The World Economy, Global Trade Policy 2004 (World Economy Special Issues)

The World Economy The World Economy Global Trade Policy 2004 Edited by David Greenaway University of Nottingham © ...

Author: David Greenaway

15 downloads 704 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

The World Economy

The World Economy Global Trade Policy 2004

Edited by

David Greenaway University of Nottingham

© 2005 by Blackwell Publishing Ltd BLACKWELL PUBLISHING 350 Main Street, Malden, MA 02148-5020, USA 108 Cowley Road, Oxford OX4 1JF, UK 550 Swanston Street, Carlton, Victoria 3053, Australia The right of David Greenaway to be identified as the Author of the Editorial Material in this Work has been asserted in accordance with the UK Copyright, Designs, and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher. First published 2005 by Blackwell Publishing Ltd Library of Congress Cataloging-in-Publication Data ISBN 1-4051-2988-3 (paperback) A catalogue record for this title is available from the British Library. Set in 11/13pt Times by Graphicraft Limited, Hong Kong Printed and bound in the United Kingdom by TJ International, Padstow Cornwall The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com

Contents Foreword Contributors

vii viii

REGIONAL FOCUS 1

2

Economic Partnership Agreements Between Sub-Saharan Africa and the EU: A Development Perspective ............ LAWRENCE E. HINKLE and MAURICE SCHIFF

1

What Role for the EPAs in the Caribbean? ............ MICHAEL GASIOREK and L. ALAN WINTERS

15

WTO TRADE POLICY REVIEWS 3

4

5

Trade Policy in Burundi: Reform Without Political Stability.................................................................. CHRIS MILNER

43

Regionalism Within Multilateralism: The WTO Trade Policy Review of Canada ....................................... KEITH HEAD and JOHN RIES

57

Trade Policy Making in a Small Island Economy: The WTO Review of the Maldives ................... PREMA-CHANDRA ATHUKORALA

81

SPECIAL FEATURE 6

7

Trade Reform and the Southern Mediterranean ................................................................................... MICHAEL GASIOREK

101

Beyond Border Barriers: The Liberalisation of Services Trade in Tunisia and Egypt ....................... DENISE EBY KONAN and KARL E. KIM

109

v

vi 8

CONTENTS Rules of Origin and the EU-Med Partnership: The Case of Textiles .................................PATRICIA AUGIER, MICHAEL GASIOREK and CHARLES LAI-TONG

129

Does the Quality of Institutions Limit the MENA’s Integration in the World Economy?.......................................... PIERRE-GUILLAUME MÉON and KHALID SEKKAT

155

Index ........................................................................................................................

179

9

Foreword Each year The World Economy publishes an issue dedicated to developments in global trade policy. That issue always includes a number of ingredients: evaluations of a range of WTO Trade Policy Reviews; a regional feature; and a ‘special feature’ devoted to a current issue. Because of the interest in this particular issue of the journal, it is always published also as a stand alone book. Global Trade Policy 2004 includes Trade Policy Reviews of Burundi, Canada and The Maldives; the regional focus is on Economic Partnership Arrangements; and the special feature on trade reform in the Southern Mediterranean. I am grateful to contributors for the timely delivery of manuscripts and to Blackwell for the expeditious publication of this volume. David Greenaway, Director, Leverlulme Centre for Research on Globalisation and Economic Policy, University of Nottingham.

vii

Contributors

Prema–Chandra Athukarola – Patricia Augier – Michael Gasorick

–

Keith Head –

Australian National University Université de la Mediterranée Sussex University and GREQAM University of British Columbia

Lawrence Hinkle

–

The World Bank

Karl Kim

–

University of Hawaii

Denise Eby Konan

–

University of Hawaii

Charles Lai-Tony

–

CNRS

Chris Milner

–

University of Nottingham

Pierce-Guillaume Méon

–

University Robert Schuman

John Ries

–

University of British Columbia

Maurice Schiff –

The World Bank

Khalid Sckkat –

University of Brussels and The European Commission

L. Alan Winters –

The World Bank and Sussex University

viii

ECONOMIC AGREEMENTS: SUB-SAHARAN AFRICA & THE EU

1

1

Economic Partnership Agreements Between Sub-Saharan Africa and the EU: A Development Perspective Lawrence E. Hinkle and Maurice Schiff

1. INTRODUCTION

HE EU is Sub-Saharan Africa’s (SSA) largest single trading partner, buying on average 31 per cent of its merchandise exports and providing 40 per cent of its merchandise imports. The trade relationship between SSA and the EU is important for the region’s development, and proposed free trade agreements (FTAs) between the EU and SSA could have a significant impact on most of SSA. The Cotonou Agreement, signed in June 2000 between the EU and SSA (and other ACP1) countries, provides for negotiation of Economic Partnership Agreements (EPAs) between them. The negotiations are to be completed by December 2007, and the EPAs are to go into effect in January 2008. The Cotonou Agreement provides for replacing the unilateral trade preferences that the EU accords to the ACP countries under the Lomé convention with EPAs involving reciprocal obligations. Not only will the EU provide free access to its markets for ACP exports, but ACP countries will also have to provide free access to their own markets for EU exports. In addition to reciprocity, a second principle of the Cotonou Agreement is that of differentiation, whereby ACP LDCs are to be treated differently from ACP non-LDCs. This means that LDCs are unlikely

T

The authors would like to thank Ousmane Badiane, Nancy Benjamin, Paul Brenton, Uri Dadush, Luis de Azcarate, Caroline Freund, Alan Gelb, Alberto Herrou-Aragon, Bernard Hoekman, Christiane Kraus, Jeff Lewis, Dorsati Madani, Marie Francoise Marie-Nelly, Will Martin, John Nash, Manuel de la Rocha, Salomon Samen, Claire Thirriot, Fahrettin Yagci and Gianni Zanini for helpful comments on an earlier, longer draft on which this paper draws. The views expressed in this paper are those of the authors and are not necessarily those of the World Bank, its Board of Directors, or the governments they represent. 1

For a list of abbreviations see the Appendix.

1

2

LAWRENCE E. HINKLE AND MAURICE SCHIFF

to have to reciprocate and open their markets to EU exports as much as the non-LDCs in order to maintain their preferential access to EU markets. As is shown in Section 2, differentiation creates some complications with respect to regional integration within SSA. The planned EPAs are intended to restructure trading arrangements between the EU and the SSA countries to make them more effective in promoting EUSSA trade, more supportive of broader development goals, and more compatible with World Trade Organisation (WTO) rules. They also provide for possible coordination between the EU aid programme and the EPAs. The EU has stated its desire to use the EPAs as instruments of development, and this paper analyses them from a development perspective. The main element of the EPAs is to be the establishment of FTAs between the EU and each of the regional EPA negotiating blocs. The latter are self-determined groupings of the 77 ACP countries, which are encouraged to form regional blocs for pursuing regional integration and negotiating EPAs with the EU. The EPAs offer considerable potential benefits to SSA countries. It is argued that these potential benefits might be realised if the EPAs are amended as recommended in this paper. However, EPAs also pose a number of policy, administrative and institutional challenges for SSA countries, including: replacing forgone tariff revenues, avoiding serious trade diversion, appropriately regulating liberalised service industries, and liberalising internal trade within SSA’s regional trade blocs. To achieve pro-development outcomes through the EPAs, it is essential both that the SSA countries and organisations concerned be well prepared for these negotiations and that the EU adopt a benevolent, development-oriented approach to them, subordinating its commercial interests to the development needs of the SSA countries when necessary. Two points need to be made here. First, there are doubts whether, unless they are modified, the trade EPAs will actually prove beneficial to SSA. This paper does not take a position on whether or not SSA should enter into EPAs with the EU. Rather, it starts from the notion that the process of forming EPAs is unlikely to be reversed and examines the conditions that will maximise SSA’s benefits from the EPAs. Second, the EPAs also apply to non-SSA ACP countries. This paper focuses on SSA, though many of the recommendations may also apply to non-SSA ACP countries. Note that Pascal Lamy, the EU Trade Commissioner, made a proposal at the May 2004 G-90 summit in Dakar that might lead to a change in the EPA process. He proposed that the G-90, a group consisting of ACP and non-ACP LDC countries, should not have to make concessions at the WTO Doha Round of multilateral trade negotiations, i.e., he proposed a ‘free round’ for the G-90. This proposal opens the door to the possibility that the same might apply to the ACP countries in the EU-ACP negotiations and that the EPA process might be reversed. This paper, which draws on Hinkle and Schiff (2003), (available from the authors upon request), is organised as follows. Section 2 deals with the


3

Everything-but-Arms (EBA) initiative. Section 3 examines the relationship between EPAs and SSA regional trade blocs. Section 4 analyses the necessary accompanying reforms in SSA and Section 5 looks at the relationship between EPAs and the WTO’s Doha Round. Section 6 deals with liberalisation of trade in services, and Section 7 concludes by examining the conditions for the EPAs to benefit SSA and accelerate its integration in the global economy.

2. THE EU’S EVERYTHING-BUT-ARMS (EBA) INITIATIVE

In addition to providing trade preferences to ACP countries under the Cotonou Agreement, the EU now provides full quota-free, tariff-free access to its market to LDCs, including those in SSA, under the Everything-but-Arms (EBA) Initiative, which it adopted in 2001. When full market access is phased in for bananas, rice and sugar by 2009, the EBA Initiative will grant full, unrestricted market access to LDCs including the commodities currently subject to the EU’s commodity protocols with the ACP countries. Although the EBA Initiative is definitely beneficial for the LDCs, it has somewhat complicated the EPA process by creating different trading environments and negotiating incentives for the LDCs and nonLDCs in SSA, some of which are members of the same customs union or FTA. One of the stated objectives of the EPA process is to promote regional integration among SSA countries. However, as mentioned in Section 1, the principle of differentiation, whereby LDCs may not have to open their markets to EU exports in order to maintain preferential access for their own exports to the EU, has complicated the process of regional integration within SSA. Customs Unions (CUs) such as UEMOA and CEMAC have both LDC and non-LDC member countries. If the former do not liberalise their trade policy with respect to imports from the EU while the latter do, it will mean the end of the common external tariff (CET) and of the CUs. Provision of EBA access to LDCs has exacerbated the problem because it has been provided unconditionally, without the EU asking for any concession in return. Thus, with EBA access in their pocket, it is unlikely that LDCs will provide improved access to imports from the EU unless they receive other benefits for doing so. The best way to reduce these complications and facilitate regional integration in SSA would be for the EU to provide EBA market access to all SSA (and other ACP) countries signing EPAs. Provision of EBA access to the non-LDCs in SSA under the EPAs would also give them unrestricted access to the markets now governed by the EU’s commodity protocols once any transitional arrangements have expired and would effectively resolve the question of the future of the protocols. All developing countries and LDCs are recognised as legitimate groups for special and differential treatment (Part 4 of the GATT and the Enabling Clause) while entities such as SSA and the ACP are not. Thus, providing unilateral EBA

4


access to all SSA is not WTO compatible. On the other hand, providing EBA access would be WTO compatible if EBA access were an integral part of FTAs between the EU and SSA regional trade blocs. The principle of reciprocity in the Cotonou Agreement does imply the establishment of FTAs between the EU and each of SSA’s regional EPA negotiating blocs. Non-LDC countries that choose not to enter into these EPAs will lose their Cotonou preferences in EU markets and are thus likely to enter into EPAs. As for LDCs, the EU is likely to have to provide them with additional incentives to convince them to enter into EPAs. Thus, providing EBA access to all SSA countries will only be WTO compatible if SSA countries enter into EPAs (as part of regional EPA negotiating blocs) with the EU. The same applies to the other ACP countries. However, as noted above, with the proposal made by Pascal Lamy at the G-90 Ministerial at Dakar in May 2004, the above might not hold, especially if the G-90 becomes a legitimate group under the Enabling Clause. The rules of origin (RoO) under both the Cotonou Agreement and the EBA Initiative are complex and restrict SSA exports. Those under the EBA Initiative are the same as under the EU GSP and are the more restrictive. Unless RoO are liberalised and simplified, the benefits of even full tariff-free, quota-free market access under EPAs will be limited. In fact, Brenton (2003) shows that SSA LDCs make little use of EBA preferences and prefer to export to the EU under Cotonou preferences. The reason is that, even though Cotonou preferences are clearly less generous than EBA preferences, the RoO that apply under Cotonou are sufficiently less restrictive to make exporting under the latter more attractive to most SSA exporters. A simpler, less restrictive rule of origin for the EPAs would be to give SSA exporters the choice between meeting a ‘change of tariff heading’ rule or a uniform 25 per cent value-added rule, together with liberalisation of the general tolerance rule to permit 25 per cent imported inputs and full cumulation among all ACP countries, other LDCs, South Africa and the EU.2 Independent reviews of the EU’s import safeguard measures (such as anti-dumping) and sanitary and phyto-sanitary (SPS) regulations (such as those on ground nuts) should also be undertaken to see if more development-friendly provisions concerning these can be included in the EPAs.

3. EPAS AND REGIONAL INTEGRATION IN SSA

An important objective of the EPA process is to promote outward-oriented regional integration among the SSA countries and to limit the ‘hub and spoke’

2

Liberalisation of EBA RoO might provide an incentive for LDCs to join EPAs with the EU.


5

effect that bilateral free trade just between the EU and individual SSA countries could have. Hence, the SSA countries are expected to form themselves into selfdetermined regional groups for negotiating the EPAs. Determining the nature of the EPA groupings in SSA while also supporting existing regional trade agreements (RTAs) has been no simple matter. A major reason is the high degree of heterogeneity in SSA’s RTAs, which include a number of overlapping PTAs, FTAs and customs unions (CUs) with different structures, operational rules and implementation levels. As mentioned in Section 2, one problem is that of CUs with both LDC and non-LDC members. This is the case for UEMOA and CEMAC, and is also true for SACU (with Lesotho being an LDC). Another problem is that a number of Eastern and Southern African countries are members of both COMESA and SADC. However, the interaction of the EPA process and the political support for regional integration in SSA has provided a dynamic impetus for rationalising this situation, and the composition of the four regional EPA negotiating groups in SSA has now largely been resolved. In Eastern and Southern Africa, the regional trade group ESA encompasses most of COMESA as well as the SADC countries that decided to join the ESA negotiating group with the COMESA countries. The non-SACU three SADC countries that decided not to join in the ESA are Angola, Mozambique and Tanzania. The first two may join SACU. Tanzania is a member of the EAC, together with Kenya and Uganda. These two have joined the ESA. There are at present four negotiating groups in SSA: ECOWAS + Mauritania in West Africa; CEMAC + São Tomé in Central Africa; ESA in Eastern and Southern Africa; and the SADC group in Southern Africa. The latter consists of Botswana, Lesotho, Namibia and Swaziland (known as BLNS, or SACU minus South Africa) + Angola, Mozambique and Tanzania. Two issues are left to be resolved, related to the EAC and SACU. The EAC is expected to become a CU and it therefore seems to make little sense for Tanzania not to join Kenya and Uganda in the ESA. In SACU, the question is what to do about South Africa. This is examined at the end of this section. Significant barriers to intra-regional trade still remain within ‘free’ trade areas, and even within customs unions, in SSA. The ECOWAS FTA has not yet really been implemented (the main problem being Nigeria3), and some countries in COMESA still maintain barriers to free intra-regional trade. Even though CEMAC and UEMOA are ‘customs unions’, substantial obstacles to internal free trade and country deviations from the common external tariff remain in both. The EAC customs union is not yet operational but is likely to face similar problems. Only SACU is a fully functioning customs union with full internal free trade, a 3

Nigeria’s economy is larger than that of all of the other ECOWAS countries put together, and its trade policy is much more restrictive than that of the UEMOA.

6


common external tariff observed by all of its members, and common administration of the external tariff and pooling of the revenues from it. Barriers to intra-SSA trade within CUs and FTAs may not, in themselves, be a problem in terms of negotiating EPAs as long as the CUs have a common external tariff in place. However, further intra-CU and intra-FTA liberalisation of trade in both goods and services are likely to benefit their members. There are several steps which should be taken in the context of the EPAs to strengthen regional integration in SSA. First, those member countries of customs unions that have not fully implemented their CU’s common external tariff (CET) need to do so as soon as possible. As long as the CET is not fully implemented by all member countries, RoO are going to be needed. Note that the RoO within SSA regional trade blocs are complex and restrictive (particularly in SADC) and there are protectionist pressures to increase their restrictiveness.4 Second, full intra-CU and intra-FTA free trade in goods and services, as well as free movement of labour, needs to be implemented in most CUs and FTAs. And, third, SSA countries should take advantage of the good practice precedent that would be set with regards to liberalised RoO in the EPAs to adopt similar standardised, liberal RoO for the various regional trade areas in Africa. Intra-CU and intra-FTA trade liberalisation in SSA should preferably be accompanied by a reduction in MFN tariffs. The reason is that without the latter, the former is likely to result in costly trade diversion, with increased intra-bloc trade resulting in a reduction in non-member exports to the bloc that liberalised its intra-bloc trade. MFN trade liberalisation will help to minimise this problem. The issue of MFN liberalisation is further discussed in Section 4. The situation with the EPAs for the SACU countries is particularly complex. Four of the five SACU countries (Botswana, Lesotho, Namibia and Swaziland, or BLNS) are ACP countries and are participating in the EPA process. However, the fifth and by far the largest SACU member, South Africa, is eligible neither for the trade benefits of the Cotonou Agreement nor for an EPA. Instead, it has a separate FTA with the EU, and given that South Africa sets SACU’s trade policy, many aspects of the FTA with the EU apply de facto to the other SACU countries. To rationalise this situation, SACU may eventually need to be included as one of the customs unions which signs an EPA outside the framework of the Cotonou Agreement (given that South Africa is not eligible). The four SACU member countries eligible for EPAs should ask the EU to give them EBA market access with improved, less restrictive rules of origin. In addition, if South Africa

4

RoO are prevalent not only in SSA FTAs but in CUs as well (except for SACU), and not only because the CET has not been fully implemented. In the UEMOA and CEMAC, tariff revenues are not pooled. Thus, RoO are needed to ascertain whether imports by a landlocked member country originated in a non-member country and therefore should pay the CET, or whether they originated in a member country and should be imported free of tariffs.


7

takes the lead in gradually lowering SACU’s peak tariffs on an MFN basis to a maximum of 20 per cent as recommended for other CUs and SSA countries (for reference to this, see last part of Section 4 below), a step which would probably be very beneficial for the less developed BLNS countries, South Africa should also receive improved EBA access to the EU market, even if this access might require a longer transition period for the EU than in the case of the other SACU countries.

4. PREFERENTIAL REDUCTIONS IN TARIFFS ON MERCHANDISE IMPORTS FROM THE EU AND ACCOMPANYING REFORMS IN SSA

a. Restructuring Indirect Tax Systems An important fiscal issue raised by the EPAs is the impact on government revenues of preferential reductions in SSA tariffs on merchandise imports from the EU. Revenues from tariffs still amount to 2 per cent of GDP in the median SSA country, and some countries depend even more heavily on tariff revenues, with these amounting to 4 to 6 per cent of GDP. Since the EU is the largest source of imports for most SSA countries, supplying 40 per cent of total imports on average, some countries are likely to lose significant tariff revenue from reducing tariffs on imports from the EU. For example, even assuming no trade diversion, an average country, in which tariff revenues are 2 per cent of GDP and where 40 per cent of imports come from the EU, would lose tariff revenues equivalent to close to 1 per cent of GDP (equivalent to 7 to 10 per cent of government revenues) from eliminating tariffs on all imports from the EU. The revenue losses would be significantly greater in countries that are highly dependent on tariff revenues. To protect their fiscal positions and maintain macroeconomic stability, the SSA countries will need to reform their indirect tax systems so that revenues from the VAT and non-discriminatory excise taxes levied at equal rates on imports and domestic products replace the forgone tariff revenues. Countries that face particularly large revenue losses may need to consider additional measures such as strengthening other components of their tax and revenue systems or curtailing low priority or inefficient expenditures. b. Problematic Partial Liberalisation Many SSA countries still have high and distorted MFN tariff structures, and preferential reduction of tariffs on merchandise imports from the EU under EPAs will need to be accompanied by MFN tariff reductions if the preferential reduction in tariffs is to be beneficial for SSA. Preferential tariff reductions under

8


EPAs can, in the presence of high MFN tariffs, lead to costly diversion of trade from low cost (non-EU) to high cost (EU) foreign suppliers, as well as transfers of tariff revenues from SSA governments to EU exporters.5 Note that the non-EU suppliers also include the SSA countries that are not members of a specific SSA regional bloc. Thus, preferential tariff reductions under EPAs in the presence of high MFN tariffs is likely to divert trade away from these countries by favouring imports from the EU. In addition, as a result of the lobbying of protected domestic SSA producers, the EPA negotiation process can also lead to a number of exceptions, with a tariff regime that maintains protective tariffs for all import-competing domestic industries while eliminating the revenues from tariffs on imports from the EU that do not compete with domestic production. Subsequent negotiation of similar such partial free trade agreements with additional OECD countries (as, for example, South Africa is currently doing with the US) would aggravate these problems. c. Simultaneous MFN Tariff Reductions The World Bank advocates a pro-development outcome of the Doha Round that would include developing countries’ achieving average tariffs of 5 per cent for manufacturing, with a maximum tariff of 10 per cent, and average tariffs of 10 per cent for agriculture, with a maximum of 15 per cent. In view of the uncertainty about the outcome of the Doha Round at this point, we suggest a maximum MFN tariff rate of no more than 20 per cent, the maximum rate currently in effect in UEMOA. The lower rates suggested by the World Bank as objectives for the Doha Round would be even better for those countries that can achieve them. Because of the likely losses of revenue from eliminating tariffs on imports from the EU, the required MFN tariff reductions would need to be carried out in a revenue-neutral fashion, for which there is ample scope in most SSA countries due to the prevalence of extensive tariff exemptions and of prohibitive (minimal revenue generating) peak tariffs. The MFN liberalisation should be completed before, or at the same time as, the preferential tariff reduction with the EU takes place so that trade diversion is minimised and EU suppliers do not have a chance to sell into highly protected domestic SSA markets and thus obtain large implicit transfers of tariff revenues until the MFN tariff is reduced. 5

In the case of homogeneous goods, as long as SSA countries continue to import some units from non-EU sources after liberalising their trade policy with respect to EU imports, domestic prices remain unchanged. Then, EU exporters can capture the tariff revenue previously collected on imports from the EU by raising their prices in SSA markets (which amounts to a worsening of SSA’s terms of trade), and SSA countries lose these tariff revenues to the EU exporters. In the case of heterogeneous goods, SSA domestic prices of imports from the EU will fall, resulting in greater imports from the EU. This will be accompanied by a reduction in imports from other sources and a loss of tariff revenues previously collected on these imports.


9

5. RELATIONSHIP OF EPAS TO THE WTO’S DOHA DEVELOPMENT ROUND

The EPA negotiations between the EU and the SSA countries overlap with the multilateral negotiations of WTO’s Doha Development Agenda, and there are several important interrelations between the two sets of negotiations. A prodevelopment outcome in WTO’s Doha Round would amplify the benefits from the EPA process. Many of the most difficult trade liberalisation issues, particularly agricultural ones, need to be addressed in the Doha Round. Hence, it would be desirable if the Doha Round could be successfully completed before key decisions have to be taken concerning the EPAs.6 Any reductions in the EU’s MFN tariff rates in the Doha Round will reduce the preference margin that SSA (and other ACP) countries will have under the trade EPAs, and some of these countries are worried about potential preference erosion. Long-term competitiveness and sustainable development cannot be built solely upon special preferences allowing SSA countries to benefit from transient distortions in world trade policies, and adjustment to preference erosion is going to be necessary sooner or later in any case. The EPA process may provide an opportunity to commence this adjustment in favourable circumstances as the financial cooperation with the EU could be used for providing adjustment assistance for countries facing significant transitional costs from preference erosion. The World Bank and IMF have also offered to help with financing any transitional costs of adjusting to preference erosion.

6. LIBERALISATION OF TRADE IN SERVICES AND RELATED REFORMS

Because of the underdeveloped nature of the export service sector in SSA and the constraints on expanding the employment of temporary workers in the EU, SSA’s major gains from liberalisation of trade in services are likely to come on the import side. Liberalisation of imports of services should be included as an integral part of the EPAs as these could have an important impact on productivity and growth in SSA. The priority sectors for liberalisation appear prima facie to be transportation, telecommunications and finance; and sub-sectoral studies of these need to be carried out for each sub-region in order to formulate plans for doing so. Imports of services should be liberalised on both an MFN and an intra-SSA regional bloc basis (including intra-bloc labour flows) at the same time as they are liberalised vis-à-vis the EU in order to attract investment by the most efficient service 6

Pascal Lamy has also recently proposed that the EU eliminate export subsidies on the condition that others, such as the US, follow suit. It is likely that what will happen in the Doha Round with this proposal will not be decided before the July 2004 deadline. If export subsidies are not eliminated in general in the Doha Round, they might still be eliminated later on for SSA.

10


providers and benefit from economies of scale. Accompanying reforms in the regulatory environment will also be needed, and the timing of the liberalisation in various service sectors should be determined by the capacity of the SSA countries to implement the required accompanying regulatory reforms in these sectors. The EU is not only interested in investment in the service sector but is likely to be interested in the FDI question in general as a component of the EPAs. Thus, even though the G-90 is against inclusion of investment as a topic of negotiation in the Doha Round, the EU claim is envisaged in the Cotonou Agreement, and SSA countries should be aware that the investment issue may come back in the EPA negotiations. On the export-side, the EU will need to take a more generous approach towards services in EPAs than in previous trade agreements with developing countries for the SSA countries to benefit significantly from improved access to the EU market. Even then, the gains in the near term from increased SSA exports of services are likely to be smaller than those from liberalising imports of services; but SSA should, nevertheless, explore what scope there is for increasing the numbers of temporary workers (Mode IV of the GATS) in the EU and for assisting the tourism sector through EPAs.

7. CONCLUSION: EPAS AS AN OPPORTUNITY TO ACCELERATE TRADE INTEGRATION IN SSA

Assuming the recommended policy changes are implemented, the EPA process can offer a favourable opportunity for SSA countries to integrate into the global economy, to strengthen regional integration in Africa, and to accelerate their reform programmes. The EPAs’ negotiating schedules and deadlines create a useful dynamic impetus for global and regional integration in a situation where progress would otherwise be halting, at best. Partial preferential liberalisation of trade between SSA and the EU under EPAs would, by itself, be problematic, possibly even disadvantageous on balance, for most SSA countries because of the transfer of tariff revenues to EU exporters and the probability of significant costly trade diversion. However, the necessity for the SSA countries to liberalise their imports from the EU in order to preserve and improve upon their preferential access to the EU market also provides an opportunity for them to accelerate the overall trade liberalisation process and the opening up of their economies. The fundamental condition for realising the potential benefits of the EPAs is to actually utilise the EPAs as instruments for development. Doing so will pose challenges for both the EU and SSA. For the EU, the challenges will be to subordinate its commercial interests in the EPAs to the development needs of the SSA countries and to effectively coordinate trade and development assistance. In particular, since much greater tariff reductions will have to be made under the

ECONOMIC AGREEMENTS: SUB-SAHARAN AFRICA & THE EU 11 EPAs by SSA countries than by the EU, the EU will need to include generous treatment of SSA exports in the trade components of EPAs in non-tariff areas such as liberalising its rules of origin, extending market access for everything but arms to the non-LDCs in SSA, eliminating agricultural export subsidies and decoupling agricultural production support in products of particular interest to SSA, and admitting larger numbers of temporary workers from SSA under EPAs. Adequate technical and financial support for the implementation of the EPAs also needs to be provided. In particular, financial assistance should, in principle, increase in real terms from the Cotonou level by at least the amount needed to finance the transitional costs associated with the EPAs and related reforms in the Doha Round. In addition, a pro-development outcome in WTO’s Doha Round would amplify the benefits from the EPA process, and it would be desirable if the Doha Round could be successfully completed before key decisions have to be taken concerning the EPAs. Because of its central role in the whole EPA process, if the EU cannot rise to these challenges, the EPAs may end up doing more harm than good for SSA, at least in the medium term until the SSA countries are able to implement the necessary reforms on their own. The reality is that very few SSA countries seem to want to reform. However, for those who do, the key requirement is a commitment to utilise the EPAs to accelerate reforms that are necessary in the long term for integrating with the global and regional economies; that is: the restructuring of indirect tax systems, MFN tariff reductions, liberalisation of services imports on an MFN basis and related regulatory reforms in the services sector, and liberalisation of trade in both goods and services (including labour flows) within the regional trade blocs in SSA. Moreover, there are many obstacles to expanding production of tradable goods in Africa besides those that will be addressed under the EPAs and accompanying reforms. The supply response to the EPAs reforms will be much greater if these trade reforms are followed by actions in other areas. Particularly important are (a) a competitive exchange rate policy that encourages the expansion and diversification of exports; (b) other trade facilitation measures such as reforms in customs administration; (c) improvements in the investment climate to encourage a positive response from the private sector; (d) implementation of an effective domestic competition policy; and (e) infrastructure investment. Finally, it should be noted that Pascal Lamy, the EU Trade Commissioner, made a proposal at the May 2004 G-90 summit in Dakar that might lead to a change in the EPA process. He proposed that the G-90, a group consisting of ACP and non-ACP LDC countries, should not have to make concessions at the WTO Doha Round of multilateral trade negotiations, i.e., he proposed a ‘free round’ for the G-90. This proposal opens the door to the possibility that the same might apply to the ACP countries in the EU-ACP negotiations and that the EPA process might be reversed.

12

LAWRENCE E. HINKLE AND MAURICE SCHIFF APPENDIX

Abbreviations ACP AGOA BLNS CAP CEMAC

African, Caribbean and Pacific countries Africa Growth and Opportunity Act Botswana, Lesotho, Namibia and Swaziland Common Agricultural Policy Communauté Economique et Monétaire de l’Afrique Centrale (Central African Economic and Monetary Community or CAECM) CET Common External Tariff COMESA Common Market for Eastern and Southern Africa CU Customs Union EAC East African Community EBA Everything-but-Arms EC European Commission ECOWAS Economic Community of West African States EPA Economic Partnership Agreement ESA Eastern and Southern Africa EU European Union FTA Free Trade Agreement GATS General Agreement on Trade in Services GSP Generalised System of Preferences LDC Least Developed Country MFN Most Favoured Nation PTA Preferential Trade Agreement RoO Rules of Origin RTA Regional Trade Agreement SACU Southern African Customs Union SADC Southern African Development Community SPS Sanitary and Phyto-Sanitary SSA Sub-Saharan Africa UEMOA Union Economique et Monétaire Ouest-africaine (West African Economic and Monetary Union or WAEMU) WTO World Trade Organisation

ECONOMIC AGREEMENTS: SUB-SAHARAN AFRICA & THE EU 13 REFERENCES Brenton, P. (2003), ‘Integrating the Least Developed Countries into the World Trading System: The Current Impact of EU Preferences under Everything But Arms’, World Bank Policy Research Working Paper No. 3018 (World Bank. Washington, DC, April). Hinkle, L. and M. Schiff (2003), ‘Economic Partnership Agreements between Sub-Saharan Africa and the EU: A Development Perspective on its Trade Component’ (Mimeo, Africa Region, World Bank).

WHAT ROLE FOR THE EPAS IN THE CARIBBEAN?

15

2

What Role for the EPAs in the Caribbean? Michael Gasiorek and L. Alan Winters

1. INTRODUCTION

REFERENTIAL trading relations between EU members and the Caribbean have a long history stretching from previous colonial regimes, through successive rounds of the Lomé Convention starting in 1975, to the Cotonou Agreement, signed in 2000. The main feature of the Lomé agreements between the EU and a diverse group of 77 countries from the African, Caribbean and Pacific (ACP) was non-reciprocal duty-free access to the EU for most ACP exports. However, non-reciprocal arrangements are WTO incompatible, so, under the Cotonou Agreement of 2000, the parties agreed to negotiate WTO-compatible European Partnership Agreements (EPAs) by 2008. In comparison with the preceding arrangements two key features of the EPAs stand out: first, and in order to ensure WTO compatibility, the agreements will be more reciprocal in nature, and second, the agreements will not be between the EU and the ACP states as a group but between the EU and a number of different regional groupings. In considering a possible EPA between the EU and the Caribbean region, three initial conditions in the Caribbean region are significant. First, the Caribbean is more distant from Europe than are many African countries, and closer to America than any other ACP countries. Second, and relatedly, for many Caribbean economies the USA is a more important trading partner than the EU. This is unique within the ACP group and has important implications for the economics of the EPA. Third, Caribbean countries are small or very small both in terms of their

P

The authors are grateful to Caglar Ozden for advice, to Audrey Kitson-Walters for logistical help, and to Phai Tam Hung for excellent research assistantship, but exempt them from responsibility for the paper’s shortcomings. The findings, interpretations and conclusions expressed in this paper are entirely those of the authors and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent.

15

16

MICHAEL GASIOREK AND L. ALAN WINTERS

physical size and their GDP. Their economies are typically highly specialised in production and provide only very small markets. They are vulnerable to climatic and economic shocks, and typically have limited institutional and financial capacity to deal with the processes of change and transition. This paper considers the possibilities for an EPA between the EU and the Caribbean – given the stated objectives of the EPAs and the particularities of the Caribbean region. The objective of the EPAs are those of ‘. . . reducing and eventually eradicating poverty consistent with the objectives of sustainable development and the gradual integration of the ACP countries into the world economy’.1 Specifically, it is intended that the EPAs should encompass the liberalisation of trade in goods and services; foster and promote greater regional integration, in particular among small island and landlocked economies; provide development finance and cooperation; assist the development of appropriate institutional capacities, and promote simple and transparent rules for business. Throughout the Cotonou Agreement there is reference to the need for flexibility to allow for the particular circumstances of different ACP economies, and also to the need for special and differential treatment when circumstances require. The question we pose is ‘what form of EPA might be pro-development and pro-poor for the Caribbean region?’ And, perhaps more fundamentally ‘to what extent is it in the interests of the Caribbean countries to sign an EPA?’ Our primary focus is on the trade-related aspects of the EPA, but we also discuss the broader issues inherent in the EPA negotiations, and the linkages between these and the trade-related aspects. In the next section we identify the key features of the proposed EPA agreements. In Section 3 we discuss the Caribbean economies focusing on those aspects that we consider to be central to evaluating the EPAs. In Section 4 we analyse the likely effects of the EPA on Caribbean trade performance and hence, indirectly, on development and poverty eradication.

2. THE EPAS

The negotiation of the EPA agreements was originally intended to have two phases (ACP-EC EPA Negotiations, 2003). In the first, negotiations between the EU and the whole bloc of ACP countries were to focus on six thematic areas of common ACP interest. In the second, and on the basis of an agreement on the six thematic areas, negotiations were to take place at the regional level, where the composition of the regional groupings was to be determined by the ACP states themselves. In practice, agreement on the first phase has not been reached and the EU now requires that first and second phase negotiations should continue in

1

Article 1, Partnership Agreement, Official Journal of the European Communities, 15.12.2000.


17

parallel. The ACP countries would prefer to have agreement first on the crosscutting issues, but in practice it appears that a pragmatic approach is being taken with parallel negotiations taking place. In the first phase of negotiations the principle was established that the focus of the EPAs should be on promoting growth, integration of the ACP states into the world economy and the eradication of poverty. This was to be achieved through enabling appropriate structural transformations, increasing supply and production capacities, and via regional integration. While these are laudable general objectives, agreement on the precise modalities, coverage and transitional arrangements has been hard to achieve. The six thematic areas identified for the first phase were: 1. Market access: This includes negotiations over rules of origin, customs procedures, trade facilitation, safeguards, WTO compatibility, product coverage and discussions over the nature of any transitional arrangements. While both sides have accepted the need for flexibility and asymmetry in terms of timing, product coverage, and possibly also rules of origin, the actual length of the transition periods and the degree of product coverage are still under discussion. On the latter, the EU would like 90+ per cent coverage and the ACPs less, while on the former, the ACP states have suggested a five-year moratorium on liberalising their own tariffs on conclusion of the EPAs, which the EU has not accepted. 2. Agriculture, including fisheries: There is agreement on the importance of agriculture, of export diversification on the side of the ACP states and of tackling SPS issues in the negotiations. However, there is a lack of consensus on sequencing. The EU would like negotiations on agricultural trade liberalisation and on the concomitant provision of support for agriculture in the ACP countries to proceed in parallel, whereas the ACP states think assistance should precede negotiations. 3. Trade in services: While services are mentioned in the Cotonou agreement, there is no obligation for liberalisation of trade in this sector. In principle however, both sides have agreed to include services liberalisation in the EPAs, based on the positive list approach and taking into account the specific circumstances of countries/regional groupings. The ACP countries have requested more access to the EU market under Mode 4 of the GATS (temporary movement of workers), which the EU has agreed to consider. 4. Trade-related issues: Both sides recognise the importance of non-tariff issues (competition policy, IPRs, technical regulations and standards, trade and the environment, and trade and labour standards) but differ with regard to matters of both coverage and sequencing. For the ACP states there appear to be two prior conditions: first, the need to improve legal and

18

MICHAEL GASIOREK AND L. ALAN WINTERS institutional capacities before embarking on negotiations, and second the need to reach multilateral (WTO) agreement prior to EPA-based negotiations. The EU’s position is that these issues are already in the Cotonou Agreement, and that capacity building and negotiations should proceed simultaneously. 5. Development cooperation: The need to support capacity building and infrastructure development is recognised by both sides. The difference is over the levels of support. 6. Other/Legal issues: e.g. dispute settlement. This includes a range of other issues such as dispute settlement, the legal status of the Agreements, as well as institutional matters.

The second phase of negotiations with individual countries/country groupings is now getting under way in a serious fashion. In this phase commitments can differ across EPAs. Hence the relevance of a paper focused just on one region.

3. THE CARIBBEAN ECONOMIES

The Caribbean region contains both independent countries and nonindependent states. Of these, the ACP and Cotonou agreements include 16 Caribbean economies. Within this grouping there are two principal regional trading arrangements in force. First, CARICOM, which comprises 15, largely English-speaking, countries. Originally established in 1973, CARICOM provides for free trade between the member states. Within CARICOM there is the Organisation of Eastern Caribbean States (OECS), established in 1981 by the Treaty of Basseterre. Here the main objective was on coordinating and harmonising OECS states’ external relations, and secondarily on increasing ‘functional cooperation’ on a range of other economic, social and political issues.2 The latter objective has, in principle at least, been significantly advanced since 2001 with the agreement to form the OECS Economic Union. In addition there is the Association of Caribbean States established in 1995 with the aim of promoting cooperation and integration between a wider grouping of 25 Caribbean states.3

2

CARICOM comprises: Antigua and Barbuda, Dominica, Grenada, Montserrat, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Bahamas, Barbados, Belize, Guyana, Haiti, Jamaica, Suriname, and Trinidad and Tobago. In addition Anguilla, British Virgin Islands, Bermuda, Cayman Islands, Turks and Caicos Islands are associate members. The countries in italics are members of the OECS. 3 The ACS comprises CARICOM less Montserrat plus Colombia, Costa Rica, Cuba, Dominican Republic, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama and Venezuela.


19

a. Economic Structure and Performance There is considerable diversity in economic performance amongst the Caribbean economies (see Table 1). The largest in terms of GDP are Trinidad and Tobago, Jamaica, Bahamas and Haiti, and in terms of population Haiti and Jamaica. Looking at GDP per capita the poorest Caribbean economy by far is Haiti ($430) followed by Guyana ($921), with per capita income levels being the highest in the Bahamas ($16,498) and Antigua and Barbuda ($10,618). Five economies (Barbados, Dominica, Grenada, Haiti and St. Lucia) experienced negative growth in 2001 and in 2002. For many of the economies rates of inflation are comparatively low, but not for Suriname, Haiti or Jamaica. All of the economies except Haiti and Suriname are pretty open to international trade, with export/GDP ratios of 40 to 90 per cent. Import/GDP ratios are generally higher, at least in part because these economies receive flows of remittances from citizens working abroad. Haiti’s low openness ratio presumably just reflects its generally poor performance in terms of security, governance and infrastructural investment. An interesting feature to which we will return is the importance of services in total export earnings for many of the countries. Turning to structural issues, Table 2 details the breakdown of economic activity, while Table 3 considers the composition of goods trade for sub-sets of Caribbean economies. A common feature in Table 2 is the large role of services, which provides over 50 per cent of GDP for all but Guyana and Grenada, and is around 80 per cent for Antigua and Barbuda and Barbados. Service activities encompass data processing, banking, and, of course, tourism which is a key industry for most of these economies. It is perhaps worth noting that while tourism is clearly an area of comparative advantage, it is also an industry which can experience sharp fluctuations. Tourism receipts dropped significantly after the events of September 2001, and natural disasters such as volcanic activity in Montserrat or hurricanes can similarly have a devastating impact. Guyana has the highest share of agriculture in GDP (46 per cent), but for all the other economies this sector comprises less than 25 per cent and for seven less than 10 per cent. For only three countries is agriculture the second most important sector in terms of its contribution to GDP. However, while agriculture may not be the most significant sector in terms of economic activity, it is nevertheless a key sector in terms of exports, especially for some of the small Caribbean economies; see Table 3. The main export products include bananas (e.g. for St. Lucia, St. Vincent and the Grenadines, Dominica and Grenada), and beverages and tobacco (St. Kitts and Nevis). For none of the Caribbean economies does manufacturing constitute more than 20 per cent of GDP, and for eight of the economies it is less than 10 per cent. For only one economy (Jamaica) is manufacturing the second most important sector. The structure of manufacturing is fairly similar across the region, with significant

1893 7346

0.866 9.372 2.7 2.7

2.1 0.7 −1.8 3.5 −3.6 −0.5 1.1 −0.9 1.5 0.8 −0.5 1.1

GDP Growth (Per cent, constant prices)

28.3 4.2

2.2 2.0 1.5 1.2 −0.3 3.0 5.3 8.7 6.5 2.1 0.9 1.0

Inflation (Per cent, annual change)

419 1300

65 308 268 231 71 94 763 8511 2598 38 149 114

Population (’000) 2001

Sources: Population: ECLAC, Statistical Yearbook for Latin America and the Caribbean 2002: Macroeconomics: International Monetary Fund, World Economic Outlook Database, September 2003 at http://www.imf.org/external/pubs/ft/weo/2003/02/data/index.htm Trade data: World Development Indicators, World Bank.

Note: * Haiti’s trade data refer to 2001.

10681 16498 9346 3308 3538 4060 921 430 2914 7270 3815 3219

0.721 5.058 2.443 0.838 0.254 0.414 0.709 3.553 7.701 0.356 0.660 0.361

Antigua and Barbuda Bahamas, The Barbados Belize Dominica Grenada Guyana Haiti* Jamaica St. Kitts and Nevis St. Lucia St. Vincent & the Grenadines Suriname Trinidad and Tobago

GDP Per Capita ($, current prices)

GDP (billion $, current prices)

21 48

60 n.a. 52 52 54 47 93 9 39 46 53 48

Exports as Per Cent of GDP

TABLE 1 The Caribbean Economies 2002: An Overview

9 13

91 n.a. 81 39 65 69 26 47 59 61 80 75

Services as Per Cent of Total Exports

45 44

68 n.a. 55 66 62 57 106 36 60 71 57 59

Imports as Per Cent of GDP

25 17

28 n.a. 33 24 32 25 31 20 32 27 25 25

Services as Per Cent of Total Imports

20 MICHAEL GASIOREK AND L. ALAN WINTERS


21

TABLE 2 Breakdown of Economic Activity by Sector Country

Antigua & Barbuda Bahamas Barbados Belize Dominica Grenada Guyana Haiti Jamaica St. Kitts & Nevis St. Vincent & Grenada St. Lucia Suriname Trinidad & Tobago

Agriculture

Manuf.

Construction

Services

Mining & Quarrying

1995

2001

1995

2001

1995

2001

1995

2001

1995

2001

3.88 3.20 5.76 20.55 17.69 10.13 48.79 23.56 9.05 5.33 14.09 12.31 15.96 1.89

3.50 n.a. 5.44 24.13 14.64 7.44 46.21 18.86 7.51 4.55 10.07 6.62 n.a. 1.27

2.31 3.50 6.64 13.74 6.33 6.59 3.85 19.19 16.86 10.78 8.37 8.88 15.43 16.83

2.32 n.a. 5.54 13.14 6.50 7.72 3.24 16.93 15.09 10.78 6.14 7.81 n.a. 22.55

10.36 n.a. 4.74 6.54 7.99 8.09 4.19 9.72 12.41 12.17 11.26 9.69 2.46 9.40

13.34 n.a. 6.06 6.80 7.49 9.09 4.71 11.18 10.62 17.31 10.36 10.41 n.a. 11.12

81.80 n.a. 82.29 58.54 67.19 74.71 26.12 47.39 54.82 71.37 65.94 68.52 56.99 56.90

79.09 n.a. 82.34 55.31 70.64 74.92 26.51 52.91 58.90 66.99 73.19 74.52 n.a. 54.51

1.66 n.a. 0.57 0.63 0.80 0.48 17.05 0.14 6.86 0.34 0.33 0.61 9.16 14.98

1.74 n.a. 0.62 0.63 0.73 0.84 19.33 0.12 7.88 0.37 0.24 0.64 n.a. 10.55

Source: Data on Anguilla from the Anguilla Statistic Office, National Accounts Publications 2002, and 2001. Shares are calculated from GDP data at constant 1990 prices; Data on Bahamas and Montserrat from the CARICOM Secretariat, CARICOM Selected Economic Indicator 1985, 1995–1999. Shares calculated from GDP data at current market prices; Data on other countries from the ECLAC, Statistical Yearbook for Latin America and the Caribbean 2002. OECS economies in italics.

shares in food processing, clothing and textiles, light manufacturing, cement and cement products. Food processing contributes importantly to exports of SITC 0 (Food and Live Animals) in Table 3, but otherwise no clear pattern is obvious in manufactures: for several countries – mostly the smallest – one of the manufacturing sectors provides an important share of exports, e.g. chemicals in the Bahamas and in Dominica, and machinery and transport equipment, which comprised 70 per cent of St. Kitts’ exports in 2001. While the share of construction in GDP is always less than 20 per cent, for seven of the economies this is the second most important sector. This again suggests a good deal of volatility in these economies. For all but four of the economies the share of mining and quarrying is less than 2 per cent, and in 2001, is most significant for Guyana (19.33 per cent), and Trinidad and Tobago (10.55 per cent). Finally, if we compare 1995 to 2001 in Table 2, there is some evidence of structural change. For each of the economies the share of agriculture in GDP has declined, and for seven of the economies the principal sector of expansion was construction, and for five of the economies, services. For only one economy was there a significant expansion of manufacturing (Trinidad and Tobago). For all the economies imports are much more diversified than exports. There are significant food imports into most of the economies, and a clear requirement

36.38 87.23 30.08 38.74 35.09 42.58 17.92 23.17 58.54 0.62 13.44 3.57 83.29

Exports Bahamas Belize Barbados Dominica Grenada Guyana Jamaica St. Kitts & Nevis St. Lucia Montserrat Suriname Trinidad & Tobago St. Vincent

Source: World Bank WITS database.

