integration and conflict. He attempted to measure whether regional organisations created ‘islands of peace’ in world politics (1987: 3). His findings gave somewhat split conclusions from which no clear-cut peace arguments were traced. Nye’s study focused on indicators for integration and from that perspective tried to draw comparisons with indicators of conflict. Although we have a different approach to the meaning of conﬂict (see below) we will make use of his methodological contribution and relate integration and conﬂict to our third dimension, namely intervention. Nye’s diachronical approach did enable the observer to identify shifts and changes over time, also enabling us to collect data for forthcoming longitudinal studies. Furthermore, the previous studies related to a Cold War context where regionalisation was very much formed by this bi-polar structure. States did not want to merge into federations but had a fundamental need for security arrangements to strengthen their positions vis-à-vis the superpowers. The regional organisations’ agreements and statutes did not mention the establishment of a federative state. Nye also emphasised that ‘[i]n the absence of a strong political will for regional amalgamation, yet in the presence of the incentives for regionalism described above a frequent result is the formation of organisations with integration as an explicit goal’ (1987: 23). The regional organisation, where the goal was to establish some integration in different sectors in the region, functioned as a ‘security umbrella’ for the loose state structure. However, in the current globalising world the outcome can go far beyond the initial agreements that were signed in the Cold War context. Hence, Nye’s questions remain: ‘What happens to these organisations once they are founded? Do they increase interdependence among their member states?’ (1987: 23). We also add the questions: ‘Is interdependence fostering peace?’ and ‘Can regional interventions contribute to regional peace building?’.

Regional integration Regional integration in this chapter is defined as the establishment of linkages between different, and geographically near, polities in the political, economic, social and military areas. In order to operationalise the concepts, following up on Nye’s study, we break the concept of integration into four components/ sectors, developing concrete measurements at each sector. Since, integration can have such varied meanings we want to be able to measure which component matters most. •

Political integration: emergence of a regional political system: institutional integration and common policy formation. •

Indicators: distribution of information, administrative regulations, institutional resources, scope, extent and salience of common policies, regional conﬂict management mechanisms. The four sub-dimensions that Nye (Nye, 1987) measured are institutional integration (P1), policy integration (P2), attitudinal integration (P3) and security community (P4).

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Impact of ROs on peace building 241 We will, within the political integration indicator, only focus on the three former sub-dimensions. The security community measurement is placed under both the social and the military indicators below. •

Economic integration: emergence of a regional transnational economy: regional interdependence and regional welfare. •

•

Social integration: emergence of a pluralistic security community. •

•

Indicators: Price convergence, convergence of interest rates, trade integration of goods and services, business investment, import and export of goods, balance of trade, share of intra-regional trade, share in world markets (of a certain region), common external tariffs, free flow of factors, harmonisation of policies. As has been done in previous studies, we consider full economic regional integration to be in place when a regional transnational economy is established. Hence, we need to measure both the formal and behavioural aspects of economic integration (cf. Balassa, 1962; Nye, 1987). Both the trade (EI1) and the service (EI2) aspects are included.

Indicators: regional internet contacts and phone calls, tourism, and academic mobility. These indicators relate to the extent contact/networks are established between states. Nye (1987) uses the following submeasures of social integration: mass social integration (SIm) and elite social integration (SIe). Both these measures relate to transactions and meetings but with less emphasis on actual attitudes. In order to come to terms with attitudes related to identification we add an identification measure (SIi). This measure will both identify the acceptance and willingness to meet with the ‘other’, but will also measure the identiﬁcation with a regional ‘imagined community’ (Anderson, 1991).

Military integration: the emergence of a regional military capability: integration of policies, armies and intelligence services. •

Indicators: common military exercises (training and military operations), exchanges of intelligence information and experience, military agreements and protocols. Military integration is constituted of two sub-dimensions. The first relates to coordination of security arrangements (MIc), the second to the extent to which a common security system is constructed (MIs).

Peace The second variable to operationalise is peace. Peace is theoretically, and commonly, deﬁned as a situation in which there is absence of both open violence – negative peace – and structural violence – positive peace (Galtung, 1996). The difficulties in measuring peace are threefold. First, the concept of peace is not

242 Rodrigo Tavares and Michael Schulz Table 12.1 Regional integration dimensions (Nye, 1987: 49, with minor alterations) Type of integration

Sub-dimensions

Measurement

Political (PI)

Institutional (PI1) Bureaucratic

Budget and staff as percentage of budget and administrative staffs of all member countries

Jurisdictional

Supranationality of decisions; legal scope; expansion of jurisdiction

Policy (PI2)

Scope Salience Extent

Attitudinal (PI3)

Elite and mass polls probing intensity and urgency Bargaining behaviour, ﬂexibility in length of time and number of ﬁelds

Trade (EI1)

Regional exports as percentage of total exports

Services (EI2)

Expenditures on joint services as percentage of GNP

Mass (SIm)

Transactions (trade, mail, etc.)

Elite (SIe)

Intra-regional air passengers; students in neighbouring countries as percentage of total students, intensity of regional scientiﬁc networks etc.

Identiﬁcation (SIi)

Elite and mass polls of identiﬁcation with regional identity, sense of belonging, low degree of fear of ‘other’, etc. Case studies

Security arrangement (MIc)

Formal agreements and coordinated actions

Security system (MIs)

Supranationality of decisions; legal scope; expansion of jurisdiction

Economic (EI)

Social integration (SI)

Military (MI)

universally interpreted and is subject to cultural assumptions and contextual codes. Second, positive peace is a situation which may not ever be possible to achieve since human beings seemingly always try to improve their living situation. As Galtung claimed, ‘[w]hen we feel that the state of conﬂictlessness will never be achieved, it is because we feel it is contrary to basic human and social needs’

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Impact of ROs on peace building 243 (Galtung, 1978: 491). Hence, a situation in which a person, group of people or society feel completely fulfilled as human beings in terms of basic, human and material needs is not possible to ﬁnd or measure. Third, as pointed out by Western sociologists as such as Karl Marx, George Simmel, Dahrendorf, Nieburg, S.M. Lipset or Lewis A. Coser, conflict brings about change, regeneration, improvement and grouping. Conﬂict does not mean random disorder. It must allow for the formation of collectivities in pursuit of conﬂict and it assumes functional interconnectedness between institutional areas. Conﬂict, i.e. a real or perceived situation with at least two parties in an incompatibility situation over an issue, is omnipresent in all inter-human relations and can be handled in violent or non-violent ways. The measurement of peace has virtually never been successfully attained before and it is only feasible as long as we take the concept of peace as a non-static idea, i.e. as a variable whose quality can be scored. To overcome the above pitfalls our strategy would be again to break down the variable into three related sectors: (absence of) violence, structural peace and behavioural conﬂict resolution styles (Figure 12.1). We will mainly be concerned with conflict transformation, i.e. how issues are transformed from violent conflict handling to a more peaceful way. At the end it should be possible to place peace on a continuum where no society can reach ‘total positive peace’, but rather it is a goal for which humans may strive. The first component of peace to be measured is (the absence of) physical violence. The classical measurement for intensity of violence and warfare is to calculate the number of battle-related casualties. The literature is abundant in terms of codiﬁcations of violence and its aggregation in databases. In two of the most well-known databases, the Correlates of War Project in Michigan (Small and Singer, 1982; Singer and Small, 1994) as well as in the conﬂict project in Uppsala (Harbom and Wallensteen, 2005; Wallensteen and Sollenberg, 2000) an armed conﬂict is considered to be ‘minor’ when there are at least 25 battle-related deaths per year and fewer than 1000 battle-related deaths during the course of the conﬂict; ‘intermediate’ when there are at least 25 battle-related deaths per year and an accumulated total of at least 1000 deaths but fewer than 1000 in any given year; and ‘war’ when 1000 battle-related deaths happen during one particular year. This measure is the most commonly used (with smaller variations) in most of the larger existing conflict databases worldwide.13 Moreover, Haas’ collective security database14 provides the way to operationalise the regional conﬂict intensity and the extent of warfare in each region. He has coded the intensity by aggregating the battle-related casualties to the regional level. Drawing from these databases, conﬂicts are coded for each state, and thereafter aggregated on a regional level as well as on regional organisations, i.e. all measured casualties within a particular group of states. In addition, a categorisation of the type of conﬂict must be added, clarifying for instance whether it is an identity conﬂict, resource conﬂict, rebellion, etc. The second component in peace to be measured is its structural pillar. Already existing research and databases on relations between trade and peace between

