asean.elibrary.imf.org€¦ · © 2005 international monetary fund production: imf multimedia...

The IMF’s Approach to Capital AccountLiberalization

I n t e r n a t i o n a l M o n e t a r y F u n d • 2 0 0 5

ECIFF

ON

OITA

ULAVETNEDNEPEDNI

Evaluation Report

©International Monetary Fund. Not for Redistribution

© 2005 International Monetary Fund

Production: IMF Multimedia Services DivisionFigures: Wendy Arnold

Typesetting: Alicia Etchebarne-Bourdin

Cataloging-in-Publication Data

Price: US$25.00

Please send orders to:International Monetary Fund, Publication Services

700 19th Street, N.W., Washington, D.C. 20431, U.S.A.Tel.: (202) 623-7430 Telefax: (202) 623-7201

E-mail: [email protected]: http://www.imf.org

recycled paper

The IMF’s approach to capital account liberalization: evaluation report/[pre-pared by a team headed by Shinji Takagi and including Jeffrey Allen Chel-sky . . . [et al]—[Washington, D.C.]: International Monetary Fund, 2005.

p. cm. Includes bibliographical references.ISBN 1-58906-415-1

1. Capital movements. 2. International Monetary Fund—Evaluation. I.Takagi, Shinji, 1953– II. Chelsky, Jeffrey Allen. III. International MonetaryFund, Independent Evaluation Office.

HG3891.I43 2005


http://www.imf.org

Foreword vii

Abbreviations and Acronyms ix

THE IMF’S APPROACH TO CAPITAL ACCOUNT LIBERALIZATION

Executive Summary 3

1 Introduction 9

The Scope of the Evaluation 10Sources of Evidence 13Organization of the Report 15

2 General Policy and Analysis 16

The Legal Basis 16General Operational Approach 18Multilateral Surveillance 23Assessment 29

3 Advice to Member Countries, 1990–2002 30

Capital Account Liberalization 30Macroeconomic and Structural Policies to Manage Capital Inflows 35Temporary Use of Capital Controls 43Assessment 47

4 Ongoing Country Dialogue on Capital Account Issues 49

Capital Account Liberalization 49Temporary Use of Capital Controls 53Assessment 55

5 Major Findings and Recommendations 57

Major Findings 57Recommendations 60

Boxes

1.1. The Debate on the Benefits of Capital Account Liberalization 101.2. Effectiveness of Market-Based Capital Controls: Evidence

from Chile 111.3. Capital Account Liberalization in Indonesia and Korea 142.1. The Proposal to Amend the Articles of Agreement 19

Contents

iii


Underline

Parimalam

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Parimalam

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Parimalam

Underline

Underline

Underline

Underline

Underline

CONTENTS

2.2. Mexico: Capital Account Liberalization and the Crisis of 1994–95 262.3. Chile: The IMF’s Views on the Use of Market-Based Controls

on Inflows 283.1. Latvia 333.2. Outflow Controls in Thailand (1997) and Malaysia (1998) 464.1. India 514.2. Technical Assistance 524.3. The IMF’s “Integrated” Approach to Capital Account Liberalization 55

Figures

1.1. Private Capital Flows to Developing Countries 91.2. Countries with Capital Controls 121.3. Countries with Capital Controls 121.4. Average Capital Account Openness in 12 Sample Countries 152.1. Net Capital Inflows to Developing Countries 233.1. Fiscal Performance and the IMF’s Fiscal Policy Advice in

Selected Countries, 1990–2002 393.2. IMF Support for Exchange Rate Flexibility in Selected Countries,

1990–2002 41

Tables

2.1. Notable Events Affecting Capital Account Issues, 1991–2004 203.1. The Nature of the IMF’s Advice on Capital Account Liberalization

in Selected Countries 313.2. The IMF’s Advice on Managing Large Capital Inflows, 1990–2002 363.3. Fiscal and Other Macroeconomic Indicators in Selected Countries

with Large Capital Inflows, 1990–2002 383.4. The IMF’s Position on Temporary Use of Capital Controls,

1991–2001 444.1. Policy Environments for Capital Account Liberalization in Selected

Countries 50

Appendixes

1 A More Detailed Assessment of Some Country Cases 63

The Czech Republic: Liberalization in a Transition Economy 63Colombia: Market-Based Controls on Inflows 67Venezuela: Use of Capital Controls on Outflows 70Tunisia: Gradual Liberalization 71

2 Staff Research 75

Early Work on Capital Controls and Capital Flow Management 75Later Work on Capital Controls 76Work on Sequencing 76More Recent Work 77

3 Public Communications 78

4 Cross-Border Capital Transactions and Capital Account Openness in Selected Countries, 1989–2002 80

5 Capital Account Openness in 12 Sample Countries, 1990–2002 82

6 List of Interviewees 84

iv


Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Parimalam

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Parimalam

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Parimalam

Underline

Contents

Appendix Box

A1.1. Hungary 66

Appendix Figures

A1.1. Czech Republic: Capital Account Openness and Net Private Capital Inflows 64

A1.2. Colombia: Capital Account Openness and Net Private CapitalInflows 67

A1.3. Venezuela: Capital Account Openness and Net Private Capital Inflows 70

A1.4. Tunisia: Capital Account Openness and Net Private CapitalInflows 72

Appendix Table

A1.1. Colombia: The Unremunerated Reserve Requirement, 1993–2000 68

References 86

STATEMENT BY THE MANAGING DIRECTOR, IMF STAFF RESPONSE,IEO COMMENTS ON MANAGEMENT/STAFF RESPONSE,

AND SUMMING UP OF IMF EXECUTIVE BOARD DISCUSSIONBY THE ACTING CHAIR

Statement by the Managing Director 93

IMF Staff Response 94

IEO Comments on Management/Staff Response 98

Summing Up of IMF Executive Board Discussion by the Acting Chair 100

v

The following symbols have been used throughout this report:

– between years or months (e.g. 2003–04 or January–June) to indicate the years ormonths covered, including the beginning and ending years or months;

/ between years (e.g. 2003/04) to indicate a fiscal (financial) year.

“Billion” means a thousand million.

Minor discrepancies between constituent figures and totals are due to rounding.

Some of the documents cited and referenced in this report were not available to the publicat the time of publication of this report. Under the current policy on public access to theIMF’s archives, some of these documents will become available five years after their is-suance. They may be referenced as EBS/YY/NN and SM/YY/NN, where EBS and SMindicate the series and YY indicates the year of issue. Certain other documents are to be-come available ten or twenty years after their issuance depending on the series.


Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Underline

Parimalam

Underline

This page intentionally left blank


This report reviews the IMF’s approach to capital account liberalization and re-lated issues, drawing on evidence from a sample of emerging market economies overthe period 1990–2004. The role of the IMF in capital account liberalization has been atopic of major controversy. An independent evaluation of the IMF’s advice on capitalaccount issues is therefore both timely and appropriate.

The evaluation seeks to contribute to transparency by documenting what in prac-tice has been the IMF’s approach to capital account liberalization and related issuesand to identify areas where the IMF’s instruments and operating methods might beimproved, in order to deal with capital account issues more effectively. The reportdeals not only with capital account liberalization per se but also with capital flowmanagement issues, including particularly the temporary use of capital controls.

Capital account liberalization is an area where there is little professional consen-sus, making it difficult to evaluate the IMF’s policy advice against some universal setof criteria. Moreover, the IMF Articles of Agreement give the IMF only limited juris-diction over the capital account, with the result that the IMF had no formal policy onmost capital account issues during the period under review. For these reasons, theevaluation assesses the IMF’s actual approach to these issues, identifying what policyadvice the IMF gave in the context of a specific country at a specific point in time.