13.92 12.71 12.06 16.73 16.17 12.07 14.03 15.57 20.01 15.05 13.37 7.29 24.10

Imports Bahamas Belize Barbados Dominica Grenada Guyana Jamaica St. Kitts & Nevis St. Lucia Montserrat Suriname Trinidad & Tobago St. Vincent

Food & Live Animals

13.05 0.00 7.18 0.67 0.31 0.73 4.72 3.60 20.81 0.01 0.05 2.00 3.44

2.61 0.85 2.50 2.64 2.01 1.39 0.87 3.00 3.40 5.82 3.08 0.57 2.05 6.88 1.45 0.38 4.30 0.04 18.23 58.77 0.04 0.40 0.02 67.06 0.16 0.30

3.04 0.86 3.29 2.36 2.80 0.83 1.62 2.73 2.96 3.82 1.42 2.42 2.87 0.00 0.00 8.62 0.00 0.00 0.00 0.29 0.00 0.01 0.00 5.17 65.28 0.00

10.36 16.95 11.52 9.58 8.76 22.28 18.37 7.60 9.35 11.57 6.71 32.30 9.63 0.00 0.00 0.56 0.00 0.00 0.19 0.01 0.00 0.04 0.00 0.00 0.08 0.04

0.19 0.44 0.37 1.67 0.21 0.40 0.61 0.46 0.16 0.21 1.22 0.31 0.38 42.66 0.13 12.52 54.21 1.36 0.76 5.30 0.07 1.19 0.27 0.63 17.33 2.06

9.15 10.34 9.71 12.26 6.57 11.43 10.66 7.51 8.22 6.30 10.62 7.88 10.05 0.52 0.19 16.06 0.95 3.74 5.99 0.53 0.78 3.33 4.96 0.36 8.89 7.80

18.53 17.97 17.25 17.38 16.90 16.27 13.69 19.53 17.70 17.57 17.27 13.24 19.05 0.06 0.17 12.77 0.09 58.24 1.57 0.08 69.54 5.93 50.08 0.65 1.19 1.74

26.30 28.72 27.82 26.19 32.90 27.04 23.35 28.06 24.55 30.58 36.30 30.59 18.96

0.45 10.83 10.37 1.03 1.22 3.32 12.38 2.80 9.74 43.99 0.19 1.48 1.35

14.10 10.73 15.26 11.20 13.67 8.17 14.06 15.54 13.66 8.25 9.34 5.12 12.90

0.00 0.01 1.44 0.00 0.00 26.62 0.00 0.00 0.00 0.04 12.44 0.02 0.00

1.81 0.45 0.22 0.00 0.01 0.13 2.74 0.00 0.01 0.84 0.67 0.28 0.00

Beverages & Crude Mineral Animal/ Chemical Manufactured Machinery/ Misc. Commodities Tobacco Materials Fuels Veg. Oils Products Goods Transport Manufactures n.e.s. Equipment

TABLE 3 Share of Trade by Product



23

to import finished manufactures and machinery and equipment. There is also some evidence of imported inputs – e.g. petroleum products into Trinidad and Tobago and the general significance of SITC 6 (Manufactured Goods). b. Caribbean Trading Arrangements Three sets of trading arrangements and trading ‘partners’ are significant for the Caribbean economies: the EU, the US and the Caribbean region itself. There are, of course, agreements with other countries as well – for example, the region is part of the Free Trade Area of the Americas process, and has also signed agreements with Canada, Venezuela, Colombia, Dominican Republic and Cuba – but these are smaller than the three listed. Table 4 details the share of trade by geographic origin for 14 Caribbean economies. The principal trading partner, by far, for each of the economies (except Suriname) in terms of imports is the US, almost always followed by the other CARICOM countries. The share of the EU in imports is around 13 per cent. In contrast, the EU is a much more important destination market for these OECS economies’ exports. It is the principal export market for Suriname and St. Lucia, and accounts for over 20 per cent for 11 of the 14 economies. The US and the EU together account for the majority of OECD24 trade with the Caribbean countries, and OECD and the Caribbean almost always for a very high proportion of trade. The share of trade, both imports and exports, with Mercosur, the rest of Latin America, as well as with Central America is almost always negligible – and hence not included here. The overall picture therefore is one in which the majority of both imports and exports are with the other countries of CARICOM, the US and the EU, and in which the US is clearly relatively more important on the imports side, whereas the EU’s significance tends to be on the export side and in particular for several of the small OECS economies (Dominica, Grenada, St. Lucia and St. Vincent and the Grenadines). We have already discussed trading relations with the EU. Trading relations with the US are dominated by the Caribbean Basin Initiative (CBI), which refers to the successive United States Acts designed to provide preferential access to the US market as well as various forms of technical and institutional assistance for the Caribbean economies. These include the Caribbean Basin Economic Recovery Act (1983), the Caribbean Basin Economic Recovery Expansion Act (1990), and more recently the US-Caribbean Trade Partnership Act (2000). The latter was designed to address some of the asymmetries of market access between the Caribbean and Mexico arising from NAFTA, in particular in sensitive sectors such as textiles and clothing. The key feature of the CBI is the non-reciprocal preferences under which the 24 Central American and Caribbean economies can export a range of products duty-free to the US. It is thus analogous to the US Generalised System of

Antigua and Barbuda Bahamas, The Barbados Belize Dominica Grenada Guyana Jamaica Montserrat St. Kitts and Nevis St. Lucia St. Vincent and the Grenadines Suriname Trinidad and Tobago

1999 2001 2001 2001 2001 2001 2001 2000 2001 2001 2001 2001 2000 2001

Year

12.58 0.87 19.67 4.05 27.86 25.16 17.98 12.36 17.65 17.86 24.20 27.35 20.30 2.74 10.02 2.07 16.67 8.01 15.67 11.78 12.07 8.33 10.22 12.67 15.89 21.22 28.96 17.91 73.51 88.24 68.81 57.95 61.95 67.27 56.70 65.52 79.36 77.36 66.46 63.76 65.33 61.08 49.46 83.30 42.07 46.07 36.45 46.89 37.73 45.46 59.97 50.52 41.76 35.31 26.58 34.43 76.61 0.18 47.95 6.62 59.66 17.45 13.57 3.30 60.43 2.85 29.55 60.18 5.87 19.54

CARICOM Per Cent

USA Per Cent

CARICOM Per Cent OECD24 Per Cent

Exports

Imports EEC15 Per Cent

TABLE 4 Share of Trade by Source

4.94 27.81 20.47 24.94 29.27 38.32 24.97 30.12 7.55 22.82 54.04 35.62 32.20 8.61

EEC15 Per Cent

11.48 98.77 43.39 83.81 35.23 66.81 80.65 92.07 34.88 95.58 69.08 37.06 90.32 58.06

OECD24 Per Cent

6.05 68.04 18.29 50.63 5.51 40.10 32.97 39.40 27.33 72.69 14.59 1.16 21.08 42.33

USA Per Cent



25

Preferences, which is also open to the Caribbean economies. The main differences lie in the CBI’s wider product coverage, the applicable rules of origin, and mechanisms for technical assistance or trade and investment financing. In practice, exports under the CBI appear to have been dominated by a few countries (the Dominican Republic, Costa Rica, Honduras, Guatemala, Haiti and Jamaica). Moreover, as is being increasingly recognised for GSP schemes, the true degree of preferential access is small because sensitive products are frequently excluded, and complex quota systems, rules of origin or certification requirements make it difficult for developing countries to prove that they satisfy the requisite criteria (e.g. Brenton and Machin, 2002). Hence, when in March 2004 the US government announced that Antigua and Barbuda was now too affluent to qualify for US GSP preferences, the government of Antigua and Barbuda responded that this would make little difference as 90 per cent of their exports to the US entered without preferences, and preferential exports to the US amounted to only 0.6 per cent of total exports.4 In terms of regional integration, while CARICOM has been established over thirty years, comprehensive integration has proved hard to achieve. Hence, even though in principle there is a common external tariff, in practice there are many national exemptions and variation of tariff rates across countries. In 1992 a simplified CET structure was agreed and plans laid for a reduction in the CET (to maxima of 40 per cent for agricultural products, and 20 per cent for manufactures) over five years, but actual implementation is taking somewhat longer (WTO, 2001 and 2002). In 1989 it was agreed to pursue Caribbean regional integration through both deepening and widening à la European Union. Deepening was to be achieved through the creation of a CARICOM Single Market and Economy (CSME), by eliminating tariff and quota barriers between countries and the removal of a range of non-tariff barriers. There have also been discussions on the development of a single currency, and for mechanisms allowing for the pooling of foreign exchange reserves.5 On widening, the discussion was of extending membership to have non-Anglophone countries and to the whole group engaging in broader arrangements such as the FTAA (Nicholls et al., 2000). In addition to CARICOM, the OECS also proposed deeper integration between its members. In 2001 it agreed to create the OECS Economic Union within a two-year period. This was in recognition that in order to achieve greater regional integration the OECS needed to move away from a model of integration based on inter-governmental cooperation, towards a model based more on relevant supranational bodies (OECS 2003 and 2004). The objective was to create a

4 5

Government of Antigua and Barbuda, Ministry of Foreign Affairs, Press Release, 5 March, 2004. See Nicholls et al. (2000) for a more detailed discussion.

26

MICHAEL GASIOREK AND L. ALAN WINTERS single economic space which facilitates the free movement of people, goods, services and capital and as a result economic diversification and growth, greater export competitiveness and more employment and human resource development (OECS website, http://www.oecs.org/ union.htm).

In reality progress has been much slower and negotiations on a range of ‘single market’ issues are on-going. The overall picture is thus one of a group of small island economies, most of which are specialised in services (banking and tourism) and for whom neither agriculture nor manufacturing are the principal sectors of economic activity. Rates of per capita income growth have been disappointing and many of these economies experience high rates of unemployment. The economies currently have preferential market access to the EU, the US and Canada, and are engaged in various processes of intra-regional integration (CARICOM, OECS, ACS). It is to this cocktail that the EPAs will be added.

4. WHAT ROLE FOR THE EPAS?

This section seeks to clothe the discussion above in some numerical estimates. We can quantify the effects of only a small part of what the EPA is intended to achieve, but the results raise some serious issues. As outlined in Section 1, the EPAs are concerned with issues of trade liberalisation and market access in manufacturing, agriculture and services, with both traditional and deeper forms of integration and also with development assistance. The relationship between the EPAs and regional integration between ACP states also has an explicit objective, where the aim is to use the EPAs to encourage greater regional integration, and to focus ‘first on building larger ACP markets and only then on opening those to EU products and services’.6 The objectives here are to help create larger markets (through regional integration), and through this to expand the productive capacity of the ACP economies and to encourage diversification in the ACP economies. Merely dismantling the Lomé Convention’s special protocols on rum, beef and veal, sugar and bananas – which seems likely eventually – will help the latter objective. A further important feature of the EPAs is captured by the notion of ‘development-oriented reciprocity’, whereby, to the extent that WTO rules allow, the EPAs are non-reciprocal in terms of coverage, timing and rules (e.g. of origin). Viewed purely as a regional trade agreement the key aspects of the EPAs are the changes in market access on both the import and export side that they imply. On exports, the Caribbean economies have long enjoyed non-reciprocal trade 6

Joint Report: ACP-EU Negotiations on Economic Partnership Agreements, 3rd all-ACP-EU Ambassadorial level meeting, February 2003.


27

preferences with the EU, the United States and Canada. The re-negotiation with the EU requires that either they sign up for an EPA (with reciprocal free trade) or that they adhere to the EU’s existing GSP concessions.7 The EU has agreed that any country signing an EPA will be given the same level of market access as is currently offered under the Everything-but-Arms (EBA) initiative.8 Particularly if this is accompanied by improved access for agricultural exports, or for natural persons under GATS Mode 4, this is likely to be better than non-EPA countries would get. However, since we do not know the parameters of such benefits, we cannot quantify them here. On the import side the net benefits of an EPA can be assessed in terms of the traditional balance of trade creation and trade diversion relative to an anti-monde of no change in import policy. There are certainly hopes of both wider and deeper integration, extending to services and a variety of regulatory matters. Deep integration with the EU and locally, and the creation of a larger local market, are likely to be important for economic performance. But they are sufficiently uncertain or imprecisely defined at present that we are obliged to confine our quantitative analysis to an analysis of the import regime for goods, for which we have both data and a clear policy outcome to analyse. a. Conceptual Framework for Imports To analyse the possible impact of an EPA on Caribbean import behaviour and welfare it is sufficient to distinguish four sources of supply – the EU, the US, the region itself and the (usually minor) rest of the world. Both Winters (2001) and Greenaway and Milner (2003) observe that the outcome of any analysis will differ according to whether we assume that these sources provide a homogeneous good or competing differentiated products. Greenaway and Milner assume the differentiated good model in their quantitative analysis using assumed values of the elasticities of substitution between varieties. So too do the various Computable General Equilibrium modelling analyses of the EPAs (e.g., Rutherford and Martinez, 2000). Here we take the opposite tack, following Winters (2001), and assume homogeneous goods. We do not hold this to be literally true, but feel that it is an appropriate assumption for highlighting the key issues. Despite using quite different methodologies, these previous studies both conclude that multilateral liberalisation is likely to be significantly more welfare enhancing for the Caribbean region than any plausible EPA. Central to these conclusions are the likelihoods of trade creation and trade diversion under the different arrangements. 7

Only one of the Caribbean economies (Haiti) would currently qualify for EBA preferences. Although Brenton and Machin (2002) suggest that, similarly to the operation of GSP schemes, the utilisation of EBA preferences is low. 8

28


To illustrate these we discuss a number of possible scenarios, with different implications for trade and welfare in the Caribbean. In the discussion below we focus on the import of goods by a target Caribbean economy, which we assume to have no domestic production. While this last assumption is clearly not universally true, both the share of imports in GDP and the pattern of imports across sectors suggest that it is reasonably accurate for a large number of products. For simplicity, in the graphical analysis below we distinguish two suppliers – the EU and the rest of the world (ROW), which contains all the countries for which tariffs into Caribbean markets are unchanged, including both those for which tariffs are zero (other Caribbean countries) and those facing positive tariffs. The welfare effects of the EPA depend on which group is the principal supplier, the elasticities of supply and demand, the height of the tariff and on the extent to which tariff reductions translate into price reduction. Five different possible scenarios should be identified:9 (i) A sole supplier to the domestic market: 1. Suppose the EU is the sole supplier of a given good (Figure 1a): If the EU supply curve is horizontal the initial quantity supplied would be q1. Assuming the tariff reductions translate into price reductions this would result in a new quantity supplied at a lower price, and hence pure trade creation and a net welfare gain. The amount of trade creation is given by shaded triangle ‘b’ in the figure. If the EU supply curve to that target economy is upward sloping (Figure 1b) there will be both trade creation and a terms of trade loss as the increase in domestic demand pushes up the price of the imported good. In this case the welfare impact is ambiguous. Trade creation is once again given by the shaded triangle, while the terms of trade loss is given by the dotted rectangle, ‘c’. The flatter is the EU supply schedule the more likely it is that trade creation will dominate. If the tariff reductions do not translate into price changes, as could arise, for example, where the export market is imperfectly competitive, there would be a welfare loss arising from the loss of tariff revenue – the rectangle ‘a’, in Figure 1a or areas ‘a + c’ in Figure 1b.10 2. Alternatively suppose that the rest of the world (ROW) is the initial sole supplier with a horizontal supply curve (Figure 2): If SROW represents the US tariff-inclusive supply, and if the reduction in tariffs on EU imports results in the replacing of ROW imports by EU imports (SEU), then there will be both trade creation and trade diversion as given by the shaded 9

Panagariya (2001) provides a broader menu of possibilities. A monopoly supplier of imports would drop his price slightly because his costs would be reduced by his tariff exemption. 10


29

FIGURE 1a

FIGURE 1b

FIGURE 2

triangle (‘b’) and dotted rectangle (‘c’) respectively. If there is simply a switch in import supply and no change in the domestic price then there would be pure trade diversion (‘a + c’) represented by the consequent loss of tariff revenue. It is thus the extent to which the EU tariff-free price is below the ROW tariff-inclusive price that determines trade creation, which in turn depends on the cost differences between the EU and the ROW, the initial height of the tariff, and the nature of competitive interactions in the market in question.

30

MICHAEL GASIOREK AND L. ALAN WINTERS If, on the other hand, SROW represents CARICOM exports, which face no tariff in the target economy, there is now a redirection of imports towards a more efficient supplier, in this case the EU. This undoes previous trade diversion arising from CARICOM preferences which we refer to as trade reorientation.11 In this case there is a welfare gain corresponding to areas ‘a + b’.12 If the EU supply curve is upward sloping then the same effects as outlined above will be present, but there will also be negative terms of trade effects. (ii) Both the EU and ROW supply the domestic market 3. Suppose the ROW supplied the market with an upward sloping supply curve and the EU supply curve is horizontal (Figure 3). In the initial situation q1ROW is supplied by the rest of the world, and q1ROW − q1EU by the EU. In this case the marginal import price is determined by EU suppliers. Any reduction in this arising from tariff reductions would result in a welfare gain either from trade creation or from trade reorientation, although there may also be losses from trade diversion. The net outcome will depend on whether the ROW supplier is subject to tariffs (the US), or not (CARICOM). In the former case we have losses because the pre-tariff US price was below the pre-tariff EU price for the displaced trade (q1ROW − q2ROW); the welfare loss from that trade diversion is the lost revenue (b + c). However, there are consumer surplus gains from the lower price (area ‘a’), as well as trade creation gains (area ‘e’), hence the net welfare gain is ambiguous. If FIGURE 3

11

Following Gasiorek et al. (2002), we use trade reorientation to mean the reverse of trade diversion. Hence it is where the supply of imports switches from a less efficient supplier (in this case CARICOM) to a more efficient supplier (the EU). Greenaway and Milner refer to this as ‘source substitution’. 12 Of course, in this case it is a producer elsewhere in the Caribbean who loses the market, and if he was making any surplus on these sales, his welfare would decline – see Chang and Winters (2001).


31

SROW emanates from the Caribbean and so includes no tariffs, we have trade reorientation with gains of ‘a + b + c’ as the price paid for the first q1(ROW) of import falls. Area ‘a + b’, however, was producer surplus accruing to another Caribbean producer, so the net gain to the region is area ‘c’ as well as trade creation gains represented by area ‘e’. Again, if the EU producers did not pass the tariff changes through to consumers, there would be a welfare loss from foregone tariff revenue equal to area ‘d’. 4. Suppose now that it is the EU supply curve which is upward sloping, while the ROW supply schedule is horizontal (Figure 4a). Now the marginal import price is fixed by world markets, and tariff liberalisation with respect to the EU will not change the import price but simply switch imports away from the rest of the world to the EU. When the ROW supply is subject to tariffs (the US) this results in the loss of tariff revenue (‘a + b + c + d’) of which ‘a + b’ is foregone on imports that initially came from the EU and (c + d) is from trade diversion. Where the ROW supply is tariff-free (i.e. from CARICOM) the revenue loss is just ‘a + b’, for imports (q2EU − q1EU) did not pay tariffs in the first place. A variant of the preceding is where, following FIGURE 4a

FIGURE 4b

32

MICHAEL GASIOREK AND L. ALAN WINTERS the tariff reduction, the EU can supply the entire domestic market, and ROW imports are eliminated (Figure 4b). As well as the previous effects there is now some trade creation (‘e’), which depends on the extent to which the new domestic price is below the initial price. Case 4 is the one used by Winters (2001). 5. Both the EU and ROW have upward sloping supply curves and share the market between them. This case combines cases 3 and 4 with a terms of trade change, resulting in a combination of trade diversion, creation and reorientation (if ROW includes CARICOM), revenue loss and terms of trade change. A graphical exposition is given in Winters (1991). Qualitatively, revenue loss will be greater the larger the EU initial share and trade creation and trade diversion will be larger and the terms of trade losses smaller the more elastic is the EU supply curve.

In summary welfare gains are much more likely where: • the EU is the sole supplier in the Caribbean market and the more elastic is the EU supply schedule (cases 1 and 3), as there is then the possibility of trade creation; • the sole supplier is CARICOM (case 2) which gives rise to the possibility of trade reorientation; • the market is shared between the EU and the rest of the world, but where the marginal import price is determined by the EU. Once again this allows for the possibility of either trade creation and/or trade reorientation. In contrast, welfare losses are more likely via trade diversion, revenue loss or terms of trade effects where: • the ROW supplier is subject to tariffs and is the sole or principal supplier (case 2) and • the marginal import price is determined by an elastic ROW supply schedule but the EU has a larger share (case 4). In all cases, the greater the target country’s tariff, the greater the effects are likely to be. b. Empirical Analysis In order to make a preliminary assessment of the likelihood of these different possible outcomes, it is necessary to examine the underlying patterns of trade by product and geographical source. For concreteness the discussion below focuses on the imports of six Caribbean economies: Trinidad and Tobago, the Bahamas, Jamaica, St. Kitts and Nevis, St. Lucia and Antigua and Barbuda, the last three of which are small and members of OECS.


33

Table 4 showed that while the EU is an important supplier for Caribbean economies, the other CARICOM economies and the US are much more so. The high proportion of imports coming from the US (and facing tariffs) suggests considerable scope for trade diversion, while the moderate levels of (duty-free) imports from CARICOM suggests scope for trade reorientation. From the preceding discussion it is clear that on the import side the welfare effects of any proposed EPA will depend in good measure on who is supplying given goods to the market. For our six Caribbean economies and for the year 2002, we have calculated the share of imports for each HS six-digit product from each of four sources – CARICOM, EU15, USA and the Rest of the World. In Table 5 we report the number of product categories for which each of the suppliers has a share of 20–80 per cent, 80–90 per cent and over 90 per cent. We also give the percentage of imports falling in each category and, in column (1), the total number of six-digit headings in which imports are recorded. We interpret 20–80 per cent shares as indicating a shared market and over 90 per cent as being effectively a sole supply. We do not try to interpret the 80–90 per cent range, but report it so that the reader may treat it as he sees fit. Hence the first row of Table 5 shows that for Trinidad and Tobago, the US supplies between 20 and 80 per cent of the market for 1,256 products, which account for 10.8 per cent of total imports. For 775 products the US is virtually the sole supplier with over 90 per cent of all imports, and flows of these goods account for 6.41 per cent of Trinidad and Tobago’s total imports. There are several important messages which emerge from this table. First, it is only for Antigua and Barbuda that the EU is a significant ‘sole’ supplier, where for 27.21 per cent of total imports the EU supplies more than 90 per cent of the market. In fact this figure is dominated by two import categories – motor-boats and sail-boats – which between them comprise 53 per cent of all EU exports to Antigua and Barbuda. For the remaining economies the number of headings for which the EU is the ‘sole’ supplier ranges from 8 (covering 0.03 per cent of the total trade) for the Bahamas to 196 (0.83 per cent) for Trinidad and Tobago; the largest share of trade with the EU as sole supplier is 4.8 per cent for St. Kitts and Nevis. These numbers suggest that there is little scope for pure trade creation arising via case 1. The three small OECS economies report a significant proportion of total trade for which the CARICOM countries supply more than 90 per cent of the imports: St. Kitts and Nevis 8.6 per cent; St. Lucia 4.29 per cent and Antigua and Barbuda 6.21 per cent. Thus there may be a little more scope for trade reorientation. For the remaining CARICOM economies, however, the regional share is very low. It may be that in the OECS we are observing the transhipment from a CARICOM location of small consignments from elsewhere, but it might be that in the absence of direct transport links to the EU and USA, CARICOM output looks competitive in these economies.

34

MICHAEL GASIOREK AND L. ALAN WINTERS TABLE 5 Share of CARICOM Imports by Supplier – 2002

Country

Per Cent

Detail

Trinidad & Tobago

20–80 80–90 > 90 20–80 80–90 > 90 20–80 80–90 > 90 20–80 80–90 > 90 20–80 80–90 > 90 20–80 80–90 > 90

No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total No. of products Per cent of total

Bahamas Jamaica

St. Kitts and Nevis

St. Lucia Antigua & Barbuda

trade trade trade trade trade trade trade trade trade trade trade trade trade trade trade trade trade trade

CARICOM

EEC15

USA

ROW

61 0.570 13 0.216 37 0.110 11 0.606 1 0.000 2 0.000 128 9.220 5 0.079 15 0.300 227 7.378 18 0.221 78 8.600 282 13.805 28 2.500 84 4.290 204 2.009 34 1.270 469 6.210

733 9.192 58 0.827 196 0.830 82 0.800 5 0.003 8 0.030 498 5.516 34 0.438 106 2.480 308 4.388 26 0.880 96 4.800 530 11.900 55 1.075 167 1.300 279 10.131 48 0.805 352 27.210

1256 10.803 272 3.045 775 6.410 198 14.859 160 5.440 3743 62.150 1469 24.998 376 6.276 1445 10.120 814 21.787 283 8.480 1368 18.450 1090 27.519 215 4.281 987 8.910 297 5.309 79 3.097 403 6.050

1028 8.276 171 2.234 801 52.320 93 12.031 3 0.004 12 0.650 1048 14.327 109 5.544 218 11.930 394 7.714 40 5.875 97 2.250 495 10.124 37 0.657 155 4.450 188 27.033 25 0.689 266 3.560

Perhaps more importantly, if we look at imports from either the USA or ROW we see much higher numbers of products where these are dominant suppliers. For the US the figures for the share of total trade so covered range from 6.05 per cent for Antigua and Barbuda to 62.15 per cent for the Bahamas (and in this case there are no particular industries which dominate). Given the high proportion of headings, and the high share of trade covered by those headings this would suggest considerable scope for trade diversion arising from an EPA. We showed earlier that where the market is shared by more than one supplier, there is the possibility of either welfare gains or welfare losses depending on the


35

determination of the marginal import price. We cannot shed light directly on this, as we do not have the requisite supplier and price information. The table does indicate, however, the number of headings and percentage of trade where each supplier accounts for between 20 and 80 per cent of the market. Here it can be seen that the EU supplies between 20 and 80 per cent of the market for 733 products for Trinidad and Tobago, ranging down to 82 of the products for the Bahamas. This would suggest that relatively little tariff revenue is at stake in these shared markets (via the EU’s initial share), but the division of the trade effects between trade reorientation, diversion or creation cannot be calculated without information on elasticities of supply. Table 6 cross-classifies some of the information in Table 5. Classifying headings by reporting country share between 40 and 80 per cent, we ask what are the typical shares of other suppliers. Thus if we take the first row, the products for which the EU share lies between 40 and 80 per cent account for 12.41 per cent of total imports by Trinidad and Tobago. Of this, the EU accounts for 7.29 per cent, CARICOM 0.03 per cent, the US 3.55 per cent and the ROW 1.54 per cent. That is, where the EU has a ‘material share’ of Trinidad and Tobago’s markets, it is almost exclusively sharing with tariff-paying suppliers. The same applies to the other countries. Thus for these industries, trade diversion is more likely than trade creation, and, because of the small Caribbean share, there is little prospect TABLE 6 Shares of Trade for those Industries where the Reporting Country Share is Between 40 and 80 Per cent Country

Reporting Country

CARICOM

EEC15

Trinidad and Tobago

EU US CARICOM

0.03 0.23 0.30

7.29 1.65 0.03

Bahamas

EU US CARICOM

0.00 0.63 0.34

Jamaica

EU US CARICOM

St. Kitts

US

ROW

Total

3.55 6.57 0.07

1.54 2.37 0.09

12.41 10.81 0.49

0.18 1.23 0.04

0.18 9.10 0.22

0.01 2.42 0.01

0.37 13.38 0.61

0.02 6.36 0.23

1.97 3.97 0.23

1.05 20.31 1.16

0.49 6.38 0.49

3.53 37.02 2.11

EU US CARICOM

0.07 3.46 5.49

1.90 3.61 0.33

1.36 18.62 2.07

0.20 4.29 1.35

3.53 29.98 9.24

St. Lucia

EU US CARICOM

0.40 2.51 11.98

5.33 5.45 1.05

2.05 19.80 5.24

1.35 5.30 0.79

9.13 33.06 19.06

Antigua and Barbuda

EU US CARICOM

0.38 0.72 1.56

2.97 0.96 0.33

0.80 3.96 0.49

0.83 0.60 0.30

4.98 6.24 2.68

36

MICHAEL GASIOREK AND L. ALAN WINTERS TABLE 7 Finger-Kreinin Indices of Import Similarity

Jamaica Bahamas Trinidad and Tobago St. Lucia Antigua and Barbuda St. Kitts and Nevis

EU-US

EU-CAR

EU-ROW

US-CAR

US-ROW

CAR-ROW

0.399 0.269 0.057 0.384 0.111 0.281

0.158 0.131 0.165 0.190 0.071 0.129

0.296 0.135 0.079 0.304 0.257 0.221

0.219 0.112 0.119 0.273 0.132 0.215

0.320 0.239 0.264 0.343 0.096 0.291

0.098 0.108 0.339 0.157 0.128 0.187

for trade reorientation. Conversely, where the USA has a material share, the EU typically supplies a quarter or less of the US amount. In a differentiated good model based on the ‘Armington’ assumption, in which substitution possibilities are proportional to market shares, this would be taken as evidence that trade diversion was unlikely to be serious, but in our homogeneous goods model such a comforting conclusion would not be warranted. In fact, the data suggest that liberalising trade with the USA may be better than an EPA because the relative importance of local Caribbean supplies in the USA’s ‘material-share’ products opens up the prospect of trade reorientation. A further approach to the same issue is to consider the similarity of the bundles of imports from the four groups of suppliers. Table 7 gives FingerKreinin indices of import similarity for the same six Caribbean economies calculated at the HS-6 level. Formally the index is defined as: FKab =

∑ [min(Sia, Sib)] i

where Sia denotes the share of commodity i in country a’s imports and Sib the share of commodity i in country b’s imports. The index ranges from 1 (complete overlap in imports) to 0 (no overlap). From the first row we see that the highest level of similarity in exports to Jamaica is between the EU and the US, with an index of 0.399, followed by USROW and EU-ROW. The EU data suggest a fair degree of similarity or market sharing, reinforcing fears of trade diversion. The 0.158 index for EU-CAR at least admits the possibility of trade-reorientation as well, however. There is a similar story for St. Lucia. In contrast, the EU-USA index for Trinidad and Tobago is very low at 0.057, as is the EU-ROW index. This greater difference in exports from the EU from those from elsewhere indicates a greater possibility of trade creation. However, it is also consistent with there being very little competition in supply and hence raises the question of the extent to which tariff reductions might be translated into price reductions for consumers/users. Overall the EU-USA and EU-ROW similarity indices are of comparable levels (though with a different distribution), while the EU-CAR index indicates less


37

similarity with indices ranging from 0.071 for Antigua and Barbuda, to 0.19 for St. Lucia. Thus trade diversion seems much more likely than trade-reorientation which could have led to welfare gains. The previous tables have provided different but complementary ways of assessing the possible impact of an EPA by looking at summary statistics, but they do not reflect sectoral variation. To give an indication of this we look at broad sectoral patterns for four further OECS economies for which we cannot conduct the detailed analysis above. Thus Table 8 reports, for each one-digit SITC section, the sector’s share in each supplier’s exports to the target country and each supplier’s share of total imports of the sector. Thus for Dominica the table shows that the principal imports from the other CARICOM countries are in Food and Live Animals, Mineral Fuels, and Manufactured Goods (SITC6). This pattern of CARICOM exports is also true for the other economies shown, although it is worth noting the importance of Beverages and Tobacco for Montserrat. Two of these three sectors also form a significant portion of EU exports to these OECS economies suggesting possible reorientation. For both the US and the EU the principal export sector is Machinery and Transport Equipment, with Chemicals and Miscellaneous Manufactured Articles also figuring prominently. This again reflects the similarity of US and EU export patterns, and the dangers of trade diversion. The bulk of this section has explored Caribbean import patterns to try to predict the effects of the proposed EPA. There does not appear to be much ground for optimism concerning substantial amounts of trade creation or trade reorientation, but, on the contrary, significant scope for trade diversion and revenue loss. There are many sources of uncertainty in our analysis, but we should underline here that a critical factor is the extent to which the tariff reductions will actually translate into price reductions, which, in turn, depends on which bloc supplies the marginal units, the market structures in place in the Caribbean economies and the extent to which their imports are dominated by oligopolistic or monopolistic firms and/or distribution channels. In highly imperfectly competitive markets the preferential reduction of tariffs on EU sales could simply lead to a switch in supply (away from other CARICOM suppliers, or the US) but little change in price. This would make the agreements highly trade diverting. Finally, it is worth noting that this analysis is consistent with the conclusions reached by Greenaway and Milner (2003) who predict welfare losses for Caribbean economies as a result of EPAs, but welfare gains where these economies engage in multilateral trade liberalisation. The welfare losses they report range from −1.9 to −4.5 per cent of GDP depending on the economy, while the welfare gains range from 0.3 to 0.8 per cent of GDP. Policy implications The analysis of this section suggests that by granting preferential access to their domestic markets to the EU, the Caribbean countries face a considerable likelihood of both trade diversion and the simple transfer of tariff revenue from

38

MICHAEL GASIOREK AND L. ALAN WINTERS TABLE 8 Share of Trade by Region and Sector Sectors as Shares of Source’s Total

Sources as Shares of Sector’s Total

CARICOM

EU

US

CARICOM

EU

US

Dominica 0 Food and live animals 1 Beverages and tobacco 2 Crude mater. ex. food/fuel 3 Mineral fuel/lubricants 4 Animal/veg. oil/fat/wax 5 Chemicals/products n.e.s. 6 Manufactured goods 7 Machinery/transp. equipmt. 8 Miscellaneous manuf. arts.

22.11 6.72 1.02 29.53 1.65 7.27 20.61 2.60 8.49

27.85 4.15 0.48 0.03 0.09 17.67 10.24 27.34 12.15

14.50 0.51 3.87 1.04 3.29 15.96 13.30 31.92 15.61

35.62 64.97 15.74 95.40 23.73 17.26 37.47 3.61 20.48

20.93 18.71 3.49 0.04 0.64 19.58 8.69 17.68 13.67

28.10 5.99 72.02 4.05 56.98 45.59 29.11 53.26 45.31

Grenada 0 Food and live animals 1 Beverages and tobacco 2 Crude mater. ex. food/fuel 3 Mineral fuel/lubricants 4 Animal/veg. oil/fat/wax 5 Chemicals/products n.e.s. 6 Manufactured goods 7 Machinery/transp. equipmt. 8 Miscellaneous manuf. arts.

16.93 4.08 1.81 29.73 0.51 8.35 22.98 4.34 11.29

22.86 2.29 0.25 0.10 0.53 8.30 13.70 38.20 13.73

17.24 1.78 2.30 4.22 0.25 7.92 16.46 32.31 17.47

24.01 47.31 17.93 79.91 42.01 28.20 33.70 4.65 21.01

16.14 13.24 1.25 0.13 21.97 13.94 10.00 20.38 12.72

39.81 33.56 37.10 18.46 33.18 43.50 39.29 56.38 52.90

Montserrat 0 Food and live animals 1 Beverages and tobacco 2 Crude mater. ex. food/fuel 3 Mineral fuel/lubricants 4 Animal/veg. oil/fat/wax 5 Chemicals/products n.e.s. 6 Manufactured goods 7 Machinery/transp. equipmt. 8 Miscellaneous manuf. arts.

24.07 14.57 7.03 3.21 0.00 13.23 28.54 2.77 6.57

4.77 17.73 1.89 0.00 0.00 8.94 19.89 33.59 12.12

12.55 1.86 3.11 31.33 0.33 5.24 16.02 20.88 8.29

20.22 28.94 24.71 1.66 0.00 22.25 18.07 1.20 8.11

5.30 46.57 8.78 0.00 0.25 19.89 16.65 19.21 19.79

63.78 22.36 66.02 98.31 99.75 53.31 61.34 54.61 61.88

St. Vincent 0 Food and live animals 1 Beverages and tobacco 2 Crude mater. ex. food/fuel 3 Mineral fuel/lubricants 4 Animal/veg. oil/fat/wax 5 Chemicals/products n.e.s. 6 Manufactured goods 7 Machinery/transp. equipmt. 8 Miscellaneous manuf. arts.

18.88 5.79 0.67 29.44 1.00 12.09 20.12 3.99 8.03

24.20 2.67 2.12 0.16 0.17 10.84 12.43 36.19 11.24

25.44 0.59 5.54 0.37 0.19 8.47 17.97 28.42 13.00

24.32 68.82 5.57 96.27 72.58 34.35 30.94 4.81 19.89

16.93 17.20 9.52 0.28 6.61 16.72 10.37 23.69 15.11

47.26 10.16 66.16 1.75 19.55 34.71 39.83 49.42 46.47


39

TABLE 9 Import Taxes as a Percentage of Fiscal Revenue, in 1990 –1999

Antigua and Barbuda Anguilla Bahamas Barbados Belize Dominica Guyana Jamaica St. Kitts and Nevis St. Lucia St. Vincent Trinidad and Tobago

1990

1992

1994

1996

1998

1999

52.08 n.a. 65.94 13.21 51.54 17.84 11.39 n.a. 53.49 51.92 51.09 8.17

54.57 67.84 55.62 8.08 47.82 17.45 9.50 13.71 48.29 50.01 48.71 9.36

51.48 54.92 53.65 8.63 49.70 14.67 12.82 10.89 49.13 48.34 45.91 7.71

51.08 58.25 52.77 8.08 34.41 13.99 11.67 10.83 45.33 47.89 43.60 5.20

50.36 65.57 49.79 9.35 33.50 13.53 12.06 10.60 42.04 48.61 42.75 7.22

48.08 63.18 52.67 9.57 34.77 14.61 n.a. 10.00 43.57 47.08 43.26 7.23

Source: ECLAC database and Peters (2002).

Caribbean governments to EU suppliers (if the latter fail to reduce their prices and just pocket the tax that they would no longer have to pay under the EPA). The straight revenue losses are not huge because in general the EU has only a relatively low share of Caribbean imports. However, the trade diversion, which also shows up as revenue-loss to the importing countries could be large. For many of the Caribbean economies tariffs provide a significant share of government revenue, as can be seen from Table 9. For several economies the share is well over 40 per cent, while for only one (Trinidad and Tobago) it is below 10 per cent. Thus signing an EPA will need to be accompanied by serious efforts to mobilise alternative sources of government revenue. Indeed, given that the EPAs are likely to involve significant structural changes, the demand for government revenue is likely to increase as its supply falls. This is doubly unfortunate. First, for very small countries in which very high shares of consumption are imported, import tariffs are arguably an efficient form of levying a consumption tax (Winters and Martins, 2004). To be sure, this argument calls for uniform, or at least relatively uniform, tariffs, quite unlike the current Caribbean regimes (e.g., WTO, 2001 and 2002), but such tariffs are still precluded by the EPA. Second, the traditional antidote to trade diversion is to reduce tariffs on non-preferred partners – Schiff and Winters (2003). This entails a greater loss of revenue but at least by maintaining a non-discriminatory trade policy most of the revenue is transferred to local consumers rather than foreign firms. The high revenue shares make this solution more challenging. Finally we note that to the extent that trade diversion dominates the outcome, the EPA will not impose great adjustment shocks on local firms. This is precisely why governments responding to lobbies gravitate towards trade-diverting regional arrangements – Grossman and Helpman (1995). But to the extent that governments

40


seek to make welfare gains by stressing trade creation (on the producer side) or trade-reorientation away from other OECS/CARICOM countries, adjustment costs are inevitable for these economies. As is often observed, reaping the gains from trade requires changes in the activities an economy undertakes, and this, in turn, frequently entails adjustment costs.

5. CONCLUSIONS

The choices facing the OECS and the Caribbean economies are difficult. The choice is not simply one of either signing an EPA or staying with the status quo. If the economies do not sign an EPA with the EU, they will be left with the EU’s GSP preferences, and they will also probably lose out on the range of corollary measures such as aid disbursements and technical assistance which have formed part of the Lomé and now Cotonou packages. They would need to make little direct change to their import regimes, but would see a considerable erosion of the margins of preference that they currently receive on their exports to the EU. Given the importance of the EU markets for Caribbean exports, and given the concentration of those exports in certain sectors, such a policy choice is likely to have serious long-run distributional implications for many of these economies. The opposite choice – that of signing an EPA – keeps the Caribbean close to the top of the EU’s league table of market access preferences (although there is still likely to be some preference erosion as trade liberalisation proceeds), but at the expense of distorting further their import regimes. Addressing these distortions is feasible but will entail significant adjustment costs and critical efforts to maintain government revenue. In thinking about the choices, several priorities need to be established. First, these economies clearly need to increase their supply capacities and degrees of diversification. Being reliant on a small number of sectors either in terms of domestic production (e.g., tourism), or in terms of exports (e.g., bananas) makes them highly vulnerable to economic conditions elsewhere and to natural disasters. Second, in any future set of trading arrangements it is important to minimise levels of trade diversion, and to maximise trade creation. Closely linked to this, of course, is ensuring the competitiveness of domestic markets. Third, any such process is likely to be difficult; given the relatively weak financial and institutional capacities of these economies there is a clear need to recognise the importance of technical, institutional and financial assistance, as well as the need to think carefully about the appropriate length of any transitional periods. With regard to increasing the supply capacities and diversification it is possible that improved regional integration and the creation of larger regional markets could help. However, the keys will clearly lie in domestic policy and integration with the wider global economy. Also important will be levels of


41

assistance and aid. There are two principal ways in which the EPAs could be pro-poor and pro-development. First, through genuinely maintaining/increasing levels of market access to the EU: hence the importance of the EBA preferences, and of lengthy transitional arrangements (though these can also discourage change). Second, through the levels and appropriateness of assistance and aid. What for the EU may represent small sums of financial assistance, may well represent large sums for small island developing states. Assuming that the EPAs are pro-development in the sense discussed above, then there is a tangible benefit to the OECS and Caribbean economies in signing up. The challenge then becomes to minimise trade diversion. The obvious solution is to liberalise trade on an MFN as opposed to preferential basis. Thus while signing an ‘appropriate’ EPA the Caribbean governments should simultaneously and in parallel reduce their MFN tariffs. This may in turn lead to greater adjustment problems, hence again the need for appropriate transitional periods and assistance. To put it another way, if the economies are going to liberalise trade multilaterally, as they should, why not also have some of the non-trade benefits from the EPA?

REFERENCES ACP-EC EPA Negotiations (2003), Joint Report on the all-ACP–EC Phase of EPA Negotiations (Brussels). Brenton, P. and M. Machin (2002), ‘Making EU Trade Agreements Work: The Role of Rules of Origin’, CEPS Working Document No. 183. Chang, W. and L. A. Winters (2001), ‘Preferential Trading Arrangements and Excluded Countries: Ex-Post Estimates of the Effects on Prices’, The World Economy, 24, 6, 797–807. Economic Commission for Latin America and the Caribbean (1999), ‘Trade Policy in CARICOM: Overview of the Main Trade Policy Measures’ (United Nations). Gasiorek, M., et al. (2002), ‘Study on the Economic Impact of Extending the Pan-European System of Cumulation of Origin to the Mediterranean Partners’, part of the Barcelona Report to DG Trade, European Commission. Greenaway, D. and C. Milner (2003), ‘A Grim REPA’, University of Nottingham Internationalisation of Economic Policy Research Paper, 2003/30. Grossman, G. and E. Helpman (1995), ‘The Politics of Free-Trade Agreements’, American Economic Review, 85, 4, 667–90. Nicholls, S., A. Birchwood, P. Colthrust and E. Boodoo (2000), ‘The State of and Prospects for the Deepening and Widening of Caribbean Integration’, The World Economy, Global Trade Policy, 23, 9, 1161–94. OECS (2003), ‘Towards Developing a Model of Governance for Economic Union in the OECS: A Case Study of the European Union’ (Organisation for Eastern Caribbean States, October). OECS (2004), ‘The Treaty of Basseterre and OECS Economic Union’ (Organisation for Eastern Caribbean States). Panagariya, A. (2001), ‘Preferential Trade Liberalization: The Traditional Theory and New Developments’, Journal of Economic Literature, American Economic Association, 38, 2, 287– 311. Peters, A. (2002), ‘The Fiscal Effects of Tariff Reduction in the Caribbean Community’ (Economic Intelligence and Policy Unit, CARICOM Secretariat).

42


Rutherford, T. and J. Martinez (2000), ‘Welfare Effects of Regional Trade Integration of Central American and Caribbean Nations with NAFTA and MERCOSUR’, The World Economy, 23, 6, 799–825. Schiff, M. and L. A. Winters (2003), Regional Integration and Development (Oxford University Press for World Bank). Winters, L. A. (1991), International Economics (Routledge, London). Winters, L. A. (2001), ‘Post-Lome Trading Arrangements: The Multilateral Alternative’, in J. von Hagen and M. Widgren (eds.), Regionalism in Europe: Geometries and Strategies after 2000 (Kluwer, Dordrecht), 221–60. Winters, L. A. and P. Martins (2004), ‘When Comparative Advantage Is Not Enough: Business Costs in Small Remote Economies’ (mimeo, forthcoming, World Trade Review, November). World Trade Organisation (2001), Trade Policy Review: Barbados (WTO, Geneva). World Trade Organisation (2002), Trade Policy Review: OECS-WTO Members 2001, Vols. 1 & 2 (WTO, Geneva).

TRADE POLICY IN BURUNDI

43

3

Trade Policy in Burundi: Reform Without Political Stability Chris Milner

1. INTRODUCTION

HE WTO’s 2003 Trade Policy Review for Burundi (WTO, 2003) reports on a country that has been trying to reform its trade and other macroeconomic policies since the mid-1980s against the background of continuous socio-political tensions and period outbreaks of social conflict. Having undertaken research on Burundi’s trade policies before the mid-1980s (Greenaway and Milner, 1990; and Milner, 1993), it is of interest to consider what trade policy reform has been achieved and what it has achieved in the intervening years. Given the considerable scope and external pressure for reform, the expectation at the outset of this re-visit was that there had been some significant rationalisation and simplification of trade policy, but that the returns to these reforms in terms of export growth and diversification and improved economic growth had been very limited. Even in conditions of relative socio-political stability the returns to trade policy reform may be muted if institutional and infrastructural capacity is inadequate. But the conditions of political instability and internal crisis experienced by Burundi are hardly ones to induce the human and physical capital investment decisions required to bring about the structural transformation in response to a more liberal trade policy environment.