244 Rodrigo Tavares and Michael Schulz states could be used in order to aggregate to the regional level (cf. Mansﬁeld and Pollins, 2003). Here we can show how low inter-trade levels between states have a positive impact on conﬂict, and vice versa. These databases have, however, not aggregated the inter-trade levels of states to the various regional organisations. This will be done in our study. As peace is more profound than absence of violence, we link structural peace to welfare, improvement, well-being and fulﬁlment. Its indicators relate, thereby, to education, health, safety, economic opportunities and political legitimacy. As codified by the United Nations Human Development Reports, the indicators are, inter alia, people victimised by crime, internal displaced people, estimated earned income, education index, school enrolment ratio, GDP per capita, rate of HIV/AIDS, poverty index, illiteracy rate, income inequality measures, infant mortality rate, life expectancy at birth, literacy rate, physicians per capita, sanitation, mortality rate, unemployment and population using water. An important tier of the structural component of peace is the way people relate to their political rulers. In this context we also introduce indicators of popular legitimacy, political participation and popular acceptance. Again, these measures can be aggregated, not only the state level but also the regional level. However, also data directly collected from the regional level, for example the level of popular support for the EU, can also be measured. The third component of peace to be measured is behavioural conﬂict resolution styles. We contend that attitudes go beyond the previous violence and structural components, as peace also deals with the mental interpretation of a certain reality by an individual. There is an important difference between the real-reality and the way it is read and inferred by a person. The indicator has to deal with fear, expectation, satisfaction, trust and social capital. Hence, the actual behavioural and attitudinal hostilities of the ‘other’ conﬂicting group(s) (i.e. other groups within the own society/state, neighbouring states and within the region) are far less than are perceived by the ‘we’ group(s). The importance of the extent of networks between conﬂicting groups will be a focus. We examine three kinds of networks. In general, it is hypothesised that the more we engage in networks outside our own group, the more likely that our own ideas and opinions, due to direct communication, will receive more understanding and acceptance. Also, the own group becomes exposed to direct contact with different opinions and ideas that usually are not heard of in the inner group surroundings. These direct contacts and links will in turn have a positive effect on conﬂict resolution as well as on (re)-building social trust. The relation to the support network that exists outside the own group will be examined. These relate to networks where one can receive material and/or social assistance outside the own group. The second network relates to the discussion network. Here we are interested in the extent to which we discuss political issues outside our own group. The third network, the quotidian network, relates to the frequency of meetings with people outside our own group. When we speak of segmented versus bridging social capital we refer to the amount of networks/relations we have with people who are different to ourselves. Segmented social capital refers to the closer network within our own group (ethnic, national, religious, etc.). Bridging social capital in this study is measured with the density

Impact of ROs on peace building 245 of the quotidian network. In other words, we will chart the networks within but also across state boundaries in a particular region. Some previous studies aimed at addressing conflict transformation through a continuum between cooperation and conflict. The International Political Interactions (IPI) project (www.swisspeace.org), for instance, uses the following coding scheme for the cooperation scale. • • • • • • • • •

First – mildly positive statements about other parties. No action promised. Second – strongly positive statements about other parties, their representatives, proposals or activities. Implied or literal promises of action. Third – minor cooperative actions. Fourth – agreements to attempt to settle protracted conﬂict or relaxing minor restrictions. Fifth – relaxing government sanctions or actions designed to mitigate protracted conﬂict. Sixth – reforms; relaxing major restrictions; truces. Seventh – substantial agreements. Eighth – conﬂict termination: the parties agree to terminate the internal war but do not create new institutions for managing the underlying conﬂict. Ninth – conﬂict settlement: the internal war is terminated and the underlying conﬂict settled as a consequence of the construction of institutions that will manage future conﬂict.

By the same token, the Goldstein variable (Early Warning Unit, FAST, Swiss Agency for Development and Cooperation) has a rating for proportions of conﬂict and cooperation events in time. Each event category is assigned a Goldstein value, ranging from –10 (extreme conﬂict) to 10 (extreme cooperation).15 Our measurement of peace differs from these studies because we are interested in measuring peace and not cooperation. Cooperation is a relational variable, an indicator of the level and quality of interpenetration between two units. Peace is an inherent and variable quality of a unit (a person, a group, a state, a region) that is not necessarily dependent on a relationship with another unit. By following the above-mentioned cooperation scales we could say, for example, that Swaziland and South Africa have a very high level of cooperation. Although this is an unquestionable fact, it could not be inferred that they also share high levels of ARMED CONFLICTS

STRUCTURAL

VIOLENCE

PEACE BEHAVIOURAL

Figure 12.1 Violence and peace are part of the dyadic relationship of a continuum

246 Rodrigo Tavares and Michael Schulz peace: Swaziland still looks at South Africa with fear of being dominated (behavioural attitude) and in terms of the Human Development Index (a component of structural peace) South Africa ranks 119th (0.666 in 2002) and Swaziland ranks 137th (0.519 in 2002) (United Nations 2004).

Intervention The third key concept to be operationalised is intervention. We have previously deﬁned intervention as a particular action (political, economic, military) with the purpose of affecting the management of a civil or international conflict, be it physically or structurally violent, carried out by a third party, be it an individual person, an NGO, a state, a regional organisation or the United Nations. The main objective in measuring intervention is to create mechanisms to quantify and assess the impact of an intervention in a targeted polity. In order to operationalise this correlation, peace (as we designated it above) is taken as the dependent variable and intervention as the independent variable. We argue that the best way to assess the effects of an intervention is not administratively (measuring whether the goals laid out by the ‘intervenor’ were accomplished at the end of the intervention), nor in terms of violence rates (whether the intervention helped to reduce the intensity of the violent conﬂict). The best way to evaluate the impact of an intervention is by measuring the levels of peace (violence, structural, behavioural) in the three phases of intervention: the pre-intervention phase, the intervention phase and the post-intervention phase. In the first phase the issue is emerging and a political process starts within a particular region in which mediation, internal negotiations, internal ratifications and voting may take place. In the second phase the actual intervention has begun and the primary outcomes can be identiﬁed. In the third phase a withdrawal of the intervention is imminent and the more long-term outcomes can be identiﬁed. The second objective would be to break down ‘intervention’ in order to make all its components correlate with the dependent variable peace. First, we need to address the various forms of regional interventions (see Table 12.2). • •

•

• •

Regional unilateral intervention is when one regional actor intervenes in a conﬂict without approval from the other regional members. Regional bilateral intervention is when one actor intervenes as a result of two regional actors’ agreement but without necessarily having the support/ acceptance of all regional members. Regional plurilateral intervention is a situation where more than two regional actors intervene in a regional conflict but without necessarily having the support/acceptance of all regional members. Regional multilateral intervention is a situation where all regional members have achieved consensus on the intervention. Combined multilateral/plurilateral intervention is a situation where external actor(s), such as the UN, USA or other regional organisation has reached

Development aid

Sanctions

Reconciliation

Post-armed conﬂict reconstruction

Armed conﬂict termination

Structural conﬂict (poverty, famine, water sharing, pollution, etc.)