The report begins by reviewing the IMF’s general operational approach and analy-sis as they evolved from the early 1990s into the early 2000s. It then assesses theIMF’s country work in terms of (1) its role in capital account liberalization during1990–2002, (2) its policy advice to member countries on managing capital flows dur-ing the same period, and (3) its ongoing work on capital account issues in a group ofemerging market economies during 2003–04. The report concludes by offering twobroad recommendations. First, as noted in the original terms of reference, the evalua-tion does not seek to make a judgment on whether the Articles of Agreement shouldbe amended to give the IMF an explicit mandate and jurisdiction on capital accountissues, since this is an issue that goes well beyond the scope of the evaluation evi-dence. However, the report does conclude that greater clarity on the IMF’s approachto capital account issues is needed and makes a number of suggestions as to how thismight be achieved. Second, the report supports greater attention by the IMF’s analysisand surveillance to the supply-side factors of international capital flows, a processthat is already under way.

The report was discussed by the IMF Executive Board on May 11, 2005. In keep-ing with established practice, the report is being published as submitted to the Board,except for minor factual corrections. This volume also includes the response of IMFmanagement and staff to the evaluation, the IEO response, and the Summing Up ofthe Board discussion.

David GoldsbroughActing Director

Independent Evaluation Office

Foreword

vii


The report was prepared by a team headed by Shinji Takagi and including JeffreyAllen Chelsky, Edna Armendariz, Misa Takebe, and Halim Kucur. It also bene-fited from substantive contributions from Rawi Abdelal, Masahiro Kawai, DavidPeretz, Jai Won Ryou, and participants in a workshop held at the Institute for In-ternational Economics—William Cline, Morris Goldstein, Michael Mussa,Edwin Truman, and John Williamson. Administrative support by AnnetteCanizares, Arun Bhatnagar, and Maria S. Gutierrez, and editorial work by IanMcDonald and Esha Ray are gratefully acknowledged. The report was approvedby David Goldsbrough, Acting Director of the IEO.

viii


ADR American Depository ReceiptAREAER Annual Report on Exchange Arrangements and Exchange

Restrictions (IMF)BIBF Bangkok International Banking FacilityCAC Collective action clauseEFF Extended Fund Facility (IMF)EMU Economic and Monetary UnionERM Exchange Rate Mechanism (European Monetary System)EU European UnionFDI Foreign direct investmentFSAP Financial Sector Assessment Program (IMF and World Bank)GATS General Agreement on Trade in ServicesGATT General Agreement on Tariffs and TradeGDP Gross domestic productGFSR Global Financial Stability Report (IMF)ICM International Capital Markets Department (IMF)ICMR International Capital Markets Report (IMF)IEO Independent Evaluation Office (IMF)IMF International Monetary FundIOSCO International Organization of Securities CommissionsLEG Legal Department (IMF)LIBOR London interbank offered rateLOI Letter of Intent (IMF)MAE Monetary and Exchange Affairs Department (IMF)1

MFD Monetary and Financial Systems Department (IMF)OECD Organization for Economic Cooperation and DevelopmentPDR Policy Development and Review Department (IMF)RES Research Department (IMF)SDR Special drawing right (IMF)SDRM Sovereign Debt Restructuring MechanismSBA Stand-By Arrangement (IMF)TA Technical assistanceUNCTAD United Nations Conference on Trade and DevelopmentURR Unremunerated reserve requirementWEO World Economic Outlook (IMF)WTO World Trade Organization

Abbreviations and Acronyms

ix

1Effective May 1, 2003, name was changed to Monetary and Financial Systems Department.


The IMF’s Approach to Capital Account Liberalization


A gainst the background of highly volatile inter-national capital flows and the associated finan-

cial instability experienced by a number of majoremerging market economies in recent years, the roleof the IMF in capital account liberalization has beena topic of major controversy. The IMF’s role is par-ticularly controversial because capital account liber-alization is an area where there is little professionalconsensus. Moreover, although current account lib-eralization is among the IMF’s official purposes out-lined in its Articles of Agreement, the IMF has noexplicit mandate to promote capital account liberal-ization. Indeed, the Articles give the IMF only lim-ited jurisdiction over issues related to the capital ac-count. Nevertheless, the IMF has given greaterattention to capital account issues in recent decades,in light of the increasing importance of internationalcapital flows for member countries’ macroeconomicmanagement. In view of these facts, an independentassessment of how the IMF has addressed capital ac-count issues seems warranted.

The evaluation seeks to (1) contribute to trans-parency by documenting what in practice has beenthe IMF’s approach to capital account liberalizationand related issues; and (2) identify areas, if any,where the IMF’s instruments and operating methodsmight be improved, in order to deal with capital ac-count issues more effectively. The issues addressedin the evaluation cover not only capital account lib-eralization but also capital flow management issues,including particularly the temporary use of capitalcontrols. The evaluation, however, does not addressthe question of whether liberal capital accounts areintrinsically beneficial—on which the broader acad-emic literature has not been able to provide a defini-tive answer—or whether the Articles of Agreementshould be amended to give the IMF an explicit man-date and jurisdiction on capital account issues. Manyaspects of these issues are not amenable to evidencefrom the evaluation. However, the evaluation doesshed some light on the consequence of the lack ofexplicit mandate and jurisdiction on the IMF’s workon capital account issues.

The report begins by reviewing the IMF’s generaloperational approach and analysis as they evolved

from the early 1990s into the early 2000s. It then as-sesses the IMF’s country work in terms of (1) its rolein capital account liberalization during 1990–2002,(2) its policy advice to member countries on manag-ing capital flows during the same period, and (3) itsongoing work on capital account issues (where out-standing issues are identified for 2003–04). The reportconcludes by offering two broad recommendations.

The IMF’s General OperationalApproach and Analysis

Despite the ambiguity left by the Articles ofAgreement about its role in capital account issues,the IMF responded to the changing international en-vironment by paying increasing attention to issuesrelated to the capital account. Concurrent with theinitiatives to amend the Articles to give the IMF anexplicit mandate for capital account liberalizationand jurisdiction over members’ capital account poli-cies, the IMF expanded the scope of its operationalwork in the area. It encouraged the staff to givegreater emphasis to capital account issues in ArticleIV consultations and technical assistance and to pro-mote capital account liberalization more actively.

In multilateral surveillance, the IMF’s analysisprior to the mid-1990s tended to emphasize the ben-efits to developing countries of greater access to in-ternational capital flows and to pay comparativelyless attention to the potential risks of capital flowvolatility. More recently, however, the IMF has paidgreater attention to various risk factors, including thelinkage between industrial country policies and in-ternational capital flows as well as the more funda-mental causes and implications of their boom-and-bust cycles. Still, the focus of the analysis remainson what emerging market countries should do tocope with the volatility of capital flows (for exam-ple, in the areas of macroeconomic and exchangerate policy, strengthened financial sectors, andgreater transparency). Although the IMF has ad-dressed the moral hazard aspect of boom-and-bustcycles by encouraging greater exchange rate flexibil-ity in recipient countries and attempting to limit ac-

Executive Summary

3


EXECUTIVE SUMMARY

cess to IMF resources during a crisis, it has not beenat the forefront of the debate on what, if anything,can be done to reduce the cyclicality of capitalmovements through regulatory measures targeted atinstitutional investors in the source countries.

From the beginning of the 1990s, the IMF’smanagement, staff, and Executive Board wereaware of the potential risks of premature capital ac-count liberalization and there is no evidence to sug-gest that they promoted capital account liberaliza-tion indiscriminately. They also acknowledged theneed for a sound financial system in order to mini-mize the risks of liberalization. Such awareness,however, largely remained at the conceptual leveland did not lead to operational advice on precondi-tions, pace, and sequencing until later in the 1990s.At the same time, a subtle change was taking placewithin the institution. As preliminary evidenceemerged on the apparent effectiveness of Chile’scapital controls in the mid-1990s, some in the IMFbegan to take a favorable view of the use of capitalcontrols as a temporary measure to deal with largecapital inflows.

In the event, the proposed amendment of the Arti-cles put forward in the late 1990s failed to garnersufficient support, leaving ambiguity about the roleof the IMF. In the meantime, something of a consen-sus—the so-called “integrated” approach—hasemerged within the IMF that places capital accountliberalization as part of a comprehensive program ofeconomic reforms in the macroeconomic policyframework, the domestic financial system, and pru-dential regulations. While few would disagree withthe prudence and judiciousness of the new approach,it has proved to be difficult to apply as an opera-tional guide to sequencing because it emphasizes allof the potential interlinkages but does not provideclear criteria for identifying a hierarchy of risks.Moreover, these views remain unofficial, as theyhave not been explicitly endorsed by the ExecutiveBoard.