T

2. TRADE POLICY: THEN AND NOW

a. Trade Policy in 1986 The pre-reform (1984) structure of protection in Burundi was the outcome of a complex array of tariff and non-tariff interventions with protective and other (for example, tax revenue, foreign exchange control) objectives. 43

44

CHRIS MILNER

There was a comprehensive system of import licences in situ up to 1986. Since these licences were distributed administratively rather than auctioned the motive behind their use was clearly not revenue generation. Indeed it did not appear that the overriding objective for QRs was protective. Their coverage was much wider than required to protect the relatively narrow range of products produced by the (modern) industrial sector. Clearly there was some protective intent, but policy was not guided by the systematic and planned use of import bans and quotas designed to deliberately insulate ‘infant’ activities from foreign competition. Rather the immediate stimulus for QRs was as an instrument for rationing scarce foreign exchange. The greater priority given to imports of intermediate and capital goods than to final consumer goods meant that substantial protection did in general result for local manufacturers, but the level of protection was likely to vary between products (and over time) as local demand and foreign exchange pressures varied. Table 1 sets out the average price-raising effect of QRs (column b) for a sample of 33 products over a period of several months in 1984, and compares these with the corresponding nominal tariff rates (column a). For almost half the sample of products quantitative restrictions had a greater impact on domestic prices in Burundi than import duties and the average (unweighted) level of tariff-equivalence was greater than the average tariff level. Thus the combined or nominal level of protection (column c) was very high but also very variable between products (ranging from 9 per cent for cement to 501 per cent for salt). There was no systematic relationship between nominal tariffs and the tariffequivalents of QRs, the instruments neither systematically complemented each other nor were they substitutes. The pattern was essentially random. Given the evidently unsystematic manner in which import licensing operated, this was not unexpected. Thus the policy regime in situ before reform was one under which import licences were administratively allocated on the basis of some mixture of ‘priority requirements’ and historical rights, subject to foreign exchange pressures. The price-raising effects were uncertain and therefore imposed costs on producers and consumers alike. Producers were in general more than compensated for this by higher average prices. The large (gross) consumer losses associated with these import restrictions did, however, have their corresponding income transfers. In the case of tariffs the transfer was from consumers to the government. In the case of quantitative restrictions that were not auctioned and not country-specific the transfer was to the recipients of import licences, that is a restricted number of licensed importers. Using the estimates of tariff-equivalence in Table 1 yields a potential total rent transfer of almost FB (Burundian francs) 2 billion in 1984 just for the imports of the commodities in this sample. This was equivalent to about onesixth of government revenue in 1984. It is most unlikely that quota rents were distributed in an equitable fashion. Rather the magnitude of these rents provided incentives/opportunities for the bribery and corruption of those responsible for


45

TABLE 1 Sources of Nominal Protection in Burundi (1984)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

Condensed milk Butter Cheese Salt Sugar Powdered milk Jam Margarine Biscuits Vinegar Roasted Coffee Wine Vermouth Spirits Shirt (male) Shoes (male) Shoes (female) Shoes (child) Polish Soap powder Toilet soap Tooth brush Typewriter Calculator Cement Rolled metal Corrugated iron Tyres (car) Lorry (>5 tons) Tobacco Cigarettes Tea Flour

Average

(a) Nominal Tariff (Per cent)

(b) Tariff Equivalent of Quantitative Restriction (Per cent)

(c) Nominal Protection (Per cent)

31 54 106 19 9 31 111 29 104 69 154 36 65 87 59 36 36 36 57 57 57 51 49 49 9 14 24 36 19 154 154 154 9

113 0 108 482 60 273 206 44 89 153 0 21 50 0 47 110 0 7 161 0 35 264 0 0 0 0 359 84 84 207 0 53 23

144 54 214 501 69 304 317 73 193 222 154 57 115 87 106 146 36 43 218 57 92 315 49 49 9 14 383 120 103 361 154 207 32

(60)

(91)

(151)

Notes: (a) Combined entry and fiscal tax on imported goods. (b) Nominal protection minus nominal tariff and an assumed profit margin of 30 per cent. (c) Total price-raising effect of tariff and non-tariff interventions. Source: Greenaway and Milner (1990, Table 7).

licence administration and/or for patronage on a family, social class, political party and/or tribal basis (see Milner, 1996, for further discussion of these issues). There were three separate, import-specific taxes in operation in Burundi: ‘fiscal duty’ and ‘entry tax’ at variable rates of CIF import value and an ‘administrative

46

CHRIS MILNER

tax’ of a uniform rate on CIF value (4 per cent on imports from outside the regional preferential trading area of which Burundi is a member). This diversity of tax heading may have had some costs, including for instance monitoring costs and revenue loss associated with incentives to misclassify/record imports. There were also potential costs from this source of complexity on the taxpayer; current supply decisions or future investment decisions having greater information requirements than would have applied if there had been a single import tax. But not only was there non-uniformity on tax rates between the separate taxes, but considerable non-uniformity of tax rate for each tax within and between sections of the BTN import classification. (There was also often considerable variation in rates within each two-digit category and even within four-digit groups.) As Table 2 shows, there were in 1986 fifteen different rates of the fiscal tax which may apply in Section XVI (machines, and so on), or nine different rates of the entry tax which may apply in Section IV (processed foods). Again such a tax regime was likely to impose costs simply because of its complexity. Exemptions from border taxation of various forms were also permitted under the investment code, according to status. Thus the actual custom revenue rates (that is ‘ex-post’ measures of total customs revenue collected relative to CIF import value) were invariably substantially lower than might be expected from the posted or ex-ante rates of duty. Compare columns (1) and (4) in Table 2. Consistent or predictable production, consumption and import effects are less likely where the policy instruments are complex, administered in a discretionary and non-transparent manner. There is little evidence in the summary of the tariff schedule provided in Table 2 to suggest that consistent criteria were being applied. In Section IV (processed foods) nominal (posted) tariffs on six-digit product lines varied from 4 per cent to 154 per cent (inclusive of the statistical tax), with a rate of over 100 per cent in the majority of cases. The average rate of duty collection was, however, only 24.1 per cent in 1984! Similarly, in Section XVI (machines and instruments) the nominal rates varied from 9 per cent to 109 per cent (including statistical tax). In this case the modal combined rate (24 per cent) was much lower, but since the customs collection rate was only 9.5 per cent there is again clear evidence that this product group was not treated in a consistent manner by the tariff schedule. In the case of competing industrial imports, the treatment by the tariff schedule was again far from consistent. There were considerable inter-sectoral, intra-sectoral, and even intra-firm variations in nominal tariffs on final products. Consider Table 3 which gives weighted average tariff rates for imports competing with a sample of domestic manufacturers. There was substantial divergence between the average rate by sector and between the posted rates within sectors. Production or resource allocation effects depend upon the effective protection rates rather than nominal rates, that is upon the net effect of tariffs on both final and intermediate products. We have seen how tariffs and the price-raising effects


47

TABLE 2 Border Taxation of Imports in Burundi (1984) Sections of the Customs Schedule

(1) Ex-post Tariff Averages

I

11.8

7

2

0–150

7.7

6

5

5–150

4.2

5

2

5–100

24.1

8

9

0–150

3.7

5

2

0–20

9.1

11

4

5–102

12.4

11

5

5–83

67.4

9

3

5–100

13.7

7

3

5–80

14.4

14

3

0–100

39.8

14

5

0–155

35.1

5

3

10–150

9.9

11

5

5–85

0.2

6

1

0–100

11.5

14

5

0–103

9.5

15

3

5–105

II III IV V VI VII VIII IX X XI XII XIII XIV XV XVI

Live animals etc. (chs 1–5) Vegetables etc. (chs 6–14) Fats and oils (ch. 15) Processed foods etc. (chs 16–24) Mineral products (chs 25–7) Chemicals (chs 28–38) Plastic products etc. (chs 39–40) Skins, leather etc. (chs 41–3) Wood (chs 44–6) Paper (chs 47–9) Textiles (chs 50–63) Shoes (chs 64–7) Stone, cement, glass (chs 68–70) Precious metals (chs 71–2) Metals (chs 73–83) Machines and instruments (chs 84–5)

(2) Number of Fiscal Rates

(3) Number of Entry Tax Rates

(4) Range of Combined (ex-ante) Rates

Source: Greenaway and Milner (1990, Table 1).

of QRs on final products varied considerably across products. Table 3 also illustrates the variability of tariffs on intermediate inputs within and between sectors. This combined variability generated considerable variation in resource allocation effects within the industrial sector. Although subject to considerable variability, the structure of protection identified for the pre-reform period was in general likely to generate high levels of (average) effective protection. There is clear evidence of tariff escalation, that is higher output tariffs (i.e. subsidy rates to domestic manufactures) than input tariffs (i.e. taxes on domestic producers), to be found even in the tariff data. But

48

CHRIS MILNER TABLE 3 Average Nominal Scheduled Tariff Ratesa

Sector

Agricultural products Food Leather and footwear Textiles Wood and paper products Metal products Chemicals Pharmaceuticals Construction goods

Competing Final Imports

Intermediate Imports

Weighted Average (Per cent)

Range (Per cent)

Range (Per cent)

143 72 32 43 46 31 28 52 37

9–154 32–154 29–36 14–44 21–69 14–69 9–69 11–57 34–44

9–44 8–43 14–39 9–59 9–24 9–29 9–54 9–14

Note: a Identified for a sample of domestic firms. Source: Greenaway and Milner (1990, Tables 3 and 4).

this information tends to understate average net protection. On the one hand, QRs were pervasive, but more importantly their incidence was greater in the case of final than intermediate goods. On the other, duty exemptions were widely obtained on intermediate and capital goods (as a result of administrative discretion or of enabling legislation for the Investment Code). This meant that the output subsidy was higher and the input tax on local manufactures was much lower than can be inferred from posted border duties.1 Albeit in an ad hoc and unsystematic manner, all the above trade policies and related industrial policies like industrial licensing, investment promotion, wage and prices controls were directed up to 1986 towards import-substitution (IS) and the regulations of import-competing (non-traditional) industries. There were no measures or policies which discriminated specifically in favour of manufactured or other exports. Import duty exemptions were available on sales to both home and export markets.2 There were no specific export subsidies or specific measures of investment promotion (tax holidays) which discriminated in favour of exports. The Investment Code allowed for the refund of any export tax paid on exports, but since there was no corresponding tax paid on domestic sales this can 1

For a discussion of the contribution of effective protection to policy reform see Greenaway and Milner (2003). 2 Duty rebate was supposed to be available on imported inputs used in the production of exports where firms benefited from the Investment Code. The system was not, however, operated. Exemption from duty was the more important incentive given, but this was available also for import-competing firms.


49

hardly be described as a discriminatory measure in favour of exporting. Indeed, outside the Investment Code, manufacturing exports would (in principle at least) have been subject to export tax, in the same way as traditional exports (such as coffee) were! It is not surprising, given the smallness of the modern industrial sector and the pattern of relative incentives, that manufactured exports remained low up to 1986. b. Trade Policy in 2003 As part of the reforms launched in 1986, Burundi had eliminated most qualitative import restrictions and taken steps to rationalise its tariff structure and make customs duties the main instrument of trade policy. (Restrictive import licences were formally abolished in 2002, with import licences now only being used for statistical recording.) The reforms to the customs schedule were suspended, however, when the internal crisis broke out in 1993, and remained largely unaltered until the start of 2003.3 The tariffs applied in the period up to 2003 were 10 rates in the 0 to 100 per cent range. This represented some simplification on the earlier pre-reform tariff structure, but gave considerable scope for high and variable rates of nominal and effective protection (the simple average tariff rate being 30.8 per cent). Under tariff reforms implemented in January 2003, the maximum tariff rate has been reduced to 40 per cent. The number of tariff rates has also been reduced to 8. This further simplifies the tariff structure, and reduces the variability and average levels of nominal and effective protection, the simple average tariff now falling to 23.5 per cent. (See Table 4 for a comparison of the tariff structure in 2003 and 2002.) The current trade policy environment is still not as simple or transparent as the initial picture painted above. First, there is still considerable scope for both automatic and discretionary exemptions from customs duty under the Tax Code, Investment Code, Law on Export Promotion and Free Zone Law. These are estimated to give exemptions equivalent to about 15 per cent of the value of total imports. This is likely to increase the complexity and lack of transparency of the tariff structure considerably, and to increase the scope for inter- and intra-sectoral variability in the domestic incentives to produce. Second, there remain a number of other taxes or charges on imports that have discriminatory effects. These include relatively high inspection fees (pre-shipment inspection being mandatory on all non-small imports), a service tax (6 per cent on the customs value of imports from all origins) and a transaction tax (charged at lower rates on local 3

Further over the period 1996–2003 the earlier liberalisation of non-tariff barriers or QRs was reverted. A negative list of prohibited or controlled imports expanding, as foreign exchange pressures on the Government increased.

50

CHRIS MILNER TABLE 4 MFN Tariff Structure in Burundi, 2002–03 Compared (Percentage)

Bound tariff lines (percentage of all tariff lines) Duty-free tariff lines (percentage of all tariff lines) Non-ad valorem tariffs (percentage of all tariff lines) Tariff quotas (percentage of all tariff lines) Non-ad valorem tariffs with no ad valorem equivalent (percentage of all tariff lines) Simple average bound rate Agricultural products (HS01–24) Non-agricultural products (HS25–97) Overall standard deviation of bound rates Simple average applied rate Agricultural products (HS01–24) Non-agricultural products (HS25–97) Domestic tariff ‘peaks’ (percentage of all tariff lines)a International tariff ‘peaks’ (percentage of all tariff lines)b Overall standard deviation of applied rates

2002

2003

21.6 0.1 0.0 0.0 0.0

21.6 0.1 0.0 0.0 0.0

68.3 94.0 37.5 41.5 30.8 63.9 25.6 12.1 42.9 28.9

68.3 94.0 37.5 41.5 23.5 35.4 21.6 0.0 42.9 14.4

Notes: a Domestic tariff ‘peaks’ are tariffs exceeding three times the simple average rate applied (Indicator 8). b International tariff ‘peaks’ are tariffs exceeding 15 per cent. Source: Adapted from Table III, 1 WTO (2003).

than on imported goods in the case of agricultural and fisheries products).4 A special surcharge is also applied on textile imports. Again these increase the complexity of the discriminatory and distortionary nature of border taxation. Given the scope for discretionary exemptions on intermediate input tariffs, this also increases the scope for increasing effective protection on import-substitute activities; indirectly reducing the incentive to produce for export markets. Finally, it should be noted that there are a number of measures directly affecting exporting, which affect the incentive to export in opposite ways. Although the Government plans to eliminate export taxes, most export products remain subject to a 5 per cent export tax (with higher or lower rates applying on specific primary commodities).5 There are also some quantitative restrictions on specific (traditional) exports. By contrast, for non-traditional (i.e. manufactured) exports there are a number of instruments of export promotion in place, including duty drawback, foreign currency retention, reduced profits tax on export profits, free zone status and (since 2000) access to preferential interest rate loans from an export promotion fund. 4

‘Small’ refers to imports of less than US$5,000 or US$3,000 in the case of food, chemical and pharmaceutical products. Petroleum products are exempt from the requirements. 5 Up to 15 per cent, for example, on fresh vegetables. In the case of green coffee beans there is a 31 per cent rate, though this has not been collected in recent years.


51

3. TRADE STRUCTURE AND ECONOMIC PERFORMANCE: THEN AND NOW

a. The Burundi Economy in 1986 Burundi was a small low-income developing country in 1986. Average GNP/head was only $238 in 1985 (see Table 5) for a total population of 4.7 million people. Population density was, however, high by African standards at 168 people per square kilometre. Burundi did not, therefore, have a relative abundance of land. Burundi could not also be described as a financial or physical capital-abundant economy, even in sub-Saharan African terms. Equally it was not well endowed with human capital. As Table 5 shows (section d) only 4 per cent of the relevant age group were enrolled in secondary schools and only 1 per cent in higher education. We are confronted therefore with a predominantly agrarian subsistence economy, producing a single cash crop (coffee) which was the country’s main export (86.5 per cent of exports in 1983). Indeed, primary sector employment accounted for 93 per cent of total employment. Burundi is a relatively remote, land-locked and inaccessible country in central Africa. In fact the terrain and climate are pleasant and conducive to a range of

TABLE 5 Key Structural Characteristics (1983) (a) Production

Per Cent

(b) Employment

Per Cent

Primary (agriculture) Secondary (modern industrial) Tertiary

55 (46) 16 (.

THE WTO TRADE POLICY REVIEW OF CANADA

59

FIGURE 1 Canada’s Imports from Major Countries and Tariff Treatments, 1990–2003

market access with poor countries, we also show imports from countries to whom Canada grants Generalised Preferential tariffs (GPT) and Least Developed Country tariffs (LDCT).3 Given their rising prominence, we consider GPT countries Mexico (diamonds) and China (inverted triangles) separately from other GPT (triangles) origins. Trade is plotted on a log scale and thus the slopes of the lines correspond to annual growth rates. The figure portrays a general growth in imports from 1990–2000. The growth rates are particularly strong for China and Mexico in the period. After 2000, imports from the US and Japan fall a bit while imports from China, LDCT and other GPT continue to rise. Particularly noteworthy is the near doubling of imports from LDCT countries in 2003, rising from C$586 million in 2002 to C$1.0 billion the following year. This surge corresponds to Canada’s decision to eliminate tariffs and quotas on LDCT imports. Imports from Bangladesh rose from C$164 million in 2002 to C$350 million in 2002. The post-2000 growth rates indicate that, while Canadian trade remains strongly oriented to the US, recent trends have reduced this dependency somewhat. To assess how open a country is to trade, the conventional metric is to examine ‘openness’, defined as the sum of imports and exports divided by GDP.4 As noted 3

Canada grants LDCT to 48 countries. See, for example, Andrew Rose’s recent paper (2004) that argues that WTO members are no more open than non-members. 4

60

KEITH HEAD AND JOHN RIES

in the Report (p. vii), Canada scores an impressive 80 per cent on this metric, more than double the openness of the United States. This measure suffers from at least two important defects. First, there is a systematic tendency for large countries (as measured by national income) to have low measured openness relative for any given level of trade barriers. Intuitively, this is because large economies account for a large share of the world’s aggregate supply and demand and this makes them appear more insular. Second, because bilateral trade is inversely related to distance, remote countries tend to have lower multilateral trade levels. We construct a benchmark of multilateral trade that accounts for size and distance between trading partners. Let imports (M) from country i to country j be given by M ij = π ij Yj, where Yj country j’s gross national income (GNI) and π ij is the share of j’s demand met by country i producers. This share depends on the size of country i (because larger economies produce more of the goods that consumers demand) and the distance between i and j. Following evidence from estimation of gravity equations we assume bilateral imports are proportional to market size and inversely proportional to distance. To assure that π ij is bounded between zero and one and that market shares reflect size and distance to alternative suppliers, we assume:

π ij =

Yi / dij , ∑ lYl / dlj

(1)

where ᐉ indexes the countries (including j ) from whom j consumers purchase goods. Summing j’s imports, Mij, across foreign sources i gives an expression for j’s total imports Mwj. Summing across j, leads to total world imports Mww. The variable of interest is country j’s share of world imports: Mwj /Mww. Using greatcircle distances and GNI data for 2002 from the World Bank’s World Development Indicators (WDI), we calculate what Mwj /Mww should be for each country and compare it to actual shares of world imports (also obtained from WDI). Figure 2 plots on a logarithmic scale actual import shares against the theoretical benchmark for the set of countries with at least 0.5 per cent of world imports. If all countries imported according to the theory, the points would line up on the 45-degree line. Points below indicate actual imports less than the theoretical prediction. We observe that the theory fits quite well for large countries such as the United States (us), Great Britain (gb) and Germany (de). The same is true for France (fr), China (cn) and Italy (it). India (in) falls far short of the benchmark. Contrary to conventional wisdom, Japan (jp) imports more (5.5 per cent of world) than the benchmark predicts (4.1 per cent). In 2002, Canada accounted for 3.7 per cent of world imports, a much larger share than Canada’s share of GNI that year of 2.2 per cent, but less than the 3.9 per cent predicted by the benchmark. The benchmark takes into consideration Canada’s small size and proximity to the large US economy and expects higher import shares than simple GNI shares


61

FIGURE 2 Canada’s Imports Relative to the Gravity Benchmark

would indicate. Thus, while Canada has increased imports at a steady rate, this analysis suggests that Canada is no more open to imports than other large countries or what should be expected given its economic characteristics. An examination of trade by product reveals that Canadian trade is characterised by two-way trade in industrial goods, machinery and automotive products. Table 1 contains two panels showing imports and exports from 1999–2003. The first two rows of each panel report the value of trade whereas the other rows show trade shares by major product.5 In 2003, industrial goods and materials, machinery and equipment and automotive products accounted for 60.7 per cent of exports and 70.3 per cent of imports. The shares of these products have remained fairly stable over time. Canada’s overall trade surplus is accounted for by trade in forest products and energy. Products sold by poor countries, agriculture and apparel and footwear, have small import shares in Canada – 6.3 per cent for agriculture and 2.5 per cent for apparel and footwear. These industries are highly protected and we will see later on, that even for these products, most trade is with the United States. 5

These data come from Canadian Statistics available at <www.statscan.ca>. Unlike the trade figures reported above, total trade includes ‘special transactions trade’ and ‘unallocated adjustments’.

62

KEITH HEAD AND JOHN RIES TABLE 1 Canadian Trade Shares by Product 1999

2000

2001

2002

2003

Exports Agricultural and fishing products Energy products Forestry products Industrial goods and materials Machinery and equipment Automotive products Apparel and footwear Other consumer goods Special transactions trade Unallocated adjustments

$369.0 6.9 8.1 10.9 16.2 24.0 26.4 0.8 3.0 2.0 1.7

$430.0 6.4 12.4 9.9 15.8 25.6 22.8 0.9 2.8 1.9 1.6

$421.5 7.4 13.2 9.5 16.1 24.4 22.0 0.9 3.1 1.9 1.5

$414.3 7.5 12.0 9.0 17.0 23.5 23.4 0.9 3.4 1.9 1.6

$401.2 7.3 15.3 8.6 16.6 22.2 21.9 0.8 3.5 1.8 2.0

Imports Agricultural and fishing products Energy products Forestry products Industrial goods and materials Machinery and equipment Automotive products Apparel and footwear Miscellaneous consumer goods Special transactions trade Unallocated adjustments

$327.0 5.4 3.3 0.8 19.0 33.1 23.2 2.1 9.2 1.9 1.9

$362.2 5.1 4.9 0.8 19.1 33.9 21.4 2.1 9.0 1.8 1.8

$350.6 5.8 5.1 0.8 19.5 32.1 20.7 2.4 9.9 2.0 1.8

$356.5 6.1 4.6 0.9 19.3 29.7 22.8 2.4 10.6 1.7 1.8

$341.4 6.3 5.7 0.9 19.1 28.8 22.4 2.5 11.1 1.5 1.8

Note: First row in each panel are in billions of Canadian dollars and the other figures represent share of total product trade. Source: Statistics Canada.

Trade patterns across goods and countries are partly determined by differential tariffs that Canada applies. Figure 3 shows how Canada’s import tariffs vary across two-digit HS product categories by displaying the mean and standard deviation of tariffs within each sector. The vertical dotted line in Figure 3 shows that the average most-favoured nation (MFN) tariff is 6.8 per cent. Sixteen out of 21 sectors have average tariffs below that mean. Tariffs are higher low-skill intensive areas like footwear and textile articles. They reach their peak (314 per cent is the maximum) in the agri-food sectors (01, 03, 04). The fact that the standard deviation of tariffs in sector 01 (Live Animals and Products) is larger than the mean is remarkable since tariffs are non-negative. It shows the extreme skewness of the tariff distribution in this sector. Canada extends most-favoured nation tariffs (MFN) to almost every country. However, ‘most-favoured nation’ is very much a misnomer: many countries qualify for ‘preferential’ rates. Canada currently has four regional trade agreements in force. Tariffs were phased out under the North American Free Trade Agreement


63

FIGURE 3 Canada’s 2002 Tariffs Vary by Commodity

(NAFTA) over a 10-year period beginning in 1995. The Canada-Chile Free Trade Agreement (CCFTA) was implemented in 1997 and calls for the complete elimination of tariffs between the countries by 2003. The Canada-Costa Rica Free Trade Agreement (CCRFTA) came into force on 1 November, 2002, and tariffs are to be completely eliminated by 2011. Finally, the Canada-Israel Free Trade Agreement (CIFTA) eliminated tariffs when it was implemented in 1997. Canada also unilaterally extends tariff preferences to developing countries through its General Preferential Tariff (GPT) and Least-Developed Country Tariff (LDCT).6 Other preferential tariff arrangements include the Commonwealth Caribbean Countries Tariff and the New Zealand and Australia Tariff. Figure 4 displays average Canadian tariffs in 2002 by tariff treatment.7 The figure also indicates the dates of the phasing out of tariffs under regional trade agreements. At 6.8 per cent, the MFN average is highest (aside from the general tariff rate of 35 per cent applied to goods from Libya, Iraq and North Korea). Tariffs are lowest for regional trading partners and poor countries. Recall that in 6

In February, 2004, Canada announced it would extend these two tariff programmes for another ten years. 7 These rates are simple averages across tariff lines as reported in Table III.2 of the review.

64

KEITH HEAD AND JOHN RIES FIGURE 4 Canada’s 2002 Tariffs Vary by Country of Origin

2003 Canada eliminated tariffs on goods from the least-developed countries. Even though Canada fully phased in its CUSFTA tariff reductions by 1998, the average tariff level on the US remains positive because not all commodities qualified for tariff elimination. In 2002 Canada imposed no duties on US-origin goods for 99 per cent of its tariff items. For the remainder, mainly agricultural products, tariffs were very steep, averaging 225 per cent.

3. WELFARE EFFECTS OF CANADA’S RTAS

In addition to the four existing RTAs – NAFTA, CCFTA, CCRFTA and CIFTA – Canada is negotiating others. Discussions have been launched for a Free Trade Area of the Americas and for free trade with the Andean Community (Bolivia, Colombia, Ecuador, Peru and Venezuela), the European Free Trade Association, and the Central America Four (El Salvador, Guatemala, Honduras and Nicaragua) and Singapore. Canada is not alone in its interest in regional trade agreements. The WTO’s World Trade Report 2003 (WTO, 2003b), reports that 176 RTAs were in force in


65

2003 with only four WTO members – Hong Kong, Macao, Taiwan and Mongolia – not party to an RTA. The term used by the WTO, ‘regional trade agreements’, has become a misnomer. The World Trade Report 2003 uses the term crossregional RTAs to refer to free trade areas between countries situated in different geographical areas. From 1996 to 2002, the WTO received notification of 25 cross-regional FTAs and estimates that one-third of the RTAs currently being negotiated are cross-regional (World Trade Report 2003, p. 51). Of course, just because other countries are forming RTAs does not necessarily make it right for Canada to form them. Most economists agree that the first-best trade policy is multilateral free trade. However, in the early 1990s, with the prospects for a completed Uruguay Round looking doubtful, economists debated the merits of the regional agreements that seemed to be proliferating. The debate was about the second-best: was a multilateral framework with positive tariff levels better than one where some countries eliminated tariffs on a bilateral basis through regional trade agreements? The debate centred around the question of whether the trade creation generated by reducing tariffs between members outweighs the costs of trade potential being diverted away from efficient thirdcountry producers. Most RTAs also apply rules of origin. These rules, considered in detail in the next section, undermine the ability of exporters to benefit from preferential tariff treatments because compliance with the rules is costly and confusing. In today’s policy context, when many countries again embrace bilateral agreements and the completion of the current (Doha) multilateral trade round again seems uncertain, it seems worthwhile to revisit that literature and the conclusions it reached. It is now possible to add a consideration of the relevant empirical results. Out of the academic debate arose the concept of ‘natural’ trading blocs. The idea, developed in Krugman (1991), is that continental trading blocs are natural in the sense that they reinforce trade patterns that emerged as a consequence of non-tariff trade costs. Lower tariffs then lead to much trade creation and little trade diversion. If trade costs within continents were low relative to trade costs between continents, then there will be more intra-continental trade than intercontinental trade. A continental free trade agreement will promote the already substantial trade within continents without much consequence for trade outside the continent (since there was little to start with). Frankel, Stein and Wei (1998) develop a model based on Krugman (1991) in which RTAs improve welfare if inter-continental trade costs exceed a critical level. They use a gravity equation to estimate these trade costs and find that they are not high enough to allow for welfare-improving RTAs. Trade creation and diversion fundamentally depend on price elasticities. If goods are substitutes and cross-price elasticities between products produced by countries within trading blocs are high, then tariff reductions will generate

66


substantial trade. On the other hand, trade diversion will be large if cross-price elasticities between member and non-member country goods are high. This basic point is addressed in Krishna (2003) who uses estimated price elasticities to predict trade creation and diversion associated with hypothetical free trade agreements between the United States and various trading partners. Interestingly, he finds no significant relationship between the welfare effects of free trade agreements and distance from the United States. He concludes that free trade agreements between geographically proximate countries are not ‘natural’. Two studies have examined trade creation and diversion effects of the CUSFTA and the NAFTA. Clausing (2001) uses US import data for highly disaggregated commodities to relate import changes over the 1989–1994 period to tariff changes mandated by the NAFTA. She estimates separate tariff effects for imports from Canada and imports from the rest-of-world (ROW). Positive tariff effects on Canadian imports indicate trade diversion, and negative tariff effects on ROW trade reflect trade diversion. Figures 5 and 6 relate the magnitude of tariff reduction under the FTA to the growth in Canadian exports and imports during the implementation period.8 We group Canadian three-digit manufacturing industries FIGURE 5 Canadian Export Growth During CUSFTA Implementation

8

These figures appear in Head and Ries (2003, p. 187). They were inspired by a similar figure in Clausing (2001) that relates growth in US imports from Canada and rest of world from 1989–1994 to different tariff groups.


67

FIGURE 6 Canadian Import Growth During CUSFTA Implementation

into quintiles according to 1988 tariff levels. Then we sum imports and exports to and from the US and rest-of-world (ROW) by quintile groups and calculate the percentage change from 1988–1998. Consistent with Clausing’s results, Canadian export growth to the US rises sharply with initial tariffs, suggesting substantial trade creation. No relationship is apparent for exports to ROW. Clausing finds that CUSFTA tariff reductions explain over one-half of the increase in Canadian shipments to the US over the period 1989–1994. Romalis (2004) uses US and EU commodity-level import data to examine trade creation and diversion due to the CUSFTA and the NAFTA. His study differs from Clausing’s in a number of ways. Most importantly, he uses a difference in differences estimation technique where changes in EU trade patterns serve as a control when evaluating changes in US trade patterns. His study also includes Mexico and covers a longer time period, 1988–2000. Like Clausing, he finds substantial trade creation. In contrast to her results, his study reveals large trade diversion effects of the RTAs, amounting to about one-third of the trade creation. He explains the failure of Clausing to find trade diversion as a consequence of omitted variable bias. The reason why the US tended to have high growth from ROW in high-tariff industries was that emerging economies such as China increased their exports to all markets during this period. By differencing out the EU import pattern, Romalis intends to control for overall growth in exports of these countries.

68

KEITH HEAD AND JOHN RIES 4. RULES OF ORIGIN

Canada’s trade agreements have never resulted in common external tariffs. As a consequence, they always include rules of origin. In order to qualify for a particular tariff treatment, it is necessary to meet specified rules of origin. The ostensive purpose of rules of origin is to prevent an exporter from using a lowtariff ‘backdoor’ to introduce goods from countries that would otherwise have been subject to higher tariffs. In practice they have the consequence of limiting the ability of producers in one country to benefit from the tariff preferences built into the agreement. The obvious way to determine origin in a world where intermediate inputs might be sourced from many different countries is to specify a minimum ‘regional value content’. Specifically an agreement could require that some minimum share of the production cost (or ex-factory price) involve domestic inputs including labour. To qualify for tariff treatments other than free trade agreements, Canada requires that a certain percentage of the ex-factory price of the good originate in beneficiary countries. Additionally, it insists that the good be directly shipped from a beneficiary country. The critical value content varies by treatment: it is 50 per cent for MFN treatment, 60 per cent for GPT and 40 per cent for the LDCT. Canada’s free trade agreements eschew minimum content rules that apply to all goods and instead impose product-specific input requirements. We will consider the NAFTA rules in some detail because they are complex and potentially impose significant distortions. Qualifying as a North American originating good is at once very simple and exceedingly complex. The simple part is the NAFTA ‘Certificate of Origin’, a one-page form that must be presented at customs. In it the exporter certifies that the goods to be imported originate within the three parties to the agreement (United States, Canada and Mexico). He also agrees to provide proof of his claim if audited. He further indicates the criterion he used to determine origin. Article 401 of the NAFTA sets out four possible criteria that can be used to establish North American origin.9 Criteria (a) and (c) are the most stringent since they prohibit the use of inputs produced outside of North America. Criterion (d) is intricately worded but seems like it would rarely apply (the good must arrive in disassembled form but nevertheless its assembly must account for over 50 per cent of cost). The main criterion used for products incorporating nonNorth American inputs (‘non-originating materials’) is reproduced verbatim in the box below:

9

.


69

(b) Each of the non-originating materials used in the production of the good undergoes an applicable change in tariff classification set out in Annex 401 as a result of production occurring entirely in the territory of one or more of the Parties, or the good otherwise satisfies the applicable requirements of that Annex where no change in tariff classification is required, and the good satisfies all other applicable requirements of this Chapter.

Annex 401 is a 243-page document that specifies for every tariff item (eightdigit classifications) what inputs must be sourced from within North America in order for the final good to be deemed of North American origin. For some products, such as passenger cars, Annex 401 mandates a regional value content of not less than 50 per cent.10 For most products, Annex 401 specifies particular tariff classifications of inputs that, if non-originating, make the final good nonoriginating. To illustrate, we consider the example of televisions. This case is important because Japanese TV makers such as Sony have been eager to supply the North American market with assembly factories located in northern Mexico. There are seven different tariff items corresponding to different types of TVs. The most common is probably 8528.12.bb, single picture tube, non-projection, non-high definition, televisions with display diagonals exceeding 14 inches. The rules of origin are shown below:

A change to tariff item 8528.12.bb from any other heading, except from tariff item 8529.90.dd or 8540.11.aa or more than one of the following: tariff item 7011.20.aa, tariff item 8540.91.aa.

Tariff item 8529.90.dd comprises combinations of the following parts: (a) video intermediate amplifying and detecting systems; (b) video processing and amplification systems; (c) synchronising and deflection circuitry; (d) tuners and tuner control systems; (e) audio detection and amplification systems. Tariff item 8540.11.aa is the most important part of the TV, the cathode ray tube. Tariff items 7011.20.aa and 8540.91.aa are, respectively, the glass cone and the phosphorescent glass panel onto which the cathode emits electrons. Thus for a TV to be deemed North American origin, its most important part cannot be imported, nor can one import two of the important components of that part, nor can one import a set of other specified parts. The rule against importing both cone and glass panel is actually redundant because the rule of origin for cathode ray tubes imposes this same requirement.

10

Even in this case, the actual rule is not what it appears to be. A footnote in Annex 401 refers to Article 403 which notes that ‘Notwithstanding Annex 401’ the regional content must rise to 62.5 per cent in 2002.

70


The NAFTA rules of origin also contain a de minimus rule, stating that a good containing non-originating materials that do not satisfy the Annex 401 rules may still qualify as long as those parts constitute less than 7 per cent of the final good’s cost. This would certainly not help for a cathode ray tube, though it might matter for the 8529.90.dd list. The existence of such stringent rules suggests a few possible options. First an exporter could certify itself as North American and hope that it is not called upon to prove the case. Second, a vertical industry cluster can form within the region, allowing for adherence to the rules of origin. Finally, the exporter could opt to just pay the MFN tariffs. In the case of televisions, the MFN tariff in Canada is 5 per cent. Mexico appears to be forming a vertical cluster. James Gerber (2000) reports that ‘between 1995 and 2000, the number of cathode-ray-tube manufacturers increased from 2 to 5, television manufacturers grew from 8 to 12, and glass manufacturers from 0 to 2’. Trade data show that the US, which used to source large volumes of cathode ray tubes from Japan, now mainly imports them from Mexico. Rules of origin not only create administrative costs, they may also generate a secondary source of trade diversion. Consider cathode ray tubes which have an MFN tariff of 6 per cent. Since NAFTA allows North American-made goods to enter Canada duty free, the NAFTA can cause trade diversion from a non-member country producing the good, say Japan, to a member country, say Mexico. Suppose, however, that Japan is the low-cost producer even when the tariff is added and thus based on tariff-inclusive prices, Japanese cathode ray tubes are preferable to Mexican ones. Trade diversion may still occur because of the vertical linkage between cathode ray tubes and television. As discussed above, in order to qualify for duty-free imports of televisions under the NAFTA, TVs must use cathode ray tubes of North American origin. Thus, a TV producer may choose to purchase Mexican-made cathode ray tubes even if their tariff-inclusive price is higher than Japanese tubes. In this way, trade diversion can be compounded by rules of origin. The upper bound on the costs associated with complying with rules of origin is the MFN tariff rate. Firms will not spend more on compliance than they gain from qualifying for low preferential tariffs relative to MFN tariffs. In the extreme where compliance costs exceed their benefits, goods enter at the MFN rate and the RTA neither creates nor diverts trade.

5. SAFEGUARDS, ADDS AND CVDS

Contingency measures authorised under WTO agreements include safeguards, anti-dumping duties (ADDs) and countervailing duties (CVDs). The picture that will emerge is that Canada has targeted these measures at one sector – steel. Canada has never used safeguard measures allowed under the WTO Agreement on Trade in Agriculture and the WTO Agreement on Textiles and Clothing.


71

Only once – for nine steel products in 2002 – has Canada initiated a safeguard investigation. Likewise, countervailing duty investigations do not occur frequently. According to WTO statistics, Canada initiated 12 CVD investigations from 1995 to 2003 (there were 168 initiations in the world over this time period). The Review notes that there has been an increase in CVD investigations in Canada since the last review (6 of 12 cases since 2000). Seven of these investigations resulted in countervailing duties. Six of the seven are on product sector XV ‘Base Metals and Articles of Base Metals’ that includes steel products. The application of anti-dumping duties is of greater concern. According to the WTO, Canada has launched 122 of the world’s 2,416 investigations since the WTO’s inception in 1995. During the same period it was targeted in 25 cases. Canada is clearly a large net user of the measures. This raises the question of whether Canada is an over-user, or under-targeted, or both. Figures 7 and 8 depict anti-dumping initiations relative to trade cumulated over the period 1995–2002.11 The idea is that if all countries had the same FIGURE 7 Who (Over-)Uses ADD? 1995–2002

11

Trade data are only available through 2002. The figures include only countries that account for more than 0.5 per cent of world imports. They are drawn using a log scale so that the two largest economies (the US and European Union) do not compress the other countries into the lower left corner.

72

KEITH HEAD AND JOHN RIES FIGURE 8 Who is (Over-)Targeted? 1995–2002

propensity to use ADDs and if targets were chosen in proportion to exports, then ADD use should be proportionate to imports and targeting should be proportionate to exports. The straight line in Figure 7 corresponds to equal shares of ADD use and import shares, whereas the line in Figure 8 equates targeting shares and export shares. Countries above the line in Figure 7 may be thought of as over-users. Canada initiates only slightly more anti-dumping cases than one would expect based on its imports. The real outliers turn out to be South Africa (za) and India (in). On the other side, Japan almost never uses ADD despite accounting for over 5 per cent of the world’s imports. Figure 8 shows that Canada is notably under-targeted relative to its exports. In part this reflects a general tendency of high-income countries to be under-targeted. For example, Switzerland (ch) is hardly ever (4 out of 2,206 cases) targeted despite generating 1.4 per cent of the world’s exports. It would also seem that Canada’s high propensity for exporting to the US, which Figure 7 shows to be an under-user, also helps. There is little awareness of this in the Canadian press, which depicts Canada as a victim of US antidumping policies. At the time the Review was written, Canada had 91 measures in place. This number exceeds the 72 new measures instituted since 1995 due to the frequent long duration of ADDs in Canada. The Review notes that 9 per cent of the


73

FIGURE 9 Canada’s Anti-dumping Duties at the End of 2001, by Target

measures had been in place for 10 years or more. Another concern of the Review was the increase in anti-dumping usage – 33 new measures were imposed in 2000 and 2001, although this fell off considerably to zero and five in 2002 and 2003. Sixty-four of the 91 measures (70 per cent of the total) in place were on steel products. It should be noted that Canada is not the only country that tends to apply ADDs to steel: 528 of the 1,511 positive determinations globally between 1995 and 2003 are in sector XV. The portion of the base metal sector’s measures is 134 of 205 for the US, 71 of 187 for the EU, and 32 of 62 for Mexico. Which countries bear the brunt of these measures? Figure 9 is a re-expression of data derived from Chart III.4 in the Review showing the existing 91 antidumping duties in force in Canada on 31 December, 2001, by country and sector. The figure reveals the prevalence of steel cases that account for virtually all measures in place for countries other than China and the United States. With 16 measures, the EU is the number one target. While the frequent use of anti-dumping duties is a concern to the WTO, there are a couple of factors that mitigate their damage. First, according to the Review, the Canadian International Trade Tribunal estimates that anti-dumping measures affect less than 1 per cent of imports (p. 47). Second, Canada operates a ‘prospective system’ that allows targeted exporters to raise their export prices to the ‘normal’ price and avoid import duties. While in most circumstances, exporters

74


are worse off when they are required to price at a higher level (at the normal price), at least under Canada’s system they are able to pocket the difference rather than have a tariff applied that accrues to the host government.

6. AGRI-FOOD

Agriculture policies in Canada are important to study for a couple of reasons. First, this sector was a main source of the impasse in the Cancun meetings where a block of developing countries led by Brazil, China and India, representing twothirds of the world’s farmers, rejected the EU and US proposed agricultural framework. They objected to the continuation subsidies and trade barriers. Moreover, the limited response by the US and EU on the needs of four Western African countries – Benin, Burkina Faso, Chad and Mali – for freer trade in cotton, was seen as a lack of concern for the poor. Lacking quid pro quo concessions on trade and agriculture, poor countries refused to advance the Singapore issues. Canada, like the EU and US, maintains a number of market-distorting programmes in agriculture that impede exports from developing countries. A second reason to study agriculture is the sector’s importance to Canada. According to the Review, the agri-food sector, which includes unprocessed, semiprocessed and fully processed farm products, accounts for 4 per cent of Canadian GDP, while the related distribution, food retail and food services sector contributed another 4.3 per cent of GDP. The sector is Canada’s third largest employer. While trade in agricultural products is small relative to that of manufactures, trade is important to the industry. In 2003 exports totalled C$24.3 billion and imports C$20.6 billion,12 making Canada the world’s third largest exporter and fifth largest importer. The top exports that year were wheat and oilseeds. Imports were more diversified with wine being the largest item. As is the case with overall trade, the United States takes the largest share – 64 per cent of exports and 62 per cent of imports. Japan was the number two destination, with imports of C$2.3 billion. A few poor countries succeeded in selling relatively small amounts of agri-food into Canada: C$510 million, C$436 million and C$287 million for Mexico, Brazil and China, respectively. The Review states that 25 per cent of agri-food output was exported in 2001 (up from 15 per cent in 1996), while imports were 15 per cent of output that year. The WTO Agreement on Agriculture endeavours to lower certain types of agricultural support. It classifies support into three categories. Support that does not distort trade is called ‘green box’ and is exempt from reductions. This includes some types of direct income support for farmers, environmental protection and

12

The 2003 data comes from .


75

regional development programmes. Subsidies that are accompanied by production limits are termed ‘blue box’ and are also exempt. Support measures that distort trade such as subsidies and price supports are labelled ‘amber box’. Amber box supports are measured as Aggregate Measures of Support (AMS) and WTO members are required to reduce them if they exceed de minimis levels set at 5 per cent of agricultural output for developed countries and 10 per cent for developing countries. Canada committed to reduce total AMS from C$5.2 billion in 1995 to $4.3 billion in 2000. According to the review, Canada last notified the WTO about domestic support in 1999 when it was well below its mandated WTO ceiling. The Review acknowledges that Canada’s farm support is low relative to that of other OECD countries. Canada’s support relative to GDP is 0.7 per cent, about half the OECD average. Producer support as a per cent of farm receipts was 17 per cent in 2001 compared to the OECD average of 31 per cent. The Review also notes that Canada is shifting away from commodity-specific support and towards income ‘safety nets’ for farmers. In June 2002, Canada announced C$8.2 billion in new spending under the Agricultural Policy Framework with the goal of improving product quality and business performance in the agri-food sector. A significant amount of the spending will be green box. Total government spending supporting the agri-food sector was about C$6.2 billion in 2001/02 and the Review notes an upward trend in support. The largest expenditures are federal and provincial income stabilisation programmes (C$2.5 billion) and crop insurance (C$451 million). Both are reported to the WTO as amber box. Price support programmes raise incomes to farmers without requiring direct government expenditures (the costs are borne by consumers). These are widespread in Canada’s dairy sector and have resulted in producer prices for milk being double the world price. Prices for eggs, chicken and turkey are also set by authorities rather than the market. These commodities account for 25 per cent of total farm receipts. Price supports are complemented by supply management and import restrictions. Producers must purchase production quotas to participate in the domestic market. Import quotas apply to supply-managed commodities subject to small ‘in-quota’ tariffs. These import quotas often amount to less than 5 per cent of domestic consumption. Hence, the quotas have high ‘fill rates’. High (often prohibitive) tariffs apply to imports beyond quota levels. The Review lists complicated in-quota tariffs, out-of-quota tariffs and fill rates for agri-food products in Table IV.1. The tariffs vary according to type of importer and the table reveals out-of-quota tariffs exceeding 200 per cent for milk, chicken and eggs. Free trade partners and some other partners are granted special privileges on agri-food sales to Canada. In the case of wheat and barley, US, Mexican and Chilean producers can export unlimited amounts to Canada duty free. Some

76


partners are given reserved access to the quotas. As described earlier, these preferential arrangements create trade but also generate inefficiencies associated with trade diversion. The Review notes the important role the Canadian Wheat Board (CWB) plays in Canadian trade. The CWB is a state-trading enterprise that has exclusive authority to handle all wheat, durum wheat and barley sold for domestic consumption or exported in Canada. With export sales of C$4.2 billion in 2000–01, it is one of Canada’s top five exporters. The CWB provides farm support through advance payment and other financing programmes. This support is reported as amber box. The Government of Canada guarantees to the CWB the principal and interest of all credit receivables and lends money to the CWB at lower than market interest rates. This raises the concern that Canada is subsidising exports. Indeed in 2003, the US took Canada to WTO dispute arbitration, contending that the Canadian Wheat Board discriminated against imports and subsidised exports. On 6 April, 2004, the WTO Panel agreed with the US that CWB rules concerning grain handling and transportation violate national treatment. However, it also ruled that the CWB activities conformed with WTO state trading obligations to operate on a commercial and non-discriminatory basis. The US is appealing the decision. An area that has drawn much controversy is health standards that restrict food imports. Canada appears to be in full compliance with the WTO’s Agreement on Sanitary and Phytosanitary Standards (SPS) which covers food safety measures– Canada has never been a defendant in an SPS dispute at the WTO. Nevertheless, the Review expresses concern about: the strict use of sanitary and phytosanitary measures by Canada, which could result in barriers or increased costs for exporters from other countries (p. ix).

The Review notes that Canada made 62 notifications to the SPS committee between 2000 and mid-2002, an increase from the 40 reported in the previous Review. Many were new regulations concerning maximum residue levels for chemicals contained in food. Canada enacted emergency measures in 2000, suspending imports of live animals in countries with outbreaks of foot-and-mouth disease. Imports of live ruminants require a full risk assessment and Canada recognises, when possible, studies done by other NAFTA countries and meat inspection systems of other countries. Canada faces competing pressures concerning agricultural liberalisation. As a country with an abundance of arable land and a slight net exporter of agricultural products, it stands to gain from worldwide agricultural liberalisation. Canada has been active in demanding adherence to existing WTO agreements: it has been the complainant numerous times in SPS disputes at the WTO, including disputes with the EU concerning hormone-treated beef and genetically modified food. On the other hand, Canada can only push liberalisation along gradually as many


77

agri-food sectors have enjoyed high levels of protection, and the elimination of this protection imposes significant costs on a small but politically strong constituency. On balance, Canada realises that its broad interests are served by further agricultural liberalisation and it is committed to the current round of WTO agriculture negotiations.

7. CLOTHING AND TEXTILES

Clothing and textiles account for 3.3 per cent of manufacturing GDP and 6.9 per cent of manufacturing employment. According to the Review, the industries are characterised by small establishment size and extensive outsourcing. The industries have experienced large increases in both imports and exports over the last decade with exports to shipments around 42 per cent and imports at 60 per cent of consumption in 2001. The US absorbs the lion’s share of Canadian exports – 93 per cent for textiles and 96 per cent for clothing. Imports originate primarily from China and the United States. The only least developed country that supplies Canada to any considerable extent is Bangladesh with a 3 per cent share of imports. Clothing and textile trade are governed by the WTO Agreement on Clothing and Textiles (ATC). Imports are restricted by both tariffs and quotas. The Review reports that average MFN tariffs for textiles and clothing will fall to 8.5 per cent and 14.0 per cent, respectively, according to WTO commitments. Duty remission programmes have been in place to lower the costs of imported materials and complement product lines of Canadian clothing producers. Outerwear fabrics, shirt fabrics, outerwear clothing, blouses and shirts are eligible for duty remission. One new item – high-quality fabrics – was added to the list since the last WTO review. The total remission was C$32 million in 2001. On 1 January, 2002, Canada commenced phase three of the phasing out of quotas by eliminating quotas on products representing 18 per cent of 1990 imports. The growth rate of existing quotas was raised to 27 per cent. As stipulated by the ATC, the remaining quotas (representing 49 per cent of 1990 imports), will be eliminated on 1 January, 2005, when all clothing and textiles will be integrated into the General Agreement on Tariffs and Trade. Canada has never imposed transitional safeguards on textiles and clothing authorised under the ATC. In 2002, Canada did negotiate bilateral restraint agreements with China and Taiwan following China’s WTO accession. In 2001, 49 per cent of clothing imports entered under quotas and Canada was restraining imports of textiles and clothing of 32 WTO members. Table IV.4 of the Review lists 53 country/product combinations with quota utilisation rates exceeding 90 per cent in 2001. Presumably these represent products constrained by the quota that will benefit from their elimination.