Extra-regional intervention

No

Yes

UN legitimacy when required by international law

Notes a Preventive diplomacy, preventive deployment, preventive disarmament, humanitarian action, peace building. b Mediation. c Cease-ﬁre veriﬁcation, border patrol, peace building (demobilisation, electoral monitoring, human rights protection, transitional administration, civilian policing). d Use of coercive force to maintain or restore peace.

Combined multilateral plurilateral

Armed conﬂict

Peacekeepingc

Regional plurilateral

Regional multilateral Peace-enforcementd

Conﬂict intensiﬁcation/ violence inception Inter-state conﬂict

Peacemakingb

Regional bilateral

Locale of intervention

Internal conﬂict (ethnic, Intra-regional political, territorial) intervention

Conﬂict creation

Conﬂict preventiona

Regional unilateral

Type of conﬂict that prompted the intervention

Time frame of the intervention in relation to conﬂict

Type of intervention

Forms of intervention

Table 12.2 Various forms of regional interventions

248 Rodrigo Tavares and Michael Schulz consensus with at least one intra-regional actor but without necessarily having the support/acceptance of all regional members. Second, other components should be taken into consideration such as: type of intervention, time frame of the intervention in relation to conﬂict, type of conﬂict that prompted the intervention, locale of intervention and United Nations legitimacy when required by international law, as suggested by Table 12.2.

CONCLUSION The current global regionalisation processes are seen as potentials for creating a more peaceful world. At the same time, no real attempts have been made to measure the regional peace dividend. Previous attempts, in the so-called ‘old regionalism’ context were too complex and could not give any clear-cut measures of peace. In order to make a more simpliﬁed model we propose indicators that can cover how regions cope with open conflicts. These indicators relate to regional integration and regional peace, as well as to interventions. However, the regional peace dividend cannot merely be quantitatively measured but must also include some qualitative evaluations concerning the way that regional interventions are related to time, scope and goal. Also, it should be evaluated how regional formations see conﬂicts as ‘shared problems’ rather than security concerns of particular states. However, we have also developed a method of measuring structural peace in order to enable an evaluation of the depth of the positive peace in various world regions. The time is ripe to test these indicators, enabling us to say more about the peace dividend argument for global regionalisation.

NOTES 1 Peace building is interpreted here as an inclusive concept that includes conﬂict prevention, conﬂict management, conﬂict resolution and post-war reconstruction and (re)conciliation. 2 Henceforth in this chapter, ‘regionalism’ is used to describe the existence of crossborder political, economic and social linkages within a region, whereas ‘regionalisation’ reﬂects the process of the establishment of those linkages (see Tavares (2004b) for a literature review on these concepts). 3 The term ‘new regionalism’ was ﬁrst introduced by Norman Palmer (1991). 4 Available online at . Last visited 1 April 2005. 5 The term ‘regional’ is not deﬁned in Article 51 of the UN Charter. When the Charter was drafted, an Egyptian attempt to deﬁne it was vetoed. The amendment put forward by Egypt was ‘There shall be considered as regional arrangements of a permanent nature grouping in a given geographical area several countries which, by reason of their proximity, community of interests or cultural, linguistic, historical or spiritual afﬁnities make themselves jointly responsible for the peaceful settlement of any disputes which may arise between them’ (cited in Russett, 1967: 4). 6 In the ﬁrst meeting, carried out under the Mexican presidency of the Security Council,

Impact of ROs on peace building 249 ‘most speakers agreed that the Security Council had the primary responsibility for the maintenance of international peace, but regional organisations played an increasingly important role in dealing with situations that threatened collective security. Cooperation between the Council and regional organisations should be based on complementarity of action, keeping in mind the differences between the regions’ (SC 4739th Meeting). In the second meeting, following a Romanian initiative, the Security Council invited ‘all members of the United Nations, and other parts of the United Nations system with relevant experience and expertise to contribute to this process’ (S/PRST/2004/27). 7 Complementarily to these series of meetings, the Security Council and the Secretary General have emphasised the role of regional organisations in several other documents. For example: • The Secretary General Koﬁ Annan’s report on ‘The Causes of Conﬂict: The Promotion of Durable Peace and Sustainable Development in Africa’ (1998, paragraphs 41–45) notes that Africa’s capacity for regional peacekeeping shall be strengthened (paragraph 45). • The Security Council 4465th Meeting where it was stated that ‘The Council underscores the importance of partnership and enhanced coordination and cooperation between the UN, OAU and sub-regional organisations in Africa’ (January 2002). In Security Council 4665th Meeting (on the situation in Liberia), the Council shows its commitment to support the efforts of regional actors to address the threat that Liberia ‘poses to international peace and security in the region’. • The Security Council 4766th Meeting where conﬂicts in Africa and the UN mechanism to promote peace and security were discussed (May 2003). In the meeting, Ibrahim A. Gambari, Under-Secretary General and Special Adviser on Africa noted that conﬂicts in Africa could not be resolved without taking their regional and global dimensions fully into account. 8 According to the Charter, regional organisations and arrangements have considerable scope for action in terms of ‘peace settlement’ (Art. 52 – 2/3). Their capacity in ‘peaceenforcement’ is, however, totally dependent on the Security Council’s endorsement and supervision (Art. 53 – 1). 9 In the meeting, Brazil said regional organisations were in a better position than the UN to detect early symptoms of conﬂict and act promptly, thus preventing intrastate differences evolving into intolerance, prejudice, hatred, killings and massive crises. Human rights monitoring was another example of a task best carried out by organisations closer to the area of observation. The primary role of regional organisations was preventive and their main efforts should be targeted at root causes, which may be region-speciﬁc. In the same vein, the United Kingdom pointed out that regional organisations can bring a great deal to the partnership. They can provide an accepted framework within which to take forward stabilisation. They have greater knowledge of the situation on the ground. In many cases they are developing appropriate, regionallyaccepted norms and standards, and they match them with the right mechanisms for monitoring and enforcement (Security Council Press Release 8153). 10 This obstacle was under discussion at the third Security Council Meeting with regional organisations. The United States pointed out that, ‘on the one hand a troop-contributing country may share language, cultural elements, and a common understanding with the host country. However, it may also have its own agenda independent of the peacekeeping agenda. We need to watch carefully for that possibility, given the goal of long-term peace and stability’ (cited in Graham and Felicio, 2004). 11 Security Council Press Release 8153. Available online at . 12 The League issued a statement that declared that ‘The Government of the Arab States,

250 Rodrigo Tavares and Michael Schulz as members of the Arab League, a regional organization within the meaning of the provisions of Chapter VIII of the Charter of the United Nations, are responsible for maintaining peace and security in the area. These Governments view the events taking place in Palestine as a threat to peace and security in the area as a whole and [also] in each of them taken separately’. 13 At the website of the Uppsala database we can ﬁnd several international links to wellknown war-correlates projects. Online at <www.pcr.uu.se>. 14 . 15 Available online at <www.swisspeace.org/fast/default.htm>.