The IMF’s Country Work

The evaluation assesses the IMF’s country workin terms of the following criteria: (1) Was there anydifference between the IMF’s general policy pro-nouncements and the advice it gave to individualcountries? (2) Was the IMF’s policy advice opera-tional and based on solid evidence? (3) How did theIMF’s advice change over time, and did this changekeep pace with available evidence? (4) Did the IMFgive similar advice to countries in similar situations?and (5) Was the policy advice on the capital accountset in a broader assessment of the authorities’ macro-economic policies and institutional framework?

Capital account liberalization

During the 1990s, the IMF clearly encouragedcapital account liberalization, but the evaluation sug-gests that, in all the countries that liberalized thecapital account, partially or almost fully, the processwas for the most part driven by the country authori-ties’ own economic and political agendas. In none ofthe program cases examined did the IMF requirecapital account liberalization as formal conditional-ity (which is understood to mean prior actions, per-formance criteria, or structural benchmarks), al-though aspects of it were often included in theauthorities’ overall policy package presented to theIMF. This is consistent with the interpretation of theArticles of Agreement, which states that the IMF, asa condition for the use of its resources, cannot re-quire a member to remove controls on capital move-ments. In the first half of the 1990s, in encouragingcapital account liberalization, the IMF seldom raisedthe issue of pace and sequencing. The staff occasion-ally expressed concern over financial sector weak-ness or macroeconomic instability, but this did nottranslate into concrete operational advice. Fromaround 1994, and more noticeably following theEast Asian crisis, the IMF began increasingly to giveemphasis to the need to satisfy certain preconditions;in general, the IMF’s approach in its country workwas quite pragmatic, especially in this later period,and often accepted the authorities’ own views on theappropriate pace and sequencing of liberalization.

Managing capital inflows

As countries experienced large capital inflows andassociated macroeconomic challenges in the 1990s,the issue of how to manage large capital inflows be-came a routine subject of discussion between theIMF and the country authorities. The staff’s policyadvice was largely in line with the policy conclu-sions typically derived from the scholarly literatureon open economy macroeconomics. To deal withlarge capital inflows, it advocated tightening fiscalpolicy and greater exchange rate flexibility. Thestaff’s position on sterilization emphasized its quasi-fiscal costs and longer-term ineffectiveness but was,to a remarkable extent, supportive of the country au-thorities’ policy choices, whatever they may havebeen. In a few instances, the staff also recommendedfurther trade liberalization, liberalization of capitaloutflows, and tightening of prudential regulation asmeasures to deal with large capital inflows. Theseand other structural measures, however, received rel-atively little attention in the IMF’s policy advice, al-though in recent years increasing attention has beengiven to strengthening the financial sector regulatoryframework, primarily in the context of the Financial

4


Executive Summary

Sector Assessment Program (FSAP). Given theIMF’s focus and comparative advantage, this wasprobably appropriate.

Temporary use of capital controls

Use of capital controls has been a controversialsubject, not only within the IMF but also in the aca-demic and official policymaking communities. It ispossible here to make a broad characterization thatthe IMF staff was in principle opposed to the use ofsuch instruments, either on inflows or outflows. Itsview was that they were not very effective, espe-cially in the long run, and could not be a substitutefor the required adjustments in macroeconomic andexchange rate policies. Even so, from the earliestdays, the IMF staff displayed a remarkable degree ofsympathy with some countries in the use of capitalcontrols. In a few cases, both before and after thecrises of 1997–98, it even suggested that market-based controls could be introduced as a prudentialmeasure. As a general rule, the IMF staff, in linewith the evolution of the institution’s view, becamemuch more accommodating of the use of capitalcontrols over time, albeit as a temporary, second-best instrument.

Ongoing country dialogue on capital account issues

In ongoing country work, as documented for theperiod of 2003–04, IMF staff has been quite accom-modating of the authorities’ policy choices whenthey have involved a gradual approach to capital ac-count liberalization or temporary use of controls. Interms of capital account liberalization, the staff hassometimes been more cautious than the authorities(as in Russia in 2003) when their preferred policyhas been to liberalize the capital account quickly. Inmost cases, the staff has taken a medium-term per-spective and has emphasized the importance ofmeeting certain preconditions, the most important ofwhich are fiscal consolidation, a sound financial sys-tem, and the adoption of a floating exchange rate(usually with inflation targeting).

In terms of advice on temporary use of capitalcontrols, IMF staff seldom challenged the authori-ties’ decision and has even supported market-basedcontrols in some cases. There was a slight differencein emphasis across countries. In a few countries (asin Russia in 2004), the staff expressed forcefully theview that capital controls, no matter how useful theymight be in the short run, could not be expected to beeffective over time and should not be used as a sub-stitute for appropriate adjustment in macroeconomicpolicies. In others (as in Colombia), the use of con-trols introduced by the authorities did not figure

prominently in policy discussions. In still other cases(as in Croatia), the staff recommended a market-based control, albeit as a last resort measure.

Overall Assessment

Throughout the 1990s, the IMF undoubtedly en-couraged countries that wanted to move ahead withcapital account liberalization, and even acted as acheerleader when it wished to do so, especially be-fore the East Asian crisis. However, there is no evi-dence to suggest that it exerted significant leverage topush countries to move faster than they were willingto go. The process of liberalization was often drivenby the authorities’ own economic and political agen-das, including accession to the Organization for Eco-nomic Cooperation and Development (OECD) or theEuropean Union (EU) and commitments under bilat-eral or regional trade agreements. In encouragingcapital account liberalization, the IMF pointed outthe risks inherent in an open capital account as wellas the need for a sound financial system, even fromthe beginning. The problem was that these risks wereinsufficiently highlighted, and the recognition of therisks and preconditions did not translate into opera-tional advice on pace and sequencing until later in the1990s (and even thereafter the policy advice hasoften been of limited practical applicability).

In multilateral surveillance, the IMF’s analysisemphasized the benefits to developing countries ofgreater access to international capital flows, whilepaying comparatively less attention to the risks in-herent in their volatility. As a consequence, its policyadvice was directed more toward emerging marketrecipients of capital flows, and focused on how tomanage large capital inflows and boom-and-bust cy-cles; little policy advice was offered, in the contextof multilateral surveillance, on how source countriesmight help to reduce the volatility of capital flows onthe supply side. In more recent years, the IMF’sanalysis of such supply-side factors has intensified.Even so, the focus of policy advice—beyond theanalysis of macroeconomic policies covering largecurrent account imbalances—remains on the recipi-ent countries.

In country work there was apparent inconsis-tency in the IMF’s advice on capital account issues.Sequencing was mentioned in some countries butnot in others; advice on managing capital inflowswas in line with standard policy prescriptions, butthe intensity differed across countries or acrosstime (with no clear rationale provided for the dif-ference); and a range of views was expressed on use of capital controls (though greater conver-gence toward accommodation was observed overtime). Policy advice must of necessity be tailored

5


EXECUTIVE SUMMARY

to country-specific circumstances, so uniformitycannot be the only criterion for judging the qualityof the IMF’s advice. Country documents, however,provide only an insufficient analytical basis formaking a definitive judgment on how the staff’spolicy advice was linked to its assessment of themacroeconomic and institutional environments inwhich it was given. While one can explain why cer-tain types of advice were offered in some individ-ual cases, no generalization is possible about theconsistency of the IMF’s overall policy advice.Even so, it appears that the apparent inconsistencyto a large extent reflected reliance on the discretionof individual IMF staff members, and not necessar-ily the consistent application of the same principlesto different circumstances.

The evaluation suggests that the IMF has learnedover time on capital account issues. This seems tohave affected the work of the IMF through two chan-nels. First, the IMF’s general approach did respond—albeit gradually—to new developments or new evi-dence. Second, independent of how the generalapproach changed, some of the learning becamemore quickly reflected in the IMF’s country workthrough its impact on individual staff members. Thelack of a formal IMF position on capital account lib-eralization and the associated partial disconnect be-tween general operational guidelines and countrywork had different consequences. On the one hand,this gave individual staff members freedom to usetheir own professional and intellectual judgment indealing with specific country issues. On the otherhand, the disconnect reflected the inherent ambiguityof this aspect of the IMF’s work and led to some lackof consistency in country work, as noted above.