78


Effective from 1 January, 2003, Canada has provided 48 Least Developed Countries (LDCs) duty-free and quota-free access to the Canadian market. While the preferential treatment extends to all goods, it is particularly relevant in textiles and clothing, important export products for poor countries. There is already evidence that concessions by Canada have paid dividends – Canadian imports from Bangladesh of clothing and textiles more than doubled from 2002 to 2003. Prior to 2002, Bangladesh was one of the few LDCs that filled their quotas. It seems likely that LDCs would have increased their exports to Canada by even more, were it not for the restrictions on input-sourcing due to rules of origin under the LDCT regime.13 Canada’s preferential agreements such as NAFTA, CCFTA and CCRFTA also liberalise trade in textiles. The CCRFTA calls for Canada to eliminate its quantitative restrictions on clothing and textiles originating in Costa Rica. However, under each of these agreements, the ‘yarn forward’ rule of origin requires that yarn used in apparel must originate within an agreement partner country. Man-made fibre sweaters must have fibres originating within the region. The Agreements allow specific quota amounts (the Tariff Preference Level) of nonoriginating clothing and textiles. The Review comments that NAFTA quotas have been mostly filled, suggesting that they have had a binding effect on commerce.

8. CONCLUDING REMARKS

In her summary observations, the Chair of the Review, Mary Whelan, states that: . . . adjustments in [Canada’s] trade policies have confirmed that its trade regime is amongst the world’s most transparent and liberal, notwithstanding barriers to imports in a few albeit important sectors (p. xvii).

This liberal trading regime has enabled Canada to enjoy high levels of imports and exports and, arguably, maintain economic growth. While Canada’s share of world imports exceeds Canada’s share of world income, our benchmarking exercise reveals that Canada’s import performance is about what might be expected given its size and location. While it is heavily oriented towards the United States, there are some signs of diversification. Since 2000, Canada’s trade with the US has declined (in Canadian dollars) while its trade with least developed and other developing countries has risen. While Canada does indeed impose policies that distort trade, it does not appear to be more protectionist than other similar countries. It has signed a number of 13 In their study of the Africa Growth Opportunity Act, Mattoo et al. (2003) estimate that US imports from Africa would have grown by nearly five times as much if the exporters had not been obliged to adhere to the strict rules of origin in that agreement.


79

regional trade agreements but so do most other countries. The evidence to date shows North American free trade has led to substantial trade creation, along with a smaller amount of trade diversion. Canada initiates only slightly more antidumping cases than one would expect based on its imports. Protection exists for agri-food products but its agri-subsidies are low relative to other OECD countries. Barriers to trade in textiles are being dismantled and the Least Developed Countries have enjoyed nearly free access to Canadian markets since 2003. In Canada’s report to the Review team, it states: The WTO is the cornerstone of Canada’s trade policy and the foundation for our relationships with trading partners (p. 152).

Canada continues to make liberalisation efforts in the context of WTO initiatives, including the Doha Round. This behaviour suggests that the recent increase in RTA formation and the continuing high level of trade integration with the United States do not signal diminished commitment to the multilateral trading regime.

REFERENCES Clausing, K. A. (2001), ‘Trade Creation and Trade Diversion in the Canada-United States Free Trade Agreement’, Canadian Journal of Economics, 34, 3, 677–96. Frankel, J., E. Stein and S.-J. Wei (1998), ‘Continental Trading Blocs: Are they Natural or Supernatural?’, in J. Frankel (ed.), The Regionalization of the World Economy (Chicago: Chicago University Press), 91–113. Gerber, J. (2000), ‘Maquiladoras in Transition: It’s Not Just Cheap Labor Anymore’, Cross-Border Economic Bulletin (September), Online access at . Head, K. and J. Ries (2003), ‘Free Trade and Canadian Economic Performance: Which Theories Does the Evidence Support?’ in R. G. Harris (ed.), North American Linkages: Opportunities and Challenges for Canada (Calgary: University of Calgary Press). Krugman, P. (1991), ‘The Move to Free Trade Zones’, Policy Implications of Trade and Currency Zones (Federal Reserve Bank of Kansas City). Mattoo, A., D. Roy and A. Subramanian (2003), ‘The Africa Growth and Opportunity Act and its Rules of Origin: Generosity Undermined?’, The World Economy, 26, 6, 829–51. Romalis, J. (2004), ‘NAFTA’s and CUSFTA’s Impact on International Trade’ (University of Chicago GSB, manuscript). Rose, A. (2004), ‘Do WTO Members Have More Liberal Trade Policy?’, Journal of International Economics, 63, 209–35. The Economist (2003), ‘The WTO Under Fire’ (18 September, www.economist.com). World Trade Organisation (2003a), Trade Policy Review: Canada 2003 (Geneva: WTO). World Trade Organisation (2003b), World Trade Report 2003 <www.wto.org> (Geneva: WTO).

THE WTO REVIEW OF THE MALDIVES

81

5

Trade Policy Making in a Small Island Economy: The WTO Review of the Maldives Prema-chandra Athukorala

1. INTRODUCTION

HE Trade Policy Review of Maldives 2003 (henceforth referred to as TPRM03) is the first review of the country under the Trade Policy Review Mechanism of the World Trade Organisation (WTO). It will be valued by the trade policy analysts not only for the purpose of assessing the Maldives’ performance in meeting multilateral rules and disciplines as a WTO member, but also as a valuable contribution to the literature on the characteristic and development issues of small developing economies. It is a highly timely study given the emphasis placed in the current policy debate in the WTO and other international fora on ‘especial’ problems faced by small economies in the world trading system. In March 2002, the WTO inaugurated a Work Programme on Small Economies (WPSE) in pursuance of a decision made at the fourth Ministerial Meeting held in Doha. The objective of WPSE is to frame policies to address trade-related issues relating to successful integration of small economies into the multilateral trading system. Reconsideration of ‘special and differential treatments’ (SDTs) for small economies in the light of the recommendations arising from the WPSE is likely to figure prominently on the agenda of the Doha Round trade negotiations. Following the Doha Ministerial Declaration, a number of multilateral and regional financial institutions and international agencies also have launched various work programmes to deal with special problems believed to be characteristic of small-state economies (Commonwealth Secretariat/ World Bank, 2000). Insights gained from in-depth country case studies like TPRM03 are vital for assessing these programmes and informing the related policy dialogue.

T

81

82

PREMA-CHANDRA ATHUKORALA

The purpose of this paper is to examine the trade policy regime and traderelated development issues in the Maldives based on the TPRM03. Section 2 provides a stage-setting overview of the salient economic characteristics of the Maldivian economy and its performance over the past two decades. Section 3 examines main facets of current trade policy regime, with emphasis on differences between reform commitments and actual implementation. Section 4 provides a critical evaluation of some selected issues central to the future course of the preform process and growth prospects of the Maldivian economy. The final section presents some concluding remarks. Throughout the paper, an attempt is made to place the Maldivian experience in the context of the existing body of knowledge on structural features and trade-related developmental issues in small developing economies.1

2. TRADE AND DEVELOPMENT IN THE MALDIVES: AN OVERVIEW

The Republic of Maldives is an archipelago of 1,192 small coral islands covering a land area of 115 square miles, located about 300 miles south of the Indian subcontinent (Table 1).2 With a population of 277,000 (in 2002), it is the eighth smallest among the 147 member countries of the WTO.3 The Maldives became an original member of the WTO on 31 May, 1995, having been a contracting party to the General Agreement on Tariffs and Trade (GATT) since 19 April, 1983. From 1990, the GDP per capita of the Maldives has persistently exceeded the minimum cut-off point (US$1,035) used by the World Bank in demarcating lower-middle income countries. However, according to the United Nations’ country classification, which forms inter alia, the basis for according ‘special and differential treatments’ under various WTO agreements, it is still considered a least developed country (LDC), on the grounds of economic disadvantages arising from the small economic size and narrow domestic production base. From 1970, the earliest year from which national account statistics are available, up to about mid-1985, the Moldivian economy remained sluggish with GDP per capita hovering around US$250. From then on, the economy has expanded rapidly, lifting GDP per capita beyond the US$1,000 mark in 1990 and doubling that level by 1997 (Table 2). Over the past two decades, the Maldives’ per capita 1

The key recent works on this subject include Alesina and Spoalare (2003), Armstrong and Read (1998), Easterly and Kraay (2000) and Srinivasan (1986). For a useful survey see WTO (2002a). The term ‘small economy’ and ‘microstate economy’ are used interchangeably in this paper. 2 The Maldives, formerly a sultanate under the British protection, became a republic briefly in 1953 and a monarchy again until a republican government returned in 1968. 3 The other seven countries in descending order (with population (thousands) circa 2000 in brackets) are Belize (240), St Lucia (156), Vanuatu (197), St Vincent and Grenadine (115), Grenada (98), Antigua and Barbuda (68), St Kitts and Nevis (41). (Source: World Bank, World Development Indicators 2002.)

US$2274 7.5 100.0 16.4 9.4 6.4 15.3 8.4 74.7 32.9

82.7 17.3 9.6 6.8

GDP per capita (2002) GNP per capita annual growth (1995–2002), per cent

Composition of GDP (2001–02 average), per cent Primary production Agriculture Fishing Industry Manufacturing Services Tourism

Composition of exports (goods + services), per cent 2000/1 Services (predominantly tourism) Merchandise exports Marine products Clothing

85 99 36

77th 29th

71 21 99

45.0 28.0 13.0 55.0 16.0 31.0

Composition of Government revenue (2000), per cent Tax revenue Import duties Tourism taxa Non-tax revenue (excluding grants) Profit transfers from SOEs Land-lease rent on tourist resorts Life expectancy at birth (2000), years Infant mortality rate, per 1,000 (2000) Adult literacy (per cent) (2000) UN Human Development Index (HDI) (2000) Ranking: Overall Among medium-HD countries School Enrolment ratio (per cent) (2000) Pre-school Primary Secondary

164.3 79.8 84.5

Trade dependence (2000/1), per cent Total trade/GDP Export/GDP Imports/GDP

Source: Compiled from WTO (2003), Table 1.1; ADB, Key Indicators of Developing Asian Pacific Countries (www.adb.org), and World Bank, World Development Indicators (www.worldbank.org).

a

Note: Revenue from the bed tax on foreign tourists (US$6 per bed-night).

300 281 88 1.9 27.5

Land area, sq. km Population, thousand; as of July 2002 Labour force, thousand in 2001 Annual population growth (1900–2002), per cent Urban population, per cent of total population

TABLE 1 Maldives: Key Economic and Social Development Indicators

THE WTO REVIEW OF THE MALDIVES 83

1,848 7.9 6.2 26 28.9 31.4 −2.5 17.7 −1.6 77.6 3.2 35.6 3.2 11 6.5 11.7

1,323 8.3 7.3 6.5 22.25 31.95 −8.3 15.3 −1.1 36.15 3.4 34.4 4.1 5 7.9 10.8

1996

31 32.3 −1.4 14.9 −6.8 99.7 3.4 33 6.9 11 6.5 11.7

2,005 9.1 7.6 23.1

1997

31.8 37.4 −5.1 16.2 −4.5 119 4.1 35.8 3.5 12 6.6 11.8

2,069 9.6 −1.4 22.8

1998

34.1 41.2 −7.3 16.7 −15.3 127 3.8 33.2 3.9 12 5.8 11.8

2,215 7.2 3 3.6

1999

34.8 43.2 −7.7 20 −9.5 123 3.8 31.8 4.2 13 8.9 11.8

2,304 4.8 −1.2 4.1

2000

35.5 44.4 −8.5 21.4 −11.2 93 2.9 31.9 4.4 12 7.2 12.2

2,272 3.4 0.7 9.1

2001

38 47.7 −9.2 18.7 −8.6 133 4.6 29.3 4.2 12 7.3 12.8

2,279 6.0 0.9 19.3

2002

Source: Compiled from WTO (2003), various tables; ADB, Key Indicators of Developing Asian Pacific Countries (www.adb.org), and World Bank, World Development Indicators (www.worldbank.org).

GDP per capita at current market price (US$) Real GDP growth, per cent Inflation (CPI), per cent Broad money (M2) growth, per cent Fiscal operation (per cent of GDP) Government revenue (including grants) Government expenditure Overall budgetary surplus/deficit Government domestic debt (per cent of GDP) Current account balance (per cent of GDP) International reserves, US$ million Months of imports (c.i.f .) External debt (per cent of GDP) External debt service ratio, per cent FDI inflows (US$ million) Per cent of gross domestic fixed investment Exchange rate, rufiyaa per US$ (monthly average)

1990–95

TABLE 2 Maldives: Economic Performance

84 PREMA-CHANDRA ATHUKORALA


85

income has been the highest among all countries in South Asia. Rapid growth has been accompanied by low inflation and rapid reduction in unemployment. There has also been a notable improvement in social development indicators (Table 1). With a literacy rate of 99 per cent, an infant mortality rate of 21 per 1,000 and an average life expectancy of 71 years, by 2000 the Maldives was in the top onethird of the UN ranked ‘medium human development’ countries. Fishing is the traditional mainstay of the Maldivian economy. With only 3,000 hectares of arable land, most agriculture is subsistence farming of mainly native crops, mostly coconuts, yam and fruits. The only significant traditional manufacturing activities are fish processing and shipbuilding. Other manufacturing activities include small-scale production of soft drinks and foodstuffs, building material and some light manufactured goods. Over the past one-and-a-half decades, tourism has gained importance as the most dynamic line of economic activity. There has also been a notable diversification of the fishing industry away from traditional dry fish production and towards fish processing for developedcountry markets. More recently, the clothing industry has become a major source of export earning and employment generation with the entry of ‘quota hoping’ investors from other countries in the region. Annual tourist arrivals reached 461,000 in 2001, almost a five-fold increase from the levels in the mid-1980s.4 By 2001/2, tourism directly accounted for over 34 per cent of GDP (up from 18 per cent in 1990) and provided employment for over one-fifth of the labour force. Tourism also has substantial links to other sectors of the economy, such as construction, transportation, telecommunication and distribution, which together accounts for over 50 per cent of GDP. The relative importance of fishing declined substantially from 30 per cent of GDP in the mid-1980s to 6 per cent in 2001/2, but this sector still accounted for 25 per cent of total employment. The share of manufacturing in GDP increased from about 6 per cent in 1990 to 9 per cent in 2001/2, due predominantly to the expansion of the export-oriented garment industry. Manufacturing accounted for over 18 per cent of total employment in 2000. As in most other small economies, a critical structural feature of the Maldivian economy is its heavy trade dependence. The ‘trade coefficient’ (total trade (including services) as a percentage of GDP) stood at 164 by 2002. Over half of the total catches of fish and over 90 per cent of clothing production are exported. Such a high degree of trade dependence generally implies a high degree of exposure of the economy to exogenous shocks and high sensitivity to developments in the global trading environment (Armstrong and Read, 1998; and World Bank/Commonwealth Secretariat, 2000). However, the level of economic activity in the Maldivian economy has been much less volatile compared to many other microstates (WTO, 2002b, Table 5A). This is presumably because the fortune 4

Unless otherwise stated, data reported in the text come from WTO (2003).

86


of tourism, the prime source of economic expansion since the mid-1980s, is remarkably less affected by the world business cycle. Given its geography, the Maldives has also been less susceptible to vagaries of the weather and natural disasters compared to other microstates. Services exports (predominantly earnings from tourism) accounts for over 80 per cent of total foreign exchange earnings. Total merchandise exports (17 per cent of total foreign exchange earnings) come from marine products (9.6 per cent) and clothing (6.8 per cent). The current account in the balance of payments has been in deficit for the past ten years, despite the widening net surplus in the services account resulting from rapid expansion in tourism. The current account deficit is mostly financed through foreign aid and concessionary borrowing from international development finance agencies, and private capital flows, mostly foreign direct investment. The narrow population base, coupled with the low educational level of the domestic labour force, has led to a high proportion of expatriates in the workforce of the country. Expatriate workers account for over 20 per cent of the workforce and they a play a key role in various high- and middle-level occupational categories as well as in semi-skilled and unskilled occupations such as domestic helpers and construction workers. Total unrequited outward transfers, which reflect predominantly outward remittances by migrant workers, increased persistently from around US$5 million (1 per cent of GDP) in the early 1990s to over US$30 million in 2001 (5 per cent of GDP). The government revenue structure of the Maldives is much in line with that of the other microstates. Taxes on foreign trade (including tourism) are the predominant source of government revenue. In 2000, import duties accounted for 28 per cent of total government revenue. Tourism-related (the bed tax on foreign tourists (13 per cent) and rentals on leased island resorts (31 per cent)) accounted for over 45 per cent. The balance consisted largely of profit transfers from public enterprises whose activities are also closely related to foreign trade. There are no taxes on personal income, capital gains, business profits (other than a bank profit tax), wealth or real estate. Given the narrow domestic tax base, rapid expansion in government development expenditure in recent years5 has resulted in a widening budget deficit. The budget deficits in the Maldives have been historically financed through recourse to foreign concessionary borrowing. But, recent years have seen an increased reliance on borrowing from the Maldives Monetary Authority (the central bank). This has reflected in a sharp increase in money supply, but domestic inflation remains rather low because of the national currency, rufiyaa (Rf.), remain in virtual fixed parity against the US dollar (see below on exchange rate policy). 5

The major contributing factor to the jump in expenditure was the government’s Hulhumale project, a large-scale infrastructure initiative to create a land mass and develop a new town on an island near the capital city of Male.


87

Foreign direct investment (FDI), mostly in tourism, fisheries and the garment industry, has played an important role in recent economic expansion. Average annual FDI inflows increased from US$5 million during 1985–1995 to $12 million during 1997–2000. The total stock of FDI as at March 2002 was US$65 million, 30 per cent of which was in manufacturing and the balance in hotels, utilities and services. About half of tourist resorts are fully foreign-owned or run as joint ventures with foreign firms. Over 84 per cent of manufacturing investment was in garments, with cement (7 per cent) and boat building (8 per cent) accounting for the balance (WTO, 2003, Table 1.5). About two-thirds of the stock of non-tourism FDI is in joint ventures with SOEs involved in public utilities (mainly in telecommunication and water/sewerage).

3. TRADE POLICIES AND PRACTICES

Given the small domestic market and the narrow domestic resource base, there is little room in the Maldives for using tariffs and other trade restrictions as a means of promoting import substitution and infant industry. However, substantially high tariffs maintained for revenue reasons and direct import restrictions that favour state trading corporations have continued to remain important features of the economic landscape of the country, notwithstanding the emphasis placed on outward-oriented, private-sector-led growth strategy since 1989. This section discusses the key elements of the current trade and investment policy regime. a. Import Policy Regime Tariffs are the main instrument of border protection in the Maldives. Except in the case of cigarettes, on which a specific duty of Rf.0.30 per stick (introduced in 2000 in place of a 50 per cent ad valorem tariff), all other duties are ad valorem tariffs levied on CIF (cost, insurance and freight) import value (Table 3). The present structure of applied ad valorem tariffs (in 2002) has ten bands: duty fee, 5 per cent, 10 per cent, 15 per cent, 20 per cent, 25 per cent, 35 per cent, 50 per cent, 100 per cent and 200 per cent. The simple average tariff is 20.8 per cent,6 with a simple average degree of dispersion (measured by the coefficient of variation) of 60 per cent. The three main staple foods – flour, rice and sugar – and all imports intended for commercial re-export (including inputs used in export production) enter the country free of duty. Only 3 per cent of the total tariff lines have rates exceeding three times the simple average tariff rate (domestic tariff peaks). These high tariffs apply, inter alia, to automobiles, beverages, textiles

6

This is the second lowest average tariff rate in South Asia after Sri Lanka (World Bank, 2003).

Simple Average 18.0 18.8 24.4 20.8 20.8

0–50 15–25 10–25 0–200 0–200

Range

Source: WTO (2003), Table III.1, Chart III.2 and Table AIV.

b

a

Notes: At the eight-digit level of the harmonised system (HS). Domestic tariff peaks are defined as those exceeding three times the overall simple average NFN rate. c International tariff peaks are defined as those exceeding 15 per cent.

Agriculture Fishing Mining and quarrying Manufacturing All sectors

Applied tariff rate by major economic sectors

5

4

Number of tariff lines Bound tariff (per cent of lines) Simple average bound rate Agricultural, mining and quarrying Manufacturing Distribution of applied rates by tariff bands

1 2 3

33.3 26.1 12.3 62.0 60.1

CV

TABLE 3 Tariff Structure of Maldives (MFN Tariff)a, 2002

Domestic Tariff Peak b 0 0 0 1.5 1.4

Tariff (Band) 0 5 10 15 20 25 35 50 100 200

65.0 59.9

International Tariff Peaksc

5,321 100 39 41 39 Per Cent of Total Tariff Lines 0.1 3.9 14.6 21.5 11.7 43.1 3.4 0.2 1.3 0.1

88 PREMA-CHANDRA ATHUKORALA


89

and clothing. Peak/prohibitive rates (50 per cent, 100 per cent, 200 per cent) are used largely for environmental and/or human safety purposes (such as on imported plastic bags and motor vehicles). As part of its WTO commitments, the Maldives notified bound rates on all tariff lines in 1995. Some 3 per cent of tariff lines (mainly meat, alcoholic beverages, tobacco products, plastic bags, passenger motor vehicles, buses, motorcycles and their components) were bound at 300 per cent, and the rest mostly at 30 per cent. The simple average bound rate is 39 per cent, with rates for agricultural products and manufactured goods averaging, respectively, to 41 per cent and 39 per cent. The Maldivian bound tariff rates are on average much lower than the overall South Asian levels, with the sole exception of those in Sri Lanka (World Bank, 2003). However, the Maldives has so far failed to keep applied rates within bound levels. Applied rates for 149 tariff lines exceed the bound rate of 30 per cent by between 5 and 170 percentage points, with the majority of differences clustering at the upper end (TPRM03, p. 36). In other cases, bound rates are much higher (on average by 18 percentage points) than the applied rates. This gap imparts a degree of uncertainty to the tariff by giving policy makers ample room for arbitrary changes. The Government of the Maldives makes extensive use of import duty concessions as part of the government policy to promote domestic and foreign investment in export-oriented manufacturing and tourism. Under the Maldivian Customs legislations, the President may grant duty-free import concession for any ‘economically productive activity’ for a period of up to ten years. He may also allow items to be imported duty free under ‘special circumstances’ when it would benefit the nation. There are no precise estimates, but it is believed that these concessions amount to well over one-third of potential tariff revenue (WTO, 2003, pp. 36–37). Fishing and tourism sectors have been the major beneficiaries of import duty concessions. Apart from the revenue losses involved, these concessions seriously undermine the transparency of the tariff structure. As regards the sectoral distribution of tariffs, there are no significant differences in tariff rates applicable to agriculture, other primary production and manufacturing (Table 3). Moreover, data on simple average tariffs by stage of processing at the two-digit level of the International Standards Industry Classification (ISIC) (WTO, 2003, Table III.2) do not reveal a clear pattern of tariff escalation (that is, a tendency to tax inputs to the production process at lower rates and final goods at higher rates). Only in two of the nine ISIC industries is the tariff on fully processed products somewhat higher (by about 2 to 5 percentage points) than first-stage processed or semi-processed products. In the remaining cases tariff rates are remarkably similar or indicate the reverse pattern.7 Both these features 7

Thus the inference made in TPRM03 that ‘[T]he Maldives’s tariff structure has a degree of builtin escalation’ is not consistent with the data reported therein.

90


of the tariff structure – the absence of significant difference in tariff rates across sectors/industries and the absence of clear evidence of tariff escalation – are consistent with the fact that tariffs in the Maldives are determined primarily, if not solely, by revenue considerations, not by considerations of industry protection. b. Non-tariff Barriers Importation of a small number of items (alcoholic beverages, pork, used cars and motorcycles) are prohibited or restricted for security, environmental or religious reasons. Import quotas apply to the three major staple foods (rice, flour and sugar) which together account for around 20 per cent of the total import value. Around 70 per cent of these imports are reserved for the state-owned State Trading Organisation (STO). The (officially declared) objective of this policy is to ensure adequate supply of these products at affordable prices. Importation of all other goods requires an Open General Licence (OGL). However, OGLs are issued automatically for a certain value, and permission for additional amounts is normally granted. c. Export Policy Regime Since 1996, there have been no export taxes other than a 50 per cent duty on exports of ambergris8 and export controls on timber. However, garment manufacturers pay a royalty on exports which is akin to an export tax (as it does not apply for domestic sales). Until recently, the royalty rate was 3 per cent of export turnover. It was replaced in August 1999 with specific rates ranging from US$0.025 to US$0.10 per piece, depending upon the clothing category, and in a few cases depending on the market destination and the material used (i.e. cotton or nylon) (WTO, 2003, Table II.3). The foreign firms, which dominate the garment industry, have established production plants in the Maldives with a view to exploiting ‘quota rents’ in developed-country markets arising from import restrictions under the Multifibre Arrangement (MFA). The quota rents are believed to be very high compared to the minute royalty charged by the Maldivian authorities. Thus the existing royalty presumably has only a marginal effect on export profitability. But, complexity of the royalty rate schedule, which has 30 different rates applicable to exports to different markets and clothing produced using different material in an overlapping fashion, seems to impose unnecessary transaction costs to exporters. A royalty of 2 per cent is charged on the sales turnover of both local and foreign companies involved in fishing within the 200-mile exclusive economic

8

A wax-like substance present in the intestines of whales, used as a fixative in perfume.


91

zone (EEZ) of the country. This is a resource rent rather than an export tax (which applies to both exports and local sales). A fixed tax of US$6 per bed-night is levied on foreign tourists. As discussed, this has become an important source of government revenue in recent years. d. Trade in Services Relating to trade in services, the Maldives’ sectoral commitments under the GATS cover very few sub-sectors in business services (i.e. certain professional services, computer and related services). There is no specific market access or national treatment commitments affecting services such as accountancy, auditing, bookkeeping, data processing, database, software implementation and consultancy services relating to hardware installation. The Maldives did not participate in the WTO negotiations on basic telecommunication service (Fourth Protocol) or the extended WTO negotiations on financial services (Fifth Protocol). e. Other Measures Affecting Production and Trade (i) Exchange rate and the external payments regime The Maldives became a member of the International Monetry Fund (IMF) in 1978. Since then its national currency, rufiyaa, has been pegged to the US dollar. The peg has been adjusted marginally from time to time to avert major misalignments,9 but overall the Maldivian exchange rate regime can be described as a de facto fixed rate regime. Given the high degree of openness of the economy and the ‘thin’ currency markets, which constrain smooth economic adjustment under a flexible exchange rate, the current exchange rate regime appears to be the optimal choice for the Maldives (Corden, 2002). The rufiyaa is convertible on both current and capital account transactions. Both residents and non-residents can freely hold foreign currency accounts. Foreign investors are permitted to repatriate profits and investment proceeds freely. There are no restrictions on offshore borrowings by both local and foreign firms. (ii) The FDI regime The policy regime for FDI in the Maldives is very liberal. The Law Governing Foreign Investment (No. 25/79), promulgated in 1979, covers foreign investment in business enterprises. The Law on Doing Business in the Maldives by Foreign Nationals, 1979 (No. 4/79), also promulgated in 1979, covers foreign investment in trade (imports and exports) and related services. The Ministry of Trade and Industry, through its Foreign Investment Services Bureau (FISB) handles 9

The latest adjustment was in July 2001 when the rufiyaa was devalued by 8.5 per cent against the US dollar.

92


registration of foreign investment in all sectors, except tourism, which is the responsibility of the Ministry of Tourism. Full foreign ownership of investment projects is allowed in all sectors and activities. As already noted, foreign investors are permitted to repatriate capital and profits in full. Employment of foreign workers in approved projects is permitted when workers with required skills are not locally available. Most investment incentives apply equally to foreign and domestic investors. The incentives include duty-free imports of machinery, construction material and capital goods for the first year, and for two years on large projects (above US$200,000), dutyfree access to raw material used in export production (over half the total tariff revenue forgone from duty concessions accrues to foreign investors (US$11.8 million in 2001)). The Maldives became a member of the World Bank’s Multilateral Investment Guarantee Agency (MIGA) on 26 March, 2002, and is aiming to ratify the Convention for the Settlement of Investment Disputes. (iii) Privatisation policy State involvement is pervasive in the Maldivian economy. Twenty state-owned enterprises (SOEs) (twelve of which are fully state owned and the others have minority private sector participation) operate in many important sectors of the economy, including trade, banking, construction, insurance, fishing and utilities. They employ around 8 per cent of the total labour force. Privatisation of SOEs is a key element of the new private-sector-oriented development strategy in the Maldives. A major privatisation programme was launched in 1999. The programme aims to restructure and corporatise SOEs under the Companies Act of 1996, with a view to eventual privatisation. Some SOEs have already been partially divested and implicit subsidies to some SOEs have been reduced. Seven state-owned enterprises, including Air Maldives, have been opened to minority private participation. There is a plan to abolish the telecommunication monopoly in 2008. Exclusive trading rights of the State Trading Organisation (STO), the largest SOE in terms of sales turnover and also by far the largest importer in the country, were abolished in the late 1990s. As a further step to reduce STO trading monopoly, government procurement of goods and services above Rf.5,000 are now made through open competitive tender. The fish processing and exporting monopoly of the Maldives Industrial Fisheries Company (MIFCO) monopoly was abolished at successive stages starting in 2000.10 Greater competition following privatisation has resulted in an increase in export volume and diversification of exports into fresh/chilled tuna, including

10 Until 1990, fish exports remained under the exclusive monopoly of MIFCO. In that year private sector firms were allowed to trade in dried/salted tuna. This was followed by liberalising trade in yellow-fin tuna in 1996 and trade in skipjack in 2000.


93

exports to the Japanese and East Asian sashimi markets. Some SOEs, including MIFCO and Air Maldives, have been a drain on the government budget. Despite these notable initiatives, the implementation of the privatisation programme has fallen short of the original expectations. For instance, over 30 per cent of total import trade is still handled by the STO. As already noted, 70 per cent of import quotas of rice, wheat flour and sugar is reserved for the STO. The government purchases over 80 per cent of its goods and services requirements from the STO and other SOEs. Despite recent attempts to deregulate fish processing and export trade, the Maldives Industrial Fisheries Company (MIFCO) is still the dominant player in this sector. The delayed and cautious approach to privatisation reflects resistance arising from concern about the possible anti-competitive effects of privatised firms in key activities, which would dominate the small domestic market. It also reflects the need to strengthen the judiciary and the legal system as well as to broaden the tax base (WTO, 2003, pp. 40–41). Another, and perhaps the most important, explanation (which is not discussed in TPRM03) is the importance of SOEs as a vehicle for dispensing political patronage. In particular, most SOEs provide employment to those whom the government wants to reward on political grounds. (iv) BTAs, FTAs and GSP The Maldives has a bilateral trading agreement (BTA) with India. This BTA, signed on 31 March, 1981, is basically a ‘Most Favoured Nation’ (MFN) agreement, which has so far had little impact on trade between the two countries. The Maldives is a founding member of the South Asian Association for Regional Cooperation (SAARC) Preferential Trading Agreement, signed in April 1993 and came into effect in December 1995. SAPTA was notified to the WTO in 1997 under the Enabling Clause. In principle, it covers all products – manufacturers and commodities – in their raw, semi-processed and processed forms. Tariff preferences under SAPTA apply to 390 tariff lines, covering mainly vegetable oils, leather or textile products, timber and timber products, footwear, headwear and umbrellas. The overwhelming majority of these tariff lines have preference margins of 7.5 per cent and the balance of 10 per cent and 15 per cent, of the MFN related rates. Most of these 390 products have MFN rates of 25 per cent in the Maldivian tariff schedule. TPRM03 has not undertaken any analysis of the impact of these trade preferences on the Maldives’ trade performance, but the actual impact is unlikely to be substantial because most of the significant preferential rates relate to commodities which are rarely traded between the Maldives and the other South Asian countries. The EU and Japan provide the Maldives with trade preferences under the Generalised System of Preferences (GSP). The USA, the major export market, does not extend GSP privileges to the Maldives. Owing to rules of origin and product exclusions, the main export from the Maldives that benefits from GSP is

94


processed (canned) fish exports. The other major export product, clothing, does not qualify because of the low domestic input content (less than 30 per cent in most cases). The current average tariff on fish product imports to the EU is 4.5 per cent. Fish products from the Maldives are allowed at the preferential rate of 1.6 per cent.

4. PROBLEMS OF GLOBAL INTEGRATION

As we have seen in the previous section, the trade and investment policy regimes in the Maldives have undergone noteworthy reforms, but much remains to be done to set the stage for economic expansion through global economic integration. What are the constraints faced by the Maldivian authorities in translating the promised reforms into action? Is there any need for reconsidering reform priorities in the light of recent changes in the international economic scene? How can the international development community assist the Maldives in the reform process? A full treatment of these issues was beyond the terms of reference of the TPRM03, but it has drawn attention to a number of related themes. a. Budgetary Implications of Tariff Reforms As in many other developing countries, in the Maldives government revenue implications are a major concern often voiced as an argument against further reduction/rationalisation of import duties. It is argued that import duties are an important source of government revenue, and the speed with which customs duties are reduced needs to be determined in accordance with the speed and effectiveness of domestic tax reforms aimed at overhauling the revenue structure to compensate for the lost revenue. Otherwise the anticipated economic gains from tariff reforms could well be erased by adverse budgetary implications arising from revenue shortfall. The discussion on the future direction of tariff reforms in the TPRM03 is much in line with this conventional view. It simply emphasises the need for broadening the internal tax base to reduce the government’s heavy reliance on border taxes which would facilitate further tariff reductions, without examining (at least analytically) revenue effects of tariff reduction. Broadening the internal tax base to reduce the government’s heavy reliance on border taxes would, of course, facilitate further tariff reductions. In that sense, tariff and domestic tax reforms are intertwined. Moreover, as has been argued in the TPRM03, unlike tariff on foreign products, a true consumption tax, such as the conventional value-added tax would be non-discriminatory and tax both imports and domestically produced goods equally. In particular, a broad-based consumption tax or a value-added tax (VAT) with the usual adjustments to exclude exports would


95

ensure neutrality with regard to exports, and the potential economic benefit in improved efficiency from such a switch may be substantial. However, the argument that tariff reforms need to wait for domestic tax reforms ignores the important point that higher tariffs do not necessarily imply higher revenue (Pritchett and Sethi, 1994). There are strong reasons to believe that tariff reduction could result in less than one-to-one reduction in revenue collected, and in some instances even in an increase in revenue. First, high tariff levels are usually prohibitive and inhibit imports, so that bringing them down into the range where imports materialise has the potential to increase revenue. Second, the tariff rates being reduced are subject to exemptions that make the reductions effectively less than what they appear to be. We have already noted that the revenue loss resulting from various duty concessions could be as high as one-third of the realised revenue. Third, higher tariffs generally encourage attempts by importers to get their imports classified under lower duty slabs within the given commodity category through various means in order to minimise duty payments. This has become a lucrative practice because of the ongoing process of ‘product fragmentation’ (splitting of manufactured goods, in particular electronic and electrical goods, into separate components) which enables importers to import certain final products as separate components (which are generally subject to lower duty compared to the whole product). Fourth, tariff cuts also have the potential to increase revenue collection through reducing incentives to smuggling. b. Institutional and Financial Constraints on Meeting Reform Commitments The TPRM03 contains an informative discussion on the problems encountered by the Maldives in the reform process, particularly in the implementation of its reform commitments as a WTO member. The Maldives has not yet fully implemented the provisions of the WTO Customs Valuation Agreement. It invoked the five-year transitional period available to developing countries until 31 May, 2000, and subsequently obtained an extension until 31 May, 2002. But the implementation process had not yet been completed (at the time of writing the TPRM03 in mid-2003). There is no competition law or other laws that regulate the activities of enterprises and protect consumer interest which are needed to provide the necessary setting for speedy implementation of the privatisation programme. Intellectual property legislations are yet to be promulgated. As many other small developing countries, the Maldives does not have permanent representation at the WTO in Geneva. Consequently, it has not been able to participate effectively in the rule-making process and to promote and defend its interests in the WTO in accordance with its national policy priorities. The key inference arising from this discussion is that, in most cases, lack of human and institutional capabilities and the financial constraint, rather than lack

96


of reform commitments on the part of the authorities, are the main reason for delays in the implementation process. This inference corroborates the emerging consensus in related policy debate that technical assistance and the capacitybuilding initiative to assist small developing countries in the implementation of reform commitments should be a core element of the development dimensions of the multilateral trading system (Finger and Schuler, 2002; and Hoekman, Michalopoulos and Winters, 2004). It is important to note that, as part of the economic transition through global integration in the post-war era, most developed countries have already put in place institutions and policies to participate effectively in a rule-based world trading system. There is, therefore, a strong case for a well-coordinated international initiative (involving the WTO, international aid agencies such as the World Bank, the IMF, regional development banks and individual donor countries) to assist developing countries in putting in place similar institutions and policies, if we expect them to become equal partners in the process of removing impediments to global economic integration (Bhagwati, 2003). An important limitation of the otherwise comprehensive treatment in TPRM03 of constraints of the implementation of reforms is the absence of any discussion on the implementation issues relating to the WTO agreement on Sanitary and Phytosanitary Standards (SPS Agreement). This should be a pivotal theme of any discussion on the Maldives’ participation in the multilateral trading system, given the heavy reliance of the economy on fish product exports. The Maldives’ success in the expansion of fish product exports depends critically on its ability to meet increasingly more stringent food safety standards imposed in developed countries. Not only are these standards typically much higher and costly to meet than those required for selling on the domestic market, but they are also subject to frequent changes. Moreover, and perhaps more importantly, as traditional trade barriers such as tariff and quantitative restrictions continue to decline, protectionist interests are likely to make increasing use of food safety regulations and other technical barriers to block trade. Meeting international safety standards is far more complicated and costly in the case of processed food than in primary agricultural products. Thus, these standards can retard exports by developing-country producers even when they are imposed on genuine health and safety considerations because of the limited availability of compliance resources. The purpose of the Sanitary and Phytosanitary (SPS) Agreement and the related dispute settlement procedures of the WTO are to strengthen multilateral discipline in the implementation of food-safety standards (SPS standards) in agricultural trade, with a view to achieving the objective of protecting consumers while regulating the use of these standards as a means of non-border trade protection. The developing countries have so far failed to effectively participate in the implementation of the Agreement as equal partners (Finger and Schuler, 2002;


97

and Athukorala and Jayasuriya, 2003). Unlike conventional trade policy reforms, SPS regulations cannot be implemented simply through legislative declaration. There is a need for a global framework to support national capacity building and improve the design of international standards. Quite apart from helping countries when they confront a dispute, there seems to be a great need for international initiatives to educate developing-country exporters and policy makers about the new legislation and on how to comply with internationally adopted food standards. This is certainly an area where there is ample room for fruitful ‘aid for trade’ initiatives by the international development community. The SPS Agreement itself tries to facilitate effective participation of the developing countries in its implementation by encouraging developed-country members to provide technical assistance and accord special and differential treatment to developing countries (Articles 9 and 8). But so far developed countries have failed to take any serious steps in this direction. International organisations, such as the UNCTAD, the ITC and the World Bank have to begin to provide this kind of technical assistance. But, these initiatives are still in their early stages and the technical and financial support provided so far falls below what is required. c. Graduation from LDC Status With the sustained increase in GDP per capita passing beyond the LDC threshold, the Maldivian authorities have become seriously concerned about the implications of losing preferential access to developed-country markets and other ‘special and differential treatment’ associated with the LDC status. The TPRD recognises the imminent graduation from LDC status as a ‘major challenge for the Maldives’, but it stops well short of assessing the potential adverse impact. However, based on the information provided in the report relating to trade patterns and the actual coverage of existing trade preferences one could argue that the fear of losing trade preferences is vastly exaggerated. Trade preferences enjoyed by the Maldives have been by and large limited only to fish (mostly canned and preserved tuna) exports to the EU. However, the share of exports to the EU in total fish product exports from the country has declined sharply in recent years (from 21 per cent in 1997 to 12 per cent in 2001). This suggests that sheltering exports from world competition through preferential tariffs per se is unlikely to ensure export success. What is important is to focus on redressing supply-side constraints on export performance. The exclusive monopoly enjoyed until recently by MIFCO is considered a major factor behind the Maldives’ poor export performance in fish products. Until recently MIFCO’s long-standing monopoly in fish processing and exporting depressed raw fish processing below the levels in world markets, penalising local fishermen and acting as a disincentive for expanding fish catch (WTO, 2003, p. 45). There are indications that gradual relaxation (since 1996) and the final

98


abolition (in 2000) of MIFCO’s monopoly has begun to boost production and improve efficiency in the industry. d. Trade and Environment Given its fragile ecology and peculiar bio-diversity, and the heavy dependence on two environmentally sensitive industries (fishing and tourism), environmental issues are central to the national development policy of the Maldives. Rapid expansion of reef fisheries has already led to over-fishing of certain species, such as grouper, snapper and sea cucumbers. There is anxiety about the long-term sustainability of tourism, particularly given the effects of pollution on the coral on which the island stands (Farmer, 1993, p. 169; and Armstrong and Read, 1998, p. 570). In both areas, trade and foreign investment policy needs to be appropriately linked with national resource management policy and an enhanced regulatory framework for greater surveillance and control of environmental degradation. Unfortunately, the environmental dimension of trade and investment policy is entirely missing in the TPRM03.

4. CONCLUSION

Given the narrow resource base and small domestic market, openness to foreign trade and investment remains the ‘natural’ policy choice for the Maldives. Since the late 1980s the Maldivian authorities have made considerable progress in implementing policy reforms driven by this conviction. The reform process is far from complete, however. High import tariffs maintained predominantly on revenue considerations, a large direct role played by the public sector in foreign trade and some key sectors of the economy, the lack of transparency in duty concessions and other investment incentives, failure to incorporate environmental concerns as part of the national development policy, and delays in the implementation of multilateral reform commitments under the WTO are among the key items of the unfinished reform agenda. The TPRM03 provides a useful framework for examining these and related issues, and designing appropriate policies and strategies by the international development community for helping microstate economies like the Maldives in global economic integration in an orderly fashion. The report is particularly commendable given the severe staffing and financial constraints under which the Trade Policy Review mechanism of the WTO operate. There are, however, many areas in which further work is needed in order to inform the policy debate. These include the political economy of SOE reforms – in particular, the relative importance of genuine economic reasons versus political patronage in determining the nature and speed of the reform process; revenue implications of tariff reforms;


99

the implications for Maldivian fish exports of the SPS agreement and the related institutional mechanisms for monitoring international food safety standards, and the impact of graduation from LDC status for export performance.

REFERENCES Alesina, A. and E. Spoalare (2003), The Size of Nations (Cambridge, MA: MIT Press). Armstrong, H. W. and R. Read (1998), ‘Trade and Growth in Small States: The Impact of Global Trade Liberalisation’, The World Economy, 21, 4, 563–85. Athukorala, P. and S. Jayasuriya (2003), ‘Food Safety Issues, Trade and WTO Rules: A Developing Country Perspective’, The World Economy, 26, 9, 1395–416. Bhagwati, J. (2003), In Defence of Globalization (New York: Oxford University Press). Commonwealth Secretariat and the World Bank (CM-WB) (2000), Small States: Meeting Challenges in the Global Economy, Report of the Commonwealth Secretariat – World Bank Joint Task Force on Small States (Washington, DC: World Bank, www.worldbank.org). Corden, W. M. (2002), Too Sensational: On the Choice of Exchange Rate Regimes (Cambridge, MA: MIT Press). Easterly, W. and A. Kraay (2000), ‘Small States, Small Problems? Income, Growth and Volatility in Small States’, World Development, 28, 1, 2013–27. Farmer, B. H. (1993), An Introduction to South Asia (2nd edn, London: Routledge). Finger, J. M. and P. Schuler (2002), ‘Implementation of WTO Commitments: The Development Challenge’, in B. Hoekman, A. Mattoo and P. English (eds.), Development, Trade and the WTO: A Handbook (Washington, DC: World Bank), 493–503. Pritchett, L. and G. Sethi (1994), ‘Tariff Rates, Tariff Revenue, and Tariff Reforms: Some New Facts’, World Bank Economic Review, 8, 1, 1–16. Srinivasan, T. N. (1986), ‘The Cost and Benefits of Being a Small, Remote, Island, Landlocked or Mini-State Economy’, World Bank Research Observer, 1, 2, 197–202. World Bank (2003), ‘Trade Policy in South Asia: An Overview’, Study prepared under the Poverty Reduction and Economic Management Project, South Asian Region (Washington, DC: World Bank, mimeo). WTO (2002a), ‘Small Economies: A Literature Review’, Committee of Trade and Development WT/COMTD/SE/ W/4 (Geneva: WTO, www.wto.org). WTO (2002b), ‘Trade and Economic Performance: The Role of Economic Size?’, Committee of Trade and Development WT/COMTD/SE/W/5 (Geneva: WTO). WTO (2003), Trade Policy Review of Maldives 2003 (Geneva: WTO).

TRADE REFORM AND THE SOUTHERN MEDITERRANEAN

101

6

Trade Reform and the Southern Mediterranean Michael Gasiorek HE countries of the Southern Mediterranean region are diverse with varying economic structures, institutional features and socio-political frameworks, and therefore with correspondingly varying needs, economic priorities and policies. The countries also vary greatly in size and in their resource endowments. While there are clear differences, there are also common elements. In particular, over the past couple of decades rates of growth of per capita GDP have been disappointingly low, levels of foreign direct investment as a proportion of GDP have also been disappointing, and the pace of domestic institutional reform has been comparatively slow. In terms of policy imperatives possibly the central issue facing many of these economies is the dramatic and ongoing rise in the population and consequently labour force. For example, over 1993–2001 the Egyptian labour force grew by nearly 25 per cent, and over 1995–2002 in Morocco by over 100 per cent.1 In consequence, the growth of per capita GDP over the last two decades for many of these economies has been low and, in more recent years, even negative. Without higher rates of growth many of these economies will experience even higher rates of unemployment as well as any attendant social problems. Achieving higher rates of growth is clearly therefore a central imperative for these economies. One of the key elements in that strategy would increasingly appear to be the pursuit of more outwardly-oriented trade strategies coupled with domestic institutional reform. This reorientation of policy is being followed by many of these countries though with varying degrees of commitment and enthusiasm. On the trade side the policies typically involve a combination of multilateral liberalisation of tariff and non-tariff barriers, as well as increasing involvement in bilateral and regional free trade agreements. In particular, and following the Barcelona Declaration in 1995, twelve countries (Algeria, Cyprus, Egypt, Jordan, Israel,

T

1

Source: ILO Laborstat Database.