REFERENCES Adler, E. and Barnett, M. (1998) Security Communities, Cambridge: Cambridge University Press. Adler, E., Crawford, B., Bicchi, F. and Del Santo, R. (eds) (2006) The Convergence of Civilizations: Constructing a Mediterranean Region, Toronto: Toronto University Press. Anderson, B. (1991) Imagined Communities, London: Verso. Balassa, B.A. (1962) The Theory of Economic Integration, London: Allen and Unwin. Buzan, B. and Wæver, O. (2003) Regions and Power: The Structure of International Security, Cambridge and New York: Cambridge University Press. Deutsch, K. et al. (1957) Political Community and the North Atlantic Area, Princeton: Princeton University Press. Farrell, M., Hettne, B. and Van Langenhove, L. (eds) (2005) Global Politics of Regionalism: Theory and Practice, London and Ann Arbor: Pluto Press. Francis, D. (2001) The Politics of Economic Regionalism: Sierra Leone in ECOWAS, Aldershot: Ashgate. Galtung, J. (1978) Peace and Social Structure: Essays in Peace Research, 3, Oslo: PRIO. Galtung, J. (1996) Peace by Peaceful Means: Peace and Conflict, Development and Civilisation, London: Sage Publications. Geeraerts, G. and Stouthuysen, P. (eds) (1999) Democratic Peace: Myth or Reality for Europe, Brussels: VUB University Press. Ghebremeskel, A. (2002) ‘Regional Approach to Conflict Management Revisited: The Somali Experience’, The Online Journal of Peace and Conﬂict Resolution, 4(2). Graham, K. and Felicio, T. (2004) Regional Security and Global Governance: A Study of Interaction Between Regional Agencies and the UN Security Council with a Proposal for a Regional–Global Security Mechanism, Bruges: United Nations University (UNUCRIS). Grugel, J. and Hout, W. (eds) (1999) Regionalism Across the North–South Divide, London and NY: Routledge. Haas, E. (1969) Tangle of Hopes: American Commitments and World Order, Englewood Cliffs, New Jersey: Prentice-Hall. Harbom, L. and Wallensteen, P. (2005) ‘Armed Conﬂict and Its International Dimension’, Journal of Peace Research, 42 (5). Hettne, B. (2001) ‘Regionalism, Security, and Development: a comparative perspective’, in Hettne et al., 2001. Hettne, B. (2002) ‘Globalisation, Regionalisation and Security: The Asian Experience’, The European Journal of Development Research, 14 (1). Hettne, B., Inotai, A. and Sunkel, O. (1999–2001) The New Regionalism Series, 1–5, London: Macmillan.

Impact of ROs on peace building 251 Kacowicz, A.M. (1998) Zones of Peace in the Third World: South America and West Africa in Comparative Perspective, New York: State University of New York Press. Kivimäki, T. (2001) ‘The Long Peace of ASEAN’, Journal of Peace Research, 38(1). Laakso, L. (2002) European Integration for Conflict Prevention and Peace Building, Helsinki: Helsinki University. Lake, D.A. and Morgan, P. (eds) (1997) Regional Orders: Building Security in a New World, Pennsylvania: Pennsylvania State University. Lepgold, J. (2003) ‘Regionalism in the Post-Cold War Era: Incentives for Conflict Management’ in P. Diehl and J. Lepgold (eds) Regional Conﬂict Management, Lanham, Boulder, NY and Oxford: Rowman & Littleﬁeld Publishers. Leurdijk, J.H. (1986) Intervention in International Politics, Leeuwarden: Eisma B.V. Publishers. Maclean, S. (1999) ‘Peace Building and the New Regionalism in Southern Africa’, Third World Quarterly, 20(5). Mansfield, E.D. and Pollins, B. (2003) Economic Interdependence and International Conﬂict, Ann Arbor: University of Michigan Press. Marshall, M.G. and Gurr, T.R. (2003) Peace and Conflict 2003, Maryland: Centre for International Development and Conﬂict Management. Merrit, R.L. and Russett, B.M. (eds) (1981) From National Development to Global Community – Essays in Honour of Karl W. Deutsch, London: George Allen & Unwin. Mitrany, D. (1966) A Working Peace System, Chicago: Quardrangle Books. Mitrany, D. (1975) The Functional Theory of Politics, London: Martin Robertson & Company. Nye, J. (1968) International Regionalism: Readings, Boston: Little Brown. Nye, J. (1987; 1971) Peace in Parts: Integration and Conﬂict in Regional Organisation, Boston: Little, Brown and Company. Palmer, N.D. (1991) The New Regionalism in Asia and the Paciﬁc, Lexington: Lexington Books. Rosenau, J.N. (1971) The Scientiﬁc Study of Foreign Policy, New York: The Free Press. Russett, B. (1967) International Regions and International System: A Study in Political Ecology, Chicago: Rand McNally &Company. Schulz, M., Söderbaum, F. and Öjendal, J. (2001) Regionalisation in a Globalising World: A Comparative Perspective on Forms, Actors and Processes, London: Zed Books. Singer, J.D. and Small, M. (1994) Correlates of War Project; International and Civil War Data, 1816–1992, ICPSR 995, Ann Arbor: MI. Singer, M. and Wildavsky, A. (1993) The Real World Order: Zones of Peace, Zones of Turmoil, New Jersey: Chantam House Publishers. Small, M. and Singer J.D. (1982) Resort to Arms: International and Civil Wars, 1816–1980, Beverly Hills, CA: SAGE. Tavares, R. (2004a) ‘The Contribution of Macro-Regions to the Construction of Peace: A Framework for Analysis’, Journal of International Relations and Development, 7(1). Tavares, R. (2004b) ‘The State of the Art of Regionalism: The Past, Present and Future of a Discipline’, UNU-CRIS e-Working Paper, United Nations University/Comparative Regional Integration Studies’ (W-2004/10) [www.cris.unu.edu]. Wallensteen, P. and Sollenberg, M. (2000) ‘Armed Conﬂict 1989–99’, Journal of Peace Research, 37(5): 635–49. Young, O.R. (1968) ‘Intervention and International Systems’, Journal of International Affairs, Vol. XXII(2).