In more recent years, somewhat greater consis-tency and clarity has been brought to bear on theIMF’s approach to capital account issues. For themost part, the new paradigm upholds the role ofcountry ownership in determining pace and sequenc-ing; takes a more consistently cautious and nuancedapproach to encouraging capital account convertibil-ity; and acknowledges the usefulness of capital con-trols under certain conditions, particularly controlson inflows. But these are still unofficial views, nomatter how widely they may be shared within the in-stitution. While the majority of staff members nowappear to accept this new paradigm, some continueto feel uneasiness with the lack of a clear position bythe institution.

Recommendations

The evaluation suggests two main areas in whichthe IMF can improve its work on capital account issues.

Recommendation 1. There is a need for moreclarity on the IMF’s approach to capital account is-sues. The evaluation is not focused on the argumentsfor and against amending the Articles of Agreement,but it does suggest that the ambiguity about the roleof the IMF with regard to capital account issues hasled to some lack of consistency in the work of theIMF across countries. This may reflect the lack ofclarity in the Articles, but with or without a changein the Articles it should be possible to improve theconsistency of the IMF’s country work in otherways. For example:

• The place of capital account issues in IMF sur-veillance could be clarified. It is generally un-derstood that while under current arrangementsthe IMF has neither explicit mandate nor juris-diction on capital account issues, it has a re-sponsibility to exercise surveillance over certainaspects of members’ capital account policies.However, much ambiguity remains on the scopeof IMF surveillance in this area. The cleareststatement of the basis for surveillance of capitalaccount issues is embodied in the 1977 Execu-tive Board decision calling for surveillance toconsider certain capital account restrictions in-troduced for balance of payments purposes, butthe qualification limiting the scope to balance ofpayments reasons is too restrictive to cover therange of capital account issues that surface inthe IMF’s country work. On the other hand, thebroader statement of the IMF’s surveillance re-sponsibility, found in the preamble to Article IV,is too wide to serve as an operational guide tosurveillance on capital account issues. Therewould be value if the Executive Board were for-mally to clarify the scope of IMF surveillanceon capital account issues. Such a clarificationwould recognize that capital account policy isintimately connected with exchange rate policy,as part of an overall macroeconomic policypackage, and that in many countries capitalflows are more important in this respect thancurrent flows; capital controls can be used tomanipulate exchange rates or to delay neededexternal adjustment; and a country’s capital ac-count policy creates externalities for other coun-tries. Capital account policy is therefore of cen-tral importance to surveillance.

• The IMF could sharpen its advice on capital ac-count issues, based on solid analysis of the par-ticular situation and risks facing specific coun-tries. Given the limited evidence that exists inthe literature on the benefits or costs of capitalaccount liberalization in the abstract, the IMF’sapproach to any capital account issue must nec-essarily be based on an analysis of each case.

6


Executive Summary

For example, if a capital control is involved, theIMF must ask in the context of a specific coun-try what objectives the control is designed toachieve; if it is accomplishing them; andwhether there are more effective or less distor-tionary ways of achieving the same objectives.Such assessments need to be set in an overallconsideration of the macroeconomic policyframework and whether controls are being usedas a substitute for, or to seek to delay, necessarychanges in such policies. The evaluation indi-cates that this is already done in some, but notall, cases. If a capital control measure is judgeduseful to stem capital flight under certain cir-cumstances, the IMF should ask what support-ing policies are needed to make it more effectiveor less distortionary (for example, setting up asystem of monitoring external transactions). Interms of providing advice on capital account lib-eralization, just to spell out all the risks inherentin opening the capital account is of limited use-fulness to countries seeking IMF advice. To as-sist the authorities decide when and how to openthe capital account, the IMF should providesome quantitative gauge of the benefits, costs,and risks (and, indeed, practicality) of moving atdifferent speeds. Admittedly, this is not an easytask. Drawing on the well-established literatureon welfare economics, the IMF must ask suchquestions as: What distortions are being createdwhen one market is liberalized but not another?What is the nature of risks being borne by resi-dents when capital account transactions are lib-eralized only for nonresidents? And what are thecosts to the economy (in terms of investmentflows) of allowing equity inflows but not debtinflows?

• The Executive Board could issue a statementclarifying the common elements of agreement oncapital account liberalization. At present, thereremains considerable uncertainty among manystaff members on what policy advice to provideto individual countries. This has led to hesitancyon the part of some within the staff to raise capi-tal account issues with country authorities. TheExecutive Board could provide clear guidance tostaff on what the IMF’s official position is. Thisis not to say that the Executive Board must comeup with a definitive statement on all aspects ofpace and sequencing. Given the lack of full con-sensus, one should not expect such a definitiveview from the Board. However, Board guidanceon what are the minimum common elements onwhich there is broad, if not universal, agreementwould be useful to the staff and member coun-tries. Although the details are for the Board to de-

cide, such a statement might include some or allof the following elements: (1) that in a first-bestworld there would be no need for controls overcapital movements (though financial marketsmay not always operate accordingly); (2) thatcontrols should not be used as a substitute for adjusting macroeconomic or structural policies;(3) a broad (as opposed to unnecessarily com-plex) framework of sequencing based on the con-sensus in the literature on the order of economicreforms; (4) the importance of taking country-specific circumstances into account; and (5) thatrisks can never be totally eliminated, so theyshould not be used as a reason for permanentlydelaying liberalization.

Recommendation 2. The IMF’s analysis and sur-veillance should give greater attention to the sup-ply-side factors of international capital flows andwhat can be done to minimize the volatility of capi-tal movements. The IMF’s policy advice on manag-ing capital flows has so far focused to a consider-able extent on what recipient countries should do.While this is important, it is not the whole story. Asdiscussed in the evaluation report, the IMF’s recentanalyses have given greater attention to supply-sidefactors, including the dynamics of boom-and-bustcycles in emerging market financing. The IMF hasalso established an International Capital MarketsDepartment (ICM) as part of an effort to better un-derstand global financial markets; it participatesactively in the work of the Financial StabilityForum, which was established to monitor potentialvulnerabilities in global financial markets; and ithas proposed a Sovereign Debt RestructuringMechanism (SDRM), encouraged the use of collec-tive action clauses (CACs), and has attempted toplace limitations on countries’ access to IMF re-sources in a crisis, in an effort to reduce the per-ceived moral hazard that may have led capital mar-kets to pay insufficient attention to the risks ofinvesting in developing countries and contributedto the boom-and-bust cycles of capital movements.These are important and welcome initiatives, butthe IMF has not yet fully addressed issues of what,if anything, can be done to minimize the volatilityof capital flows by operating on the supply side—as yet, little attention seems to be paid to supply-side risks and potential mitigating actions in the in-dustrial countries that are home to the major globalfinancial markets. The IMF could usefully providemore input into advanced country financial super-vision and other financial market policy issuesglobally. Are current global supervision guidelinesdesigned to help create stability? What if any actioncould be taken on the supply side to reduce cycli-cality and herd behavior? Admittedly, this is a diffi-

7


EXECUTIVE SUMMARY

cult topic on which little professional consensusexists. Yet, this is an area where a significant debatehas taken place in the academic and policymakingcommunities and to which the IMF could con-tribute further. Indeed, one of the broad themesidentified as potential priorities for the IMF’s re-

search program over the medium term—on institu-tions and contractual mechanisms that can helpprotect countries from external volatility—goessome way in this direction, but should not focusonly on policies in countries that are recipients ofcapital inflows.

8


The 1990s witnessed a large swing in global pri-vate capital flows (Figure 1.1). Net private flows

to developing countries, for example, grew from lessthan $100 billion in 1990 to well over $200 billion in1995. The subsequent years, however, saw an equallysubstantial reversal of these inflows, which causedseveral emerging market economies to experience se-vere capital account crises. The volume of privatecapital flows to developing countries remained sub-dued through the early 2000s.