101

102

MICHAEL GASIOREK

Lebanon, Malta, Morocco, the Palestinian Authority, Syria, Tunisia and Turkey) are involved in a process of trade liberalisation with their largest and nearest trading partner – the European Union.2 In addition to these Euro-Med agreements several of these countries have recently signed bilateral agreements with the US, for example, Morocco (2004) and Jordan (2002); and are also engaged in regional economic cooperation amongst themselves, such as the Pan-Arab Free Trade Agreement (launched in 1998), and much more recently the Agadir Agreement (2004) which establishes a free trade agreement between Egypt, Jordan, Tunisia and Morocco by 2005. It is thus on different aspects of the role and importance of trade for the Southern Mediterranean region that the three articles in this mini-symposium focus. The article by Konan and Kim focuses on Egypt and Tunisia, and empirically evaluates the potential benefits from different processes of trade liberalisation, and in particular analyses the importance of services as well as goods trade liberalisation. The article by Augier, Gasiorek and Lai-Tong assesses the potential importance of rules of origin and their cumulation in trade between countries party to overlapping free trade agreements. They include a range of Southern Mediterranean countries and focus on textiles, which is a key industry for many of these economies. Finally, Méon and Sekkat, assess the role of the quality of domestic institutions in impacting on both trade and foreign direct investment more broadly in the Middle East and North Africa (MENA). Before discussing some of the key issues and conclusions from the articles it is worth highlighting in a little more detail some key features of the region. Hence, Table 1 presents certain principal economic indicators for a subset of these countries, for 1998 and 2001. First, it can be seen that there are substantial differences in size, with Egypt being the largest economy by far in terms of both GDP and population, and Jordan being the smallest. Each of the economies is classified by the World Bank as lower-middle income where the average GDP per capita in 1998 and 2001 was $1,450 and $1,390 respectively. For the four Southern Mediterranean countries here, income per capita is highest in Tunisia and then Jordan, and for each of these it is higher than the average for the lowermiddle income country group. GDP per capita is lowest in Morocco and then Egypt and for each of these lower than the average for the lower-middle income country group. From the table it can be seen that rates of economic growth ranged from 7.7 per cent for Morocco to 3 per cent for Jordan in 1998, and from 4.9 per cent for Jordan to 1.9 per cent for Tunisia in 2001. Only for Jordan did the rate of growth increase. For each of the other economies there was a significant decrease in growth rates. At the same time over this period each of these economies saw high rates of population growth thus impacting on per capita income levels which fell for Morocco and Tunisia. The high rates of population 2

See the article in this volume by Augier, Gasiorek and Lai-Tong for more details.

61.6 1,270 82.1 4.5 17.4 31.7 50.9 16.8 26.6 26.8 57.3† 1,100 1.3 8.2 9.2

66.4 1,470 89.8 3.0 16.8 35.0 48.1 18.2 23.4 22.0† 4.6 1,590 7.9 3.0 3.0 25.5 71.5 44.8 64.3 22.0 n.a. 310.0 3.9 5.2 1,760 9.3 4.9 2.0 23.4 74.6 45.1 71.7 14.7 n.a.

2001 27.8 1,250 35.8 7.7 17.2 31.6 51.2 27.8 31.9 36.7 58.3† 333.1 0.9 13.7

1998

Morocco

12.5

29.6 1,190 37.3 4.5 16.1 31.1 52.8 29.9 35.9 33.4

2001

9.3 2,050 19.9 4.8 12.5 28.3 59.2 43.0 46.3 30.6 54.0† 649.8 3.3

1998

Tunisia

15.0

9.8 2,000 21.2 1.9 10.4 29.1 60.5 45.5 50.4 34.0

2001

Notes: 1998 data (except on unemployment rates are from WDI 2003), in which: * Per capita income is GNI per capita (formerly GNP per capita) is the gross national income, converted to US dollars using the World Bank Atlas method, divided by the midyear population; ** Inflation as measured by the annual growth rate of the GDP implicit deflator, which is the ratio of GDP in current local currency to GDP in constant local currency; *** Foreign direct investment is the net flows of sum of equity capital, reinvestment of earnings, other long-term capital and short-term capital as shown in the balance of payments. For Israel, 2000, 2001 data calculated from http:// www.cbs.gov.il/engindex.htm. † data for 2002. ‡Non-tariff Measures (NTMs) is calculated as frequency ratio in per cent of all HS two-digit product categories over 1989–94, and includes licensing, prohibitions, quotas and administered pricing. (Source: Constantine Michalopoulos, ‘Trade Policy and Market Access Issues for Developing Countries’, World Bank PRWP No. 2214, October 1999.)

Population (million) Income per capita (current US$)* GDP (current billion $) GDP growth (annual per cent) VA in industry (per cent of GDP) VA in services (per cent of GDP) VA in agriculture (per cent of GDP) Exports of good and services (per cent of GDP) Imports of goods and services (per cent of GDP) Average tariff rate (unweighted) Non-tariff measures‡ Foreign direct investment (current million US$)*** FDI as a per cent of GDP Unemployment

1998

1998 2001

Jordan

Egypt

TABLE 1 Basic Economic Indicators

TRADE REFORM AND THE SOUTHERN MEDITERRANEAN 103

104

MICHAEL GASIOREK

growth over this period alone range from 5.3 per cent for Tunisia, to 13 per cent for Jordan. As can also be seen, these economies have relatively high rates of unemployment with, in 2001, the lowest being in Egypt at 9.2 per cent, and the highest in Tunisia at 15 per cent. Another feature, which emerges from Table 1, is the high proportion of agriculture in GDP. This is particularly so for Jordan at over 70 per cent, with the share for each of the other three economies being around 50 per cent. The share of industry in Jordan is by far the lowest (2–3 per cent), while being 10–12 per cent for Tunisia, and approximately 17 per cent for each of Egypt and Morocco. The importance of agriculture for these economies underlines the importance both of the process of agricultural restructuring taking place and which is likely to be induced through lower trade barriers, and the importance of access to the EU and other developed country markets for agricultural products. Turning now to the role of exports and imports, it can be seen that the most open of the economies here is Jordan where exports of goods and services constitute in the region of 45 per cent of GDP in both 1998 and 2001, while imports as a proportion of GDP rose from 64.3 per cent in 1998 to 71.7 per cent in 2001. The share of exports for Tunisia is very similar at 43 per cent in 1998 and 45.5 per cent in 2001, and the share of imports is somewhat lower at 46.3 and 50.4 per cent respectively. In contrast both Egypt and Morocco, who have been slower to liberalise their respective trade regimes, are clearly less open economies. Egypt’s exports as a proportion of GDP were only 16.8 per cent in 1998 and 18.2 per cent in 2001, while for Morocco the corresponding figures are 27.8 and 29.9 per cent. On the import side the figures for Egypt are 26.6 and 23.4 per cent, and for Morocco 31.9 and 35.9 per cent. The table also gives some information on the presence of trade barriers. Average tariffs are high in all countries and particularly so in Morocco and Tunisia, followed by Egypt and then Jordan. Tariff rates declined substantially for Jordan over 1998–2001, marginally for Egypt and Morocco, and appear to have risen slightly for Tunisia. For 1989–94, the table also gives the frequency ratio for nontariff measures. The frequency ratio gives the percentage of two-digit HS categories for which there is a non-tariff measure in place and is therefore only a crude measure of such barriers as it does not capture the degree of restrictiveness. Nevertheless, the ratio ranges from 54 per cent for Tunisia to 58.3 per cent for Morocco, and compares to an average frequency ratio for low income countries of 37 per cent, and for high income countries of 3.1 per cent. Hence the height of the tariffs, and of the NTM measure confirms the fairly high levels of protection of these economies. It is worth noting, however, that while Tunisia’s MFN barriers are high, trade has been significantly liberalised with regard to its principal trading partner – the EU. Finally, looking at foreign direct investment for Jordan and Tunisia, FDI as a proportion of GDP at 3.9 and 3.3 per cent respectively is slightly higher than the


105

average for lower-middle income countries which is 2.2 per cent. In contrast, for both Morocco and Egypt the levels of FDI as a proportion of GDP are disappointingly low at 0.9 and 1.3 per cent respectively. Interestingly, FDI is lowest in the countries which have liberalised their trade regimes the least. In summary, clearly there is some considerable heterogeneity but there are also some common features. The Southern Mediterranean economies are typically characterised by low growth rates, high levels of population growth, and a high reliance on both agriculture and services, and many of the economies have high tariff and non-tariff barriers to trade, and correspondingly levels of FDI tend to be low. As discussed earlier many of these economies are starting to engage more in process of regional and multilateral trade liberalisation. While the empirical literature on the relationship between trade and growth is not unambiguous, much of the evidence would appear to suggest a positive and largely causal correlation between the two.3 There would appear to be little doubt therefore, that trade and trade policies, will play an important role in the future evolution of the Southern Mediterranean. The aim of this mini-symposium is thus to focus on key aspects of the potential role of trade liberalisation. The paper by Konan and Kim uses computable general equilibrium modelling to focus on the potential gains to Egypt and Tunisia from different possible trade liberalisation scenarios. Where typically CGE evaluations tend to focus on the impact of goods trade liberalisation, this paper allows for both goods and services trade liberalisation scenarios. This is important for two reasons. First, as can be seen from Table 1, the share of services in GDP for these economies is approximately twice as important as the share of manufacturing. Secondly, there are important interactions between economic structure and the liberalisation of the service and manufacturing sectors. Hence Konan and Kim show that the underlying structural features of the economies of Egypt and Tunisia imply that the gains from trade liberalisation are greater for the latter than the former. MFN liberalisation leads to a potential welfare gain for Tunisia of 4.27 per cent of GDP, whereas for Egypt a much more modest gain of 0.46 per cent. The central explanation for these differences (reflected also in Table 1), is that Tunisia is a much more open economy importing a wide range of manufactured goods, and with exports concentrated largely in textiles. In contrast, Egypt is a much more closed economy both in terms of imports and exports, and hence benefits less from the process of goods trade liberalisation. Most interestingly also, while Tunisia gains from participation solely in the Euro-Med agreements, Egypt suffers a small welfare loss. Again this can be understood with reference to the underlying structural features of the two economies. For Tunisia the EU is by far the most important trading partner, whereas for Egypt this is far less the 3

See, for example, Dollar (1992), Edwards (1998), Rodriguez and Rodrik (2000), or Winters (2004) for useful discussions on this.

106

MICHAEL GASIOREK

case. Trade liberalisation with the EU thus implies much more trade diversion for Egypt, hence the welfare loss. When the authors allow for both goods and services trade liberalisation the welfare gains for Tunisia and Egypt rise to a potential 8.32 and 8.35 per cent of GDP, which in turn suggests that structural impediments in services are greater in Egypt than in Tunisia. In addition Konan and Kim show how including services liberalisation tends to generate more balanced growth, and thus from a political economy perspective perhaps an easier transition process. This more balanced growth arises from the importance of services as in input across a wider range of economic activities. As mentioned earlier the Southern Mediterranean countries are increasingly becoming involved in a range of overlapping regional trade agreements. An issue of increasing concern to academics, practitioners and policy makers is the role of the underlying rules of origin in impacting upon trade flows between countries within a preferential trading agreement, and across agreements. This issue is addressed by Augier et al. who focus on the role of the cumulation of rules of origin in the context of EU preferential trading arrangements with a focus on the textile industry and the countries of the Southern Mediterranean. First the authors explain how rules of origin, depending on how they are formulated, can serve to divert or suppress trade between those countries who are members of a PTA and non-member countries. Secondly, using a cross-section estimating equation they show how the lack of cumulation of rules of origin appears to restrict trade between non-cumulating countries by between 73 and 81 per cent in aggregate and for certain Southern Mediterranean economies by even more. These are important results because they suggest that rules of origin may serve to significantly divert trade away from non-PTA partner countries. This is important, therefore, not just for those excluded countries but also for the members of the preferential trading area. The traditional approach to evaluating the welfare gains from regional integration is in terms of trade creation and trade diversion. The work of Augier et al. suggests that constraining rules of origin are likely to increase the likelihood of trade diversion and concomitantly trade suppression. With the proliferation of overlapping regional trade agreements it is thus becoming increasingly important to understand the impact of the underlying rules of origin. Their work is also important in pointing to the importance of the possibility of cumulating rules where regional agreements do overlap. In this context the decision under the Agadir Agreement between Egypt, Jordan, Morocco and Algeria to adopt the pan-European rules is to be welcomed as it greatly enhances the possibilities for such cumulation in the context of trade relations with the EU and its partner countries. Finally, the paper by Méon and Sekkat provides an illuminating empirical analysis of the importance of the quality of domestic institutions in promoting trade and foreign direct investment. Using both panel data and cross-section estimates the authors augment the more traditional explanatory variables (such as


107

exchange rates and growth rates) for trade and foreign direct investment with several different variables each of which attempts capture the quality of the underlying institutional framework. They show that both trade and foreign direct investment are positively correlated with the underlying quality of those domestic institutions. The authors then provide simulations for selected Middle East and North African economies indicating the increase in both exports and FDI that could arise from improvements in institutional quality. For example, depending on the institutional indicator used and on the underlying estimation methodology the increase in the exports to manufacturing ratio for Morocco ranges from 18.5 to 78.9 per cent. While the results for FDI are slightly less robust econometrically, overall they suggest that FDI is likely to be impacted by an improvement in the quality of the institutions by a similar order of magnitude. The paper thus provides important further confirmation of the importance of the interrelationship between trade liberalisation, foreign direct investment and the underlying structural features of the institutional and political landscape. The three papers together complement each other by each focusing on an important aspect of the reform and transition process facing the Southern Mediterranean. They suggest that there are indeed potentially large gains to be had from liberalising trade. This clear message emerges from the Konan and Kim paper. However, if those gains are to be realised it is likely to be important not simply to liberalise goods trade but also services trade as these are highly complementary. The paper also highlights, however, that the welfare gains from regional trade liberalisation are not assured and this arises from the presence of trade diversion. Augier et al. complement this by showing that restrictive rules of origin are likely to heighten the possibility of trade diversion between PTA member and non-member countries. Finally, the focus by Méon and Sekkat on institutions provides important evidence that the ultimate success of the process of adjustment through a more outwardly oriented strategy is likely to equally depend significantly on internal reform and change.

REFERENCES Edwards, S. (1998), ‘Openness, Productivity and Growth; What Do We Really Know?’, Economic Journal, 108, 383–98. Dollar, D. (1992), ‘Outward Oriented Developing Countries Really Do Grow More Rapidly’, Economic Development and Cultural Change, 40, 523–44. Rodriguez, F. and D. Rodrik (2000), ‘Trade Policy and Economic Growth: A Skeptic’s Guide to the Cross-national Evidence’, in B. Bernanke and K. S. Rogoff (eds.), Macroeconomics Annual 2000 (Cambridge, MA: MIT Press for NBER), 261–324. Winters, L. A. (2004), ‘Trade Liberalisation and Economic Performance’, Economic Journal, 114 (February), pp. F4–F21.

LIBERALISATION OF SERVICES TRADE IN TUNISIA & EGYPT

109

7

Beyond Border Barriers: The Liberalisation of Services Trade in Tunisia and Egypt Denise Eby Konan and Karl E. Kim

1. INTRODUCTION

HILE the welfare implications of the liberalisation of goods trade are widely understood, the benefits associated with reform of trade in services have not been adequately assessed. This is due, in part, to the more complex character of services trade and the more recent attempts to negotiate multilaterally through GATS (General Agreement on Trade in Services), and other agreements, the development of appropriate mechanisms for services trade liberalisation. Although services account for an ever increasing share of production and employment in countries across the world and the share of services in global trade and investment has increased significantly, efforts to liberalise trade in services have been limited by several factors. There are important differences between goods and services trade. While much of trade theory and empirical work has developed around the trading of goods, less has focused on trade in services. Traditional trade policy analysis involves tinkering with tariffs, or related border barriers. While some services (e.g. software, architectural drawings, consultant reports, etc.) are tangible, visible and storable and are thus like many other goods, many services require direct interaction between producers and consumers (Mattoo, 2000). This proximity requirement means that for many services, factors of production (labour and capital) must be mobile for international transactions to occur. For this reason, the real barriers to increased trade in services are not simply tariffs, but rather, beyond

W

For instructive comments on an earlier version of this paper, the authors thank Aristomene Varoudakis and the participants of the 2003 Middle East Economics Association Annual Meeting in Saint Quentin en Yvelines, France.

109

110

DENISE EBY KONAN AND KARL E. KIM

those at the border. Domestic regulations and conditions within various countries related to market entry, ownership and operations of foreign firms are crucial to understanding reform of trade in services. Trade in services also differs from trade in goods in another fundamental way. Deardorff (2001) theorises how liberalisation in services trade can stimulate trade in goods. Clearly removing barriers to trade in fundamental services such as transportation, communications, banking, accounting, information processing, etc. will help to facilitate economy-wide gains. It should be noted, that unlike the trade in goods, many services can be fragmented in terms of production. Whether referred to as outsourcing, subcontracting or multi-stage production, the point here is that different parts of the service production process can be located in different places allowing for overall efficiency gains. This ‘brave new world’ of trade, according to Feenstra (1998) would result in the ‘integration of trade and the disintegration of production’ on a global scale. There are also opportunities with services reform to link up with strategies that focus on urban infrastructure, regional development and growth triangles (Kim and Wu, 1998). The research reported in this paper is focused both on the specific impacts of alternative trade reform policies on the welfare of individual economies as well as the broader implications and consequences of shifting from goods to services trade. At one level the research is about two very different countries (Tunisia and Egypt) within the MENA (Middle East and North Africa) region. At another level, the analysis contributes to the debates and deliberations arising out of the Doha Round regarding the services trade liberalisation. The analysis builds on two earlier studies. In Konan and Maskus (2004), an applied general equilibrium model of services liberalisation in Tunisia is developed. Konan (2003) provides a comparison of Tunisia and Egypt trade liberalisation scenarios, most notably the prospects for greater Arab integration, using data from earlier years. GATS has defined trade in terms of four different delivery modes: Mode 1: cross-border – in which the services (e.g. software, consultant report, accounting or bookkeeping, etc.) from one country are supplied to another by mail, delivery or electronic means (many of these services are quite similar to traditional goods); Mode 2: consumption abroad – in which the consumer (e.g. tourism, educational services, health care, etc.) travels from one country to another to receive the service; Mode 3: commercial presence – typically involves foreign direct investment such that a business or enterprise (e.g. bank, insurance, engineering firm, etc.) is established in another country as a way to provide services to consumers; and Mode 4: presence of natural persons – services are provided by nationals of one country who move into another country. Clearly, the movement of factors of production raises questions and challenges for traditional trade theory which has focused more on the movement of goods across borders. For this reason, cross-border (Mode 1) services may be easiest to consider, while the presence of natural persons (Mode 4) may be the most


111

challenging, if not analytically or theoretically, then certainly from a political perspective. Mode 2 (consumption abroad) represents a special case in that consumers do not permanently relocate to the country providing service. Perhaps the most important service delivery mode to examine is Mode 3 (commercial presence), not just because it is growing in terms of investment flows, but also because the service providers must interact closely with local authorities and are subject to various local regulations and rules as well as international policies. It is against this background that the impact of various trade policies in the MENA region will be examined. The need to examine trade policy in this part of the world arises from the fact that this region relative to others in the world has experienced weak economic performance. While world trade has increased by approximately 8 per cent a year, trade in this region has grown by less than 3 per cent annually. In addition, the region is losing export market share and experienced only a trickling volume of foreign direct investment. Indeed, if oil exports are not considered, the total volume of international trade activity for the MENA region would be very small. As has been noted earlier, transaction costs, structural impediments, non-tariff barriers and other frictional costs of doing business in the region appear to play a bigger impediment to trade and investment than border barriers (Zarrouk, 2003; and Hoekman and Konan, 2001a). While most of the MENA nations have initiated trade reforms, the experiences across the region are markedly different. Some countries have established bilateral trading agreements. Others have joined regional trading blocs, such as the Greater Arab Free Trade Agreement. Several have signed a European Mediterranean Agreement. Countries in the region have also implemented policies to reduce tariffs preferentially as well as on an MFN (Most Favoured Nation) basis. Several countries, including both Tunisia and Egypt, have joined the WTO. Others have launched privatisation initiatives as well as programmes to stimulate greater FDI. Still, after all is said and done, the MENA region still trades much less than expected (Nugent, 2002). In this paper the impact of different trade policies on Tunisia and Egypt are examined using Computable General Equilibrium (CGE) models. Using countrylevel input-output data, the behavioural relationships between producers and consumers are captured and the various barriers to trade in both goods and services are simulated. As trade liberalisation may entail the removal of tariffs on imports, domestic tax policy instruments are adjusted to maintain government expenditure neutral reform packages. After formulating a baseline model of economic conditions, alternative scenarios based on the following different trade policies are developed: (1) the elimination of tariffs on goods imported from the European Union (European Mediterranean Agreement); (2) the elimination of all tariffs on a Most Favoured Nation (MFN) or non-discriminatory basis with all trading partners; (3) the elimination of border barriers on services trade; (4) elimination of services investment barriers; (5) joint impact of border and investment

112


services liberalisation; and (6) the elimination of all goods and services barriers. Country-wide impacts on welfare, output, income distribution and sector effects are estimated in both Tunisia and Egypt. While there has been speculation as to the consequences of these developments, there has been much less formal analysis of the impact of these reforms on both trade patterns and on the macroeconomic conditions within the region. There is clearly a need to analyse the effects of alternative policies on the trading of both goods and services. There is also need for research on the effect of these policies on specific countries within the region.

2. A TALE OF TWO COUNTRIES

The approach taken in this paper is to focus closely on two countries within the MENA region – Tunisia and Egypt. In some ways they are typical of countries in the region. In other ways they are not. Data on conditions for the two countries are presented in Table 1 (Socio-economic Indicators, Tunisia and Egypt, 2001). Tunisia is smaller than Egypt both in land area and population size. With more than a million square kilometres of land area and a population of 65.2 million persons, Egypt is one of the largest in the region. Of the two countries, Tunisia is more urbanised, with 66 per cent of the population living in urban areas, compared to only 43 per cent in Egypt. While Egypt’s poverty rate (23 per cent) is much greater than in Tunisia (8 per cent), relative to others in the MENA region neither country is extremely wealthy nor extremely poor. The fertility rate in Tunisia (2.1 births per woman) is much lower than in Egypt (3.2 births per woman). Life expectancies are higher in Tunisia (72.4 years) than in Egypt (68.3 years). The illiteracy rate is much higher in Egypt than in Tunisia. While the GDP in Egypt (US$99 billion) is much larger than in Tunisia (US$20 billion), the annual growth rate in GDP is higher in Tunisia (4.9 per cent) than in Egypt (2.9 per cent). Another major difference between these two countries is TABLE 1 Socio-economic Indicators, Tunisia and Egypt, 2001

Land Area (square km) Population Per Cent Urban Population Poverty Rate (1995) Fertility Rate (births per woman) Life Expectancy (years) Illiteracy Rate (15 per cent and above) GDP GDP Annual Growth Trade in Goods (share of GDP)

Tunisia

Egypt

163,000 9.7 million 66% 8% 2.1 72.4 27.9 $20 billion 4.9% 80.8%

1 million 65.2 million 43% 23% 3.2 68.3 43.9 $99 billion 2.9% 17.1%


113

TABLE 2 Composition of the Economy, Tunisia (1995) and Egypt (1997) Output

Agriculture Petroleum Processed Food Clothing Other Manufacturing Utilities Construction Transport, Communication Trade, Finance, Insurance Hotel and Restaurant Other Services TOTAL

Exports

Imports

Tunisia

Egypt

Tunisia

Egypt

Tunisia

Egypt

8.38 4.91 11.09 12.32 20.44 2.12 6.93 5.79 8.78 5.13 14.09 100.00

16.80 4.36 10.11 9.06 17.59 1.87 7.80 7.81 15.02 2.00 7.58 100.00

1.89 8.41 6.32 41.65 26.66 0.00 0.00 11.19 3.88 0.00 0.00 100.00

2.58 14.47 3.00 8.48 15.40 0.00 0.00 25.26 17.06 11.64 2.11 100.00

6.98 6.17 5.70 20.07 52.44 0.02 0.00 3.79 4.72 0.00 0.11 100.00

6.84 1.10 7.57 3.07 44.90 0.06 0.96 5.23 16.31 0.00 13.97 100.00

the share of the GDP that involves trade in goods (80 per cent for Tunisia versus 17.1 per cent for Egypt). Table 2 provides the composition of production, exports and imports in the economies of Tunisia (1995) and Egypt (1997). It is evident that Egypt’s economy is far more dependent on agriculture as almost 17 per cent of output is in agriculture in Egypt, compared to 8.38 per cent in Tunisia. This is reflected in the fact that more of Egypt’s population lives in rural areas. It also serves to highlight some of the challenges facing Egypt which has a much larger, younger, poorer and more unskilled workforce than Tunisia. Interestingly, while agriculture is an important component of output it is not a major export activity, with the possible exception of cotton. Both countries are significant importers of manufactured goods. It is important to note that manufacturing comprises 44 per cent of Tunisia’s output compared to 36 per cent of output in Egypt. Tunisia also exports and imports a significant volume of clothing. More than 20 per cent of Tunisia’s imports are for clothing, principally as intermediate goods for the robust clothing and textile industry. In Tunisia, 75 per cent of exports are for manufactured goods, with 42 per cent of this in clothing. While Egypt has a much less developed manufacturing sector, it has certain natural resource advantages evident in output and export. Petroleum, transportation services related to the Suez Canal, and touristic antiquities are all significant components of Egypt’s economy. The trade, finance and insurance sectors are significant in terms of output (15 per cent), exports (17 per cent) and imports (16 per cent) in Egypt. A total of 57 per cent of the exports in Egypt are services-related compared to 15 per cent in Tunisia (Tohamy, 2001). Tunisia, however, maintains a large social services sector which includes education, health care and other public services comprising more than 14 per cent of output.

114


The statistical data do not adequately describe some other important differences between these two countries. In 1995, Tunisia became the first country in the region to sign a partnership agreement with the EU (European Union). It has maintained close trading relations with Europe, particularly with France. Since independence from France, Tunisia has experienced steady improvement in social and economic conditions. Egypt, on the other hand, has experienced less progress and has remained more detached from the world economy. While Tunisia has been more open to trade as measured by the ratio of imports and exports to GDP, the Egyptian index of openness has actually declined from 14 per cent in 1990 to 10 per cent in 1999 (Hoekman and Kheir-El-Din, 2000). In both Tunisia and Egypt there are relatively high levels of fiscal, administrative and regulatory restrictions in place. While liberalisation efforts are under way, the regulatory environments in each country are widely thought to pose obstacles to international trade and investment. As much has been written on this topic for Tunisia, Egypt and the region, a brief summary is provided. Interested readers are referred to Hoekman and Konan (2001a and 2001b), Hoekman and Kheir-El-Din (2000), Konan and Maskus (2003) and Zarrouk (2003). Table 3 provides ad valorem taxes and regulatory barriers. Both countries maintain significant tariff rates on manufacturing imports. The weighted average tariff rate in Tunisia was 11 per cent and in Egypt was 18 per cent. Certain import-competing manufacturing sectors are also heavily protected through the tariff structure, clothing tariffs are 21.6 per cent in Tunisia and 36.2 per cent in Egypt. Table 3 also provides the primary domestic tax, which in Tunisia is the value-added tax (VAT) and in Egypt is the goods and services tax (GST). In both countries, tax rates vary across sectors, often favouring agriculture at the expense of manufacturing. TABLE 3 Ad Valorem Taxes and Regulatory Barriers Border Barriers

Agriculture Petroleum Processed Food Clothing Other Manufacturing Utilities Construction Transport, Communication Trade, Finance, Insurance Hotel and Restaurant Other Services

Domestic Barriers

Investment Barriers

Tunisia

Egypt

Tunisia

Egypt

Tunisia

Egypt

13.0 19.5 18.5 21.6 15.2 0.0 0.0 50.0 30.0 0.0 5.0

8.5 5.0 11.0 36.2 22.3 0.0 3.0 50.0 6.0 3.0 3.0

6.0 7.3 21.4 22.3 16.3 8.2 17.0 6.3 6.0 6.0 6.0

3.4 0.4 9.6 4.9 8.6 2.5 10.0 10.0 10.0 5.0 5.0

0.0 7.3 30.0 5.0 3.0

3.0 15.0 15.0 5.0 15.0


115

Regulatory and administrative barriers in both countries, and throughout the MENA region, are quite substantial (Zarrouk, 2000). According to the survey results of Zarrouk (2003) regulatory border barriers add 10.6 per cent to producer costs. Customs clearance, public sector corruption, inspection and conformity certification, and transshipment regulatory measures were viewed as significant barriers. For example, within the region it generally takes two to five days to release goods imported by air freight from customs, and up to ten days for sea shipment. The global standard is less than six hours for air freight and 24 hours for sea freight customs clearance. In analysis of various trade integration efforts for Egypt, Hoekman and Konan (2001a and 2001b) found that the gains from ‘deep’ integration through the liberalisation of non-tariff barriers yield gains that are significantly higher than available through shallow integration alone. In part this is due to the resource-using nature of many non-tariff barriers. The barriers to trade in services are generally quite different from those involved in goods trade. As many cross-border services transactions do not involve tangible products, customs and port clearance procedures are generally not relevant. Table 3 provides our best estimates for ad valorem tariff-equivalent estimates for cross-border trade and for foreign investment in services. These measures, discussed in more detail in Konan and Kim (2003), indicate high levels of production for regulated industries including air and maritime transportation, and telecommuncations. As has been discussed, for many services there is a need for proximity between the producer and client. Hence, barriers to the movement of factors, commercial presence and consumers become relevant. Of concern are foreign investment measures, capital controls, competition policies, product specifications, labour movement restrictions, and other domestic regulatory and administrative restrictions. In the case of Tunisia and Egypt, these barriers may not only raise transactions costs but they may be entirely prohibitive to some or all modes of international services transactions. Data on services trade and the barriers thereto are presently sparse and under construction. The interested reader is referred to Findley and Warren (2000) for a summary of the issues. Within the MENA region, regulatory policies which reduce competition in service industries are quite common (Zarrouk, 2000). In Tunisia, exclusive commercial agency laws in distribution services limit foreign entry in distribution and sales, protect domestic monopolists and prevent parallel imports (Maskus and Lahouel, 1999). While the Tunisian market is closed to foreign franchising, Egypt has recently permitted franchisors to operate popular chain restaurants, hotel management and other services. In both countries, and indeed in the region, state monopolies (Tunisie Telecom and Egypt Telecom) provide basic telecommunications services. This has resulted in costly and inefficient services, with waits for new telephone main lines averaging 1.9 years in Egypt and 0.9 years in Tunisia in 2000. Galal (1998) found that a more competitive regulatory regime for telecommunications in Egypt may increase

116


welfare by 1.2 per cent of GDP, or $800 million. Commercial banking and insurance services are non-competitive in Tunisia and Egypt. There are 14 private commercial banks in Tunisia, four of which are foreignowned. Of the 30 commercial banks in Egypt, only six are foreign owned accounting for only 7 per cent of total commercial bank assets. Nonetheless, Egypt’s financial system is widely considered to be the most open system in the MENA region (Tohamy, 2001). Air and maritime transportation services are also highly regulated and dominated by inefficient public monopolies. Mohieldin (1997) has found that fees charged by public companies providing port services for handling and storage of goods are some 30 per cent higher than in neighbouring countries. The paper by Mattoo, Ruthindran and Subramanian (2001) constructs indices of liberalisation for financial and telecommunication services based on the existence of regulations on competition, foreign equity restrictions and capital controls. These policy-based indices of liberalisation (ranging from 1 to 9, with higher values indicating greater liberalisation) are created for 57 countries. Tunisia is ranked as a highly restrictive country, with a financial liberalisation score of three and a telecom liberalisation score of one. Egypt is found to provide a more liberal environment with rankings of eight in financial services and five in telecommunications. While these measures are imperfect, the study found relatively strong cross-country econometric evidence that openness in these service sectors contributes significantly (about 1.5 percentage points) to overall economic growth. Liberalisation reform efforts are under way in both Tunisia (World Bank, 2000) and Egypt (Tohamy, 2001). The Tunisian government has made GATS commitments in three sectors: telecommunications, financial services and tourism. The government of Egypt has set forth GATS commitments in construction, transportation, financial and tourism services. While appearing to be a commitment to gradual reform, the GATS schedule provides some indication of the direction of liberalisation for each country. Egypt and Tunisia, therefore, each face unique conditions and opportunities in the international trade arena. In the next section, the various scenarios involving different levels of trade reform applied to both goods and services at the border and beyond are deregulated and applied to Tunisia and Egypt.

3. EVALUATION OF ALTERNATIVE POLICIES

a. Model Description In order to assess the effects of the alternative trade policies, numerical applied general equilibrium models are developed for Tunisia and Egypt. A full description of the data as well as the technical features of the models is described in


117

detail in Konan and Kim (2003). To summarise, a social accounting matrix is assembled which describes the flow of goods, services and factors through each economy in a baseline year. Data from 1995 on the commodity structure in Tunisia includes 36 sectors: an agricultural sector, 17 manufacturing sectors, petroleum, mining and 15 services sectors. The Egypt social accounting matrix is based on 1997 data and consists of 32 sectors: three agricultural, petroleum, mining, 20 manufacturing, and seven services sectors. To permit a comparison, the results of the two models are aggregated into 11 sectors, common to both Tunisia and Egypt, as presented in Table 2. For each production sector, the purchases of intermediate inputs and primary factors (labour and capital) are provided. Demand in each sector is a combination of intermediate demand and final expenditures by households, government, exporters and investors. Border barriers (tariffs on goods and non-tariff barriers on services), domestic barriers (the value-added tax (VAT) for Tunisia and the goods and services tax (GST) for Egypt), and services investment barriers are given in Table 3. The data are presented at a more disaggregated level in Konan and Kim (2003). Two independent models of economic behaviour are then developed for Tunisia and Egypt. In each model, producers maximise profits in their decisions regarding the purchase of inputs and the level output. Households make optimal purchasing decisions based on the prices of good and services and the income that is received from a variety of sources, including returns to labour and capital as well as transfers from the public sector. The demand for imports and the supply of exports are assumed to respond to fixed international prices, following an assumption that each economy is too small to influence their terms of trade with world trading partners. Traded goods and services are imperfectly substitutable for locally produced commodities, the Armington assumption. The public sector collects a variety of trade and domestic taxes and also purchases goods and services. An important assumption is that any policy change will be revenueneutral in terms of government expenditures. Domestic taxes (the VAT for Tunisia and the GST for Egypt) are automatically adjusted to offset any changes in the collections of tariffs. Within this framework, a tariff drives a wedge between the international price and the internal (domestic) price of an import good. While a tariff will generate tax revenue for the government, it also imposes real economic costs. Production will inefficiently expand in import-substituting sectors, while export-oriented production contracts. Consumers are also harmed in that they pay higher prices for both imports and the domestic import substitute. Complicating matters further, tariffs generally vary widely across sectors, creating additional distortionary effects (Konan and Maskus, 2000). It is also important to note that piecemeal tariff liberalisation is second-best. Removing or lowering a subset of tariff barriers, while maintaining others, can distort trade and production patterns in complicated ways and even lower social

118


welfare in some cases. The European Mediterranean Agreements are examples of preferential trade reforms whereby Tunisia and Egypt have, independently, agreed to eliminate tariff barriers more rapidly with the EU than with other trading partners. Such an agreement may expand trade, but will also alter the pattern of trade in a manner that discriminates against non-member countries. In the simulations conducted in the present paper, it is assumed that non-tariff goods barriers are maintained at baseline levels and are unaffected by reductions in tariff rates. The models in the current paper provide a special treatment of international trade and investment in services sectors using a methodology developed by Konan and Maskus (2004). See Dee, Gardin and Holmes (2000) for a review of the use of computable general equilibrium models on services liberalisation. Table 3 provides estimates of border and investment barriers in services. As there is sparse evidence on the actual level of services barriers, those assumed in the present analysis are conservative estimates of those expected in Tunisia and Egypt; see Konan and Kim (2003). In terms of cross-border services trade the primary policy instruments are regulatory in nature. Border barriers on services involve the creation of ‘deadweight costs’ in that they do not generate tax revenues or economic rents but rather simply increase the cost of the international transaction. Multilateral negotiations are also ongoing to provide a more liberal environment for foreign investment or the establishment of a foreign presence in services sectors. It is important to note that for many of these sectors, there is little or no inward foreign direct investment in the benchmark year thus precluding marginal analysis of FDI inflows. In the models, existing regulatory policies have two effects on services markets. The first is an inefficiency effect, as firms are not generally selected on a competitive basis. This is particularly important in Tunisia and Egypt as foreign firms are excluded from key markets. Local services producers may not be employing best practices by world standards and may be limited in the range and variety of the products they provide. The second effect involves the creation of imperfectly competitive markets, in which domestic firms earn real economic rents by limiting production and setting prices above marginal costs. The analysis of Tunisian services liberalisation by Konan and Maskus (2004) illustrates how important it is to distinguish between investment barriers due to imperfect competition, rent-generating, and those due to the use of inefficient technologies and are thus resource-using. The impacts of six trade reform scenarios are presented in Table 4 (Trade Liberalisation in Tunisia), and Table 5 (Trade Liberalisation in Egypt). Goods liberalisation is represented by two reform scenarios. The European Mediterranean Agreement scenario is represented as the elimination of tariff barriers on European imports coupled with a relatively conservative estimate of the improvement in the EU terms of trade of 2 per cent on clothing and 1 per cent on all other exports of goods to the EU. The MFN scenario involves the unilateral


119

TABLE 4 Trade Liberalisation in Tunisia Goods Liberalisation

Services Liberalisation

EU

MFN

Border

Macroeconomic Indicators Household welfare (EV) Output, real Returns to capital, real Returns to labour, real

3.80 6.12 1.29 10.77

4.27 6.91 1.80 11.73

1.05 0.86 1.24 0.60

Output Share Agriculture (Base = 19.5) Manufacturing (Base = 32.8) Petroleum (Base = 7.0) Services (Base = 40.7)

15.0 44.4 5.2 35.4

15.2 43.4 5.2 36.2

20.1 31.4 7.0 41.4

Investment 3.60 4.80 6.89 3.46 21.9 27.1 6.5 44.4

Combined Joint

G&S

4.85 5.84 8.06 4.48

8.32 13.17 12.94 14.29

22.1 26.5 6.5 44.9

21.0 30.4 5.4 43.3

TABLE 5 Trade Liberalisation in Egypt Goods Liberalisation

Services Liberalisation

EU

MFN

Border

Macroeconomic Indicators Household welfare (EV) Output, real Returns to capital, real Returns to labour, real

−0.16 0.35 0.18 1.66

0.46 0.82 0.76 2.81

0.78 1.07 0.76 0.66

Output Share Agriculture (Base = 16.5) Manufacturing (Base = 17.3) Mining (Base = 4.3) Services (Base = 43.0)

16.6 16.7 4.5 43.2

16.8 16.1 4.4 43.4

16.6 17.3 4.4 42.9

Combined

Investment

Joint

G&S

6.90 11.85 10.73 9.48

7.66 12.91 11.45 10.11

8.35 14.79 12.77 14.41

15.9 16.5 3.3 45.5

15.9 16.5 3.3 45.4

16.0 15.5 3.4 45.8

Notes for Tables 4 and 5: Goods Liberalisation – EU: involves preferential goods trade liberalisation under the Euro Med Agreement. Tariffs on imports of goods from the EU are eliminated, and the terms of trade on exports to the EU improve. Goods Liberalisation – MFN: involves the removal of all goods tariffs on a non-discriminatory, or mostfavoured nation (MFN) basis. Services Liberalisation – Border: involves the removal of cross-border barriers in services trade. Services Liberalisation – Investment: involves the elimination of internal barriers to foreign investment in services sectors. Services Liberalisation – Joint: considers the combined effects of border and investment liberalisation. Combined – G&S: involves the joint liberalisation of goods and services barriers.

120


elimination of tariffs on goods on a most-favoured nation or non-discriminatory basis. Tariff rate data for both countries are obtained from customs collections data, reported in Table 3. Services liberalisation is represented in three scenarios. The border scenario eliminates border barriers on Mode 1 services transactions, the rates of which are given in Table 3. As noted above, these barriers are regulatory in nature and have the effect of imposing resource-using transactions costs. The second scenario, services liberalisation – investment, eliminates the regulatory investment measures placed on foreign investment in the domestic provision of services, again in Table 3. This involves removing a price wedge that is due to a combination of producer markups and cost inefficiencies, which for purposes of analysis, are assumed to be evenly split. b. Goods Trade Liberalisation In this section, the effects of goods trade liberalisation scenarios are discussed. Consider first the impact of goods trade liberalisation in Tunisia, Table 4. The first column, Goods Liberalisation – EU, indicates the percentage change in various macroeconomic indicators under the European Mediterranean Agreement. It is estimated that household welfare, measured as Hicksian-equivalent variation, would improve by 3.80 per cent while real output would increase by 6.12 per cent. The benefits accrue disproportionately to labour with the price of labour increasing by 10.77 per cent in real terms. In contrast, real returns to capital increase by a modest 1.29 per cent. This export-led growth involves the movement of resources into manufacturing. The share of the economy in manufacturing increases from 32.8 per cent in the benchmark to 44.4 per cent. Economic activity in agriculture, petroleum and mining, and services decline accordingly. The next column, Goods Liberalisation – MFN, describes the outcome of tariff liberalisation on a non-discriminatory basis. Household welfare increases by 4.27 per cent and real output by 6.91 per cent. Returns to labour improve by 10.77 per cent and returns to capital improve by 1.8 per cent. Following Tunisia’s comparative advantage, production increases in manufacturing, largely clothing, and declines in all other sectors. In the case of Tunisia, the European Mediterranean Agreement offers nearly the same benefits as a complete liberalisation of manufacturing goods trade (Table 4). Economic impacts in terms of GDP growth, household income and returns to factors are slightly lower when tariffs on manufacturing goods are eliminated with the EU (Table 4, column 1) rather than with all trading partners (column 2). In large part, this is due to the strong trading relationship that Tunisia presently has with the EU. Tunisia’s exports are strongly concentrated in the clothing sector, 95 per cent of which are destined to Europe. About 76 per cent of Tunisia’s chemical exports are to the EU. Imports to Tunisia are also


121

primarily from the EU. Tunisia imports from Europe 92 per cent of clothing, 76 per cent of chemicals and 77 per cent of machinery. Even Tunisia’s petroleum imports are routed through Europe (63 per cent). The only significant import that is not dominated by the EU is the agricultural sector. Given this high benchmark trade share with the EU, a comprehensive regional agreement approximates a move to free trade. It is rather striking that in both goods liberalisation scenarios there is a substantial expansion of the manufacturing sector. As evident by Table 3, Tunisia maintains high tariffs to support domestic production of import-substituting manufacturing goods. Indeed, tariff liberalisation results in a reduction of nonexport-oriented manufacturing activities such as food processing. At the same time, clothing manufacturing and exports expand significantly for several reasons. Tunisia’s revealed comparative advantage is strongest in clothing and an enhanced ability to import frees factor resources for use in clothing production. Clothing exports rely on imported intermediate goods which are cheaper following a reduction in manufacturing tariffs. Finally, in the case of the EU Mediterranean agreement, export terms of trade in clothing are expected to increase by 2 per cent. The impacts of goods liberalisation in Egypt are given in Table 5. As shown in the first column, Goods Liberalisation – EU, the European Mediterranean Agreement imposes trade diversionary effects which lowers household welfare by 0.16 per cent and increases output by only 0.35 per cent. MFN tariff reduction (column 2 of Table 5) implies rather small gains for Egypt. This nondiscriminatory elimination of goods tariffs would increase Hicksian household welfare by 0.46 per cent and real output by 0.82 per cent. Real returns to labour increase more significantly by 2.81 per cent and real returns to capital improve by 0.76 per cent. In either liberalisation scenario, the composition of production changes very little. It is apparent that Tunisia and Egypt differ in their ability to gain from goods trade liberalisation, either through the European Mediterranean Agreement or through broader MFN tariff liberalisation. The reasons are informative. Under the European Mediterranean Agreement, consumers expand their purchase of EU products at the expense of imports from the rest of the world which remain subject to tariffs. The subsequent fall in tariff revenues is not completely compensated for by improved consumer surplus. As discussed in Konan and Maskus (2003), Egypt’s trading patterns are rather diversified on a regional basis. While the EU is an important trading partner, MENA and the US also maintain significant trade shares in key sectors. Egypt’s economy would experience a slight decline in welfare (Table 4) upon entry into the European Mediterranean Agreement. While the EU is an important trading partner, Egypt also exports a significant share of manufacturing goods to MENA. The US is an important supplier of agricultural goods. The majority of

122


Egyptian imports of transportation equipment, clothing and appliances are from Asia. A preferential trade agreement with Europe would divert imports from other trading partners in favour of more highly priced (but untaxed) European products. The impact is realised in a loss of tariff revenues, not only from Europe but elsewhere (Konan and Maskus, 2003). Depending on the mechanism by which the government recovers these lost revenues, the social welfare effects may be higher or lower than indicated in Table 5 (Konan and Maskus, 2000). In a related paper using older data Konan (2003) considers the implications for Tunisia and Egypt of various trade liberalisation scenarios including integration in the context of the Greater Arab Free Trade Agreement (GAFTA). Not surprisingly, welfare in Tunisia would decline in response to a shallow trade liberalisation agreement that involves only the MENA region. For Egypt, the GAFTA implies an estimated increase in welfare of 0.2 per cent, with a 2.1 per cent increase in GDP. It is shown that Egypt would gain more from closer integration with MENA than with the EU.1 Tunisia, in contrast, benefits greatly from EU integration but loses in terms of both GDP and welfare from a MENA agreement. A cointegration study of the Arab Maghreb region indicates that these nations do share robust ties in their macroeconomics, financial markets and monetary policies (Darrat and Pennathur, 2002). Thus, there may be inherent benefits if Arab economic integration were to be staged in concert with efforts toward global integration. Next, consider the impact of non-discriminatory goods trade reform, Tables 4 and 5, column 2. In these scenarios, goods tariff barriers are eliminated unilaterally. Tunisia gains nearly 7 per cent in GDP growth from a multilateral elimination of manufacturing tariffs, with household income increasing by 4.3 per cent. An equivalent reform in Egypt is estimated to yield modest growth rates of less than 1 per cent in either household welfare or output. Why are Tunisia’s gains from traditional goods trade liberalisation so high, and Egypt’s so low? Sensitivity analysis helps to identify possible explanations and rule out others. Weighted average tariffs are quite similar between Tunisia and Egypt over the baseline period. Differences in the effective rate of protection at the sector level, or the distribution of tariffs across industries, also appear not to explain this effect. There is no evidence that Tunisia is initially more protectionist than Egypt. It is the case that Egypt’s domestic tax structure is highly distorted through the goods and services tax (Konan and Maskus, 2000). However, reform scenarios simulations that involve joint domestic and international

1

Note that Konan (2003) assumes that government revenue neutrality is maintained through an endogenous and proportional adjustment in the goods and services tax for Egypt and the valueadded tax for Tunisia, while the present paper relies on adjustments in a non-distortionary lumpsum tax. The Egypt data in Konan (2003) are for the 1990 baseline year. Thus, simulation results differ somewhat with those of the present paper.