Index

Abbott, K.W., et al. 194; and Snidal, D. 194 Abramovitz, M. 146 accountability, assessment of 223–5; concept 215–17; constructing indicators for 226–8; embeddedness of indicators 226; existence indicators 226; global level 217–19; good practice in 229–30; and governance 186, 187, 191, 210; methodological issues 227–8; need to plug gaps in multilevel governance 221–2; quality of indicators 226; reasons for measuring global accountability 225–6; regional gaps 219–21; stakeholder view of 216–17; traditional understandings of 216; ways of plugging gaps 222–3 ACP countries 21 Adams, R. et al. 84, 94, 96 Adler, E. et al. 233; and Barnett, M. 195, 233 African Mission in Burundi (AMIB) 239 African Peer Review Mechanism (APRM) 188 African Unity (AU) 188, 238 AGIL scheme 54 Agreement on Sanitary and Phytosanitary Measures (WTO) 121 Agreement on Technical Barriers to Trade (WTO) 121 Ahmad, E. and Craig, J. 135, 141 Alesina, A. and Spoalore, E. 133, 134 AMIB see African Mission in Burundi Andean Community 201 Andean Employers’ Consultative Council 201 Andean Integration System 201 Andean Labour Consultative council 201 Anderson, G. and Tollison, R. 144

Anderson, J.E. and van Wincoop, E. 97 Anderson, K. and Norheim, H. 82, 83 Anderson, V. 11 Annan, K. 239 Anselin, L. 157, 160 ANZCERTA see Australia and New Zealand APR Heads of State Forum 188 ASEAN see Association for South-East Asian Nations ASEAN Fund 141 Association of Caribbean States 143 Association Councils 124 Association of South East Asian Nations (ASEAN) 75, 78, 80, 81, 101, 189, 212n24 AU see African Unity Australia and New Zealand (ANZCERTA) 123 Baier, L.S. and Bergstrand, J.H. 163, 178 Balassa, B.A. 13, 15, 17, 67, 83, 241 Baldwin, R. 109, 123 Barnett, M. and Crawford, 233 Barro, R. and Sala-i-Martin, X. 146, 147, 151, 155 Baumol, W. 146 Becker, W.E. and Kennedy, P.E. 178 Bekemans, L. et al. 17 Bensidoun, I. and Chevallier, A. 30 Benz, A. 219, 220 Bernard, A. and Durlauf, S. 153, 154 Best, E. 16, 19, 23, 24, 29, 184 Bhagwati, J. et al. 108, 110 bilateral trade intensities 95–6; advantages of ﬁxed effect model 99; gravity model 96–7, 99; RTAs as determinants of trade ﬂows 97–101; and sectoral specialisation 98

Index 253 Bird, G. and Rajan, R.S. 177 BIS 223 Bond, E.W. and Syropoulos, C. 128 Boutros-Ghali, B. 237 Bowen, H.P. 83 Boyle G.E. and McCarthy, T.G. 152 Brennan, G. and Buchanan, J. 131, 132 Britz, G. and Schmidt, M. 198 Broadway, R. 143 Brown, A.J. 67 Buchanan, J. and Goetz, C. 143 Buitelaar, 38 Bureau, D. and Champsaur, P. 144 Buzan, B. and Wæver, O. 233 CACM see Central American Common Market Campos, W. 204 Canada–US FTA 94 Caribbean Community (CARICOM) 204, 211n18, 239 CARICOM see Caribbean Community Casella, A. 134–5; and Frey, B. 135 CCJ see Central American Court of Justice CEC 195, 199 Central America 200–1, 203, 206–7, 208, 212n23 Central American Common Market (CACM) 200 Central American Consultative Committee 205 Central American Court of Justice (CCJ) 200 Central American Integration System (SICA) 200–1, 205 Central American Parliament (PARLACEN) 200 Central and Eastern European countries (CEECs) 88 CEPII index of relative trade intensity 89–94 Chatterji, M. 151, 159; and Dewhurst, J.H. 152, 158, 159 Checkel, J. 196 CHELEM-CEPII database 86, 98 Cheng, I.-H. and Wall, H.J. 86, 87, 96, 99 Cheshire, P. and Carbonaro, G. 159 civil society organisations (CSOs) 218 Cliff, A. and Ord, J.K. 160 CMs see common markets Cold War 162, 166, 171, 173, 174, 177, 189, 232, 237, 240 Collier, P. et al. 18

Comecon see Committee for Mutual Economic Assistance COMESA 20, 22, 33–4 Committee for Mutual Economic Assistance (Comecon) 89, 99, 189 Committee on Regional Trade Agreements (CRTA) 110, 128n4 Common Labour Market (Nordic countries) 195 Common Market of the South (Mercosur) 188, 202, 207–8 common markets (CMs) 165 communicative indicators, conceptual de-/reconstruction 53–4; and contextual understanding 52; and cultural-political correctness 58; and cultural/ life-world knowledge 59; and deﬁnition of integration 53; and ideology 59; and importance of participation/conceptual complementarity 53; inclusion/ exclusion perspective 53, 58; and integration as process/outcome 56–7; and local discourse 52; and local ownership/community empowerment 60; normative/life-world perspective 57, 60–1; and perceptual dimension of integration 52–3; positive suggestions 61; and prerequisites of discourse 57–8; and result-orientated process 59–60; and structural-functionalism 54–5, 56, 61; and sustainable peace building 55–6, 58, 60 Compensation and Development of the Common Market for Eastern and Southern Africa 141 confederation of public universities (CSUCA) 201 conﬂict resolution see peace building Convention on Civil Society 199 Cooper, R. 15, 37, 133, 134 Corkery, J. 185 Correlates of War Project 243 Corruption Index 171 Cotonou Agreement 9, 10, 20, 37n5 Council of Europe 187, 188 Country Review Team 188 CRTA see Committee on Regional Trade Agreements CSOs see civil society organisations CSUCA see confederation of public universities CUs see customs unions

254 Index customs unions (CUs) 162, 163, 165 Dahl, R. 190 Dalgaard, C.J. and Vastrup, J. 159 Dalum, B. et al. 69 De Lombaerde, P. 37, 146 Deardoff, A.V. 83, 96 decentralisation, contribution to economic development 133; and ﬁscal federalism 130–5; ﬁscal indicators 135–6; and minimisation of of overall size of public sector 132; theorem 131–2 Delors, J. 199 democracy 168–9, 188–9 Deutsch, K.W. and Savage, I.R. 89 Devlin, R. and Ffrench-Davis, R. 28, 29, 38 DG Development 21, 22, 30, 32–3 Diarra, C.O. 238 Dickey, H. 159 discourse theory 52, 57–8 Dobson, W. 14 Dorrucci, E. et al. 10, 17, 22 Drysdale, P.D. and Garnaut, R. 67, 83, 89 ECB see European Central Bank economic convergence, approaches to 147n5; background 146–7; cross-section (b-convergence) 150–3, 159n9–n13; debates on 146; deﬁnition 146–7; distribution dynamics 155–7, 160n15; Gini coefﬁcient 148–9, 159n8; Gini index 147–8; nominal/real 147, 159n2, n4; and per capita income 147, 159n3; and polarisation 149–50; as reduction of inequality 147–50; spatial analysis 157–8, 160n16–n18; Theil index 149, 159n7; time-series (stochastic convergence) 153–4, 159–60n14 Economic Freedom of the World (EFW) 167, 178n8 economic integration 9, 65–179, 241, 242; and common markets 24; and customs unions 24; and free trade areas 24; gravity models 27, 38n24; and institutional integration 22, 37–8n19; monitoring 21–2; and preferentialism 24; stages approach to 17, see also regional integration; system of indicators of regional integration (SIRI) Economic and Social Consultative Forum 202 economic union 165