Against this background, there has been a majordebate over the actual and potential role of the IMFin encouraging countries to open their capital ac-counts1 and any possible associated increase in theirvulnerability to crisis. Within the broader debate overthe increasing importance of international capitalflows in the world economy,2 some have alleged thatthe IMF, in concert with some major shareholdergovernments, had encouraged member countries toliberalize their capital accounts prematurely withoutensuring that adequate institutions and prudentialregulations were in place.3 Others argue that rapidliberalization, with insufficient attention to sequenc-ing and establishing the appropriate preconditions,has been responsible for much of the financial insta-bility and economic distress experienced by manyemerging market countries.4

The role of the IMF has been particularly contro-versial because capital account liberalization is anarea where there is little professional consensus (seeBox 1.1). In this context, Eichengreen (2001) hasnoted that the views favoring liberalization emergedwith a surprising degree of certitude in advance of(and in the absence of) definitive evidence. The IMF’srole has been controversial for another reason: Al-though current account liberalization is among theIMF’s official purposes outlined in its Articles ofAgreement, it has no explicit mandate to promotecapital account liberalization. Indeed, the Articlesgive the IMF only limited jurisdiction over the capitalaccount (see Chapter 2, “The Legal Basis,” for de-tails). Nevertheless, the IMF has given greater atten-tion to capital account issues in recent decades, given

Introduction

9

CHAPTER

1

1Since the fifth edition of the IMF’s Balance of Payments Man-ual was published in 1993, the term used for statistical purposeshas been the “capital and financial account.” However, this reportfollows the established practice, both within the IMF and in theacademic literature, of using the term “capital account” to de-scribe the subset of the balance of payments that covers all non-current international transactions.

2The broad international interest in capital account issues that ex-isted during the 1990s can be seen, for example, in the coveragegiven by successive issues of the UNCTAD’s Trade and Develop-ment Report (see, in particular, Chapter 5 of the 1999 report).

3While such a policy was often referred to as part of the “Wash-ington consensus,” full capital account liberalization was in factnot one of the 10 policy reforms that Williamson (1990) consid-ered as forming the Washington consensus. The presumed con-sensus was not on liberalization of capital flows in general, butrather more specifically on that of foreign direct investment.

4Academic proponents of these views are Desai (2003), Stiglitz(2000, 2002, and 2004), Wade (1998–99), and Wade and Veneroso(1998).

Figure 1.1. Private Capital Flows to DevelopingCountries1

(In billions of U.S. dollars)

Source: IMF database.1Portfolio investment flows, other private investment flows, and foreign

direct investment to all developing countries, Israel, and Korea. Excludes government borrowing.

–400

–300

–200

–100

0

100

200

300

400

Inward direct investment

Changes in liabilities

Changes in assets

0220009896949290

Outward direct investment

Net flows


CHAPTER 1 • INTRODUCTION

the increasing importance of international capitalflows for macroeconomic stability and exchange ratemanagement in many countries. In view of thesefacts, an independent assessment of how the IMF hasaddressed capital account issues seems warranted.

The Scope of the Evaluation

The evaluation seeks to (1) contribute to trans-parency by documenting what in practice has been theIMF’s approach to capital account liberalization andrelated issues; and (2) identify areas, if any, where theIMF’s instruments and operating methods might beimproved, in order to deal with capital account issuesmore effectively.5 The issues addressed in the evalua-tion cover not only capital account liberalization butalso capital flow management issues, including partic-ularly the temporary use of capital controls. We evalu-ate the IMF’s actual approach to these issues, not nec-essarily its official policy. Indeed, as will become

clear, it is difficult to argue that the IMF had a firmformal policy on the issues we address—at least notduring the period covered by the evaluation. In evalu-ating the IMF’s approach, we rely primarily on coun-try-based analysis. We will try to identify, for exam-ple, what policy advice the IMF gave in the context ofa specific country at a specific point in time. Althoughcontext is important, the focus remains on the role ofthe IMF. We make no judgment on the underlyingpolicies adopted by country authorities.

Given the lack of consensus in the academic andofficial policymaking communities, there is no uni-versal set of criteria against which the IMF’s ap-proach to capital account issues can be assessed.Rather, we take a pragmatic approach to evaluationby asking the following questions about the IMF’spolicy advice:

(1) Was there any difference between the IMF’sgeneral policy pronouncements and the adviceit gave to individual countries?

(2) Was the IMF’s policy advice operational? Wasit based on solid evidence?

(3) How did the IMF’s advice change over time?Did this change keep pace with available evi-

10

Box 1.1. The Debate on the Benefits of Capital Account Liberalization

The theoretical rationale for capital account liberal-ization is based primarily on the argument that freecapital mobility promotes an efficient global allocationof savings and a better diversification of risk, hencegreater economic growth and welfare (Fischer, 1998).An opposing view has held that there is considerableinformation asymmetry in international financial mar-kets, so that free capital mobility—especially when sig-nificant domestic distortions exist—does not necessar-ily lead to an optimal allocation of resources (Stiglitz,2000 and 2004). Between these two opposing positionsis the view that, while there are benefits to be gainedfrom liberalization, the magnitude of the gains is rela-tively small.1 While the idea that free capital mobilityenhances economic welfare is an appealing concept tomany economists, there has been surprisingly little em-pirical evidence to date to either support or refute con-clusively such a view.

Recent empirical work has addressed this issue fromthe standpoint of the effect of capital liberalization oneconomic growth (see Edison and others, 2002, for asurvey). Unfortunately, the debate remains inconclu-

sive because such empirical studies inherently involvea joint test of the effect of liberalization on growth andthe particular method of quantifying the degree of lib-eralization or effectiveness of capital controls. Thisproblem is common to all empirical studies in thisarea.2 As it turns out, empirical results are sensitive notonly to the quantitative measure of capital controls butalso to the choice of sample and methodology. For ex-ample, while Quinn (1997) finds a positive associationbetween capital account liberalization and economicgrowth, Grilli and Milesi-Ferretti (1995) and Rodrik(1998) fail to find any such relationship. This ambigu-ity may reflect the role of institutions (for example, therule of law), macroeconomic stability, and other factorsin determining the effect of liberalization on growth(Arteta and others, 2001; Eichengreen and Leblang,2002).3 On the other hand, studies that have more nar-rowly focused on stock market liberalization havefound a positive impact on growth (for example, Henry,2003).

1For example, Gourinchas and Jeanne (2004) use a cali-brated neoclassical model to show that, for a typical develop-ing country, the welfare gains from switching from financialautarky to perfect capital mobility is about 1 percent perma-nent increase in domestic consumption.

2Another common problem is the endogeneity of capitalcontrols, which makes it difficult to disentangle the effect ofcapital controls per se from that of the macroeconomic and in-ternational environments within which they are introduced.

3Prasad and others (2003) also consider the effects of finan-cial integration on consumption smoothing and find little evi-dence to indicate the benefits of liberalization. See Stiglitz(2004) for commentaries on this work.

5The IMF’s instruments include surveillance, technical assis-tance, and IMF-supported programs.


Chapter 1 • Introduction

dence—that is, did the IMF learn as new evi-dence emerged (see, for example, Box 1.2)?

(4) Did the IMF give similar advice to countriesin similar situations?

(5) Was the policy advice on the capital accountset in a broader assessment of the authorities’macroeconomic policies and institutionalframework?

In asking the last two questions, in particular, weare not seeking to assess the IMF’s policy advice inindividual countries against a specific yardstick of“appropriateness,” since, as already noted, there is nosuch agreed measuring rod in many circumstances.Rather, the aim is to collect evidence related to twocommon (and, to some extent, contradictory) criti-cisms of the IMF’s approach, namely that (1) itadopted a “one size fits all” approach in its policy ad-vice, and (2) it was “inconsistent” by giving different

policy advice to countries in broadly similar situa-tions. In making such judgments, we faced a numberof limitations—most notably that the rationale forparticular policy advice was not always spelled out inthe relevant staff reports.