123

tax liberalisations do not improve economic growth for Egypt substantially. It is thus reasonable to conclude that the tax policy structures of the two countries do not explain why Tunisia is more able to gain from goods trade reform than Egypt. A clearer explanation appears to lie in the structure of production and trade. In many respects Tunisia is a classic small open economy. Tunisia is heavily reliant on a wide variety of manufactured imports, while exports are highly concentrated in clothing. As shown in Table 1, goods trade accounts for 80.8 per cent of GDP in Tunisia. The reduction of tariffs accelerates trading patterns that are already deeply embedded in the economy. In contrast, Egypt is relatively more self-sufficient for reasons that appear to go beyond simple tariff protections. Indeed, goods trade comprises only 17.1 per cent of GDP (Table 1). The volume of trade in Egypt is quite low in most manufactured goods. Exports also appear to be relatively low overall, and not highly concentrated in any set of industries. This may be due to the traditional availability of remittances from foreign aid, tourism, and repatriation of income from guest Egyptian workers overseas, which provide alternative sources of foreign exchange. It may be due to inadequate infrastructure, regulation or informal barriers that suppress trade in Egypt. Survey evidence and industry studies appear to verify the existence of structural impediments to trade in Egypt (Cassing et al., 2000; Fawzy, 1999; Kheir-El-Din, 2000; and Tohamy, 1999). The strongly closed context of Egypt’s trading environment appears to offer a realistic explanation for why the gains from traditional trade liberalisation in Egypt are so low. Another explanation for the unusually low volume of trade in Egypt is the pervasiveness of a variety of regulatory and administrative barriers to trade. These barriers are presumed by many to be substantial. Extensive firm-level survey analysis by Jamel Zarrouk (2003) indicates that non-tariff barriers (NTBs) in the MENA region range from 8 to 20 per cent of the value of imports across industries, with a trade-weighted average barrier of 11 per cent. Ranked in the order of importance by MENA firms, these barriers include customs duties, domestic taxes, customs clearance, public sector corruption, inspection and conformity certification costs, transshipment regulatory barriers and entry visa restrictions. The present paper does not consider scenarios whereby regulatory and other non-tariff barriers to goods trade are liberalised as this has been addressed in previous analyses. Konan (2003) and Hoekman and Konan (2001a and 2001b) consider scenarios of Egyptian goods trade liberalisation which allow for a reduction in red tape and find that the gains from ‘deep’ integration outweigh substantially that of ‘shallow’ liberalisation of tariff barriers. While there is no doubt that structural impediments in the MENA region are significant, most especially for Egypt, less clear is the process by which the impediments might be eliminated.

124


c. Services Trade Liberalisation Trade reform in services sectors are decomposed into scenarios involving border liberalisation and investment liberalisation. Consider the role of services liberalisation in Tunisia (Table 4). The impact of cross-border reform in services comprises less than a quarter of overall services gains, with household welfare increasing by 1.05 per cent and real output by less than 1 per cent. More significant are the gains attributed to the liberalisation of foreign investment in services sectors, amounting to increases in household welfare by 3.60 per cent and real output by 4.80 per cent. Investment and border trade reform in services sectors together increase household welfare by 4.85 per cent and real output by 5.84 per cent. Table 5 provides macroeconomic estimations of services liberalisation for Egypt. The elimination of border barriers improves household welfare by 0.78 per cent and real output by 1.07 per cent. Investment liberalisation leads to welfare gains of 6.90 per cent and real output gains of 11.85 per cent. Thus, the overall gains from services liberalisation are primarily due to foreign investment in the services sector. Taken jointly, border and investment liberalisation improves real output by 12.9 per cent. In comparing the results for these two countries, several interesting observations can be made. First, for both Tunisia and Egypt, the gains from services liberalisation are substantial. Services liberalisation provides improvements in welfare of 3.6 per cent in Tunisia and 6.9 per cent in Egypt. The growth rate of real output in Tunisia is 4.85 per cent and in Egypt is 12.91 per cent on an annual basis. Second, the gains to the economy are rather evenly distributed across factor owners. In Tunisia, services liberalisation translates into factor return increases of 8.06 per cent for capital and 4.48 per cent for labour. Egypt experiences increases in real returns to capital of 11.45 per cent and to labour of 10.11 per cent. Third, while services liberalisation increases overall output productivity, both in level terms and in value, the composition of production changes very little under services liberalisation. In Tunisia, the output share of services expands from 40.7 per cent to 44.9 per cent of the economy. Growth is most notable in communications, trade and financial services, insurance and the visitor industry. While most manufacturing sectors decline with services liberalisation, agriculture and clothing production actually increases. Clothing is heavily traded and liberalisation of services helps to reduce transactions costs in this sector. In Egypt, services output increases from 43 per cent of the economy to 45.4 per cent. Declines in the GDP share of non-services sectors are moderate. That services liberalisation provides for balanced growth is strikingly different to the experience of goods liberalisation, in which production becomes more specialised following a country’s comparative advantage. This is due to several


125

factors. First, services play important intermediate functions in production. Improvements in services sectors tend to enhance productivity throughout the economy. Second, services are an important component of the transactions cost of international trade, and thus goods that are exported benefit from improvements in the provision of services. Third, foreign investment and ownership essentially transfers an improved technology for services provision, and these services are produced using domestic labour and capital. Rather than expanding production primarily to service the external market, as is the case with goods liberalisation, investment in services sectors tends to be oriented to domestic or regional markets.

4. ALTERNATIVE PATHS TO PROSPERITY

To summarise, the combined effect of goods and services liberalisation are substantial for both countries. In Tunisia, household welfare increases by 8.32 per cent and in Egypt by 8.35 per cent. However, Egypt and Tunisia differ in terms of their predispositions to gain from trade liberalisation. The simulation results presented here imply distinct pathways to prosperity for these two MENA countries. For Tunisia, traditional trade liberalisation (the reduction of tariff barriers on manufacturing goods) has the potential to generate significant improvement in economic growth and the standard of living. In particular, wages increase dramatically as the demand for workers rises in export-oriented sectors such as clothing. Moreover, the lion’s share of the gains from goods trade liberalisation is achievable through the European Mediterranean Agreement. Extending the liberalisation reform package to include trade and investment in services sectors leads to further economic growth and improvement in household living standards. In the case of Egypt, the reduction of manufacturing tariffs offers extremely modest economic gains when reformed in isolation. Egypt’s economy is rather closed and tariff liberalisation does not sufficiently expand the volume of trade. Interestingly, initial tariff barriers are roughly comparable between Egypt and Tunisia in both their nominal and effective rates of protection. Yet Egypt appears to suffer from significant structural impediments that limit goods trade expansion. Moreover, tariff reduction in the context of the European Mediterranean Agreement would actually erode the standard of living of a typical Egyptian household as the preferential nature of the reform would discriminate against other trading partners in the MENA region, the US and elsewhere. The clear alternative for Egypt is to focus liberalisation efforts on beyond border reforms, particularly of services sectors. The issue of services liberalisation has become increasingly prominent in recent years. It is not just that the value of global services production has been growing but, perhaps more relevant to Tunisia and

126


Egypt, services reform may provide an important new pathway to overall prosperity. Future research should focus more on providing improved estimates of the structural impediments to goods and services trade and investment, identifying and modelling the obstacles to liberalising trade in services, as well as estimating the effects of technological change on economic growth.

REFERENCES Cassing, J. H., S. Fawzy, D. Gallagher and H. Kheir-El-Din (2000), ‘Enhancing Egypt’s Exports’, in B. Hoekman and J. Zarrouk (eds.), Catching Up with the Competition: Trade Policy Challenges and Options for the Middle East and North Africa (Ann Arbor: University of Michigan Press). Darrat, A. F. and A. Pennathur (2002), ‘Are the Arab Maghreb Countries Really Integratable? Some Evidence from the Theory of Cointegrated Systems’, Review of Financial Economics, 11, 79–90. Deardorff, A. V. (2001), ‘International Provision of Trade Services, Trade and Fragmentation’, Review of International Economics, 2, 233– 48. Dee, P., A. Gardin and L. Holmes (2000), ‘Issues in the Application of CGE Models to Services Trade Liberalization, 2000’, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications (London and New York: Routledge). Fawzy, S. (1999), ‘The Business Environment in Egypt’, in S. Fawzy and A. Galal (eds.), Partners for Development: New Roles for Government and Private Sector in the Middle East and North Africa (Washington, DC: The World Bank). Feenstra, R. C. (1998), ‘Integration of Trade and Disintegration of Production in the Global Economy’, Journal of Economic Perspectives, 12, 4, 31–50. Findley, C. and T. Warren (2000), Impediments to Trade in Services: Measurement and Policy Implications (London and New York: Routledge). Galal, A. (1998), ‘Towards More Efficient Telecommunications Services in Egypt’, Policy Viewpoint No. 2 (January). Hoekman, B. and D. E. Konan (2001a), ‘Deep Integration, Nondiscrimination and EuroMediterranean Free Trade’, in J. von Hagen and M. Widgren (eds.), Regionalism in Europe: Geometries and Strategies After 2000 (Boston/ Dordrecht/ London: Kluwer Academic). Hoekman, B. and D. E. Konan (2001b), ‘Overlapping Free Trade Agreements and the Middle East and North Africa: Economic Incentives and Effects on Egypt’, in J. Devlin, S. Dessus and R. Safadi (eds.), Towards Arab and Euro-Mediterranean Regional Integration (Paris: OECD). Hoekman, B. and H. Kheir-El-Din (eds.) (2000), Trade Policy Developments in the Middle East and North Africa (Washington, DC: The World Bank). Kheir-El-Din, H. (2000), ‘Enforcement of Product Standards as Barriers to Trade: The Case of Egypt’, in B. Hoekman and H. Kheir-El-Din (eds.), Trade Policy Developments in the Middle East and North Africa (Washington, DC: World Bank). Kim, K. and C.-T. Wu (1998), ‘Regional Planning’s Last Hurrah: The Political Economy of the Tumen River Regional Development Plan’, GeoJournal, 44, 3, 239–47. Konan, D. E. (2003), ‘Alternative Paths to Prosperity: Trade Liberalization in Egypt and Tunisia, in A. Galal and B. Hoekman (ed.), Arab Economic Integration: Between Hope and Reality (Centre for Economic Policy Research in Europe (CEPR) and Brookings Institution). Konan, D. E. and K. E. Kim (2003), ‘From Here to There: Trade Reform in Tunisia and Egypt’, Background Working Paper for the Middle East Region Group (World Bank). Konan, D. E. and K. E. Maskus (2000), ‘Joint Trade Liberalization and Tax Reform in a Small Open Economy: The Case of Egypt’, Journal of Development Economics, 61, 365–392. Konan, D. E. and K. E. Maskus (2003), ‘Bilateral Trade Patterns and Welfare: An Egypt-EU Preferential Trade Agreement’, Manuscript.


127

Konan, D. E. and K. E. Maskus (2004), ‘Quantifying the Impact of Services Liberalization in a Developing Country’, World Bank Working Paper, 3193. Maskus, K. E. and M. E.-H. Lahouel (1999), ‘Competition Policy and Intellectual Property Rights in Developing Countries: Interests in Unilateral Initiatives and a WTO Agreement’. Paper for the WTO/ World Bank Conference on Developing Countries in a Millennium Round, 20–21 September 1999 (Geneva: Trade and Development Centre, 1999). Mattoo, A. (2000), ‘Trade in Services: Economics and Law’, World Bank Working Paper. Mattoo, A., R. Rathindran and A. Subramnian (2001), ‘Measuring Services Trade Liberalization and its Impacts on Economic Growth: An Illustration’, World Bank Working Paper No. 2655. Mohieldin, M. (1997), ‘The Egypt-EU Partnership Agreement and Liberalization of Services’, in A. Galal and B. Hoekman (eds.), Regional Partners in Global Markets: Limits and Possibilities of the Euro-Med Agreements (London: CEPR). Nugent, J. B. (2002), ‘Why Does MENA Trade so Little?’, Background Working Paper presented to the Middle East Region Group at the World Bank. Tohamy, S. (1999), ‘Tax Administration and Transactions Costs in Egypt’, in S. Fawzy and A. Galal (eds.), Partners for Development: New Roles for Government and Private Sector in the Middle East and North Africa (Washington, DC: The World Bank). Tohamy, S. (2001), ‘Egypt’s Service Liberalization, Services Barriers, and the Implementation of the GATS Agreement’, in R. M. Stern (ed.), Services in the International Economy (Ann Arbor: University of Michigan Press). World Bank (2000), Republic of Tunisia Social and Structural Review 2000: Integrating into the World Economy and Sustaining Economic and Social Progress (Washington, DC: The World Bank). Zarrouk, J. (2000), ‘Regulatory Regimes and Trade Costs’, in B. Hoekman and J. Zarrouk (eds.), Catching Up with the Competition: Trade Policy Challenges and Options for the Middle East and North Africa (Ann Arbor: University of Michigan Press). Zarrouk, J. (2003), ‘A Survey of Barriers to Trade and Investment in Arab Countries’, in A. Galal and B. Hoekman (ed.), Arab Economic Integration: Between Hope and Reality (Centre for Economic Policy Research in Europe (CEPR) and Brookings Institution), forthcoming.

RULES OF ORIGIN AND THE EU-MED PARTNERSHIP

129

8

Rules of Origin and the EU-Med Partnership: The Case of Textiles Patricia Augier, Michael Gasiorek and Charles Lai-Tong

1. INTRODUCTION

INCE the Barcelona Declaration of 1995, the EU has been pursuing an active policy of trade liberalisation with the countries of the Southern Mediterranean. The twelve countries forming part of this Euromed partnership are: Algeria, Cyprus, Egypt, Jordan, Israel, Lebanon, Malta, Morocco, the Palestinian Authority, Syria, Tunisia and Turkey.1 The objective of this process is to increase the degree of integration between the EU and the Southern Mediterranean and in the process to facilitate their economic development by encouraging the growth of competitive market economies. In practice the policy involves the signing of bilateral Association Agreements or Interim Agreements with these countries, as well as strongly encouraging moves towards greater regional integration among these countries themselves. The Agreements tend to focus on bilateral trade liberalisation through the reduction of tariffs, but also contain provisions on technical assistance and aid, as well as on the harmonisation of standards and bureaucratic procedures. The Euromed process can also be seen as part of a wider context in which there has been a dramatic upsurge in preferential trading agreements (PTAs). Hence the number of PTAs notified to the WTO is now about 250, of which

S

The authors would like to thank David Evans, Peter Holmes, Sherman Robinson and Alan Winters for extremely useful comments on earlier drafts of this work. They would also like to thank participants at a joint CGD-GDN workshop, and a joint IADB-DELTA-INRA-CEPR conference for their comments, as well as participants at seminars at Sussex and Nottingham Universities. Any errors or omissions remain the responsibility of the authors. 1

The aim of the Euro-Mediterranean Association Agreements is to establish free trade areas in place of the current system whereby the Mediterranean countries have tariff-free access to the EU, but the converse is not true. Six agreements are in place (Israel in June 2000, Jordan in May 2002, Lebanon in March 2003, Morocco in March 2000, The Palestinian Territories in July 1997 and Tunisia in January 1996). Turkey has been part of a customs union with the EU since 1996.

129

130 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG 150 have been notified since 1995. These developments encompass both NorthSouth PTAs such as the Euromed partnership, the negotiation of the European Partnerships Agreements (EPAs) with the ACP countries, or the recently signed free trade agreement between the US and Morocco, as well as South-South integration schemes such as the Agadir process between Tunisia, Morocco, Algeria and Egypt, or the process of integration in Latin America such as Mercosur. It is worth also noting that the EU regards the promotion of regional integration groupings among less developed countries as a key feature of its development strategy. All such preferential trading arrangements have detailed protocols on rules of origin (ROOs). Those rules are needed in order to determine the geographic origin of goods and thus to determine the appropriate level of customs duty which should be applied. Take the example of three countries – the EU, Morocco and Poland, where the EU and Morocco have signed a bilateral free trade agreement. Rules of origin are required to prevent trade deflection, i.e. to ensure that Poland does not, for example, export goods to Morocco via the EU. Determination of the origin of a final good becomes more complicated where imported intermediates are used. Hence, where Morocco imports an intermediate from Poland which is then used in the production of a final good exported to the EU, rules (of origin) are then required to determine whether the final good is deemed as truly originating in Morocco or not. There is a relatively small theoretical literature on the impact of ROOs on trade patterns. The literature clearly identifies, however, that such rules can easily be used to restrict/suppress trade between countries, or to divert trade away from more efficient to less efficient suppliers. On the face of it this is perhaps surprising since rules of origin are typically formulated in the context of a process of trade liberalisation and are thus needed to prevent trade deflection. However, it arises because rules of origin focus on the geographical sourcing of intermediate inputs by firms. Depending on how the rules are then formulated, flexibility in the sourcing of intermediates from the cheapest suppliers may well be restricted. Suppose, in the EU-Morocco example given earlier, that goods are deemed as originating only if all intermediates are sourced from within the free trade area. Any good that used intermediates sourced from outside the free trade area would thus be subject to tariffs. Clearly, this is an extreme example but it serves to highlight that restrictive rules of origin either lessen the real extent of integration implied by a PTA, or give strong incentives to producers to source their intermediates from within the free trade area. While the possible impact of ROOs has been identified theoretically, there is very little empirical work on this. In part this is no doubt due to the technical opacity of the underlying rules; and in part this is due to the empirical difficulties of isolating the impact of rules of origin regimes. This is because they are typically applied concurrently with the implementation of the PTAs themselves, and it is therefore very difficult to identify separately the ROO effects. Nevertheless, with the dramatic rise in regional integration in recent years, and therefore


131

the dramatic proliferation of different rules of origin regimes it is becoming extremely important to assess the extent to which ROOs are indeed impacting upon patterns of trade, and thus welfare. For both developed and developing countries it is also important to establish the extent to which the proliferation of such overlapping PTAs with different rules of origin may not simply be promoting intra-FTA trade through the liberalisation of barriers to trade, but also through the raising of effective barriers to inter-FTA trade. The aim of this paper is thus to provide such an empirical assessment of rules of origin. We do so in the context of the ‘pan-European system of diagonal cumulation’ which was introduced in 1997, and the main focus of our analysis is the impact on trade in the textile industry in the Southern Mediterranean. The introduction of the pan-European system allowed for more flexibility than heretofore in the sourcing of intermediates between the EU and a subset of the EU’s trading partners. If the EU rules of origin were indeed impacting upon trade flows, then greater flexibility in the sourcing of intermediate inputs among a subset of countries should then be reflected in patterns of trade. It is precisely this which the current paper focuses on and it does so at the sectoral level where the sector we analyse is textiles. The textile sector was chosen both because of its importance to many Southern Mediterranean economies, and because it is a sector which is frequently cited as having highly restrictive rules of origin. The remainder of the paper is divided into four further sections. In Section 2 we summarise some of the key conceptual issues which arise in considering rules of origin and their cumulation, as well as providing a discussion of some principal features of the Southern Mediterranean economies and their patterns of trade. In Section 3 we examine whether there is any evidence over time that introduction of the pan-European system of cumulation did indeed impact upon patterns of trade. The results provide a prima facie case for suggesting that the introduction of cumulation impacted upon patterns of trade and particularly with regard to intermediate goods. In Section 4 we turn to our principal empirical analysis. Here we explicitly derive an equation for estimating trade flows, on a cross-section basis at the sectoral level. This is then applied to the textile industry where the analysis once again provides strong evidence for the role of cumulation. The analysis explores the extent to which lack of cumulation may impact upon trade as well as the extent to which this is important for the Euromed partner countries. Section 5 concludes.

2. CONCEPTUAL AND EMPIRICAL BACKGROUND

a. Conceptual Background Rules of origin and the cumulation of those rules are an important element in understanding the evolution of trade between countries especially in the context

132 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG of preferential trading arrangements. Where a country is granted preferences unilaterally, such as the EU’s Generalised System of Preferences (GSP) or under Everything but Arms (EBA), the rules are there to ensure that non-beneficiary countries are not routeing their exports via a beneficiary country. In the case of a free trade agreement the rules prevent imports from non-member countries from entering the free trade area via the country with the lowest external tariff (trade deflection). In the context of the Barcelona process, for the Southern Mediterranean countries the rules serve to ensure that third countries which may have tariff-free access to the EU market under the GSP do not deflect trade towards them. Typically one or more of three criteria are used in determining the originating status of a good: (a) whether or not the transformation of the good results in a different tariff classification line, (b) whether or not the value of the imported intermediate exceeds a certain percentage (often 40 per cent) of the mill price of the final good, (c) whether a particular specified production process has been employed or not. The rules of origin thus specify which of these, or combination of these, applied to each given product. While rules of origin (ROOs) are thus formulated as an integral part of a process of trade liberalisation, there are nevertheless several reasons why a given set of ROOs may result in substantially less liberalisation than implied by the preferences which have been granted. First, there are the administrative and bureaucratic costs involved with administering rules of origin regimes. Although there is little formal empirical evidence for this, there is substantial anecdotal evidence to suggest that extent of preference take-up is less than it might otherwise be because of lack of information concerning correct bureaucratic procedures, or because of the costs of fulfilling those procedures. Secondly, and perhaps more fundamentally, the underlying restrictiveness of a given ROO is likely to impact upon trade (Krishna, 2003; Krishna and Krueger, 1995; Falvey and Reed, 2002; Burfisher, Robinson and Thierfelder, 2004; and Hoekman, 1993). Consider again an FTA between the EU and Morocco where the rules of origin specify that the final good is deemed as originating from Morocco if the value of non-originating intermediates does not exceed 40 per cent of the value of the final good. Suppose that prior to the FTA Morocco imports intermediate goods from Poland comprising 50 per cent of the value of the final good subsequently exported to the EU. Unless the Moroccan firm changes its source of supply of intermediates or raises the final price of the good exported, it will not be able to benefit from the tariff-free access to the EU market. If it changes its source of supply it can either source more (or all) of the intermediate from domestic sources or from the EU. In the case of the former we have trade suppression, and in the case of the latter trade diversion. Each of these involves using a higher cost intermediate than was the case prior to the application of the free trade agreement, which would therefore also raise the price of the final good.


133

ROOs are determined product by product at a very detailed level of disaggregation. In the EU context they are determined at the HS six-digit level which covers 5,142 products. Given the product-specific nature of the rules and the technical opacity of those rules, their formulation is potentially highly amenable to protectionist pressure (Duttagupta and Panagariya, 2002; and Grossman and Helpman, 1995). For example, the support of the automobile and the textile industry was important for the successful conclusion of the NAFTA agreement. Each of these industries appear to have been influential in the shaping of the relevant NAFTA rules of origin where perhaps not surprisingly the rules for each of these sectors are very strict. For textile imports the change in tariff classification rule is employed. However, as opposed to allowing a single change in the tariff classification line, the transformed good must have moved at least three tariff classification lines in order to be considered originating. In a similar vein, textiles in most EU agreements typically require a double transformation of tariff classification lines. There is also growing empirical evidence ROOs may be indeed seriously impacting upon trade flows (e.g., Gasiorek et al., 2002 and 2003; Estevadeordal and Suominen, 2003; and Inama, 2003).2 However, an issue which has received considerably less attention in the literature concerns the implications of cumulation, and in particular diagonal cumulation. This is best understood by returning to the earlier example of an FTA between the EU and Morocco where the ROOs specify that the final good is deemed as originating if the value of non-originating intermediates does not exceed 40 per cent of the value of the final good. Suppose also that Poland has duty-free access to the EU market with identical ROOs. Prior to the FTA Morocco imports intermediate goods from Poland comprising 50 per cent of the value of the final good. Under the ROOs applicable the good is thus deemed as not originating in Morocco and customs duties would be levied. However, if the Polish intermediate were shipped directly to the EU it could do so duty free. An obvious way of resolving this anomaly is for Polish intermediates used in Morocco (and Moroccan intermediates used in Poland) to be counted as originating. Under a system of diagonal cumulation Polish intermediates used in Moroccan goods exported to the EU could count as originating. Diagonal cumulation is thus possible where countries are linked by a series of identical agreements, with identical rules of origin. 2

There is also a growing awareness among policy makers and trade policy practitioners of the role and possible importance of rules of origin. The EU has, for example, recently published a Green Paper on Rules of Origin which raises questions about (i) the criteria for acquiring originating status and the appropriate legal framework so that the rules of origin perform better and contribute to the smooth working of arrangements that are increasingly geared to increased market access; (ii)the extent to which the rules and their application adequately protects and promotes the interests of the EU and EU traders; (iii) the establishment of procedures ensuring an optimal division of tasks and responsibilities between traders and the relevant authorities. For details see: (http// europa.eu.int/comm./taxation_customs/customs/origin/rules_origin_en.htm).

134 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG The example discussed earlier showed that preferential trade agreements with bilateral cumulation of rules of origin can in principle lead to trade diversion and/ or trade suppression. This will depend on the degree of restrictivness of the underlying rules of origin, and on the trade-off which the exporters face between taking advantage of the tariff preferences and sourcing intermediates from the cheapest supplier. Moving to a system of diagonal cumulation of rules of origin widens the possible source of intermediate suppliers to all those countries which are part of that system of diagonal cumulation. Hence, where constraining rules of origin led to either trade suppression and/or trade diversion, diagonal cumulation is thus likely to lead to trade creation and/or trade reorientation. The trade creation arises as suppliers can now increase their imports of intermediates from either one of the diagonal partners or from the rest of the world. The trade reorientation occurs as importers switch their sources of supply from a less efficient to a more efficient partner or rest of the world supplier. Trade reorientation is thus the exact converse of trade diversion. In addition to these effects there may also be trade diversion with respect to non-partner countries, and in providing more scope for final goods producers to source cheaper intermediates there may be trade expansion in final goods arising from lower costs and hence lower prices.3 The pan-European system of cumulation of rules of origin is precisely a system of diagonal cumulation. The system was introduced in 1997 with the participating countries being the EU, EFTA, Bulgaria, Estonia, Hungary, Latvia, Lithuania, the Czech Republic, Poland, Romania, Slovakia, Slovenia and Turkey (since January 1999). In March 2002 the decision was taken, in principle, to extend the pan-European system to the Barcelona group of countries, and the new protocol on rules of origin was subsequently endorsed by them in July 2003. However, in order to participate in the pan-European system the Southern Mediterranean countries will in turn need to sign appropriate free trade agreements with the other pan-European countries. Where a series of bilateral FTAs could result in a hub-and-spoke pattern with the EU, diagonal cumulation potentially may be important in making the participating countries part of a broader multilateral free trade area. b. The Euromed Economies – Patterns of Trade and Production In this section we take a brief look at some key statistics concerning the structure of production and trade of the Southern Mediterranean economies. We turn first to Table 1A, which gives the sectoral shares of value added and employment in manufacturing for six of these economies. The bottom row gives

3

See Gasiorek et al. (2002) for a fuller discussion.

135

RULES OF ORIGIN AND THE EU-MED PARTNERSHIP TABLE 1A Sectoral Shares in Value Added and Employment Sectors

Egypt 1997 VA

Israel 1996 Emp. VA

Jordan 1997

Morocco 1999

Emp. VA

Emp. VA

15.2 15.0 1.4 4.1 7.8 2.4 3.5 5.5 3.4 15.2 3.2 14.8 5.4 3.1

20.8 9.1 2.2 8.9 7.6 11.5 4.0 5.2 13.5 10.5 3.1 1.5 0.9 1.4

Food, bev. and tob. Textiles and apparel Leath. and footwear Wood and furniture Paper and printing Chemicals Petroluem Rubber and plastic Non-met. minerals Metals Mach. not elec. Machinery, electric Transp. equipment Prof., sc. and other

18.5 20.3 12.9 29.8 0.4 1.2 0.5 1.9 3.0 4.0 18.1 9.3 13.7 2.2 1.9 2.2 10.4 8.2 7.9 10.1 4.3 4.2 3.9 2.6 3.9 3.1 0.6 0.9

12.4 7.9 0.7 2.7 7.3 5.2 5.7 6.0 4.4 13.4 3.1 22.5 6.1 2.5

Total manuf. ($m)

6,767.8

14,107.7

27.9 4.4 1.3 3.8 5.8 17.2 7.9 3.6 15.4 7.4 2.3 1.5 0.8 0.7

1,066.2

Tunisia 1999

Emp. VA

Turkey 1999

Emp. VA

Emp.

32.7 20.2 15.9 38.3 1.1 3.0 1.6 2.1 4.2 3.2 15.0 6.6 0.0 0.0 2.5 2.8 8.9 7.7 5.9 5.5 4.1 3.7 3.1 3.1 4.7 3.5 0.3 0.3

17.4 14.2 27.8 41.3 5.1 4.4 6.6 3.5 2.9 3.2 9.8 3.9 6.3 0.4 2.5 3.7 7.3 8.0 4.7 5.8 0.4 2.4 5.4 4.9 2.5 3.1 1.3 1.2

13.2 16.7 16.2 31.8 0.8 1.5 1.6 2.6 3.0 3.0 10.7 5.2 15.6 0.8 3.8 3.6 7.3 6.8 10.3 10.4 4.6 5.2 5.4 4.6 6.4 6.4 1.1 1.3

5,660.8

3,935.9

31,295.7

Source: UNIDO Industrial Statistics Database, Rev.2, 2002.

total manufacturing value added, and shows that the largest economy was Turkey, followed by Israel, with the remaining economies being considerably smaller. For each of the economies the three largest sectoral shares in terms of both value added and employment are shaded. Looking at these it can be seen that Food, beverages and tobacco is one of the three largest industries in terms of both employment and value added for each of these economies. The other sector which figures prominently is that of Textiles and apparel. For Morocco, Tunisia and Turkey this is one of the most important sectors in terms of value added, and for all of the countries except Jordan it is one of the most important sectors in terms of employment shares. Indeed for Egypt, Morocco, Tunisia and Turkey this is the most important sector with the share of manufacturing employment ranging from 29.8 per cent in Egypt, to 41.3 per cent in Tunisia. Other sectors that are clearly important in terms of either value added or employment for at least three of these economies are Chemicals, Non-metallic minerals, and Metals. Overall, the table shows there are some considerable similarities in production structure for most of these economies, in particular in the emphasis on Food, beverages and tobacco, and on Textiles and apparel. Perhaps not surprisingly the country that differs the most significantly from this is that of Israel, where, for example, the most significant sector in terms of value added is Machinery. It is sometimes suggested that the structure of the Southern Mediterranean economies is so similar that there is perhaps little to be gained from regional integration amongst them. Or, in the context of this paper, that the lack of diagonal

136 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG TABLE 1B Finger-Kreinin Indices of Trade Similarity Egypt Egypt Israel Jordan Morocco Tunisia Turkey

21.10 32.36 29.49 33.84 39.42

Israel

Jordan

Morocco

Tunisia

Turkey

51.57

68.53 63.31

63.52 60.28 77.34

61.81 60.49 67.77 76.27

66.19 59.70 70.14 71.29 64.86

35.21 25.17 31.50 35.56

46.66 43.69 55.17

71.90 46.87

49.62

cumulation between, e.g., Morocco and Egypt is of little importance. Tables 1B and 1C shed some light on this, where the former provides aggregate measures of the degree of export and import similarity across the Southern Mediterranean and the latter details the key import and export industries by country. In Table 1B we have computed Finger-Kreinin indices of export and import similarity. Formally the index is defined as:

FK ab =

∑ [min(Sia, Sib )] ∗ 100, i

where Sia denotes the share of commodity i exports (or imports) by country a; and correspondingly Sib denotes the share of commodity i exports (or imports) by country b. Since the index sums only the minimum value of export shares between the two countries it ranges from 0 to 100. An index of 100 represents a complete overlap in exports, whereas an index of 0 represents no overlap. The table gives the computed FK index with respect to the countries’ total imports (dark shaded, top-right) and total exports (light shaded, bottom left). The index was calculated for 2001 on the basis of two-digit SITC trade flows. There are several features which emerge clearly from this table. First, these indices do not overall suggest a very high similarity in trade patterns for these economies, though there is a higher similarity for imports in comparison to exports. The average index for imports across all bilateral pairings is 66.7, and for exports the corresponding index is 44.1. In comparison, the average import index for all the bilateral pairings between each of the CEFTA countries and the Baltic states is 72.3, and the average export index is 48.9. There is thus greater similarity between the CEFTA and Baltic states than there is between the Southern Mediterranean economies. Clearly, however, there are cross-country differences. For imports, the highest index is between Morocco and Tunisia (76.27), and the lowest is between Egypt and Israel (51.57). On the export side most of the indices are below 50, with the lowest again being between Israel and Egypt (21.1) and the highest between Morocco and Tunisia (71.9). Overall the table does not suggest a substantial degree of overlap in the specialisation of trade.

62.08 78.74

30.81

48.60

Total (5)

Total (10)

5.75

6.96

37.59

4.74

7.04

4.90 4.47

4.88

9.73

Source:Comtrade data.

64.60

49.15

10.36 6.67

4.94

20.82

6.36

M

M X

Israel

Egypt

6.83

Fish etc. Cereals etc. Veg. and fruit Textile fibres Crude fertilisers Petroleum etc. Organic chemicals Inorganic chemicals Med. products Fertilisers Textile yarn etc. Non-metallic min. Iron and steel Power generating mach. Special indust. mach. General indust. mach. Office machinery Telecom. equip. Elec. mach. Road vehicles Clothing etc. Misc. manuf. Transactions n.e.s.

Note: M = imports; X = exports.

03 04 05 26 27 33 51 52 54 56 65 66 67 71 72 74 75 76 77 78 84 89 93

Description

77.02

63.14

3.68

13.65 11.40

31.20

3.21

X

54.88

38.40

8.95

4.71

6.25

13.51

4.98

M

Jordan

65.24

48.41

6.22 12.92

8.43

14.89

5.95

X

60.13

43.70

5.61 4.67

12.71

13.71

7.00

M

Morocco

86.61

67.34

32.78

9.23

6.35

7.44

11.54

X

TABLE 1C The Sectoral Breakdown of Trade: Shares of Imports and Exports

58.84

41.46

7.56 6.89 5.57

5.43

16.01

M

Tunisia

82.12

68.42

40.12

12.23

4.45

4.14

7.48

X

52.14

30.22

5.18 4.38

4.73

4.64

11.29

M

Turkey

72.47

55.99

7.33 21.26

7.98

12.58

6.84

X


137

138 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG We explore this further in Table 1C, which focuses on the sectoral breakdown of imports and exports. For each of the countries included we have calcululated the share of imports and exports at the two-digit SITC rev.3 level. The table then gives for each of these countries the five largest import (M) and export (X) industries. Thelast two rows give the aggregate shares of the five and ten largest industries. In aggregate the five largest import industries account for between 30 and 49 per cent of all imports, and the ten largest between 48 and 64 per cent of all imports. For exports the five largest account for between 48 and 68 per cent, and the ten largest for between 65 and 84 per cent of all exports. For each of the economies this suggests a high degree of specialisation particularly with regard to exports. Each country (except Egypt) has at least one industry which comprises more than 10 per cent of all imports, and more than 20 per cent (except Jordan) of all exports. In terms of imports there are five sectors which are one of the top five industries for three or more countries – Cereal (SITC 04), Petroleum (SITC 33), Textile yarn (SITC 65), Electrical machinery (SITC 77), and Road vehicles (SITC 78). In terms of exports there are three such sectors – Vegetables and fruit (SITC 05), Electrical machinery (SITC 77) and Clothing etc. (SITC 84). In terms of the SITC two-digit industries listed above textiles comprises most of industries 26 (Textile fibres), 65 (Textile yarn) and part of 84 (Clothing etc). Taking these in aggregate it is clear that textiles and clothing are significant import and export industries for the Southern Mediterranean countries. Textile yarn comprises 4.64 and 6.25 per cent of imports for Turkey and Jordan, and 12.71 and 16.01 per cent respectively for Morocco and Tunisia. On the export side, Textile yarn comprises 6.96 per cent of Egyptian exports, and 12.58 per cent of Turkish exports, while Clothing etc. comprises 5.75 per cent of Egyptian exports, 12.92 per cent for Jordan, 21.26 per cent for Turkey, and rising to 32.78 and 40.12 per cent for Morocco and Tunisia, respectively. These are substantial figures and emphasise again the importance of the textile industry for these economies. It is also interesting to note for example the importance of Textile yarn imports (largely intermediates) into Jordan, Morocco and Tunisia, coupled with high export levels of Clothing etc. (final goods) for the same economies. Finally, we turn to the geographical pattern of trade. Table 1D details the share of imports and exports by partner ‘country’ at the aggregate level and for textiles. For purposes of comparability with our empirical analysis later in the paper, here we have defined the textile industry more narrowly than above (i.e. excluding apparel) as defined by the ISIC rev.3 classification.4 The partner countries we consider are the EU15, the remaining members of the pan-European system (CEFTA, EFTA countries and the Baltic states) which we call the pan-EU, the Southern Mediterranean countries (Med), and the Rest of the World. In order to 4

The figures here are calculated by using the correspondence between the ISIC rev.3 and the SITC rev.3 at the SITC five-digit level of classification.


139

TABLE 1D Trade by Geographical Origin Reporter

Partner

Total Trade Imports

Textile Trade Exports

Imports

Exports

1994

2001

1994

2001

1994

2001

1994

2001

Egypt

EU15 Pan-EU Med Rest of World

37.65 5.57 2.35 54.43

32.86 4.58 3.06 59.50

44.59 2.78 13.10 39.52

30.91 1.14 13.45 54.50

29.70 3.10 5.11 62.09

17.94 0.86 5.44 75.76

66.21 2.97 6.05 24.76

57.40 1.67 2.94 38.00

Israel


50.94 7.75 0.81 40.50

45.63 6.59 1.01 46.77

30.18 3.87 1.02 64.93

28.53 3.27 1.69 66.51

58.77 5.46 4.71 31.06

39.66 4.19 6.26 49.89

57.13 1.05 1.13 40.69

49.73 1.05 3.56 45.67

Jordan


32.85 3.92 7.34 55.89

31.42 2.87 7.65 58.06

8.61 0.74 9.19 81.46

4.87 0.74 14.08 80.31

25.48 4.46 11.86 58.19

12.14 0.45 35.71 51.69

10.17 0.48 8.17 81.18

4.87 0.07 16.53 78.53

Morocco


54.92 3.94 3.29 37.85

57.51 2.51 3.43 36.56

62.55 2.06 4.42 30.97

73.49 1.55 1.90 23.05

84.70 1.23 2.01 12.05

87.30 0.36 1.81 10.54

80.68 0.56 3.49 15.28

92.02 0.15 0.21 7.61

Tunisia


72.64 2.80 3.96 20.60

70.66 2.86 3.90 22.59

77.40 1.34 5.93 15.33

80.32 0.86 3.34 15.49

93.66 0.60 0.81 4.92

91.87 0.42 1.22 6.50

81.58 0.56 1.41 16.45

90.24 0.22 0.27 9.27

Turkey


45.35 5.73 2.48 46.44

48.56 5.93 4.70 40.82

48.39 5.89 6.58 39.14

52.56 5.22 7.91 34.31

31.46 2.90 1.84 63.80

47.76 2.80 1.40 48.03

62.51 7.80 3.52 26.18

63.61 5.66 4.87 25.87

Note: For Israel the figures are for 2000. For Egypt and Jordan the figures for the textile industry are for 1999.

prevent any anomalies, arising from one year’s data we have taken the average shares over 1993–95, and 1999–2001. If we consider first total trade we can see that over 1993–95 the EU is a key supplier of imports to all these economies, and the principal supplier for Israel, Morocco and Tunisia (with the respective shares of imports being 50.9, 54.9 and 72.64 per cent). A very similar picture emerges over 1999–2001 where once again EU is the principal supplier for Morocco (57.5 per cent) and Tunisia (70.6 per cent) as well as Turkey (48.6 per cent). On the export side the importance of the EU over both time periods is for most countries even more pronounced, with Jordan being the major exception. There are also some interesting changes over time. With regard to both imports and exports, three countries – Egypt, Israel and

140 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG Jordan – see a reorientation away from the EU and towards the Rest of the World. This is most marked for Egypt where the EU import share declined from 37.6 to 32.8 per cent, and the export share from 44.6 to 30.9 per cent. In contrast, Morocco and Turkey see the EU import share rise (while Tunisia experiences a small fall), while each of these economies see a rise in the share of exports going to the EU which is most marked for Morocco (from 62.5 to 73.5 per cent). It is also worth noting the small import shares accounted for by the pan-EU countries and the other Med countries. For the latter the import share from the other Southern Mediterranean countries over 1993–95 ranges from 0.8 for Israel, 2.5 for Turkey, to 7.3 per cent for Jordan. These shares remain relatively stable over time, except for Turkey where the share rises to 4.7 per cent by 2001. With regard to exports there is greater variation in the share accounted for by the other Southern Mediterranean countries. Hence, over 1993–95, 13.1 per cent of Egyptian exports went to the Med, while for Israel the figure was 1.02 per cent. Over time the biggest change in these shares is experienced by Jordan (from 9.2 to 14.1 per cent), while Turkey also sees a rise (from 6.6 to 7.9 per cent). In contrast, Morocco sees a large fall (from 4.4 to 1.9 per cent) and Tunisia a slightly smaller one (from 5.9 to 3.3 per cent). If we look at textile trade the EU is the principal source of imports over both 1993–95 and 1999–2001 for Israel, Morocco and Tunisia – with the Moroccan and Tunisian shares being as high as 84.7 and 93.7 per cent respectively in the earlier period, and 80.1 and 81.6 per cent respectively in the latter period. The EU is also the principal export market over the period for Egypt, Israel, Morocco, Tunisia and Turkey. Once again the shares of imports and exports with the other pan-EU countries and with the Med countries is extremely low. Interesting also are the changes over time. Egypt, Israel and Jordan all see a substantial reorientation of textile trade away from the EU and either towards the Rest of the World (for Egypt and Israel) or towards the other Med countries (for Jordan). In contrast, Morocco, Tunisia and Turkey all see a reorientation of textile trade towards the EU and this is generally true of both imports and exports. Noticeable also is the decline in imports coming from other Southern Mediterranean countries for Morocco and Turkey, and the decline in exports going to these countries for Egypt, Morocco, Tunisia and Turkey. This would appear to strongly suggest that the Euromed agreements together with their rules of origin for textiles have contributed to the changing patterns of trade. Such effects can also be seen elsewhere and are no doubt important in explaining the strong regionalisation and vertical integration of the textile trade over the last fifteen years. Hence, with regard to NAFTA, Mexican exports of clothing to the US rose from 4.8 per cent in 1993 to 16.2 per cent in 1999, while at the same time the share of Asian exports to the US fell from 80 per cent in 1993 to less than 50 per cent in 1999. Similarly, US exports of textiles to Mexico rose from 17 per cent in 1996 to 34 per cent in 1999 (Avisse and Fouquin, 2003). Even though multilateral liberalisation, with the


141

Agreement on Textiles and Clothing, represents a major step forward in the process of textile integration, nevertheless underlying rules of origin within the various regional trading agreements may serve to protect, at least to a degree, regional and domestic markets.

3. TREND ANALYSIS OF THE PAN-EUROPEAN SYSTEM

The aim of this section is to establish whether there is any prima facie evidence on the impact of cumulation from the trade data following the introduction of the pan-European system in 1997. Prior to 1997 diagonal cumulation applied between the EU and EFTA. After 1997 the system was extended to the countries of Central and Eastern Europe and the Baltic states. If cumulation were to impact upon trade flows in the first instance one would expect this to be true principally within intermediate sectors than final goods sectors, and secondly one would expect a re-direction of trade towards imports from either the EU or the other pan-EU countries relative to the rest of the world. Consider Figure 1, which for SITC 773 (Electrical distribution equipment) plots Polish imports from three sources – the EU, the other Pan-European countries (labelled pan-EU) and the rest of the world (ROW). To make comparison easier for each series in the figure the imports are plotted in index form (i.e. imports relative to the value of imports in 1997). Hence EUI gives the value of Polish imports in each year from the EU relative to imports from the EU in 1997. PEI gives the index for imports from the other pan-EU countries, and RWI from the rest of the world. From the figure we can see that imports of 773 from the other pan-European countries increased sharply in the post-1997 period, both in comparison to prior to 1997, and in comparison to imports from the EU or from the Rest of the World.5 FIGURE 1 Polish Share of Imports by Source: SITC 773

5

The reader will note that the increase in pan-EU imports appeared to happen after 1998 rather than 1997. Interestingly, for the case of Poland the pan-European system was not implemented until 1998.

142 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG Of course this is but one example for one industry and one country. To examine this more formally we have taken the 20 most important SITC three-digit industries by value of world imports for the non-EU/pan-EU economies.6 We then calculated these indices for each industry and country for which we had reliable data over this time period. The countries covered include the Czech Republic, Hungary, Latvia, Lithuania, Romania, Slovakia, Slovenia, Poland, Iceland, Norway and Switzerland, and the 20 industries comprise nearly 34 per cent of the total world imports of these countries in 1999. We then take the ratios PEI/RWI, and EUI/RWI and run a simple time-trend panel regression from 1994 to 2001. Specifically the equations that we have estimated for each of the industries are: PEI = α 0 + α 1 T + α 3 D + α 4 D*T RWI

(1)

EUI = α 0 + α 1 T + α 3 D + α 4D*T, RWI

(2)

i.e. where T is a time trend, and D is a dummy which takes a value of 0 prior to 1997, and 1 subsequently, and then D*T is a standard interaction term. If lack of cumulation was indeed important then one would expect the coefficient on D*T to be positive and statistically significant. This would represent a reorientation of trade towards either the pan-EU economies or the EU itself, and in each case away from the Rest of the World. Table 2 gives the results for the 20 industries for each of the above regressions where we focus on the results for the interaction term as this is the variable of interest. The third column of the table gives the BEC classification for each of the industries which divides the industries into final-use categories, where I = intermediates, C = capital goods, and F = final consumption goods. In all but one of the 20 industries the coefficient is positive, and in 10 of these cases the coefficient is statistically significant with regard to imports (relative to imports from the Rest of the World) from the pan-European countries, and in four of these cases it is also statistically significant with regard to EU imports. In only one case do we have a negative coefficient with regard to both pan-EU and EU imports but this is not statistically significant. In other words, for 10 of the CEECs most important import industries there was a statistically significant shift towards greater imports from other panEuropean partners after 1997. In and of itself this is an interesting result which suggests that changes in the cumulation of rules of origin may be an important part of the explanation. Secondly, if we consider the relevant BEC categories for these industries we can see that six of these industries are purely intermediate 6

Except SITC 3 (fuels, lubricants, etc.) and 9 (goods not classified by kind).


143

TABLE 2 Trend Regression Results Industry

BEC

SITC

Description

781 542 764 752 784 776 772 778 759 893 699 821 641 728 713 541 684 515 874 792

Passenger motor vehicles Medicaments Telecommunications equipment Automatic data processing equip. Part, tractors, motor veh. Transistors, valves etc. Electrical switching, relay circuits Electrical machinery Parts for office machines Articles n.e.s. plastics Manuf. base metals Furniture etc. Paper and paper board Other machine parts Internal combustion engines Medicines Aluminium Organo-inorganic compounds Measure, control, instruments Aircraft associated equip.

I/F F C C I I I I/F/C I I/F I F I C/I I/F/C I I I C/ I C/F/I

EUI/RWI

PEI/RWI

D*Trend

p-value

D*Trend

p-value

0.281* −0.046 0.054 0.071 0.587* 0.064 −0.088 −0.002 0.266** 0.050 −0.076 0.049 0.184 0.220 0.332 0.250*** 0.062 0.519 0.153 10.630

0.058 0.622 0.522 0.317 0.055 0.532 0.425 0.974 0.010 0.514 0.103 0.513 0.142 0.250 0.503 0.008 0.561 0.181 0.218 0.260

2.083*** 0.153 0.915* 0.517** 0.589*** 0.852** 0.049 0.083 0.920*** 0.095 −0.016 0.182 0.268* 0.277 0.687** 0.201* 0.050 1.210* 0.644 0.695

0.000 0.244 0.061 0.036 0.007 0.043 0.739 0.402 0.000 0.440 0.842 0.240 0.052 0.126 0.017 0.056 0.677 0.088 0.127 0.772

Note: *, ** and *** denote statistical significance at the 10 per cent, 5 per cent and 1 per cent levels, respectively.

goods industries, two of them are mixed sectors, and two of them are capital goods industries. Clearly there are many factors which impact upon trade flows. Nevertheless, it is somewhat striking that after the introduction of cumulation of rules of origin there was a marked reorientation of trade towards the panEuropean economies, and partially also towards the EU in 14 of the 20 most important sectors, almost all of which are intermediate goods sectors. The evidence from this analysis would thus appear to suggest that there is indeed a prima facie case that allowing for cumulation in 1997 impacted upon patterns of trade between the EU and the non-EU pan-European countries. The next section of this paper turns to a more detailed and more explicit sectoral analysis where we focus on the textile industry and the impact of the lack of cumulation on the Southern Mediterranean economies.