ECOWAS 20, 238, 239 EDF projects 21 EFW see Economic Freedom of the World El-Agraa, A.M. 38 Ermida, O. 203 Ernst&Young 143 Esteban, J.M. and Ray, D. 149 EU see European Union EU Rule of Law Mission to Georgia 239 European Association of Small and Medium Enterprises (UEAPME) 198 European Central Bank (ECB) 10, 17–18, 21, 22, 37–8n19 European Commission 9, 21, 191, 198, 199, 203 European Convention on Fundamental Rights and Freedom 187 European Court of Justice 197 European Economic and Social Committee (ESC) 197–8, 204, 206 European Ombudsman 197 European Parliament 197, 198, 220 European Single Market 94 European Social Dialogue 197, 198–9 European Union (EU) 75, 77, 81, 99, 101, 118, 122, 123, 124, 126, 143, 168, 171, 188, 191, 205, 219, 221 European Union Accession Countries 166 FCES-Mercosur 203 Fernandez, R. and Portes, J. 162 Finer, S.E. 185 Fiscal Decentralisation Indicators (World Bank) 10 ﬁscal federalism, basic theory 131–3; and country size/economic integration 134; and decentralisation 130–3; and long-term tendency towards deeper interdependence 134; measurement tools 135–6; and policy in multilevel government structure 135; precepts as general guidelines 133; and regional integration 133–5; spatial/institutional dimensions 134–5 ﬁscal integration indicators, extra-regional funding 142; in-kind contributions 142; issues 139–43; ‘net’ contributions 141–2; proposal for 136–9; regional budgets complexity/type of regional organisation 140–1; regional integration/public ﬁnance 139–40; transparency, public accounting

Index 255 standards, data availability 142–3; vertical/horizontal imbalances 141 Flowers, M. 144 FOPREL 207 Frankel, J.A. 82, 83, 84; et al. 109 Fratianni, M. and Pattison, J. 9, 37 Free Trade Agreement of the Americas (FTAA) 206, 220 free trade agreements (FTAs) 94, 101, 108, 162, 163, 165 Freeman, R. 216 Freudenberg, M. et al. 23, 86, 87, 89, 98, 99, 102 Friedman, M.J. 151 FTAA see Free Trade Agreement of the Americas FTAs see free trade agreements Funck, B. and Pizzati, L. 159 Fund for Cooperation (S. Africa) 140 Fund for Cooperation, Compensation and Development of the Economic Community of West African States 141 Galtung, J. 241, 242, 243 GAP see Global Accountability Project GATS see General Agreement on Trade in Services GATT see General Agreement on Trade and Tariffs Gaulier, G. 83; et al. 84; and Ünal-Kesenci, D. 82 Gavin, B. and Van Langenhove, L. 9 Geeraerts, G. and Stouthuysen, P. 233 General Agreement on Trade in Services (GATS) 110, 115, 119 General Agreement on Trade and Tariffs (GATT) 110, 116, 125, 126, 169 Gilpin, R. 18 Global Accountability Project (GAP) 215, 218, 222, 223–4, 227, 228, 229; complaints/redress 225; evaluation 224–5; participation 224; transparency 224 Golub, 178 González, J.I. 144 governance 9, 24, 209–10; accountability 186, 187, 191, 210; appropriateness of regional institutional arrangements 192–209, 210; at national level 185; background 183–4; basic principles 186; coherence 191–2; deﬁnitions of 185; and effective organs of state 185; effectiveness 191, 210; and central

governments 186–7; European White Paper on 191; framework for 186–7; good 185; and governance without government 186; indicators of 186; intergovernmental systems 187–9; legitimacy requirements 190–1; and openness 186; openness 191, 210; participation 191, 193–6, 210; plugging gaps in multilevel governance 221–2; principles/common standards 188–9; promotion of democratic practice 188–9; regional integration systems 189–93; and role of citizen/civil society 185; and transparency 186 Government Purchasing Agreement (GPA) 116, 118 GPA see Government Purchasing Agreement Grant, R.W. and Keohane, R.O. 222 GRECO see Group of States against Corruption Greene, W.H. 178 Greif, A. 14 Group of States against Corruption (GRECO) 188 Grugel, J. and Hout, W. 234 Guba, E.G. and Lincoln, Y. 44–7, 52 Haas, E. 54 Habermas, J. 52, 53, 56–7, 59 Harbom, L. and Wallensteen, P. 243 Held, D. 221–2; et al. 29, 48; and Koenig-Archibugi, M. 222; and McGrew, A. 218 Heritage Index 178n8 Heston, A. et al. 167 Hettne, B. 9, 15, 29, 38, 108, 233; et al. 234; and Söderbaum, F. 16, 29, 184, 193, 195 Heyden, K. 128 Hinojosa, O.R. et al. 38 IADB 10 IADB see Inter-American Development Bank Iapadre, L. 30, 84 IGOs see intergovernmental organisations IMF see International Monetary Fund Indices of Economic Integration Effort 11 INGOs see international non-governmental organisations Inman, R.P. and Rubinﬁeld, D.L. 133 institutions 9; academic 11; common

256 Index decision-making machinery 124; impact of 125; indicators of 27–8; infrastructure 124–5; intergovernmental 124; multilateral 11; regional 11; and regional preferences 125; specialist committees 124; technical cooperation/ capacity building 124 integration see economic integration; regional integration; system of indicators of regional integration (SIRI) Inter-American Convention Against Corruption 188 Inter-American Democratic Charter 188 Inter-American Development Bank (IADB) 10 intergovernmental organisations (IGOs) 218, 223, 226 International Monetary Fund (IMF) 136, 169, 171, 221 international non-governmental organisations (INGOs) 58, 218, 220, 223, 228 International Political Interactions (IPI) 245 intra-regional trade 65, 183; and geographical neutrality 66–7, 72; measuring intensity of 66–71; propensity to 73–5, 81–2; share of 66; and trade creation/relative openness 72–5; trends in 75–80, 84n16, n18, n19 intra-regional trade intensity index 67–8, 82–3n3, n4, n7; dynamic ambiguity 69–71; and gravity model 73, 83–4n13, n14; homogeneous index 68–9, 71; range asymmetry 69; range variability 68–9; and revealed trade preferences 71; symmetric index 69, 71, 83n5; and trade introversion 71, 72 IPI see International Political Interactions Jácome, F. 204 Jelin, E. 208 Jungmittag, A. et al. 83 Kahler, M. 221 Kaufman, D. et al. 228 Keohane, R.O. 18, 216; et al. 195; and Nye, J.S. 37 Keough, N. 38 Kindleberger, C.P. 18 Kivimäki, T. 233 Kojima, K. 67

Kondo, S. 186 Kreinin, M.E. and Plummer, M.G. 23 Krugman, P. 23, 109 Kunimoto, K. 67, 73 Laakso, L. 233 Lafay, G. 98 Laffan, B. 141, 144 Lake, D.A. and Morgan, P. 233, 238 Latin America 199; and regional socialisation 204–5; representativeness/resources 2034; structural problems in making regional fora work 203–5; subregional structures in 199–202 Leamer, E.E. and Stern, R.M. 83 Lebessis, N. and Paterson, J. 203 Lepgold, J. 238 Levy, J.S. 43 Linklater, A. 53 Linnemann, H. 38 Logic of Collective Action 18 Luhman, N. 55 MacDougall Report (1977) 144 McKinnon, R.I. and Nechyba, T. 133 Maclean, S. 233 McLure, C.E. Jr 140 Mansﬁeld, E.D., et al. 163; and Pollins, B. 244 Marshall, M.G., et al. 169; and Gurr, T.R. 232 Mattli, W. 13, 168, 194 Matyas, L. 98 Meeusen, W. and Villaverde, J. 146, 159 Member States of the African Union 188 Mercosur see Common Market of the South Merrit, R.L. and Russett, B.M. 232 MFN see most favoured nation Miles, M.B. and Huberman, M. 44 military integration 241, 242 Mingst, K.A. and Warkentin, C.P. 204 MINUSTAH see UN Stabilisation Mission in Haiti Mitrany, D. 232 Monge, H. 203, 204 Monterrosa, C. de 197, 204 most favoured nation (MFN) 115, 126 Musgrave, R.M. 131, 132, 139 NAFTA see North American Free Trade Agreement