The evaluation does not address the question ofwhether capital account liberalization leads to fastergrowth (or generates other benefits)—an issue onwhich the wider body of research evidence has notreached definitive conclusions—or whether the IMFArticles of Agreement should be amended to give theIMF an explicit mandate for capital account liberal-ization and jurisdiction on member countries’ capitalaccount policies. Many aspects of these issues are notsusceptible to evidence from the evaluation. In thecase of the second issue, however, the evaluation doesshed some light on whether ambiguities about theIMF’s institutional role with regard to capital accountmatters has affected its country work in this area.

11

Box 1.2. Effectiveness of Market-Based Capital Controls: Evidence from Chile

From around 1996, there began to emerge a substan-tial body of empirical research on the effectiveness ofChile’s capital inflow control, which was introduced in1991 in the form of an unremunerated reserve require-ment (URR). We discuss available evidence to providebackground to subsequent discussions on the policy ad-vice given by the IMF on such controls.

This measure required a designated share of certaincapital inflows to be deposited with the central bank atzero interest for a designated period of time (see Box2.2 for details of how the system worked). Although dif-ferent studies came to different conclusions, by 1999,the sense of the literature—though evidence was oftenweak—was that (1) the URR allowed domestic interestrates to be somewhat higher; (2) it lengthened the matu-rity of capital inflows; (3) it had only limited effective-ness, if any, in reducing the volume of total inflows; and(4) it had little or no effect on the real exchange rate (seeNadal-De Simone and Sorsa (1999) and Gallego andothers (2002) for a review of the literature).

Most of these studies, however, contain seriousmethodological problems, making it difficult to acceptany conclusion with confidence. For example, mostused net inflows as the dependent variable, but govern-ment actions (including outflows liberalization, debtprepayment, and debt conversion programs) reducedthe net inflows independently of the URR by increasingthe outflows. Along with the liberalization of capitaloutflows, there were also changes in the administrativeregulation of capital inflows. Likewise, the operation ofthe URR itself changed over time, as the authoritiestried to close loopholes and increase effectiveness bywidening its coverage and raising its rate. The exclu-sion of certain short-term flows (such as trade credits)may also have biased the results of many of the studies,given the substitutability that existed between transac-

tions subject to the URR and those that were not. Forthese and other reasons, Nadal-De Simone and Sorsa(1999) concluded that it was “premature to point at theChilean experience as supportive of the effectiveness ofcontrols on capital inflows.”

Some of these methodological problems have beenaddressed in a more recent study by Gallego and oth-ers (2002). The study extends previous research byconsidering the endogeneity of the URR (the centralbank may tighten the URR in response to changes inthe strength of capital inflows), the effect of adminis-trative controls on capital flows, and using a muchlonger sample period covering 1989–2000. A signifi-cant contribution of the study is its consideration ofthe URR’s “effective cost,” which incorporates both“tax effectiveness” and “cost” and essentially mea-sures how binding the URR was. The findings of thisstudy broadly support the conclusions of previouswork: (1) the URR temporarily allowed domestic in-terest rates to increase relative to international rates;1

(2) it had no significant effect on the real exchangerate; (3) it significantly reduced the volume of capitalinflows, though the effect diminished over time; and(4) it unambiguously changed the composition of capi-tal inflows in favor of longer maturities. The less am-biguous effect of the URR on total inflows is new, butis supported by the findings of other recent research(see, for example, Le Fort and Lehman, 2003; Ffrench-Davis and Tapia, 2004).

1This, however, was achieved probably at the expense of in-creasing the cost of capital to smaller domestic firms withlimited access to international capital markets (Forbes, 2003;Gallego and Hernandez, 2003).



Most of the country-based analysis will use asample of emerging market economies, for whichprivate capital flows have been important. This se-lection of countries is justified by the fact thatemerging market economies received almost all ofthe private capital flows to developing countries inthe 1990s and arguably required the close attentionof the IMF. The largest 25 recipients, for example,accounted for as much as 90 percent of total cross-border investments during 1990–2002.6 Moreover,it is primarily for the possible role played in thesecountries that the IMF has been criticized.

It should be clearly stated at the outset that thechoice of emerging market economies may serve tomake the IMF’s role in capital account liberalizationappear less significant than it actually was. From1992 to 1997, for example, there was a significant re-duction in the number of IMF member countries withcapital controls, and much of this reduction was ac-counted for by low-income countries, including thosein sub-Saharan Africa (Figure 1.2).7 It is possible thatthe IMF had a more direct role in encouraging capitalaccount liberalization in some of these lower-incomecountries that relied on IMF financing.8 On the otherhand, if the number of countries is weighted by GDP,there was a sharp rise in capital account restrictive-ness in the early 1990s (reflecting the fact that a num-ber of former socialist economies joined the IMF);for the period as a whole there was little change inthe degree of capital account openness among theIMF’s developing country membership (Figure 1.3).This means that the sample, in which there is agreater representation of larger or higher-income de-veloping countries, may well be biased toward thosethat tended to maintain some capital account restric-tions (see below for the list of countries included inthe sample).

The evaluation will pay particular attention tocountry experiences with capital account liberaliza-tion (in terms of speed, sequencing, and precondi-tions) and policy responses to capital flows, includ-ing the temporary use of capital controls, and the

12

6This figure does not include foreign direct investments. Thecountries in the sample used in this report account for about 40percent of this total during 1990–97 (reflecting the exclusion ofArgentina, Brazil, Indonesia, and Korea), but the share increasesto over 50 percent during 1998–2002.

7Judgment about the presence or absence of capital controls ineach country is based on a one (controlled) or zero (not con-trolled) classification provided by the IMF’s Annual Report onExchange Arrangements and Exchange Restrictions (AREAER).It should be noted that this binary classification does not take ac-count of either the intensity or the number of controls.

8A recent econometric study by Joyce and Noy (2005) showsthat an extended IMF-supported program was statistically signifi-cant in explaining a country’s decision to remove capital controlsin the 1990s, suggesting that low-income countries often liberal-ized the capital account in the context of IMF financial support.

Figure 1.2. Countries with Capital Controls1

(In percent of total IMF membership)

Source: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER).

1Based on a one (controlled) or zero (not controlled) classification (covering all capital account transactions), as provided by the AREAER. There was a definitional change from 1997 to 1998.

55

60

65

70

75

80

85

90

Developing countries

All countries

20009896949290888684821980

Figure 1.3. Countries with Capital Controls1,2

(In percent of total developing IMF membership; GDP-weighted)

Source: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER).

1Based on a one (controlled) or zero (not controlled) classification (covering all capital account transactions), as provided by the AREAER.There was a definitional change from 1997 to 1998. GDP shares are based on 1990–2000 averages.

2The line would shift downward by about 5 percentage points if China and India were excluded.

65

70

75

80

85

90

20019895928986831980



IMF’s role and advice in these areas. In discussingcontrols on capital outflows introduced in the con-text of a capital account crisis, it must be stressedthat the focus will remain on issues specific to thecapital account, and we will not consider broadercrisis management issues (including, for example,private sector involvement and debt restructuring).Likewise, no attempt will be made to establish, inthe context of a specific country, causality betweencapital account liberalization and a subsequent capi-tal account crisis, although vulnerabilities to crisiscreated by a particular policy toward capital accountliberalization may be noted.

In discussing capital account openness in spe-cific countries, the evaluation will focus on de jure(as opposed to de facto) controls on capital transac-tions as defined by the IMF’s Annual Report on Ex-change Arrangements and Exchange Restrictions(AREAER). Clearly, when one investigates the eco-nomic impact of capital account liberalization, onemust define capital account openness in a way thathas an operational content. Edison and Warnock(2003), for example, suggest such an operationalmeasure that takes account not only of the exis-tence but also of the intensity of capital controls.9In this report, we are not asking questions about theeconomic impact of a particular control measure.Instead, we are more interested in knowing, for ex-ample, what the IMF said about removing or intro-ducing a particular control measure. We are, there-fore, focusing on de jure controls.