4. CUMULATION AND TEXTILES – A SECTORAL ‘GRAVITY’ MODEL

In this section we explore the possible impact of cumulation for a given sector – that of textiles. The estimating equation is formally derived from underlying

144 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG theory, and provides a means for a cross-section analysis of trade flows. The reader will note that the derived equation is analogous in many ways to the gravity model, which has been widely used for analysing aggregate trade flows. The standard gravity model is, however, unsuitable for sectoral-level analysis, and hence the need for an alternative.7 The remainder of this section first details the model and the underlying data, and then we turn to a discussion of our results. a. The Model We assume there is a representative consumer in each economy who faces an overall budget constraint, Mj, which is given by: Mj = Uj Ej,

(3)

where Uj is the utility function, and Ej is the unit expenditure function in country j. Each of these is represented by a Cobb-Douglas aggregator and takes the following form: Uj =

∏ ( X jk )

β kj

(4)

k

Ej =

∏ (Pjk )

β kj

,

(5)

k

where X jk represents the sub-utility function over individual product varieties produced in industry k in country j, and P jk represents the aggregate price index. We assume a two-stage budgeting procedure. In the first stage consumers maximise their utility subject to their overall budget constraint, and allocate expenditure between the k industries. The β jk’s represent the shares in total expenditure. In the second stage consumers maximise the sub-utility function over individual product varieties, given by X jk, subject to the budget allocated to that industry. Hence, expenditure is divided within each group on individual product varieties. The sub-utility function over individual product varieties is assumed to be a constant elasticity of substitution aggregator, with elasticity of substitution common to all countries denoted ε k. The sub-utility function takes the following form: ⎡ X jk = ⎢∑ ∑ ( xijlk ) ⎢ i =1 l =1 ⎣ j

7

ni

ε k −1 εk

⎤ ⎥ ⎥ ⎦

εk ε k −1

,

(6)

At the aggregate level it may be reasonable to suppose that trade between pairs of countries depends on each country’s GDP, population, distance, trade barriers etc. However, at the detailed sectoral level, one would expect that determinants of comparative advantage would also play an important role.


145

k where x ijl is the output of a single variety of a good in industry k, produced in i and sold in j by firm l. Hence consumers demand each variety produced from all sources. Each industry contains a number of firms, with n ik denoting the number of firms in industry k located in country i. In order to simplify the above it is assumed that for a particular industry and country all firms are symmetric. Equation (6) can then be simplified to:

⎡ X jk = ⎢∑ nik ( xijk ) ⎢ i =1 ⎣ j

ε k −1 εk

⎤ ⎥ ⎥ ⎦

εk ε k −1

.

(7)

Consumer demands for the aggregate quantity indices are derived (by Shephard’s lemma) by partial differentiation of the expenditure function. The sub-utility function over differentiated products can then be interpreted as a quantity index. Analogously, and dual to the quantity index, P jk, represents the aggregate price index for all varieties of industry k consumed in country j. The form of P jk is given by: 1

⎡ j ⎤ 1− ε k k k k 1 − εk Pj = ⎢∑ ni ( pij ) , ⎥ ⎢⎣ i =1 ⎥⎦

(8)

where p ijk and x ijk denote the price and quantity of a single product variety of industry k produced in country i and consumed in country j. The expenditure function over differentiated products is then given by: E jk = X jk Pjk.

(9)

Demand for individual varieties of any given product, x ijk , is again given by Shephard’s lemma, by differentiation of the expenditure function with respect to individual prices, p ijk : εk

⎛ Pk ⎞ x = X ⎜ jk ⎟ . ⎝ pij ⎠ k ij

k j

(10)

Dropping industry superscripts, using equation (9), and rearranging gives: ε −1

⎛P⎞ pij xij = E j ⎜ j ⎟ . ⎝ pij ⎠ This equation can then be rearranged to yield the following:

(11)

146 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG

Xij = ni pij xij = Qi E j

(P (1 − t ) ) ∑ E (P (1 − t ) ) j

j

ij

j

−1

ij

ε −1

−1

ε −1

,

(12)

j

where Qi is the value of production in country i in the sector, Ej is the value of expenditure on the sector in country j, tij are the (ad valorem) ‘costs’ of trade between i and j (e.g., transport costs, tariffs etc.) and Pj is the price index defined earlier. Hence, exports from one country to another country in a given sector depend on: production in the exporting country; demand in the importing country; and the price index in the importing country relative to the price index (weighted by expenditure) in all other countries; as well as the costs of trade between the exporting and the importing country, relative to the costs of trade between all other countries. b. Estimating Equation and Results Based on the above, the equation which we actually estimate, using a Tobit estimation procedure, takes the following form:8 Ln(Xij) = α 0 + α1Ln(Qi) + α 2Ln(Ej ) + α3Ln(RUVij) + α4Ln(tariffij ) + α5Ln(Distij) + α6PTAij + α 7Borderij + α8Languageij + α 9Quotaij + α10ROOij , where: represents exports of textiles from country i to country j (in millions of $). The source for this data is the UN PC-TAS database. Qi: represents total production of textiles in country i (in millions of $), where the source was the Unido Industrial Statistics Database. We would expect the coefficient on this to be positive, as the greater is the level of production the higher would be the expected level of exports. Ej: represents total apparent consumption of textiles in country j (Qj + Mj − Xj), where Mj and Xj are the total imports and exports of textiles in country j. Again we would expect the coefficient to be positive as the greater is consumption in a given country, the higher would be the expected level of imports. RUVij : are relative unit values which are used to proxy the price terms in equation (12).9 The higher is the price of a given exporter relative to Xij:

8

Strictly speaking, because trade values are bounded from below by zero a Tobit procedure is the correct one to use. In practice in most cases there is little difference in the results between using the Tobit methodology and a standard OLS procedure. 9 As is well known the use of (aggregate) unit-values to represent prices is not straightforward. For this analysis we have taken great care (and time) in constructing these data. First, we calculated the


147

the average price of all other exporters the lower would be the expected level of exports. We would therefore expect this coefficient to be negative. Tariffij: Gives the bilateral MFN or preferential average tariffs between countries. Distij : Gives the distance between the economic centre of gravity of the respective countries. As is now standard in gravity modelling, in addition to the above we supplement the above with dummy variables in order to try and capture other factors, and in particular institutional arrangements between countries which are typically expected to impact upon trade flows. These include regional trading arrangements (PTAij ), or dummies to capture other affinities between countries such as a common language (Languageij ) or a common border (Borderij). We also include a dummy in order to capture quotas between countries (Quotaij ). Following our earlier work (Gasiorek et al., 2002) we also include a dummy variable designed to capture the impact of the effects of cumulation. In particular the aim is to explore whether the lack of cumulation between countries may act as a constraint on trade between them. Specifically, the objective is to determine whether trade is lower in those cases where an importing country (e.g., Tunisia) has a PTA with the EU but there is no diagonal cumulation between that importing country and the exporting country (e.g., Poland). Note that, when considering the role of diagonal cumulation, one is considering the relationship between three countries or country groupings: the exporting country, the importing country, and those countries forming the system of diagonal cumulation (in this case the panEuropean system). Given this three-part relationship which underlies diagonal cumulation the ROOij dummy takes a value of 1, if the importing country has a preferential trading agreement with the EU without diagonal (pan-European) cumulated rules of origin with the exporting country, and a value of 0 otherwise. If cumulation impacts upon trade flows we would thus expect a negative sign on the rules of origin dummy variable. The analysis is then carried out for the textile industry, ISIC (rev.3) 17, for the years 1995 and 1999, and where our estimations are based on trade flows between unit values in each import market at the five-digit SITC level. We then examined these data both for missing values and for clearly anomalous results. Where necessary we interpolated any anomalies by taking the average unit value in any given import market at the five-digit level of aggregation. The aggregate unit values were then constructed on the basis of import weights. Hence, the formula for the relative unit value is: RUVj =

UV

∑ UVijs s

wjs

×

Xijs , where UV stands for unit value, X are ∑ Xijx s

exports, i is the exporting country, j is the importing country, w represents the world, and s represents a given five-digit SITC category.

148 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG TABLE 3A Sectoral ‘Gravity’ Regression 1995 Constant Ln(Qi ) Ln(Ej ) Ln(RUVij ) Ln(Dist ij ) PTA ij EUij EFTA ij CEFTA ij Borderij Languageij Quota ij Tariff ij ROOij Tariff – CefBal Tariff – Med Tariff – Rest ROO – CefBal ROO – Med ROO – Rest

−16.51*** 1.30*** 0.99*** −0.83*** −1.12*** 0.44***

0.39 0.66*** 0.57*** −0.05*** −1.31***

1999 −17.70*** 1.41*** 1.00*** −0.80*** −1.14*** 0.29***

0.12 0.65*** 0.18 −0.06*** −1.66***

1995

1999

−16.32*** 1.29*** 1.00*** −0.81*** −1.15***

−17.86*** 1.41*** 1.01*** −0.79*** −1.13***

0.26** 4.90*** 2.18*** 0.27 0.64*** 0.55**

0.20* 4.38*** 0.39 0.13 0.63*** 0.09

−0.06*** −0.03*** −0.05*** −1.54*** −1.94*** −0.85***

−0.03 −0.05*** −0.07*** −2.35*** −1.96*** −1.15***


37 countries – all of the EU countries, three EFTA countries (Iceland, Norway and Switzerland), the CEFTA countries, the Baltic states, five countries taking part in the Barcelona process (Turkey, Jordan, Israel, Egypt, Tunisia), as well as the US, Canada, China, Japan and Australia.10 Table 3A gives the results for four sets of regressions. In the left-hand (shaded) panel we give the results where we work with aggregate PTA, ROO and tariff matrices. In the right-hand panel we then disaggregate these into key country groupings. Consider first the left-hand panel – all our key variables have the expected sign and are statistically highly significant. The coefficients on production in the exporting country and consumption in the importing country are positive, and the coefficient on relative prices (unit values) is negative. Similarly distance is negative, and a common language is positive and in all cases statistically significant. In aggregate we can see also that the PTA and tariff dummies have the expected sign and are highly significant. The coefficient on tariffs suggests that a 1 percentage point increase in a tariff, reduces trade by 5 per cent in 1995 and 6 per cent in 1999. Interestingly, the coefficient on quotas is positive for both years, and statistically significant for 1995. In principle, one would 10

Note that there are two countries ‘missing’ here – Bulgaria and Morocco. This arises from the lack of data in the underlying data set.


149

expect quotas to restrict trade and so one might expect this coefficient to be negative. However, this will depend on a couple of factors. First, it will depend on whether the quotas materially restrict trade. It is hard to get clear information on this. The available information on quota utilisation rates (in particular with regard to the EU) suggests that in many instances these are well below 100 per cent. This suggests that quotas in place for the countries in our sample may not be restricting trade. Secondly, even if quotas do restrict trade for the countries in our sample, the quotas are in place vis-à-vis the principal textileexporting countries. Our quota dummy could then be identifying who are the principal textile exporters, and it is not surprising then that the coefficient may be positive. If we now turn to the ROO dummies we can see that this variable is negative and statistically significant in all the estimations. The size of the coefficient rises between 1995 and 1999 for each category of trade. The percentage equivalent of these dummies can be found by taking [exp(dummy) − 1]*100. Applying this suggests that in aggregate the lack of cumulation implies trade being lower by up to 73 per cent in 1995, and 81 per cent in 1999. Looking at the disaggregated results once again the coefficients have the expected sign and are significant in almost all cases. If we focus on the ROO dummies, the biggest impact of the lack of cumulation in 1995 is for the Med economies where the results suggest that trade with non-cumulating countries was up to 86 per cent lower than it would otherwise have been, with a very similar figure applying in 1999. In 1999, the lack of cumulation would appear to have the biggest impact on the CEFTA and Baltic countries, which also experience the biggest change in the ROO coefficient between 1995 and 1999. In contrast there is very little change in the Med coefficient over this period. These changes are interesting and highly consistent with the relevant changes in rules of origin. The CEFTA and Baltic countries became part of the pan-European system of cumulation in 1997, and therefore one would expect a change in the coefficient as the lack of cumulation now applies to a different set of partner countries. The rise in the coefficient suggests that the lack of cumulation is more significant with regard to the CEFTA and Baltic countries’ trade with the rest of the world and the Med countries, than with the EU and EFTA. Similarly, as there were no substantial changes in the rules of origin regimes facing the Med countries, one would expect this coefficient to be relatively stable over time. In Table 3B we provide further evidence of the possible impact of cumulation by disaggregating the ROO coefficient for each of the Southern Mediterranean countries in the model.11 Given the relative significance and stability of the PTA and tariff coefficients, and in order to focus on the role of rules of origin we work with the aggregate coefficients for these variables. From the table we can 11

Note that Jordan is not included in 1999 due to lack of data.

150 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG TABLE 3B Cumulation and the Southern Mediterranean Countries 1995 Constant Ln(Qi ) Ln(Ej ) Ln(RUVij ) Ln(Dist ij ) PTA ij Borderij Language ij Quota ij Tariffij ROO ROO ROO ROO ROO ROO ROO

– – – – – – –

CefBal Turk Jor Is Eg Tun Rest

1999

−15.82*** 1.28*** 0.98*** −0.85*** −1.17*** 0.48** 0.36 0.61*** 0.45** −0.04***

−17.10*** 1.39*** 0.98*** −0.81*** −1.16*** 0.31** 0.12 0.61*** 0.09 −0.05***

−1.48*** −1.85*** −1.12 −1.30*** −1.04 −2.82*** −0.77***

−2.12*** −1.57** −2.18*** −1.47*** −2.22*** −1.13***


immediately see that there are substantial differences across the countries. For 1995 the largest coefficients are for Tunisia and Turkey, and the smallest for the Rest and for Egypt (though the latter is not significant). In 1999 the biggest impact is again for Tunisia, Israel and Turkey and once again the smallest is for Egypt and the rest. A priori one would expect that, ceteris paribus, the greater is the degree of trade liberalisation (e.g., tariff reductions), the greater would be any constraining impact of rules of origin. This is simply because where there are high tariffs it is more likely that these will be serving to restrict trade irrespective of the rules of origin regime. This issue is explored more fully below, but the results in this table lend support to this – the largest impact on trade from lack of cumulation appears to be for those economies – Tunisia, Turkey and Israel – who have liberalised trade with the EU the most. Note, also, that there is a substantial change in the Turkish coefficient which declines from −1.85 to −1.57 from 1995 to 1999. Once again, this is highly consistent with the changes in the rules of origin regimes. Turkey joined the pan-EU system in 1999, and hence cumulation was then possible with the CEFTA, Baltic and EFTA countries. The decline in the coefficient would thus suggest that the lack of cumulation was more significant with regard to these countries for Turkey, than with regard to the Southern Mediterranean countries and the rest of the world. We now turn to two further sets of experiments where we investigate first the interaction between tariffs and the lack of cumulation, and secondly we explore

151

RULES OF ORIGIN AND THE EU-MED PARTNERSHIP TABLE 3C Tariffs and Cumulation

Tariff Threshold Constant Ln(Qi ) Ln(Ej) Ln(RUVij) Ln(Distij) PTAij Borderij Languageij Quotaij Tariffij ROOLOW ROOHIGH RESROO

The Interaction Between Tariffs and Cumulation

Trade and the 1997 Pan-EU Entrants

1995

1999

1995

1999

8.50 −16.15*** 1.30*** 0.97*** −0.82*** −1.11*** 0.44*** 0.43 0.60*** 0.78** −0.05*** −1.70*** −1.04***

4.50 −17.50*** 1.41*** 0.99*** −0.79*** −1.14*** 0.29** 0.10 0.65*** 0.41 −0.06*** −2.31*** −1.41**

−17.23*** 1.29*** 1.09*** −0.66*** −1.25*** 0.64*** 0.04 0.73*** 0.61*** −0.05***

−17.37*** 1.36*** 1.12*** −0.69*** −1.37*** 0.37** −0.18 0.68*** 0.15 −0.08***

0.06

0.46**


whether trade rose between those countries, which became part of the system of pan-EU cumulation in 1997. The results for the former are given in the left-hand (shaded) panel of Table 3C, and for the latter in the right-hand panel. As previously mentioned there are likely to be important interactions between the lack of cumulation, and the impact of tariffs – the higher the tariffs the more likely it is that these are restricting trade as opposed to the lack of cumulation. In order to focus on this we divide the ROO matrix into two sub-matrices – ROOHIGH and ROOLOW. ROOHIGH then includes all the cases where tariffs are equal to or above a certain threshold, and ROOLOW all the cases where tariffs are below that threshold. However, as we do not know the appropriate level of the threshold we proceed by using an iterative procedure. The model was estimated by fixing a threshold level for different values in a range 0.5–20 per cent with a step size of 0.5. The threshold level selected is then the one that provides the highest maximum log likelihood. The results of this are given in the left-hand panel of Table 3C. The third row gives the tariff threshold. Hence consider the first column of results. Here we see that the threshold tariff level for 1995 is 8.5 per cent. Where tariffs are less than this the impact of the lack of cumulation between countries is high (ROOLOW = −1.7), and where tariffs are greater than or equal to this the impact of the lack of cumulation is much lower (ROOHIGH = −1.04). In 1999 the threshold tariff is lower (4.5 per cent) and the pattern of results is very similar. These results

152 PATRICIA AUGIER, MICHAEL GASIOREK & CHARLES LAI-TONG suggest that the height of the tariff does indeed significantly affect the impact of the lack of cumulation – and that the higher the tariff, the smaller the impact. In the right-hand panel of the table we explore what can be deduced by formally comparing 1995 and 1999. The introduction of the pan-European system of cumulation was introduced in 1997 and if the rules for textiles were constraining trade, then one would expect trade between the newly cumulating countries to rise. We proceed, therefore, by running a regression for each of the years, where we include a dummy variable (RESROO) for all those 1995 countries who became part of the pan-European system in 1997.12 The expectation therefore is that trade between these countries would thus have risen as a result of cumulation, and that the net change in this coefficient would be positive. The results show a small positive, though insignificant, coefficient (0.06) for 1995, and a positive significant coefficient (0.46) for 1999. This would appear to indicate clearly that trade between those countries that became part of the panEuropean system of cumulation (principally the CEFTA countries and the Baltic states) rose relative to their trade with other countries between 1995 and 1999.

5. CONCLUSIONS

This paper has addressed a key issue, which is proving of increasing concern to policy makers and trade policy practitioners. That issue concerns the way in which rules of origin and lack of cumulation of such rules can serve to distort patterns of trade and production. Rules of origin are formulated, in principle, in the context of a process of unilateral, bilateral or multilateral trade liberalisation. However, the extent and impact of that process of liberalisation may well be distorted or lessened by the application of particular rules of origin. Rules of origin are by their nature sector specific and thus potentially amenable to protectionist pressure from interest groups. It is perhaps then not surprising that these rules are typically perceived as being particularly tight in certain traditionally protected sectors such as textiles. It is also interesting that the textile industry lobby has been particularly keen to promote the extension of the pan-European system of cumulation to the Southern Mediterranean. Their objective here is clearly one of attempting to maintain their competitiveness in the face of the ongoing changes in the industry. There has been a strong regionalisation of the textile trade over recent years, and while the Agreement on Textiles and Clothing is likely to have a major impact, the presence of constraining rules of origin is likely to protect regional and domestic markets. 12 Note that we do not include a ROO dummy variable here. This is because the RESROO variable is a subset of the aggegate ROO dummy for 1995, but not for 1999. Including the ROO variable makes it difficult therefore to compare the coefficients across the two regressions.


153

In this paper we have addressed this issue by providing direct evidence of the impact of rules of origin and their cumulation on trade flow. The evidence suggests that the introduction of the pan-European system of cumulation served to increase trade between the EU and the other pan-EU countries. More fundamentally in this paper we have developed a framework by which the impact of lack of cumulation of rules of origin can be applied at the sectoral level. We then apply this model to the case of the textile industry, which is a key industry in terms of production and employment for the Southern Mediterranean partner countries. The results suggest that in aggregate the lack of cumulation may restrict trade between non-cumulating countries by up to 70–80 per cent. This would suggest that the rules of origin are resulting in both trade suppression and trade diversion. It also suggests that, without the extension of cumulation arrangements, a system of overlapping free trade agreements does not simply promote intra-FTA trade by reducing intra-FTA barriers, but also by effectively raising extra FTA barriers to trade.

REFERENCES Avisse, R. and M. Fouquin (2003), ‘Trade in Textiles and Clothing: Comparing Multilateral and Regional Free Trade Agreements’, Economie Internationale, No. 94–95. Burfisher, M. E., S. Robinson and K. Thierfelder (2004), ‘Regionalism: Old and New, Theory and Practice’, IFPRI Discussion Paper No. 65. Duttagupta, R. and A. Panagariya (2002), ‘Free Trade Areas and Rules of Origin: Economics and Politics’, International Trade 0308006, Economics Working Paper Archive at WUSTL. Estevadeordal, A. and K. Suominen (2003), ‘Rules of Origin: A World Map and Trade Effects’, Presented at the workshop ‘The Origin of Goods: A Conceptual and Empirical Assessment of Rules of Origin in PTAs’, INRA-DELTA (Paris, 23–24 May). Falvey, R. and G. Reed (2002), ‘Rules of Origin as Commercial Policy Instruments’, International Economic Review, 43, 2, 393–407. Gasiorek, M. et al. (2002), ‘Study on the Economic Impact of Extending the Pan-European System of Cumulation of Origin to the Mediterranean Partners’ part of the Barcelona process, report to DG Trade, European Commission. Gasiorek, M. et al. (2003), ‘The EU and the Southern Mediterranean: The Impact of Rules of Origin’, in Regional Development: Reality or Myth, Selected papers from the 9th ERF Annual Conference, October 2002. Grossman, G. and E. Helpman (1995), ‘The Politics of Free-trade Agreements’, American Economic Review, 85, 4, 667–90. Hoekman, B. (1993), ‘Rules of Origin for Goods and Services’, Journal of World Trade, 27, 4, 81– 99. Inama, S. (2003), ‘Quantifying the Trade Effects of Rules of Origin on Preferences: The Case of GSP, AGOA and ACP Preferences’, Presented at the workshop ‘The Origin of Goods: A Conceptual and Empirical Assessment of Rules of Origin in PTAs’, INRA-DELTA (Paris, 23– 24 May). Krishna, K. (2003), ‘The Origin of Goods: A Conceptual and Empirical Assessment of Rules of Origin in PTAs’, Presented at the workshop ‘The Origin of Goods: A Conceptual and Empirical Assessment of Rules of Origin in PTAs’, INRA-DELTA (Paris, 23–24 May). Krishna, K. and A. Krueger (1995), ‘Implementing Free Trade Areas: Rules of Origin and Hidden Protection’, NBER Working Paper No. 4983.

INTEGRATION OF MENA IN THE WORLD ECONOMY

155

9

Does the Quality of Institutions Limit the MENA’s Integration in the World Economy? Pierre-Guillaume Méon and Khalid Sekkat

1. INTRODUCTION

ANY professionals attribute the disappointing growth and employment records in the Middle East and North Africa (MENA) to the low integration of the region in the world economy (see Dasgupta et al., 2002, for a discussion).1 Abstracting from oil, the region scores one of the lowest ratios of exports to GDP among all regions of the world but Sub-Saharan Africa. In term of FDI inflows it also shows a similar picture. Previous researches have identified restrictive trade and exchange rate policies among the reasons for the low exports and FDI performance of developing countries. For instance, Sachs and Warner (1995) have showed that more liberalised economies tend to adjust more rapidly from primary-intensive to manufacturesintensive exports. Sekkat and Varoudakis (2002) focused specifically on the MENA and investigated whether trade policy reforms can increase the share of manufactured exports in GDP. Their results suggest that trade policy matters for the region’s performance. A similar conclusion was reached by Achy and Sekkat (2003) regarding the impact of exchange rate policy in the MENA. However, recent research suggests that for countries to fully benefit from openness strategies the functioning of institutions might be crucial.

M

The authors thank Michael Gasiorek and an anonymous referee for very helpful comments. They also benefited from very useful discussions with seminar participants at the University of Brussels, the University Robert Schuman of Strasbourg and the University Louis Pasteur of Strasbourg. They acknowledge financial support from the FEMISE network. All remaining errors are the authors’ own. 1

On the relationship between trade and growth in general see for instance Edwards (1992), Rodriguez and Rodrik (2001) and Frankel and Romer (1999).

155

156

PIERRE-GUILLAUME MÉON AND KHALID SEKKAT

As far as trade is concerned, Anderson (2001) suggested that the ill-functioning of institutions impairs foreign trade, because it increases both the costs and risks of trading abroad. Anderson and Marcouiller (2002) accordingly observe that bad institutions reduce the volume of trade. Finally, Dollar and Kraay (2002) report a positive correlation between openness and the quality of institutions with a potential bidirectional causality between the two variables. Regarding FDI, a first study by Wheeler and Mody (1992) failed to establish a significant relationship between FDI and institutions. These results being inconsistent with popular wisdom, Wei (2000) carefully re-examined the issue using a comprehensive data set on bilateral FDI flows. The results of his estimation show the existence of a negative relationship between corruption in the host country and FDI. Henisz (2000) moreover finds that foreign firms are more likely to enter wealthier countries with large population and credible political rules.2 The present paper builds on this work to examine the impact of institutions on manufactured exports and FDI attractiveness of the MENA. Econometric specifications of manufactured exports supply and FDI inflows’ determinants are estimated including both traditional explanatory variables and indicators of the quality of institutions. Six indicators of institutions are considered. Three of them concern corruption: the Corruption Perception Index published by Transparency International, the corruption index provided by the World Bank and Wei (2000)’s indicator. The other three cover different aspects of governance: government effectiveness, the rule of law (both drawn from Kaufmann et al., 1999) and a broad index of the quality of governance published in the International Country Risk Guide. The paper is organised as follows. Section 2 discusses the MENA’s recent macroeconomic performance and its participation in the world economy. Section 3 briefly presents the literature linking institutions to trade and FDI. Section 4 discusses the econometric approaches. Section 5 presents the results and Section 6 concludes.

2. THE MENA’S ECONOMIC PERFORMANCE IN AN INTERNATIONAL CONTEXT

Over the last fifteen years, growth performance in the MENA has been disappointing relative to that of the rest of developing countries. The growth rate of per capita GDP in the region declined from around 3.45 per cent on average during the 1960s and 1970s to around 0.1 per cent on average during the 1980s and 1990s. The disappointing growth record is reflected in a high unemployment rate that amounts to around 15 per cent. While the decreasing growth 2

The literature also suggests direct effects of institutions on growth. See Mauro (1995), Hall and Jones (1999) and Olson et al. (2000).


157

rate is similar for oil-exporting and non-oil-exporting countries, the unemployment rate is much higher for non-oil-exporting countries and amounts to around 25 per cent. Research on the source of growth in the MENA have shown that trade has not been utilised as a momentum for growth (see Makdisi et al., 2000, for a recent investigation). This is due to the lack of export diversification in the region. Abstracting from energy exports, the region continues to be among the least integrated regions in the world in terms of trade and FDI. It is even losing ground in this respect relative to other regions.3 The lack of export diversification makes the growth pattern in the region highly volatile. In oil-exporting countries it is closely related to energy price cycles while in others, such as Morocco, it is highly dependent on drought conditions due to the importance of agriculture in the economy. Such a situation impacts upon these economies’ capacity to invest, import, manage foreign exchange and hence their growth. The move from primary commodity exports to manufactures and service exports can be an important factor of sustained economic growth for at least two reasons. First, the demand for such exports increases more with income increase than the demand for primary products. Hence, growth prospects for a country’s exports are higher if its exports are diversified than when it is specialised in primary products. Second, the development of the manufacturing sector induces substantial dynamic productivity gains, and growth, arising from economies of scale, learning effects and externalities among firms and industries. The MENA’s record is disappointing in terms of export diversification. Figure 1 shows that the MENA, Sub-Saharan Africa (SSA) and Latin America exhibited the lowest ratios of manufactured exports to GDP in 1990. Over the 1990s, the ratio stagnated for the MENA while it increased markedly for Latin America and slightly for SSA, thereby leaving the MENA with the lowest ratio in 1999. Inside the region, there are important differences among countries. Oil-producing countries such as Libya, Saudi Arabia, Oman, Kuwait and Algeria were not diversified at the beginning of the decade and did not improve their record. Others like Oman, Tunisia, Egypt, Syria, Jordan, Morocco and Qatar showed significant improvement. FDI inflows provide developing countries with the additional resources they need to improve their economic performance. FDI inflows are expected to increase a country’s output and productivity, to encourage local investment and to stimulate the development and dispersion of technology. The MENA’s record in terms of FDI attractiveness is also disappointing. The region’s share in global investment and private capital inflows is small and

3

While world trade has been expanding in the past decade by about 8 per cent a year, the MENA’s trade has only been growing by about 2.7 per cent a year.

158

PIERRE-GUILLAUME MÉON AND KHALID SEKKAT FIGURE 1 Manufactured Exports as Per Cent of GDP

FIGURE 2 FDI Inflows as Per Cent of GDP

falling. The ratio of FDI flows to GDP was on average around 0.83 per cent for the MENA region during the 1990s while it was 1.51 in Eastern Europe, 1.66 in SSA, 2.35 in Latin America and 2.82 in East Asia. Moreover, in contrast to most developing countries where FDI as a percentage of GDP is increasing, the ratio in the MENA remained constant over the last decade (see Figure 2). Except for Tunisia, FDI inflows to the MENA are highly dependent on the presence of energy and on political conditions. For instance, investment in Algeria, Iraq or Sudan has been frightened away by political turmoil while Saudi Arabia and Egypt attracted further FDI in the last years. The flows of FDI to


159

Tunisia reached 13 per cent of GDP, which is slightly above the average of developing countries.

3. CONCEPTUAL FRAMEWORK

Although there is empirical evidence that the MENA’s disappointing performance may be due to distorted and inward-looking policies, openness strategies may not be sufficient unless they are complemented by institutional reforms.4 The present section documents how institutions can affect export performance and FDI attractiveness and provides a conceptual framework for testing such effects. a. Institutions and Trade Studies of the relationship between institutions and trade are receiving increasing attention. Institutions can either directly affect the willingness of agents to trade abroad, or affect economic variables that may in turn lower the propensity of agents to trade.5 The direct impact of institutions on the propensity to trade runs through the reduction of the expected return of trading abroad. A theoretical analysis by Anderson and Young (1999) suggests that the lack of enforcement of contracts may act as a tariff on risk-neutral traders and therefore reduce trade.6 Rodrik (2002) points out that the main impediment to international trade may indeed be the problem of contract enforcement, which is of particular relevance in international transactions since they confront traders in countries whose legal and political jurisdictions differ. Using gravity models, Anderson and Marcouiller (2002) lend empirical support to the impact of the quality of institutions on trade. Deterioration of the former induces a positive mark-up on the price of exports that reduces foreign demand and then exports. The impact of institutions on trade may also result from their effect on the risks associated with international transactions (Anderson and Marcouiller, 1997). Insecurity may prevent trade even though it offers potential mutual gains. For instance, predation reduces trade not only because it is a direct deduction on the flow of traded goods, but also because it diverts resources from their productive 4

Elbadawi (1999) has already pointed to the inadequate institutional support for investment and private sector development in the region. 5 Some authors argue that the causality between institutions and trade may run in the other direction (Rodrik, 2002; Treisman, 2000; and Hisamatsu, 2003). This reverse causality seems, however, to be statistically fragile (Knack and Azfar, 2003). While aware of the possibility of a two-sided causality between institutions and trade, we abstract from this issue here. 6 Contract enforcement becomes crucial when traders incur significant sunk costs resulting in a hold-up problem. Robert and Tybout (1997) document that such costs may be large and have dramatic consequences on trade.

160


allocation towards the defence of property rights. It follows that good institutions that may help bar predation may also foster trade.7 In addition to their direct effects, institutions may also indirectly affect trade through their impact on other variables that determine trade flows. Investment being an important determinant of trade (Rodrik, 1995; and Elbadawi, 1998), the impact of the quality of institutions on investment (Brunetti and Weder, 1998; Mauro, 1995; and Knack and Keefer, 1995) induces their indirect impact on trade. Furthermore, deficient institutions impact both on productivity (Hall and Jones, 1999) and its growth (Olson et al., 2000). As lower productivity is an impediment to competitiveness, countries with deficient institutions are likely to meet difficulties trading abroad. b. Institutions and Foreign Direct Investment The relationship between FDI and institutions has attracted greater interest than the relationship between institutions and trade. The literature underlines, however, the same type of mechanisms: direct effects stemming from the impact on the return and the risk associated with investing abroad and indirect effects coming from the impact on other variables, such as human capital or infrastructure, that were found to influence FDI inflows. The literature suggests that the main institutional impediment to FDI may not lie in its effect on the return of investing abroad but on the excess risk that it entails. Unlike trade, foreign investment is not only subject to a risk of predation and hold-up but also, and chiefly, to a risk of expropriation and nationalisation. Harms and Ursprung (2002), focusing on democracy, argue that authoritative regimes are associated with a greater risk of policy reversals, due for example to the dictator’s own whims, the need to raise public support through populist measures, or simply coups. Furthermore, Vinod (2003) points out that the impact of institutional risk on foreign investment may be even larger than expected if agents found their decisions on value at risk methods. Indeed, those methods tend to attach an extra weight to the worst scenario and may therefore deter FDI more than standard models of economic behaviour suggest. Recently, Gastanaga et al. (1998) find that corruption, bureaucratic delays and imperfect contract enforcement are associated with lower FDI to GDP ratios. Globerman and Shapiro (2002) also observe that various measures of governance quality are related to FDI inflows. Finally, Wei (2000) describes the consequences of corruption on bilateral FDI flows as a tax on foreign investors. Other 7

In addition, defective institutions also seem to distort the geographic structure of trade, as Lambsdorff (1998 and 2000) observes. He finds that some countries, like Belgium, France, Italy, the Netherlands and South Korea, are persistently over-represented in the imports of corrupt countries and that others like Sweden and Malaysia tend to trade less with corrupt importers.


161

studies examine separately the impact of the quality of governance and the impact of the risk associated with it. Campos et al. (1999) report that the predictability of corruption is a significant determinant of the investment ratio. A similar observation was made by Brunetti et al. (1998). Lambsdorff (2003) finds that the predictability of corruption has an impact on inward capital flows that is distinct from the impact of the level of corruption. Finally, institutional deficiencies may also have an indirect effect on FDI flows through their impact on other variables. The literature has shown that FDI flows are sensitive to human capital, health of the workforce and the quality of public infrastructure (Mody and Srinivasan, 1998; and Globerman and Shapiro, 2002). It has been observed that defective institutions tend to be associated with lower adult literacy rates (Kaufman et al., 1999), larger public investment in unproductive assets (Mauro, 1998), and lower expenditures devoted to the maintenance of past projects. Hence, by encouraging unproductive public investments that result in less efficient public facilities and a slower accumulation of human capital, defective institutions also indirectly hamper countries’ attractiveness for foreign investment.

4. DATA AND METHODOLOGY

To assess the impact of the quality of institutions on manufactured exports and FDI, we rely on traditional econometric specifications of these variables to which we add measures of the quality of institutions. Various indicators are now available to proxy the quality of institutions. They include the International Country Risk Guide (ICRG) index, the Transparency International (TI) index, and a set of the World Bank (WB) indices covering various dimensions of institutions’ quality. However, apart from the ICRG index and, to some extent, the TI index, all the other indices are time invariant. While the ICRG index provides a broad measure of institutions’ quality, the others have the advantage that each is designed to measure a specific aspect of governance, which is a useful piece of information for our study. In order to take advantage of the time dimension in the ICRG and the TI indices, and of the precision in the WB indices two econometric approaches are adopted. One consists of using the whole time-series cross-section sample (including ICRG and TI) as it stands and apply panel data analysis. The other estimates a fixed effects model using basic specifications for manufactured trade and FDI and then use the estimated fixed effects as a dependent variable to be explained by institution indicators and relevant control variables.8 A detailed description of the approaches is provided below. 8

This method is described in Hsiao (1986, section 3.6.1). For an application to institutional measures, the interested reader may refer to Olson et al. (2000).

162


a. The Panel Data Approach We apply the same panel data approach to both manufactured exports and FDI. However, the relationships that we estimate are naturally different and are consequently presented separately. (i) Manufactured exports To test for the impact of institutions on manufactured exports, we extend a specification that was used by Sekkat and Varoudakis (2000). We assume the following relationship between exports of manufactures and the quality of institutions: log(XMit) = α 0i + α1 * log(Eit) + α 2 * RYPit + α 3 * log(IMit−1) + α4 * log(Inst it) + µ it,

(1)

where XM it is country i’s ratio of manufactured exports to GDP for year t. Eit stands for country i’s real effective exchange rate for year t; an increase in Eit implying an appreciation of the exporter’s currency. RYPit measures the GDP growth rates of country i’s partners. IM it−1 is year t − 1’s investment in the manufactured goods sector over GDP. Inst it stands for an index of the quality of country i’s institutions for year t, higher values of Inst it standing for a lower quality of its institutions. Finally, µ it is the error term. We use the ratio of manufactured exports as the dependent variable to correct for the differences in countries’ sizes.9 α 0i is a fixed country effect that accounts for systematic cross-country differences that are not explained by the other independent variables. It therefore accounts for unmeasured or unobservable country specificities related to countries’ socio-economic characteristics. The control variables are standard in the literature and all have a well-defined expected impact on manufactured exports. The coefficient of the exchange rate should be negative, as an increase in Eit means an appreciation of the exporter’s currency. We expect a positive coefficient for the growth rates of a country’s partners. If a country’s partners grow faster, they will increase their demands for goods produced in that country, thereby raising its exports. Finally, we complement our set of control variables by the ratio of investment in the manufactured goods sector to GDP. This is based on the presumption that investment in the manufactured goods sector should raise its capacity, resulting in greater supply of manufactured goods, 9

In so doing, we may well underestimate the impact of governance on exports of manufactures. Namely, as worse governance is associated with a lower GDP (Mauro, 1995, or Knack and Keefer, 1995), both the numerator and the denominator of the exports ratio are negatively affected by defective institutions. This suggests that the impact of governance on exports may be stronger than what our point estimates already imply. This comment also applies to the FDI ratio used below.


163

hence higher exports. We accordingly expect the exports ratio to be positively correlated with investment. As institutional indicators, we use here the two timevarying indices: the index published in the International Country Risk Guide and the Corruption Perception Index constructed by Transparency International. The International Country Risk Guide provides, among others, a synthetic indicator of the political risk associated with a country on an annual basis. That indicator is constructed so as to encompass a wide variety of political factors that may influence the risk of investing in a country. Those factors are so diverse that it is not possible to associate the ICRG political risk indicator with any specific aspect of governance.10 It is therefore best interpreted as a synthetic measure of risk, broadly defined. The original index ranges from 0 to 100, the latter corresponding to the lowest possible risk. To ensure coherence with the other indicators, we re-code it such as an increase reflects higher risk. We therefore expect our ICRG index to be negatively related to the ratio of manufactured exports. Unlike the ICRG index, the TI index focuses on a specific aspect of governance, i.e. corruption. It ranges from 0 to 8, the latter corresponding to an absence of corruption. It was also re-coded such that an increase in that index reflects higher corruption. Thus, we expect the modified index to be negatively associated with manufactured exports.11 (ii) Foreign direct investment The strategy we employ to assess the impact of institutions on FDI is quite similar to the one we use for exports of manufactured goods. Namely, we add an index of the quality of institutions to a standard set of explanatory variables (see UNCTAD, 1998). The relationship that we estimate is therefore the following: log(FDIit) = β0i + β1 * log(Ypcit) + β2 * RYit + β3 * log(Prim it) + β4 * RYPit + β5 * log(Instit) + υit,

(2)

where FDIit is country i’s FDI to GDP ratio for year t and Ypcit its current GDP per capita for year t. RYit stands for country i’s current GDP growth rate for year t. Primit measures country i’s primary school enrolment ratio for year t. υit is the error term. RYPit and Instit are defined as in (1). 10

As a matter of fact, the indicator results from the aggregation of twelve basic components, unevenly weighted. Those components respectively assess: the stability of the government, socioeconomic conditions, the government’s attitude toward investment, the degree of political violence, the potential for external conflict, corruption, the role of the military in politics, the role played by religion in politics, law and order, ethnic tensions, the government’s democratic accountability, and the quality of the bureaucracy (ICRG, 1999). 11 Note that Transparency International warns against year-to-year comparisons of its index on the ground that its composition may change for one year to another. We nevertheless decided to use that index in our panel estimations for comparison with the ICRG.

164


Here again, the dependent variable is scaled by GDP to abstract from the difference in the size of countries. We tried to use FDI in the manufactured sector to get a better measure of the attractiveness of countries but such series are available for too few countries, only two of which (Morocco and Tunisia) are part of the MENA region.12 Once these series are used with governance indicators, we end up with only one MENA country (Morocco). As in (1), β0i is country i’s fixed effect, which controls that country’s specificities that are not measured by the other explanatory variables. The control variables are, as before, fairly standard. An increase in per capita income is associated with higher purchasing power and is supposed to attract more FDI: β1 should consequently be positive. Faster GDP growth suggests dynamism of the economy and is expected to attract more FDI: β2 should also be positive. A better educated, hence potentially more productive, workforce should also attract foreign investors. We therefore expect a positive β3. As faster growth of main trading partners implies more opportunity for exports FDI may also increase: β4 should be positive. Finally, as bad institutions are assumed detrimental to investment, we expect β5 to negative. b. Cross-section of Fixed Effects Apart from the ICRG risk index and the Transparency International index, other institutional indicators exist which have the advantage of being designed to measure a specific aspect of governance. These indicators are, however, time invariant. To take advantage of the information in these indicators, we use another econometric approach. In a first step we separate the explanatory variables used before into two subsets: those which vary in time (first subset) and those which do not (or almost not). The latter subset includes the new institutional indicators, human capital and per capita GDP.13 In a second step we estimate a specification where explanatory variables are fixed country effects and the ones included in the first subset. The fixed country effects therefore capture all the cross-country differences that are not explained by the differences in, and variations of, the time-varying variables. In the third step the estimated fixed country effects coefficients are used as dependent variables to be explained by institutions indicators, the other non-time-varying variables (second subset) and other control variables. The third step therefore assesses to which extent systematic crosscountry differences are attributable to differences in measured time-invariant variables, including chiefly the quality of institutions. 12

Using total FDI may blur the impact of governance because such flows can also be driven by natural resources abundance. 13 Although per capita GDP varies in time, the variations over the sample period appear to be slow enough for its inclusion in the second subset to improve the results.


165

(i) Manufactured exports The specification to be estimated in the second step for exports of manufactures is the same as (1) except that only Eit, RYPit and IMt−1 are included among the regressors. In the third step, we used the fixed country effects estimated in the second step as a dependent variable. The specification estimated in the third step is:

α 0i = η0 + η1 * log(Insti) + η2 * log(Poti) + ε i,

(3)

where α 0i is country i’s fixed effect as estimated in the second step. Instit is an index of the quality of country i’s institutions. Poti measures country i’s market potential. εi is the error term. Four time-invariant indicators of institutions are considered. The first and second ones are indices of corruption (labelled Wei and WB respectively) and are drawn from Wei (2000) and Kaufmann et al. (1999). The two other indicators shed light on the impact of two different facets of governance: the rule of law and government effectiveness. Both draw on Kaufmann et al. (1999). Government effectiveness concerns the ‘perceptions of the quality of public service provision, the quality of the bureaucracy, the competence of the civil servants, the independence of the civil service from political pressures, and the credibility of the government’s commitment to policies’. The rule of law is ‘the extent to which agents have confidence in and abide by the rules of society’. For the sake of comparability, the two governance indicators used before (ICRG and TI) are also considered. They are simply averaged over the sample period. We re-scaled all indices so that higher levels mean lower quality of governance. Their coefficients should consequently be negative. We used a country’s market potential as a control variable in the export regressions. That variable is defined as the distance-weighted average of a country’s partners’ GDPs. It therefore measures how close a country is to other markets. That variable was computed by the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) and was downloaded from its website. Our presumption is that the closer a country is to rich economies the more it will export. We therefore expect the market potential variable to be positively correlated with a country’s exports of manufactures ratio. (ii) Foreign direct investment Regarding FDI, the specification that we estimate in the first step is similar to (2). Only RYit, Primit and RYPit are used as explanatory variables however. Once that relationship is estimated, the country fixed effect can be used as the dependent variable in the third step, using the following specification:

β0i = λ 0 + λ 1 * log(Insti) + λ 2 * log(Ypci) + λ 3 * Nfppi + λ 4 * log(Poti) + ξi, (4)

166


where β0i is country i’s fixed effect as estimated in the second step and Instit an index of the quality of country i’s institutions. Ypci stands for country i’s average per capita GDP. Nfppi is a dummy variable equal to one when country i is classified as an exporter of primary products (including oil) by the World Bank. Poti measures country i’s market potential. ξ i is the error term. The measures of institutional quality are the same as before and so are the expected sign of the coefficients. A new control variable is introduced here: Nfppi. It takes account of the fact that natural resources abundant countries can attract more FDI independently of their institutional framework: λ 3 should be positive.

5. EMPIRICAL RESULTS

In this section, we successively present the results of our computations with the panel data approach and the cross-section of fixed effects. a. The Panel Data Approach Tables 1 and 2 display the estimation results of equations (1) and (2) respectively. Both equations are estimated using a sample of cross-section and TABLE 1 Regression Results. Dependent Variable: Ratio of Manufactured Exports to GDP Explanatory Variables

Institutional Indicator Used in the Specification Political Risk (ICRG)

Exchange rate Partners’ growth Investment Institutional indicator Number of observations Adjusted R2 Fixed effects Random effects

−0.43** (2.52) 0.044*** (2.67) 0.12*** (3.92) −0.78*** (3.80) 357 0.95 108.86 χ2(3) = 27.6

Corruption (TI) −0.19 (0.7) −0.04 (1.61) −0.02 (0.75) −0.25** (2.17) 125 0.97 106.17 χ2(1) = 0.65

Notes: Absolute t-statistics are displayed in parentheses under the coefficient estimates. * Test statistic is significant at the 10 per cent level; ** significant at the 5 per cent level; *** significant at the 1 per cent level. Fixed effects are not reported. The estimates are heteroscedastic consistent.