Index 257 NATO see North American Treaty Organisation NEPAD see New Partnership for Africa’s Development New Partnership for Africa’s Development (NEPAD) 188 new regionalism 9, 27, 232, 233–4, 248n3 Newell, P. and Balour, S. 217, 223 Nice Treaty 197 Nielson, G. 211 non-governmental organisations (NGOs) 11, 58, 199, 203 Nordic Council 194, 211n10 Nordic Inter-Parliamentary Union 194 North American Free Trade Agreement (NAFTA) 75, 78, 80, 81, 99, 101, 102, 115, 118, 124, 126, 127, 207, 221 North American Treaty Organisation (NATO) 236 North–South agreements 162, 177n2 Nye, J.S. 232, 233, 237, 239–40, 241 OAS see Organisation of American States Oates, W.E. 131–2 OAU see Organisation of African Unity ODECA see Organisation of Central American States OECD see Organisation for Economic Cooperation and Development Olson, M. 18 Ondarts, G. 15, 37 One World Trust 215, 223, 225 Organisation of African Unity (OAU) 188, 237 Organisation of American States (OAS) 188, 239 Organisation of Central American States (ODECA) 200 Organisation for Economic Cooperation and Development (OECD) 109, 111, 169, 171, 193, 211n6 Osborne, D. 185 Ouro Preto Protocol 202 Paas, T. 38 Padrón, A. 203 Page, S. 37 Palmer, N.D. 234, 248 Panel of Eminent Persons 188 PARLACEN see Central American Parliament Parsons, T. 54, 56

Passport Union (Nordic countries) 195 peace building 55–6, 58, 60; and absence of physical violence 242–4; background 232–3; and behavioural conﬂict resolution 243–5; and concept of peace 241–3; and cooperation 245–6; and discussion network 244; and economic integration 241; importance of regional organisations 238–9, 249n8; and intervention 235–6, 246–8; measurement of 240–8; and military integration 241; political trade-off 238–40; and quotidian network 244; and regional action 237, 249n6; and regional integration 240–1, 248n2; regionalisation as peace tool 233–5, 248–9n5; regionalism/peace nexus 232–3, 236–8; and segmented social capital 245; and social integration 241; structural pillar 243; and support network 244; terminology 232, 248n1 Peace and Security Council 239 Pelkmans, J. 23–4, 183, 184 Penn World Tables 167, 168, 169 Perroni, C. and Whalley, J. 163 Poverty Reduction Strategy Papers (PRSPs) 221 Power Without Accountability? 223 Pöyhönen, P. 38 preferential trade agreements (PTAs) 99, 101, 108 PriceWaterhouseCoopers 143 PRSPs see Poverty Reduction Strategy Papers Prud’homme, R. 133 PTAs see preferential trade agreements Quah, D. 147, 155, 158, 160 qualitative monitoring, and consequences of integration process 48–9; data gathering 43–4; failure to accommodate value pluralism 45; four-stage history of evaluation 44–7; and interconnectedness of states 47, 48; overcommittment to positivist paradigm of inquiry 45–6; purpose 44; and stakeholders 49–50; tendency towards managerialism 45; units for analysis 44 qualitative–quantitative dichotomy 42–3 Ramírez, S. 201, 204, 206

258 Index RCRP see Regional Cooperation Review Process Rebelo, S. 159 Regional Cooperation Review Process (RCRP) 20 regional fora, appropriateness of 192–209; and commitment institutions 194, 196; and compliance by national authorities 194–5; and creation of community 193–4; and credibility of commitment 194; democratic/ integration deﬁcit 205; European model 197–9; inappropriateness of 208–9; institutionalisation without credibility 206; integration without ﬁxed borders 206–7; Latin American model 199–202; legitimacy of process 195; legitimation without commitment 205–6; legitimation/ community-building 207–8, 212n24; and principle of participation 192–6; regional decisions/parallel actions 207; and regional socialisation 204–5; and regionness 195; socialisation mechanisms 196; structural problems 203–5 regional integration, assessment/measurement of 1–4; breakpoints 15; capacities for 19; causalities/tendencies 1; concepts of 13–16; and cooperation 14, 37n7, n8; deﬁnition of 1, 13–14; as driving force 1; evaluation of policies 20; and formal preferential agreements 183; formal/informal 15, 37n10; forward-looking variables 22–3; impact of 9; and interdependence 15, 29–30, 37n11; key variables 24; key variables of complexity 35; and levels of ambition 23; and migratory ﬂows 30, 38n29; monitoring 2, 4–5, 9–10, 42–50; as multidimensional social phenomenon 17; normative tradition 17; and openness 15; and peace building 240–2; political integration, governance, peace 183–251; positive/ negative 14, 37n9; primary/secondary actors 27; processes of 9; and proximity of actors 27; real degree of 183; and real regions 15–16; and structural asymmetries 18, 37n16; structural factors 27; and sub-national (regional) actors 16, 37n12; and subsidiarity 24;

theoretical/real world approaches 1, 2, 9–62, see also economic integration; system of indicators of regional integration (SIRI) regional integration arrangements (RIAs) 19; basic model 170–1, 178n15–n17; bivariate probit estimates 173–6; business cycle uncertainty 168; common traits 163; country characteristics/income 169, 178n11, n12; and democracy 168–9; economic uncertainty 163, 166; formation of 162–3; levels of participation in 163; political uncertainty 163, 166, 167–8, 178n8, n9; price uncertainty 168; results of analysis 171–6; and reverse causality 169; trade uncertainty 167, 177–8n6, n7 regional integration levels, common markets 165; customs unions 165; data set of 164–6, 177n4, n5; economic union 165; factors contributing to 177; preferential trade agreements/areas 164–5; regional integration arrangements 164; sectoral trading agreements 164; supranational union 165 regional trade agreements (RTAs) 9, 86, 87, 94, 102, 107, 162, 163, 165; assessing impact of 107–10; and bloc formation 109; and comment on decision/regulation 120; compatibility of 110; coverage 115–17; and degree of preference 121, 123, 126; as determinants of trade ﬂows 97–101; and dispute settlement provisions 126; dynamic time path analysis of 110; economic effects of deeper integration 108–9; elements of framework 115–28; entities covered 116–17; and equivalence 122; framework for 111–14; and government discrimination 125–6; and harmonisation 121–2, 123; implementation 126–7; institutional provisions 124–5; and investment 109; and liberalisation/multilateralism 108; and mutual recognition 122, 123; nature of ‘new wave’ agreements 108; need for analytical framework 111; and notiﬁcation of statutory measures 119; numbers of 108; and open decision-making 120; participation in