Sources of Evidence

The evaluation roughly covers the period1990–2004 and uses two types of documentation.First, it will use Executive Board papers and minutesof discussions on systemic themes, including WorldEconomic Outlook (WEO) and International CapitalMarkets Report (ICMR) or Global Financial StabilityReport (GFSR) exercises,10 Occasional Papers, Work-ing Papers, and various issues of the IMF Survey (formanagement speeches). The evidence gathered fromthese sources will be used to consider how the IMFviewed capital account issues over time, includingwhether it had a consistent approach and how effec-tively it adapted this approach in light of experience.

Second, the evaluation uses the IMF’s countrydocuments, including staff reports for Article IV

consultations and program reviews, internal briefingpapers and back-to-office reports for staff missions,staff memorandums and notes prepared for particu-lar country issues, the minutes of relevant Board dis-cussions, and technical assistance reports. This evi-dence is drawn from four overlapping groups ofcountries:

• Countries for which only staff reports (and, insome cases, summings up of Board discussions)are used for the period 1990–2002. There are 15countries in this category: Bulgaria, China,Croatia, Estonia, Israel, Lebanon, Peru, thePhilippines, Poland, Romania, Russia, the Slo-vak Republic, Slovenia, South Africa, andUkraine.

• Countries for which, in addition to staff reports,confidential internal documents are used for theperiod 1990–2002. There are 12 countries in thiscategory: Colombia, Chile, the Czech Republic,Hungary, India, Latvia, Lithuania, Malaysia,Mexico, Thailand, Tunisia, and República Boli-variana de Venezuela.11

These first and second groups of countries, num-bering 27, constitute the main sample upon whichcountry analysis for 1990–2002 is primarily based(see below for the selection criteria).

• Countries that have requested technical assis-tance from the IMF on aspects of capital accountliberalization, for which technical assistance re-ports are analyzed. There are 15 countries in thiscategory: Belarus, China, Colombia, the CzechRepublic, Hungary, India, the Islamic Republicof Iran, Kazakhstan, Latvia, Lesotho, Peru,Poland, Russia, Tanzania, and Tunisia.

• Countries with ongoing capital account issuesfor which confidential internal documents areused for the period 2003–04.12 There are 14countries: Bulgaria, China, Colombia, Croatia,India, the Islamic Republic of Iran, Kazakhstan,Libya, Morocco, Romania, Russia, South Africa,Tunisia, and Venezuela.13

The 27 countries in the first and second groupsare chosen on the basis of the size of portfolio capi-tal flows (absolute or relative to GDP) during1991–2002 (see Appendix 4), our own qualitative

13

9See also Prasad and others (2003, pp. 6–8), for a discussion ofthe difference between “the existence of de jure restrictions oncapital flows” and “de facto financial integration in terms of real-ized capital flows.”

10In 2002, the Global Financial Stability Report replaced theInternational Capital Markets Report.

11Brief field visits were made to receive the views of officialsand other experts in several of these countries: Chile, Colombia,the Czech Republic, Hungary, India, Latvia, Mexico, and Tunisia.A list of interviewees is provided in Appendix 6.

12These countries have been identified on the basis of a ques-tionnaire sent to recent mission chiefs and interviews with IMFstaff. The list of countries is not meant to be exhaustive.

13Additional information on ongoing issues was obtained frominterviews with senior IMF staff.



judgment of the degree of capital account opennessin 1990, and the changes introduced during the1990s. The list includes: (1) countries that signifi-cantly liberalized the capital account during the1990s; (2) countries that either still maintain or haveuntil very recently maintained significant controls oncapital account transactions; and (3) countries thatintroduced measures to restrict capital account trans-actions over the period. The second group of coun-tries was selected from this larger group for more in-depth examinations, based on our own judgment ofthe learning potential—the important criteria in-

forming this judgment were diversity of experienceand outcome—in order to make sure that we covervaried experiences with, and different stages of, cap-ital account liberalization. Argentina, Brazil, Indone-sia, and Korea were all important recipients of inter-national private capital inflows during much of the1990s but are not included in the sample, becausethe IEO’s earlier evaluations (IEO, 2003 and 2004)have already examined their relationships with theIMF (Box 1.3). This smaller sample, for example, al-lows us to take a closer look at the role of the IMF incountries that substantially eased restrictions on cap-

14

Box 1.3. Capital Account Liberalization in Indonesia and Korea

The role of capital account liberalization in the EastAsian crisis of 1997 has been a major topic of discus-sion. An earlier IEO report (IEO, 2003) discusses theeffectiveness of the IMF’s precrisis surveillance inidentifying financial sector vulnerabilities created bycapital account liberalization in Indonesia and Korea.Drawing on this report, we briefly review the IMF’srole in these two countries. The broad message is thatIMF surveillance failed to assess fully the underlyingrisks but that it did not play a major role in formulatingthe particular capital account liberalization strategyadopted by the authorities.

IndonesiaIndonesia had removed most controls on capital out-

flows by the late 1980s, and is often cited as an exam-ple of a country that had liberalized its capital accountbefore the current account. Indonesia, however, re-tained controls on various categories of capital flowsthroughout the 1990s. Almost all the liberalizationmeasures taken from the late 1980s to the mid-1990swere related to the liberalization of direct investmentinflows. Rapid capital inflows that began in 1990 tookplace against the background of financial sector liberal-ization and domestic capital market development. Atthe time of the crisis in 1997, a considerable number ofcontrols remained on many types of capital accounttransactions.1 In response to the large capital inflows,the IMF staff advocated tight fiscal and monetary poli-cies, greater exchange rate flexibility, accelerated struc-tural and banking sector reforms, and even faster exter-nal debt repayment. The IMF did not push a particularpath or pace of capital account liberalization. However,it underemphasized the risks of short-term capital in-

flows that were vulnerable to a sudden shift in senti-ment, and did not fully appreciate the weakness of thebanking sector and the vulnerability created by thecountry’s buildup of external debt.

KoreaIn Korea, it was in the context of OECD accession

that, in 1994, a Foreign Exchange System Reform Planwas announced to achieve full capital account convert-ibility in five years, in three stages (Kim and others,2001; Cho, 2001). The process began first with the lib-eralization of capital outflows, followed by a gradualeasing of restrictions on foreign investment in the do-mestic stock market and short-term trade-related bor-rowing. Notwithstanding these measures, however,Korea’s approach to capital account liberalization re-mained cautious. At the time of its OECD accession in1996, Korea retained a number of reservations to theCode of Liberalization of Capital Movements, particu-larly regarding the liberalization of long-term capitalinflows.2 The IMF staff was aware of the weak bankingsystem but did not sufficiently appreciate the vulnera-bilities created by the buildup of short-term externalborrowing by weak, poorly regulated financial institu-tions. Its view on Korea’s particular choice of sequenc-ing was that the speed of liberalization should be accel-erated. The IEO report states that staff papers andBoard discussions on Korea were “concerned primarilywith the speed of liberalization (typically recommend-ing a faster process)” and that “[issues] of sequencingand supervision were inadequately addressed in thesurveillance process.” Further liberalization of the capi-tal account proceeded in the context of a program sup-ported under the 1997 Stand-By Arrangement.

1These included nonresident purchases of Indonesianshares; the sale or issue of money market instruments abroadby residents; the granting of commercial credits by nonresi-dents to residents; purchases of land by nonresidents; bankborrowing from abroad; and bank lending to nonresidents(Johnston and others, 1997).

2As a result, at the time of the 1997 crisis, controls of onetype or another remained on such capital account transactionsas: issues of foreign-currency-denominated securities by resi-dents; purchases of local securities by nonresidents; purchasesof money market instruments by nonresidents; external borrow-ing by banks; inward direct investments; and even some tradecredits (Johnston and others,1997; Kim and others, 2001).



ital transactions during the early 1990s, the nature ofIMF advice for countries that took a gradual ap-proach to capital account liberalization, and theIMF’s views in the context of specific country expe-riences with capital controls.