167

TABLE 2 Regression Results. Dependent Variable: Ratio of FDI to GDP Explanatory Variables


GDP per capita Growth rate Primary schooling Partners’ growth Institutional indicator Number of observations Adjusted R2 Fixed effects Random effects

1.32*** (2.98) 0.0076 (1.30) 0.42 (1.13) 0.063** (2.11) −1.91*** (6.05) 833 0.67 15.96 χ2(3) = 15.97

Corruption (TI) 3.58** (2.44) −0.02 (0.94) 3.66** (2.40) −0.054 (0.74) 0.013 (0.02) 215 0.78 9.55 χ2(5) = 32.72

Notes: Absolute t-statistics are displayed in parentheses under the coefficient estimates. * Test statistic is significant at the 10 per cent level; ** significant at the 5 per cent level; *** significant at the 1 per cent level. Fixed effects are not reported. The estimates are heteroscedastic consistent.

time-series data. All series, except institution indicators, are drawn from the World Development Indicators. The sample includes annual data (1990–1999) and covers a number of countries between 34 and 107 (see the Appendix). We use the panel data econometric methodology: tests of fixed and random effects are conducted to select the most adequate models. The estimates are heteroscedastic consistent. The fixed effects and the random effects tests support the focus on the fixed effects model. The overall quality of fit is very high for exports and high for FDI. (i) Manufactured exports In the regression with the ICRG index, all coefficients of the control variables are significant and have the expected sign. Faster growth in a country’s partners leads to an increase in its exports of manufactures; the appreciation of the exporter’s currency harms its exports. Investment in the manufactured goods sector is associated with higher exports. The coefficient of the ICRG index also has the expected sign and is significant, confirming that higher political risk disables a country’s participation in world trade. The regression where the TI corruption index is used among the explanatory variables confirms the main results obtained above: the governance index is

168


significantly correlated with manufactured exports and a reduction in the level of corruption results in an increase of such exports. However, the control variables lose their significance. This may be due to the drastic reduction in the sample’s size or to the fact that, as suggested by TI itself, the index should not be used in a temporal perspective. In a nutshell, a quick glance at Table 1 seems to confirm the importance of institutions in determining a country’s participation in international trade. One may, however, wonder what the quantitative importance of governance is and, more to the point, what that result means for MENA countries. To answer those two questions, we use our estimations to compare the observed exports ratio of the MENA countries comprised in our sample with what it might be if their governance improved.14 Given the quality of its results, we focus on the degree of political risk as measured by the ICRG index. Our estimation reveals that the elasticity of the manufactured exports to the index is 0.78. This means that if any country decreased its political risk by 1 per cent, during any particular year, its exports ratio would be expected to rise by 0.78 per cent. To grasp the true meaning of that figure, let us take the example of Morocco. In 1997, that country’s ICRG score amounted to 68.8 and its exports of manufactures to 10.4 per cent of its GDP. Let us now assume that during that year Morocco could have raised its ICRG index to Switzerland’s level, which is the country that on average ranks highest on the ICRG scale, with an index of 85.08 in 1997. The choice of the highest level score may seem excessive. The outcomes are proportional however: assuming that the gap between Morocco and Switzerland is only reduced by half will result in onehalf of our computed outcome. Hence, one can easily compute the result of different scenarios. Assuming the Swiss index will result in a 23.75 per cent increase in Morocco’s ICRG index that, according to our estimation, would have resulted in an 18.45 per cent increase in Morocco’s exports of manufactures ratio. Morocco’s exports ratio would therefore have amounted to 12.32 per cent instead of 10.4 per cent in 1997, a ratio similar to Poland’s. The impact on manufactured exports ratio is far from negligible. For instance, Sekkat and Varoudakis (2002) have found that the trade liberalisation programme adopted in Morocco since 1983 has resulted in an increase of the manufactured exports ratio by around 40 per cent. Hence, the impact of improvement in the quality of institutions could be one-half that of 20 years of trade liberalisation. We ran the same exercise for every year and every MENA country in the sample. The average difference between the observed ratio and the predicted ratio if governance improved is reported in Figure 3. This shows marked differences in 14

Due to limitations in the availability of data, there are only six MENA countries in our sample for which we could run that exercise.


169

FIGURE 3 The Expected Impact of an Improvement of Governance in MENA Countries on Their Exports Ratio

the observed trade of manufactured goods ratio from one country to another. Tunisia appears to be the most open country in the sample of MENA economies, whereas Kuwait is the least open economy. More to the point, it shows that an upgrading of the political risk indicator to Switzerland’s level would have resulted everywhere in an increase of the exports of manufactured goods ratio. However, the magnitude of the average variation differs from one country to another. For instance, Kuwait could only raise its exports of manufactured items by 0.33 percentage points while Tunisia could obtain a 5.40 points increase in that ratio. (ii) Foreign direct investment The results for the FDI ratio using the ICRG index show that a higher per capita income and a faster growth of trading partners increase FDI inflows. The domestic growth and human capital coefficients are non-significant. The ICRG index is significantly and negatively correlated with the FDI ratio. This suggests that political risk is a severe impediment to FDI. When one looks at the regression where the TI corruption index is used instead of the ICRG index, the above diagnosis is, however, qualified. It appears that the coefficient of that index is not significantly different from zero. We therefore observe no meaningful relationship between FDI and corruption, as measured by the CPI index. The results therefore underline the role of political risk in general, rather than only one aspect of bad governance, i.e. corruption, in diverting FDI from a country. A sensible question is to determine by how much it does. To provide an order of magnitude, we ran the same simulation for the FDI ratio as we did for the exports

170


FIGURE 4 The Expected Impact of an Improvement of Governance in MENA Countries on the FDI Ratio

of manufactures ratio. We therefore assessed the potential impact of an improvement of the political risk of the MENA countries in our sample to Switzerland’s level. The results of our computations, based on an estimated elasticity of 1.91, are plotted in Figure 4.15 Like Figure 3, Figure 4 displays significant differences in the observed FDI attractiveness of the MENA countries of our sample. With a ratio equal to 2.15 per cent, Tunisia is the most attractive economy, followed by Jordan, whose FDI ratio amounts to 1.60 per cent. On the other hand, Algeria and Iran’s FDI ratios, which hardly exceed 0.2 per cent, are the least attractive in the sample. The feature of interest in Figure 4 is the increase in the FDI ratio that would result if the countries in our sample reduced their degree of political risk. It appears that the improvement measured in relative terms would be greater for the FDI ratio than for the exports ratio, due to the greater estimated elasticity of FDI to political risk. Some countries, such as Turkey or Egypt, might accordingly almost double their FDI ratio. In absolute terms, Tunisia and Egypt would be the chief gainers of an improvement of their political risk. They would accordingly raise their FDI ratio by more than 1 percentage point if they upgraded their risk to the level of Switzerland. The FDI ratios of Morocco, Israel and Jordan would also increase, although to a lesser extent. Finally, the case of Algeria is worth emphasising. Our estimations imply that that country might more than double its FDI ratio if it cut down its political risk. However, as that country attracts very little FDI, the increase measured in absolute terms would appear almost negligible. 15

As FDI data are more easily available than data on exports of manufactures, we could run our simulation on a larger sample of MENA countries.


171

Those results are based on a specific definition of governance, i.e. political risk as measured by the ICRG political risk index. To get a better view of the impact of institutions, using a more refined definition of institutions would be desirable. However, as the alternative measures of institutional quality are unfortunately time-invariant, our method must be adapted accordingly. That is the aim of the following section. b. Cross-section of Fixed Effects In this section, we first study the relationship between the quality of institutions and the exports of manufactures ratio, then focus on the determinants of the FDI ratio. (i) Manufactured exports Table 3 presents the result of the relationship between various dimensions of governance and manufactured exports. Given that the sample is only crosssectional, the overall quality of fit is good. The results show that the proximity of large markets has a significantly positive impact on the exports of manufactured goods, since market potential exhibits a positive and significant coefficient. Moreover, all the governance variables appear significantly in the regressions and exhibit their predicted sign. The regressions, therefore, unanimously suggest that institutional quality has an impact on manufactured exports, regardless of which aspect of governance is taken into account. More interestingly, the regressions underline the importance of a country’s institutions on its exports of manufactured

TABLE 3 Regression Results. Dependent Variable: Fixed Effects of Manufactured Exports Ratio Explanatory Variables

Constant Market potential Institutional indicator Number of observations Adjusted R2


Corruption (TI)

Corruption (Wei)

Corruption (WB)

Rule of Law

Government Effectiveness

−9.03*** (4.29) 0.59** (2.23) −0.05*** (2.76)

−5.96*** (3.05) 0.75*** (3.61) −0.73*** (3.18)

−5.73*** (2.62) 0.77*** (3.45) −1.10** (2.32)

−4.72** (2.13) 0.64*** (2.87) −1.41*** (3.37)

−9.02*** −9.12*** (4.35) (4.61) 0.63** 0.65** (2.03) (2.18) −2.19** −2.16** (2.10) (2.24)

40 0.34

38 0.39

34 0.32

38 0.42

38 0.42

38 0.42

Notes: Absolute t-statistics are displayed in parentheses under the coefficient estimates. * Test statistic is significant at the 10 per cent level; ** significant at the 5 per cent level; *** significant at the 1 per cent level. The estimates are heteroscedastic consistent.

172


goods in spite of the inclusion of an additional explanatory variable. In other words, all the governance variables remain significantly correlated with the ratio of exports of manufactured goods and the sign of their coefficients remains unaffected. To get a clearer picture of the meaning of those results for MENA countries, we complement the insights of Table 3 by a quantitative evaluation of the impact of institutions on the exports ratio of those countries. To do so, we follow the same procedure as in the previous section and use our results to simulate the consequences of an improvement of the institutions of MENA countries for their exports ratio. Let us accordingly focus on Wei (2000)’s measure of corruption and assume that the MENA countries included in our sample were able to reduce the pervasiveness of corruption to Finland’s level. That country, which scores 1.3 on Wei (2000)’s corruption scale, is the least corrupt country in our sample according to Wei (2000). Such an improvement would for instance result in a 71.7 per cent reduction in Morocco’s corruption index. According to the point estimate of the elasticity of exports of manufactured goods to corruption, such an evolution would result in a 78.9 per cent increase in that country’s exports ratio. That would take it to 15.1 per cent, a ratio comparable to France’s. The increase here is higher than with the ICRG index and becomes comparable to the one resulting from liberalisation. We followed the same line of reasoning for the four MENA countries for which data were available. The outcome of our estimations is summarised in Figure 5. Figure 5 shows differences in the relative impact of a better control of corruption between countries. Thus, it appears that Morocco and Jordan might almost FIGURE 5 Expected Impact of an Improvement in Governance on the Exports of Manufactures Ratio in Some MENA Countries


173

double their exports of manufactured goods ratio while Egypt would only raise it by one-third. As the estimated elasticity of the exports of manufactured goods ratio is the same for all countries, these differences are due to the difference in the quality of governance in each country. Thus, according to Wei (2000), Egypt fares better than Morocco and Jordan in terms of corruption. An improvement of Egypt’s degree of corruption to Finland’s, consequently results in a smaller relative improvement than for the other countries. Regarding the absolute impact of an improvement of a country’s degree of corruption, Israel would obtain the largest gain. If Jordan succeeded in controlling corruption as well as Finland, it would obtain an exports of manufactured goods to GDP ratio equal to 21.4 per cent, which would be similar to Austria’s. It would thus rank among the top ten exporters of manufactured goods as a percentage of their GDP. (ii) Foreign direct investment Table 4 reports the results of the relationship between various dimensions of governance and FDI. The regressions lead to a somewhat more mixed impression than the regressions displayed in Table 3. First, the overall quality of fit is very low, even for cross-sectional data.

TABLE 4 Regression Results. Dependent Variable: Fixed Effects of FDI Ratio Explanatory Variables


Corruption (TI)

Corruption (Wei)

Corruption (WB)

Rule of Law

−18.77*** (4.52) Per capita GDP −0.24* (1.76) Human capital 0.04 (0.19) Primary products exporter 0.07 (0.25) Market potential −0.14 (0.58) Institutional indicator −4.07*** (3.33)

−6.38** (2.95) −0.22* (1.67) 0.43** (2.20) 0.42 (1.63) 0.26 (1.24) −0.34 (1.24)

−5.23** (2.17) −0.26* (1.98) 0.49** (2.31) 0.42 (1.62) 0.22 (0.96) −0.85* (1.94)

−7.15** (2.60) −0.09 (0.50) 0.42* (1.79) 0.46 (1.57) 0.18 (0.82) −0.14 (0.24)

−7.56*** −7.56*** (4.08) (4.01) −0.06 −0.06 (0.40) (0.35) 0.40* 0.40 (1.67) (1.55) 0.47* 0.47* (1.78) (1.70) 0.19 0.19 (0.90) (0.88) 0.84 0.00 (0.01) (0.00)

Number of observations Adjusted R2

84 0.05

76 0.08

88 0.04

88 0.04

Constant

107 0.17

Government Effectiveness

88 0.04

Notes: Absolute t-statistics are displayed in parentheses under the coefficient estimates. * Test statistic is significant at the 10 per cent level; ** significant at the 5 per cent level; *** significant at the 1 per cent level. The estimates are heteroscedastic consistent.

174


Second, there is no consistent pattern of control variables significance. When significant, they tend to exhibit the right sign. The coefficient of human capital is positive and significant in four instances out of six. The coefficient of per capita GDP is significant in three instances and is negative. This sign contradicts the one found in the cross-section time-series sample. It is, however, not necessarily wrong since in a pure cross-section (cross-country) dimension it is interpreted as a proxy for (the inverse) return on capital (Edwards, 1991). Higher per capita GDP should therefore be associated with less foreign investment. Finally, only the coefficients of the ICRG political risk index and the corruption index, as measured by Wei (2000), exhibit significant signs. The signs of those coefficients correspond to their predicted signs. This suggests that political risk or corruption tend to impede FDI. The results on corruption are, however, not confirmed by the other indicators. The result regarding the ICRG index is coherent with our previous findings. Here again, the results suggest that this is political risk in general, rather than only one aspect of bad governance, which reduces FDI to a given country. As before, we tried to illustrate the meaning of those results for the four MENA countries for which data were available. We therefore supposed that those countries were able to upgrade their institutions so as to reach an institutional quality similar to Finland’s. Using our point estimate of the elasticity of the FDI ratio to Wei (2000)’s corruption index, we could compute the predicted variation of those countries’ FDI ratios. The results of our computations are displayed in Figure 6. Figure 6 confirms that institutions matter for MENA countries. Namely, if the MENA countries in our sample could control corruption, as measured by Wei

FIGURE 6 Expected Impact of an Improvement in Governance on the FDI Ratio in Some MENA Countries


175

(2000)’s index, our estimations predict that they would all experience a rise in their FDI ratio. As before, however, one can observe marked differences in the impact of controlling corruption from one country to another. Thus, in relative terms, Morocco and Jordan would gain more than Israel and Egypt.

6. CONCLUSION

The present paper examines the extent to which ill-functioning institutions disable a greater participation of MENA countries in the world economy. It builds on recent research emphasising the importance of the quality of institutions for growth, trade and investment. The paper focuses on manufactured exports and FDI attractiveness as the main indicators of the region’s integration to the world economy. The latter is seen as an important mechanism which can enable the MENA to meet the growth and employment problems it is facing. The empirical analysis relies on basic specifications of manufactured exports supply and FDI inflows’ determinants to which indicators of the quality of institutions are added. The latter includes a broad index of political risk as well as indices targeted toward specific aspects of governance such as corruption, government effectiveness and the rule of law. The estimation is conducted on a large sample of countries over the 1990s and uses different econometric approaches in order to check for the robustness of the results. Overall, the results lend strong support to the hypothesis that the functioning of their institutions may disable the participation of MENA countries in the world economy. From an econometric point of view, the results for manufactured exports are, however, stronger than for FDI. It is found that deterioration of the quality of institutions is, in general, associated with low performance in terms of manufactured exports and FDI attractiveness. Simple simulations resting on our econometric results suggest that the impact of an improvement in the quality of institutions may result in a sensitive increase of FDI inflows and manufactured exports. Depending on the institutional indicator used, the latter can improve by an amount comparable to the one resulting from liberalisation policies. For instance, institutions improvement in Morocco is found to entail an increase of the ratio of manufactured exports to GDP at least equal to one-half of the impact of the liberalisation policy initiated 20 years ago. Hence, although institutional reforms can take time, they deserve the necessary efforts given their outcomes as compared to other reforms.

176

PIERRE-GUILLAUME MÉON AND KHALID SEKKAT APPENDIX

Countries in the Sample* Albania Algeria Angola Argentina Australia Austria Bangladesh Bolivia Brazil Bulgaria Burkina Faso Cameroon Canada Chile China Colombia Congo Republic Costa Rica Côte d’Ivoire Cyprus Czech Republic Denmark Dominican Republic Ecuador Egypt * Countries in italics

Ethiopia Kuwait Finland Lebanon France Madagascar Gabon Malawi Gambia Malaysia Germany Mali Ghana Malta Greece Mexico Guatemala Mongolia Guinea Morocco Haiti Mozambique Honduras Netherlands Hungary New Zealand Iceland Nicaragua India Niger Indonesia Nigeria Iran Norway Ireland Pakistan Israel Panama Italy Papua New Guinea Jamaica Paraguay Japan Peru Jordan Philippines Kenya Poland Korea, Republic Portugal do not enter the exports sample.

Romania Russia Senegal Sierra Leone Singapore Slovak Republic South Africa Spain Sri Lanka Sweden Switzerland Syria Tanzania Thailand Togo Tunisia Turkey Uganda United Kingdom United States Uruguay Venezuela, RB Vietnam Zambia Zimbabwe

REFERENCES Achy, L. and K. Sekkat (2003), ‘The European Single Currency and MENA’s Export to Europe’, Review of Development Economics, 7, 4, 563–82. Anderson, J. E. (2001), ‘Trade and Informal Institutions’ (mimeo). Anderson, J. E. and D. Marcouiller (1997), ‘Trade and Security, I: Anarchy’, NBER Working Paper No. 6223. Anderson, J. E. and D. Marcouiller (2002), ‘Insecurity and the Pattern of Trade: An Empirical Investigation’, Review of Economics and Statistics, 84, 2, 342–52. Anderson, J. E. and L. Young (1999), ‘Trade and Contract Enforcement’ (Boston College, mimeo). Brunetti, A. and B. Weder (1998), ‘Investment and Institutional Uncertainty: A Comparative Study of Different Uncertainty Measures’, Weltwirtschaftliches Archiv, 134, 3, 513–33.


177

Brunetti, A., G. Kisunko and B. Weder (1998), ‘Credibility of Rules and Economic Growth: Evidence from a Worldwide Survey of the Private Sector’, World Bank Economic Review, 12, 3, 353–84. Campos, J. E., D. Lien and S. Pradhan (1999), ‘The Impact of Corruption on Investment: Predictability Matters’, World Development, 27, 6, 1059–67. Dasgupta, D., M. Nabli, C. Pissarides and A. Varoudakis (2002), ‘Making Trade Work for Jobs: International Evidence and Lessons for MENA’ (The World Bank; Middle East and North Africa Region). Dollar, D. and A. Kraay (2003), ‘Institutions, Trade and Growth’, Journal of Monetary Economics, 50, 1, 133–62. Economic Research Forum (2002), Economic Trends in the MENA Region. Edwards, S. (1991), ‘Capital Flows, Foreign Direct Investment, and Debt-Equity Swaps in Developing Countries’, in H. Siebert (ed.), Capital Flows in the World Economy (Tübingen, J. C. B. Mohr), 255–81. Edwards, S. (1992), ‘Trade Orientation, Distortions and Growth in Developing Countries’, Journal of Development Economics, 39, 1, 31–57. Elbadawi, I. (1998), ‘Real Exchange Rate Policy and Non-traditional Exports in Developing Countries’, WIDER Discussion Paper, forthcoming in G. Helleimer (ed.), Growth, External Sector and the Role of Non-traditional Exports in Sub-Saharan Africa. Elbadawi, I. (1999), ‘Can Reforming Countries Perform an Asian Miracle? Role of Institutions and Governance on Private Investment’, in I. Limam (ed.), Institutional Reform and Development in the MENA Region (Arab Planning Institute and Economic Research Forum). Frankel, J. A. and D. Romer (1999), ‘Does Trade Cause Growth?’, American Economic Review, 89, 3, 379–99. Gastanaga, V. M., J. B. Nugent and B. Pashamova (1998), ‘Host Country Reforms and FDI Inflows: How Much Difference Do They Make?’, World Development, 26, 7, 1299–314. Globerman, S. and D. Shapiro (2002), ‘Global Foreign Direct Investment Flows: The Role of Governance Infrastructure’, World Development, 30, 11, 1899–919. Hall, R. and C. I. Jones (1999), ‘Why Do Some Countries Produce so Much More Output Per Worker than Others?’, Quarterly Journal of Economics, 114, 1, 83–116. Harms, P. and H. W. Ursprung (2002), ‘Do Civil and Political Repression Really Boost Foreign Direct Investments?’, Economic Inquiry, 40, 4, 651–63. Henisz, J. W. (2000), ‘The Institutional Environment for Multinational Investment’, Working Paper of the Reginald H. Jones Center (The Wharton School, University of Pennsylvania). Hisamatsu, Y. (2003), ‘Does Foreign Demand Affect Corruption?’, Applied Economics Letters, 10, 1, 1–2. Hsiao, C. (1986), Analysis of Panel Data (New York, Cambridge University Press). International Country Risk Guide (1999), ‘Brief Guide to the Ratings System’ (June). Kaufmann, D., A. Kraay and P. Zoido-Lobaton (1999), ‘Aggregating Governance Indicators’, Working Paper No. 2195 (World Bank). Knack, P. and S. Keefer (1995), ‘Institutions and Economic Performance: Cross-country Tests Using Alternative Institutional Measures’, Economics and Politics, 7, 3, 207–27. Knack, S. and O. Azfar (2003), ‘Trade Intensity, Country Size and Corruption’, Economics of Governance, 4, 1, 1–18. Lambsdorff, J. G. (1998), ‘An Empirical Investigation of Bribery in International Trade’, European Journal of Development Research, 10, 1, 40–59. Lambsdorff, J. G. (2000), ‘La Propension des Exportateurs à Verser des Pots de Vin. L’impact sur les échanges’, Revue Tiers Monde, 41, 161, 89–116. Lambsdorff, J. G. (2003), ‘How Corruption Affects Persistent Capital Flows’, Economics of Governance, 4, 3, 229–43. Makdisi, S., Z. Fattah and I. Limam (2000), ‘Determinants of Growth in the MENA Region’, Global Development Network (Prague). Mauro, P. (1995), ‘Corruption and Growth’, Quarterly Journal of Economics, 110, 3, 681– 712.

178


Mauro, P. (1998), ‘Corruption and the Composition of Government Expenditure’, Journal of Public Economics, 69, 2, 263–79. Mody, A. and K. Srinivasan (1998), ‘Japanese and US Firms as Foreign Investors: Do they March to the Same Tune?’, Canadian Journal of Economics, 31, 4, 778–99. Olson, M., N. Sarna and A. V. Svamy (2000), ‘Governance and Growth: A Simple Hypothesis Explaining Cross-country Differences in Productivity Growth’, Public Choice, 102, 3–4, 341–64. Robert, M. J. and J. R. Tybout (1997), ‘The Decision to Export in Colombia: An Empirical Model of Entry with Sunk Costs’, American Economic Review, 87, 4, 545–64. Rodríguez, F. and D. Rodrik (2001), ‘Trade Policy and Economic Growth: A Skeptic’s Guide to the Cross-national Evidence’, in B. Bernanke and K. S. Rogoff (eds.), Macroeconomics Annual 2000 (MIT Press for NBER, Cambridge, MA). Rodrik, D. (1995), ‘Getting Interventions Right: How South Korea and Taiwan Grew Rich’, Economic Policy, 10, 20, 53–97. Rodrik, D. (2002), ‘Trade Policy Reform as Institutional Reform’, in B. M. Hoekman, P. English and A. Mattoo (eds.), Development, Trade, and the WTO: A Handbook (World Bank, Washington, DC). Sachs, J. D. and A. Warner (1995), ‘Economic Reform and the Process of Global Integration’, Brookings Papers on Economic Activity, 1, 1–118. Sekkat, K. and A. Varoudakis (2000), ‘Exchange Rate Management and Manufactured Exports in Sub-Saharan Africa’, Journal of Development Economics, 61, 1, 237–53. Sekkat, K. and A. Varoudakis (2002), ‘The Impact of Trade and Exchange Rate Policy Reforms on North-African Manufactured Imports’, Development Policy Review, 20, 2, 177–89. Treisman, D. (2000), ‘The Causes of Corruption: A Cross-national Study’, Journal of Public Economics, 76, 3, 399–457. UNCTAD (1998), World Investment Report. Vinod, H. D. (2003), ‘Open Economy and Financial Burden of Corruption: Theory and Application to Asia’, Journal of Asian Economics, 13, 6, 873–90. Wei, S.-J. (2000), ‘How Taxing Is Corruption on International Investors?’, Review of Economics and Statistics, 82, 1, 1–11. Wheeler, D. and A. Mody (1992), ‘International Investment Location Decisions: The Case of U.S. Firms’, Journal of International Economics, 33, 1–2, 57–76.

INDEX

179

Index access to markets, European Partnership Agreements with Caribbean 17, 26, 41 administrative tax, Burundi 45–6 African, Caribbean and Pacific (ACP) countries Lomé convention 1 European Partnership Agreements 1, 2, 15, 130 Everything-but-Arms Initiative 3, 4 regional integration 6 WTO’s Doha Round 9 see also Caribbean; Sub-Saharan Africa Agadir Agreement 102, 106, 130 agricultural sector Canada 74–7, 79 Egypt 113, 114, 121, 124 European Partnership Agreements with Caribbean 17, 19, 21, 27 Maldives 85, 89 Middle East and North Africa 157 Morocco 157 Southern Mediterranean region 104, 105 Tunisia 113, 114, 121, 124 Air Maldives 92, 93 air services sector, Egypt 116 Algeria Agadir Agreement 106, 130 economic performance 157, 158 Euromed partnership 129 governance 170 trade reform 101 Andean Community 64 Angola 5 Anguilla 39 anti-dumping duties (ADDs) 71–2 Canada 70, 71–4

European Partnership Agreements with Sub-Saharan Africa 4 Antigua and Barbuda economy 19, 20, 21, 24, 25 European Partnership Agreements 32–7, 39 Association of Caribbean States (ACS) 18 Australia 63, 64 Austria 173 automobile industry, NAFTA 133 Bahamas economy 19, 20, 21, 22, 24 European Partnership Agreements 32–7, 39 Baltic states cumulation of rules of origin system 141, 149, 150, 152 trade and production patterns 136 see also specific countries Bangladesh 59, 77, 78 banking sector Egypt 116 Tunisia 116 Barbados economy 19, 20, 21, 22, 24 European Partnership Agreements 39 Barcelona Declaration (1995) 101, 129, 132, 134 Basseterre, Treaty of 18 Belize 20, 21, 22, 24, 39 Benin 74 bilateral trading agreements (BTAs) Maldives and India 93 Middle East and North Africa 111 see also specific BTAs Bolivia 64

180

INDEX

Botswana 5, 6, 7 Brazil 57, 74 Bulgaria 134 Burkina Faso 74 Burundi evaluation of trade policy reform 55–6 trade policy 43 1986: 43–9 2003: 49–50 trade structure and economic performance 1986: 51–2 2003: 52–5 Canada and Caribbean, trade between 23, 27, 64 Commonwealth Caribbean Countries Tariff (CCCT) 63, 64 General Preferential Tariff (GPT) 59, 63, 64, 68 Least-Developed Country Tariff (LDCT) 59, 63, 64, 68, 78 New Zealand and Australia Tariff 63 Trade Policy Review 57–8, 78–9 agri-food 74–7 clothing and textiles 77–8 rules of origin 68–9 safeguards, ADDs and CVDs 70–4 trade patterns and tariff treatments 58–64 welfare effects of RTAs 64–7 Canada–Chile Free Trade Agreement (CCFTA) 58, 63, 64, 78 Canada–Costa Rica Free Trade Agreement (CCRFTA) 58, 63, 64, 78 Canada–Israel Free Trade Agreement (CIFTA) 58, 63, 64 Canada–US Free Trade Agreement (CUSFTA) 57, 64, 66–7

Canadian International Trade Tribunal 73 Canadian Wheat Board (CWB) 76 Caribbean and Canada, trade between 23, 27, 64 Commonwealth Caribbean Countries Tariff (CCCT) 63, 64 European Partnership Agreements 15–18, 40–2 Caribbean economies 18–26 role for EPAs 26–40 Caribbean Basin Initiative (CBI) 23–5 Caribbean Community and Common Market (CARICOM) European Partnership Agreements 18, 23–5, 30–8, 40 Single Market and Economy (CSME) 25 Central African Economic and Monetary Community (CAECM, CEMAC) 3, 5 Central America and Caribbean, trade between 23 Central and Eastern Europe 141 see also Eastern Europe Central European Free Trade Agreement (CCFTA) 136, 149, 150, 152 Chad 74 Chile 58, 63, 64, 75 China 57, 59, 67, 74, 77 clothing and textiles Canada 77–8, 79 Egypt 114 Maldives 85, 86, 90 NAFTA 133 rules of origin and the EU-Med partnership 131, 152 conceptual and empirical background 135, 138–41 sectoral ‘gravity’ model 143–52 Southern Mediterranean region 106

INDEX Tunisia 105, 113, 114, 120–1, 123–4 WTO Agreement 70, 77, 141, 152 Colombia 23, 64 common external tariff (CET) Caribbean 25 European Partnership Agreements with Sub-Saharan Africa 3, 5, 6 Common Market for Eastern and Southern Africa (COMESA) 5 Commonwealth Caribbean Countries Tariff (CCCT) 63, 64 Communauté Economique et Monétaire de l’Afrique Centrale (CEMAC) 3, 5 competition policy, Sub-Saharan Africa 11 construction sector, Caribbean 21 consumption tax, Maldives 94–5 contract enforcement 159, 160 corruption and FDI, relationship between 156, 174 Middle East and North Africa 156, 160–1, 175 data and methodology 163, 165 empirical results 167–8, 169, 171, 172, 174–5 Costa Rica 25, 58, 63, 64, 78 Cotonou Agreement European Partnership Agreements with Caribbean 15, 16, 17, 18, 40 with Sub-Saharan Africa 1, 4, 6, 10, 11 cotton industry 74, 113 countervailing duties (CVDs), Canada 70, 71 Cuba 23 cumulation 133–4, 152–3 diagonal 133–4, 135–6, 141, 147 pan-European system, trend analysis of 142–3 textiles 143–52

181

current account deficit, Maldives 86 customs Burundi 46–9, 52, 54, 55 Egypt 115 European Partnership Agreements with Sub-Saharan Africa 11 Maldives 94 Tunisia 115 customs unions (CUs) 3, 5, 6 see also specific CUs Cyprus 101, 129 Czech Republic 134, 142 development cooperation 18, 26, 41 diagonal cumulation 133–4, 135–6, 141, 147 dispute settlement European Partnership Agreements with Caribbean 18 WTO 76, 96 Dominica economy 19, 20, 21, 22, 23, 24 European Partnership Agreements 37, 38, 39 Dominican Republic 23, 25 East African Community (EAC) 5 East Asia 158 Eastern and Southern Africa (ESA) 5 Eastern Europe 158 see also Central and Eastern Europe Economic Community of West African States (ECOWAS) 5 Ecuador 64 Egypt Agadir Agreement 106, 130 composition of economy 113 cumulation and textiles 150 economic performance 157, 158 Euromed partnership 129 governance 169, 170, 172–3, 174–5 liberalisation of services trade 109–27

182

INDEX

socio-economic indicators 112–13 trade and production patterns 135–40 trade reform 101, 102–5 Egypt Telecom 115 El Salvador 64 Enabling Clause, Everything-but-Arms Initiative 3, 4 entry tax, Burundi 45–6, 47 environmental issues, Maldives 98 Estonia 134 European Free Trade Association (EFTA) 64, 134, 141, 149, 150 European Mediterranean Agreements 111 Egypt 118–19, 121–2, 125 rules of origin and textile sector 129– 31, 152–3 conceptual and empirical background 131–4 sectoral ‘gravity’ model 143–52 trend analysis of the pan-European system 141–3 Tunisia 118–19, 120–1, 125 European Partnership Agreements (EPAs) 130 with Caribbean 15–18, 40–2 economies 18–26 role for EPAs 26–40 with Sub-Saharan Africa 1–3, 10–13 Everything-but-Arms Initiative 3–4 liberalisation of trade in services 9–10 regional integration 4–7 tariff reductions 7–8 WTO’s Doha Round 9 European Union agricultural framework, WTO’s Doha Round 74 aid programme 2, 41 anti-dumping duties 73 and Canada, trade between 58–9, 73, 76

cumulation of rules of origin system 134, 141, 149 and Egypt, trade between 105–6, 118–19, 121–2, 125 Generalised System of Preferences (GSP) Caribbean 27, 40 Maldives 93 rules of origin 4, 132 and Maldives, trade between 93, 94, 97 and Southern Mediterranean, trade between 102, 104, 106 and Tunisia, trade between 105–6, 114, 118–22, 125 see also European Mediterranean Agreements; European Partnership Agreements Everything-but-Arms (EBA) Initiative Caribbean 27, 41 rules of origin 132 Sub-Saharan Africa 3– 4, 6, 7 exchange rate policy European Partnership Agreements with Sub-Saharan Africa 11 Maldives 86, 91 Middle East and North Africa 155 excise taxes, European Partnership Agreements with Sub-Saharan Africa 7 expatriate workers, Maldives 86 export promotion, Burundi 50 financial sector Egypt 116 Tunisia 116 Finland 172, 173, 174 fiscal duty, Burundi 45–6, 47 fishing industry European Partnership Agreements with Caribbean 17 Maldives 85, 89–93, 96, 97–8

INDEX food, beverages and tobacco sector Canada 74–7, 79 Southern Mediterranean 135 foreign direct investment (FDI) Egypt 118 European Partnership Agreements with Sub-Saharan Africa 10 Maldives 86, 87, 91–2 Middle East and North Africa 106–7, 111, 155, 156, 175 conceptual framework 160–1 data and methodology 161–6 empirical results 167, 169–71, 173–5 international context 157–9 Southern Mediterranean 101, 104–5 Tunisia 118 France 114, 172 franchises, Egypt and Tunisia 115 free trade agreements (FTAs) between South Africa and US 8 Sub-Saharan Africa 5, 6 see also specific FTAs Free Trade Area of the Americas (FTAA) 23, 25, 64 G-90: 2, 4, 10, 11 General Agreement on Tariffs and Trade (GATT) 3, 77, 82 see also World Trade Organisation General Agreement on Trade in Services (GATS) 109 delivery modes 110–11 Egypt 116 European Partnership Agreements with Caribbean 17, 27 with Sub-Saharan Africa 10 Maldives 91 Tunisia 116 goods and services tax (GST), Egypt 114, 117

183

government effectiveness, Middle East and North Africa 165, 167–70, 171, 174–5 Greater Arab Free Trade Agreement (GAFTA) 111, 122 Grenada economy 19, 20, 21, 22, 23, 24 European Partnership Agreements 38 Guatemala 25, 64 Guyana economy 19, 20, 21, 22, 24 European Partnership Agreements 39 Haiti 19, 20, 21, 25 Honduras 25, 64 Hong Kong 65 Hungary 134, 142 Iceland 142 import licences, Burundi 44–5, 49, 55 import substitution, Burundi 48, 50, 52 India 57, 72, 74, 93 indirect taxation 7 see also value-added tax infrastructure, European Partnership Agreements with Sub-Saharan Africa 11 institutional factors, Middle East and North Africa 155–6, 175–8 conceptual frameworks 159–61 data and methodology 161–6 economic performance in international context 156–9 empirical results 166–75 intellectual property rights, Maldives 95 International Country Risk Guide (ICRG) index data and methodology 161, 163, 165 empirical results 166, 167, 168–71, 172, 174 International Monetary Fund (IMF) 9, 91, 96

184

INDEX

investment policy Egypt 124–5 European Partnership Agreements with Sub-Saharan Africa 111 Tunisia 124–5 see also foreign direct investment Iran 170 Iraq 63, 158 Israel and Canada, trade between 58, 63, 64 cumulation and textiles 150 Euromed partnership 129 governance 169, 170, 172, 173, 174– 5 trade and production patterns 135–40 trade reform 101 ITC 97 Jamaica economy 19, 20, 21, 22, 24, 25 European Partnership Agreements 32–7, 39 Japan anti-dumping duties 72 and Canada, trade between 58–9, 74 Generalised System of Preferences 93 and Maldives, trade between 93 NAFTA rules of origin 69 Jordan Agadir Agreement 106 economic performance 157 Euromed partnership 129 governance 169, 170, 172–3, 174–5 trade and production patterns 135–40 trade reform 101, 102–5 Kenya 5 Kuwait 157, 169, 170 Lamy, Pascal 2, 4, 11 Latin America and Caribbean, trade between

23

economic performance 157, 158 integration process 130 Latvia 134, 142 law, rule of, Middle East and North Africa 165, 171, 175 least developed countries (LDCs) and Canada, trade between 78, 79 Cotonou Agreement 1–2 EU’s Everything-but-Arms Initiative 3, 4 Maldives 77–8 UN definition 82 Lebanon 102, 129, 170 legal issues, European Partnership Agreements with Caribbean 18 Lesotho 5, 6, 7 liberalisation of trade Burundi 54–5 Canada 58, 76–7, 79 Egypt 105–6, 109–27 Euromed partnership 129 European Partnership Agreements with Caribbean 17, 36, 37, 41 with Sub-Saharan Africa 6, 7–8, 9–10 Morocco 168 rules of origin 132, 150 Southern Mediterranean 107 Tunisia 105–6, 109–27 Libya 63, 157 Lithuania 134, 142 lobbying, European Partnership Agreements with Sub-Saharan Africa 8 Lomé convention Caribbean 15, 26, 40 replacement by Cotonou Agreement 1 Macao 65 Maldives economic and social development indicators 82, 83

INDEX economic performance 82–5 Foreign Investment Services Bureau (FISB) 91–2 Ministry of Tourism 92 Ministry of Trade and Industry 91–2 Open General Licences (OGLs) 90 State Trading Organisation (STO) 90, 92, 93 Trade Policy Review 81–2, 98–9 global integration, problems of 94–8 trade and development 82–7 trade policies and practices 87–94 Maldives Industrial Fisheries Company (MIFCO) 92, 93, 97–8 Maldives Monetary Authority 86 Mali 74 Malta 102, 129 manufacturing sector Burundi 53 Caribbean 19–21, 23 Egypt 113, 114, 122–5 Maldives 85, 87, 89 Middle East and North Africa 155, 157, 158, 175 data and methodology 162–3, 165 empirical results 167–9, 171–3 Southern Mediterranean 105, 134–5 Tunisia 113, 114, 121–4 maritime transportation services sector, Egypt 116 Mauritania 5 Mercosur 23, 130 Mexico anti-dumping duties 73 and Canada, trade between 59, 64, 67, 74, 75 and Caribbean, asymmetric market access to US 23 clothing and textiles 140 NAFTA rules of origin 69, 70

185

Middle East and North Africa (MENA) 106–7 Egyptian trade 121, 122, 125 financial system 116 institutions as limiting factor for global integration 155–6, 175–8 conceptual framework 159–61 data and methodology 161–6 economic performance in international context 156–9 empirical results 166–75 liberalisation of services trade 111, 115 non-tariff barriers 123 Tunisian trade 121, 122 migrant workers, Maldives 86 mining and quarrying sector, Caribbean 21 Mongolia 65 Montserrat 19, 22, 24, 37, 38 Morocco Agadir Agreement 106, 130 corruption 172–3 economic performance 157 Euromed partnership 129, 130, 132, 133 FDI 164 governance 169, 170, 172–3, 174–5 ICRG index 168 trade and production patterns 135–40 trade reform 101, 102–5, 107 and US, free trade agreement between 130 most favoured nation (MFN) status Burundi 50 Canada 62, 63, 64, 68, 77 Egypt 105, 118–20, 121 European Partnership Agreements with Caribbean 41 with Sub-Saharan Africa 6, 7–8, 9 Maldives 93 Middle East and North Africa 111

186

INDEX

NAFTA rules of origin 70 Tunisia 104, 105, 118–20 Mozambique 5 Multifibre Arrangement (MFA) 90 Multilateral Investment Guarantee Agency (MIGA) 92 multi-stage production, service sector 110 Namibia 5, 6, 7 natural disasters, Caribbean 19 natural trading blocs 65 New Zealand 63, 64 Nicaragua 64 Nigeria 5 non-tariff barriers (NTBs) Burundi 55 Egypt 115 European Partnership Agreements with Caribbean 17–18 Maldives 90 Middle East and North Africa 123 Southern Mediterranean 104, 105 Tunisia 115 see also specific NTBs North American Free Trade Agreement (NAFTA) automobile industry 133 Canada 58, 62–3, 64 welfare effects 66, 67 clothing and textiles 78, 133, 140 Mexico’s access to US markets 23 rules of origin 68–70, 133 North Korea 63 Norway 142 oil industry, Middle East and North Africa 157 Oman 157, 170 Organisation of Eastern Caribbean States (OECS) Economic Union 18, 25–6

European Partnership Agreements 18, 23, 32–3, 37, 40–1 outsourcing, service sector 110 Palestinian Authority 102, 129 Pan-Arab Free Trade Agreement 102 patronage, Maldives 93 Peru 64 Poland ICRG index 168 pan-European system, trend analysis of 141–2 rules of origin 130, 132, 133, 134 political patronage, Maldives 93 political risk, Middle East and North Africa 175 data and methodology 161, 163, 165 empirical results 166, 167, 168–71, 172, 174 poverty Burundi 56 Caribbean 16, 17, 41 preferential trade agreements (PTAs) 129–30 between Egypt and EU 122 and rules of origin 130–1, 132 Southern Mediterranean 106 Sub-Saharan Africa 5 see also specific PTAs privatisation Maldives 92–3, 95 Middle East and North Africa 111 Qatar 157 quantitative restrictions (QRs) Burundi 44–5, 47–50, 55 Maldives 96 quotas Canada 75, 76, 77, 78 Maldives 90, 93 rules of origin and the Euromed partnership 148–9

INDEX regional integration in Caribbean 16, 17, 25–6, 27 Middle East and North Africa 157 and rules of origin 130–1 in Sub-Saharan Africa 2, 3, 4–7 regional trade agreements (RTAs) 64–6 Canada 57, 62, 64–7, 79 numbers and extent 64–5 Southern Mediterranean 106 Sub-Saharan Africa 5 see also specific RTAs Romania 134, 142 royalties, Maldives 90–1 rule of law, Middle East and North Africa 165, 171, 175 rules of origin (ROOs) Canada 68–70, 78 Euromed partnership and textile sector 130–1, 152–3 conceptual and empirical background 131–4, 141 sectoral ‘gravity’ model 147, 148, 149–52 trend analysis of the pan-European system 141, 142–3 European Partnership Agreements with Caribbean 17 with Sub-Saharan Africa 4, 6 regional trade agreements 65 Southern Mediterranean 106 safeguards, Canada 70–1 St. Kitts and Nevis economy 19, 20, 21, 22, 24 European Partnership Agreements 32–7, 39 St. Lucia economy 19, 20, 21, 22, 23, 24 European Partnership Agreements 32–3, 39 St. Vincent and the Grenadines economy 19, 20, 21, 22, 23, 24

187

European Partnership Agreements 38, 39 sanctions, international, against Burundi 55 sanitary and phytosanitary (SPS) regulations Canada 76 European Partnership Agreements with Caribbean 17 with Sub-Saharan Africa 4 Maldives 96–7 Saõ Tomé 5 Saudi Arabia 157, 158 service sector delivery modes 110–11 Egypt 105, 106, 109–27 European Partnership Agreements with Caribbean 17, 19, 21 with Sub-Saharan Africa 9–10 and goods trade, comparison between 109–10 Maldives 86, 91 Middle East and North Africa 157 Tunisia 105, 106, 109–27 see also General Agreement on Trade in Services Singapore 64 Singapore issues 57, 74 Slovakia 134, 142 Slovenia 134, 142 smuggling Burundi 55 Maldives 95 social services sector, Tunisia 113 Sony 69 South Africa 5, 6–7, 8, 72 South Asian Association for Regional Cooperation (SARC) 93 Southern African Customs Union (SACU) 5–7 Southern African Development Community (SADC) 5, 6

188

INDEX

Southern Mediterranean, trade reform 101–7 see also European Mediterranean Agreements Sri Lanka 89 State Trading Organisation (STO), Maldives 90, 92, 93 steel sector, Canada 70, 71, 73 subcontracting, service sector 110 Sub-Saharan Africa (SSA) economic performance 157, 158 European Partnership Agreements 1– 3, 10–13 Everything-but-Arms initiative 3–4 liberalisation of trade in services 9–10 regional integration 4–7 tariff reductions 7–8 WTO’s Doha Round 9 Sudan 158 Suriname 19, 20, 21, 22, 23, 24 Swaziland 5, 6, 7 Switzerland 72, 142, 168, 169, 170 Syria 102, 129, 157, 170 Taiwan 65, 77 Tanzania 5 tariffs Burundi 44–50, 54–5 Canada 58–60, 62–4, 66–7, 68, 75 Egypt 111, 114–15, 117–23, 125 Euromed partnership 129, 133–4, 148, 150–2 European Partnership Agreements with Caribbean 17, 28–33, 37–9, 41 with Sub-Saharan Africa 6, 7–8, 10–11 Maldives 87–90, 93, 94–5, 96 Middle East and North Africa 111 NAFTA rules of origin 69, 70 regional trade agreements 65–6

Southern Mediterranean 104, 105 Sri Lanka 89 Tunisia 111, 114–15, 117–23, 125 taxation Burundi 48–50, 52–4 Egypt 114, 117, 122–3 Maldives 86, 90, 94–5 Tunisia 114, 117, 123 see also specific taxes telecommunications sector Egypt 115–16 Tunisia 115, 116 temporary workers Caribbean 77 Sub-Saharan Africa 10 textiles see clothing and textiles tourism sector Caribbean 19 Egypt 113 Maldives 85–7, 89, 91, 92, 98 Sub-Saharan Africa 10 Transparency International (TI) Corruption Perception Index data and methodology 161, 163, 165 empirical results 166, 167–8, 169 Treaty of Basseterre 18 Trinidad and Tobago economy 19, 20, 21, 22, 23, 24 European Partnership Agreements 32–7, 39 Tunisia Agadir Agreement 130 composition of economy 113 cumulation and textiles 150 economic performance 157, 158, 159 Euromed partnership 129 FDI 158, 159, 164 governance 169, 170 liberalisation of services trade 109–27 socio-economic indicators 112–13 trade and production patterns 135–40 trade reform 102–5

189

INDEX Tunisie Telecom 115 Turkey cumulation and textiles 150 Euromed partnership 129 governance 170 rules of origin 134 trade and production patterns 135–40 trade reform 102 Uganda 5 UNCTAD 97 unemployment, Middle East and North Africa 156–7 Union Economique et Monétaire Ouest-africaine (UEMOA) 3, 5, 8 United Nations Conference on Trade and Development 97 least developed countries 82 United States of America agricultural framework, WTO’s Doha Round 74 anti-dumping duties 73 and Canada, trade between 17, 57– 60, 64, 78 agri-food 74, 75, 76 anti-dumping duties 72 clothing and textiles 77 welfare effects 66–7 and Caribbean, trade between 15, 23–5, 27, 30, 33–8 clothing and textiles 140 and Egypt, trade between 121, 125 Generalised System of Preferences 23–5 and Maldives, trade between 93 and Morocco, free trade agreement between 130 NAFTA rules of origin 70 openness, lack of 60 and South Africa, trade between 8

and Southern Mediterranean, trade between 102 and Tunisia, trade between 121 welfare effects of free trade agreements 66 value-added tax (VAT) European Partnership Agreements with Sub-Saharan Africa 7 Maldives 94–5 Tunisia 114, 117 Venezuela 23, 64 West African Economic and Monetary Union (WAEMU, UEMOA) 3, 5, 8 World Bank (WB) Doha Round, WTO 8, 9 global integration 96, 97 institutional quality proxy indices 161 lower-middle income countries 82, 102 Maldives 92 Multilateral Investment Guarantee Agency (MIGA) 92 World Trade Organisation (WTO) Agreement on Agriculture 74–5 Agreement on Clothing and Textiles (ACT) 70, 77, 141, 152 Agreement on Sanitary and Phytosanitary Standards (SPS) 76, 96–7 Agreement on Trade in Agriculture 70 authorised contingency measures 70 Burundi, Trade Policy Review 43–56 Canada, Trade Policy Review 57–79 Customs Valuation Agreement 95 dispute settlement procedure 76, 96 Doha Round 2, 8, 9, 11, 81 Cancun meeting 57, 74 European Partnership Agreements with Caribbean 15, 18, 26

190

INDEX

with Sub-Saharan Africa 2, 4, 8, 9, 11 global integration 96 Maldives, Trade Policy Review 81– 99 Middle East and North Africa 111

regional trading agreements 64–5 Work Programme on Small Economies (WPSE) 81 see also General Agreement on Tariffs and Trade; General Agreement on Trade in Services

The World Economy, Global Trade Policy 2004 (World Economy Special Issues)

The World Economy: Global Trade Policy 2009 (World Economy Special Issues)

The World Economy: Global Trade Policy 2010 (World Economy Special Issues)

Understanding the world economy