Index 259 108; policy area 115; principles 118–19; and prior notiﬁcation 120; and PTAs 108; regulatory instruments 117; and reviews/remedies 127; and right to bring a case 127; and rulemaking 108; rules of origin 118; safeguards 125–6; schedules 115–16; and secondary measures 119; structure of 109; substantive measures 121–3; systemic effects 109; and tariffs 108; terminology 128n1; and trade creation/diversion 108; transparency 119–21; transparency in 111 regionalisation, agents 234; bilateral trade intensities 95–101; CEPII’s index of relative trade intensity 89–94; direction 234; formal/real distinction 184; geographical classiﬁcation 103; and governance 184, 185–93; importance of 101–2; institutional context 94–5; natural pattern of 87; old/new 234; overview 87–95; as peace tool 233–5; and polarisation 88; and regionness/ regionhood 184, 210–11n3, n4; similarity of specialisation patterns 104; and stability of system 184; vectors/motivation 234; and welfare of populations 184 relative trade intensities (RTIs) 86, 183; in America 90; in Asia-Oceania 93; CEPII’s index of 89–94; in Eurafrica 91–2; and former communist countries 89; independent from size/openness of partners 89; neighbourhood relationships 94; as purely relative 89 Rhodes, R.A.W. 187 RIAs see regional integration arrangements Rodrik, D. 167, 173 Romer, D. 178 Romer, P. 159 Rosenau, J.N. 186, 235–6 Rosendorff, P. 169 RTAs see regional trade agreements RTIs see relative trade intensities Ruiz-Tagle, J. 207 SADC see South African Development Community Samuelson, P. 131, 132 sanitary and phytosanitary (SPS) measures 119, 121, 123, 125 Savage, I.R. and Deutsch, K.W. 82

Schiff, M. 162 Schmitter, P.C. 192 Schott, J. 108 Schulz, M. et al. 232 Schütz, A. and Luckmann, T. 52 Shah, A. 144 SICA see Central American Integration System Silva Lopes, P. 37 Singer, J.D. and Small, M. 243 Singer, M. and Wildavsky, A. 233 Slocum, N. et al. 27, 37 Small, M. and Singer, J.D. 243 Smismans, S. 198 social integration 241, 242 social science 42–3 Solow, R. 151 South African Customs Union 165 South African Development Community (SADC) 20, 140, 165, 238 South Asian Development Fund 141 Southern Common Market (Mercosur) 75, 76–7, 78, 80, 81, 99 SPS see sanitary and phytosanitary measures Stakhovitch, A. 18 Sundelius, B. 210, 211; and Wiklund, C. 211 supranational union 165 sustainability 11; peace building 55–6, 58, 60 system of indicators of regional integration (SIRI) 9; actors/structural factors 26–7, 38n22; and concepts of integration 13–16; construction/design of 10, 11; and counterfactual analysis 28; criteria for implementation of 11–13; dimensions/categories of variables 21–4; effects 28–9, 38n25; ex ante/ ex post approach 23; and governing needs/capacities 24; institutionalisation/ policy implementation 27–8; interdependence 29–31, 38n29; and international/intercultural context 13, 37n6; methodological/organisational options 18–20; and motivations 16; open/ﬂexible approach 17–18; and overlapping agreements 19, 29; positive/negative 16; proposal 24–31; quantitative/qualitative balance 29, 38n26; restrictions on 16, 37n14; role of theory, the normative, Euro-centred tradition 17–18; and trade creation/

260 Index diversion 23, 38n20; traditional (comparative) vs relative (reﬂexive) indicators 11; users/producers of 10–11, see also economic integration; regional integration Tavares, R. 233, 248 TBTs see technical barriers to trade technical barriers to trade (TBTs) 116, 119, 121, 123, 125, 126 Tegucigalpa Protocol (1991) 200 Ter-Minassian, T. 144 TERM see Transport and Environment Reporting Mechanism Thant, M. et al. 15 Tiebout, C. 131, 132 Tinbergen, J. 16, 38 TNCs see transnational corporations TOI see Trade and Openness Index Trachtman, J. 128 trade creation, cyclical/structural forces 72, 83n10; internal/external effects of integration 72, 83n11; and relative openness 72–5; and trade-to-GDP ratio 72, 83n9 Trade and Openness Index (TOI) 167 Trade Related Investment Measures (TRIMs) 115 transnational community 183, 210n1, n2 transnational corporations (TNCs) 218, 220, 223, 226, 228 Transparency International 171 Transport and Environment Reporting Mechanism (TERM) 38n26 TRIMS see Trade Related Investment Measures UEAPME see European Association of Small and Medium Enterprises UN Stabilisation Mission in Haiti (MINUSTAH) 239 UNDP 185 UNDP see United Nations Development Programme UNECA 10, 11, 19, 20, 21, 37, 37n4 UNICE see Union of Industrial and Employers’ Confederation of Europe Union of Industrial and Employers’ Confederation of Europe (UNICE) 198, 203 Union of International Associations 143 United Nations (UN) 208, 235, 236, 237

United Nations Charter 238, 249n7 United Nations Development Programme (UNDP) 58, 59, 227, 228 United Nations Human Development Reports 244 US Federal Communication Commission (FCC) 117 Van der Voort, W. 197 Van Langenhove, L. 9, 43, 210 Van Velden, L. 37 Verdoorn, P.J. and Meyer zu Schlochtern, F.J.M. 38 Villaverde, J., and Maza, A. 159; and Sánchez-Robles, B. 159 Viñals, J. 159 Vollrath, T.L. 83 Wallensteen, P. and Sollenberg, M. 232, 243 Warsaw Pact 189 WB see World Bank Weiler, J.H.H. 191 Weingast, B.R. 133 Werth, M. et al. 38 Whalley, J. 163 White Paper on European Governance (2001) 191 Winters, A. 108, 128 Woods, N. and Narlikar, A. 222 World Bank 10, 11, 19, 20, 21, 28, 36, 37, 58, 65, 108, 136, 221, 223 World Trade Organisation (WTO) 38, 84, 169, 219, 221; and common international standards 123; and dispute settlement 127; dispute settlement mechanism 220; emergence of 166; estimate of number of RTAs 108; and global free trade 9; and government discrimination 126; institutional provisions 124, 126; positive/negative listings 115–16; and RIA compatibility 19; rules 110, 116, 128; rules of 107; and schedules 115–16; sectoral coverage of 115; trade agreements notiﬁed to 86, 94; and transparency 111, 120–1, 128n7 WTO see World Trade Organisation Wu, J. 163 Young, I.M. 216 Zhang, J. and van Witteloostuijn, A. 84

eBooks – at www.eBookstore.tandf.co.uk

A library at your fingertips!

eBooks are electronic versions of printed books. You can store them on your PC/laptop or browse them online. They have advantages for anyone needing rapid access to a wide variety of published, copyright information. eBooks can help your research by enabling you to bookmark chapters, annotate text and use instant searches to find specific words or phrases. Several eBook files would fit on even a small laptop or PDA. NEW: Save money by eSubscribing: cheap, online access to any eBook for as long as you need it.

Annual subscription packages We now offer special low-cost bulk subscriptions to packages of eBooks in certain subject areas. These are available to libraries or to individuals. For more information please contact [email protected] We’re continually developing the eBook concept, so keep up to date by visiting the website.

www.eBookstore.tandf.co.uk