In order to aid the evaluation, we have created anindex of de jure capital account openness that utilizesa more detailed classification of capital account trans-actions than the simple 1/0 system. In particular, fol-lowing Miniane (2004), we assign 1 (or 0) to each ofthe 10 categories of capital account transactions as re-ported in the AREAER when a restriction is present(absent), and express the sum in percentage terms.This index has been calculated for the 12 core coun-tries to which we give closer attention (see Appendix5).14 It turns out that 7 of these countries maintainedmoderate to extensive restrictions on capital accounttransactions almost consistently during 1990–2002; 2countries had a largely open capital account; another2 countries eased restrictions significantly in the late1990s or early 2000s; and 1 gradually reduced restric-tions over the period. As a result, when we look at theaverage index of capital account openness for thesecountries, we find that the index remained relativelyhigh throughout the period, but observe a gradual de-cline in restrictiveness (Figure 1.4).

Organization of the Report

The rest of the report is organized as follows.Chapter 2 “General Policy and Analysis” reviews thelegal basis for the IMF’s work on capital account is-sues, intellectual and operational developmentswithin the IMF in this area from the early 1990s tothe early 2000s, and how the issues were viewed inthe IMF’s multilateral surveillance work. The fol-lowing two chapters present an analysis of the IMF’sapproach to capital account liberalization based onwork in individual countries. Chapter 3 “Advice toMember Countries” assesses the IMF’s specific ad-vice to member countries during 1990–2002 on cap-ital account liberalization, macroeconomic and

structural policies to manage large capital inflows,and the temporary use of capital controls. Chapter 4“Ongoing Country Dialogue on Capital Account Is-sues” provides an overview and assessment of theIMF’s latest country work on capital account issues.Chapter 5 “Major Findings and Recommendations”summarizes major findings and suggests two broadrecommendations to help improve the IMF’s opera-tions in the area of capital account issues.

The main body of the report is followed by six ap-pendixes. Appendix 1 “A More Detailed Assessmentof Some Country Cases” provides a more in-depthanalysis of how the IMF viewed capital account is-sues over time in the context of four countries withdiverse experiences: the Czech Republic, Colombia,Tunisia, and Venezuela. Appendix 2 provides anoverview of relevant staff research on capital ac-count topics during 1990–2004. Appendix 3 reviewsthe IMF’s public communications on capital accountissues, focusing on management speeches and otherpublic statements. Appendix 4 summarizes quantita-tive indicators of capital flows in the main sample of27 countries. Appendix 5 depicts the indices of capi-tal account openness for the 12 core countries. Fi-nally, Appendix 6 provides a list of people inter-viewed by the evaluation team.

15

Figure 1.4. Average Capital Account Openness in 12 Sample Countries(In percent)

Source: IEO estimates based on MFD data. See Appendix 5.

55

60

65

70

75

80

85

90

022000989694921990Le

ss o

pen

Mor

e op

en

14IMF (1999, pp. 83–96) offers a methodology of calculatingan index of capital account restrictions based on even more de-tailed transaction categories and ranks 41 industrial, developing,and transition economies for 1996. See also Johnston andTamirisa (1998). Because of data limitation, however, we insteadfollow—with some modifications—the methodology of Miniane(2004), who extends the indices to the pre-1996 period but for asmaller set of 34 countries and based on 10 categories of capitalaccount transactions.


This chapter discusses the legal basis for theIMF’s work on capital account issues, intellec-

tual and operational developments within the IMF re-lating to its policies in this area, and how capital ac-count issues were viewed by staff and the ExecutiveBoard in the context of multilateral surveillance exer-cises. In reviewing the IMF’s general policies andanalyses, we discuss, but only briefly, the backgroundto the debate in the 1990s on whether or not the IMF’sauthority in this area should be expanded and the as-sociated initiatives taken by IMF management toamend the IMF Articles of Agreement. Although theconsequences of such an amendment would havebeen significant for the IMF’s formal role, the focusof this report remains on what the IMF actually did orsaid in the course of its operational work.

The Legal Basis

The IMF’s approach to capital account issues, in-cluding capital account liberalization in particular, hasbeen a controversial topic, in part because there existslittle consensus on what the role of the IMF in thisarea should be. To give a sense of issues involved, wefirst discuss what the IMF Articles of Agreement say,including the distinction between “purpose” (or“mandate”) and “jurisdiction,” as well as between cur-rent and capital account transactions. We then discussthe evolving role of the IMF in capital account issues,as it was interpreted by the Executive Board, and thecontext in which the evolution took place.

Mandate versus jurisdiction

Within the IMF, the term “mandate” has been used,in place of the legal term “purpose,” to refer to the ob-jectives which the IMF must pursue in its operationsand activities;1 “jurisdiction,” on the other hand, refers

to the IMF’s legal authority to assess and enforcemember countries’ compliance with obligations speci-fied under the Articles.2 Article I of the IMF’s Articlesof Agreement sets out the “purposes” of the IMF and,in effect, defines the institution’s mandate, including:

• To promote international monetary cooperationthrough a permanent institution which providesthe machinery for consultation and collabora-tion on international monetary problems;

• To facilitate the expansion and balanced growthof international trade, and to contribute therebyto the promotion and maintenance of high levelsof employment and real income and to the de-velopment of the productive resources of allmembers as primary objectives of economicpolicy;

• To promote exchange stability, to maintain or-derly exchange arrangements among members,and to avoid competitive exchange depreciation;and

• To assist in the establishment of a multilateralsystem of payments in respect of current trans-actions between members and in the eliminationof foreign exchange restrictions which hamperthe growth of world trade.

Consistent with the IMF’s mandate to facilitate theexpansion and balanced growth of internationaltrade, the members of the IMF bestowed upon the in-stitution jurisdiction over restrictions on the makingof payments and transfers for current internationaltransactions. Article VIII (“General Obligations ofMembers”) stipulated that, without the approval ofthe IMF, members could not (except under “transi-tional arrangements” defined in Article XIV, Section2 or with respect to “scarce currency” provisionsunder Article VII) impose restrictions on the makingof payments or transfers for current international

General Policy and Analysis

16

CHAPTER

2

1“Mandate” is not a legal term used in the Articles of Agree-ment, but it assumed currency in place of “purpose” as used in theArticles. These two terms are often used interchangeably withinthe IMF, because an institution’s mandate is essentially the pur-poses for which it was founded or which were subsequently as-signed to it.

2How “jurisdiction” is defined here is consistent with the usageof the word as it pertains to the IMF’s responsibilities vis-à-visthe making of payments and transfers associated with current ac-count transactions.


Chapter 2 • General Policy and Analysis

transactions, such as purchases of imports. At thesame time, the IMF was not given jurisdiction overthe underlying current account transactions for whichthe payment was required.3

It should be noted that current internationaltransactions as defined in the Articles are broaderthan the standard statistical definition of currenttransactions and include some categories that arenormally considered to be capital transactions andcapital transfers. In particular, Article XXX definespayments for current transactions as “paymentswhich are not for the purpose of transferring capi-tal.” It then explicitly mentions restrictions on “nor-mal short-term banking and credit facilities” and“payments of moderate amount for amortization ofloans or for depreciation of direct investments” asbeing subject to IMF jurisdiction.

The IMF and the capital account

The evolution of the IMF’s involvement with thecapital account was different. The Articles of Agree-ment did not provide the IMF with a clear mandateto encourage capital account convertibility. The ex-clusion of most capital transactions (and the associ-ated making of payments and transfers) from IMFjurisdiction was deliberate (de Vries, 1969, p. 224).Both of the main architects of the Bretton Woods in-stitutions, John Maynard Keynes and Harry DexterWhite, argued that countries should be protectedfrom the disruptive impact of speculative interna-tional capital movements and that a world of unre-stricted capital movements was not compatible witheither a stable exchange rate system or a liberal in-ternational trading system.4 These views reflectedthe consensus, held at the time the Articles were

being drafted, that “the large short-term capitalflows of the 1920s and 1930s had led to disaster” bythreatening exchange rate stability and making it dif-ficult to achieve monetary and fiscal stability(James, 1996, pp. 37–38). Bloomfield (1946) char-acterized this consensus as a “highly respectabledoctrine, in academic and banking circles alike” andhaving “been officially crystallized in the BrettonWoods Fund Agreement.”

Consequently, unlike restrictions on the currentaccount, capital controls were seen to be a necessaryand useful instrument of economic management,particularly for giving governments autonomy fromfinancial markets and discouraging “speculative”capital and “hot money.”5 Thus, Article VI, Section 1allowed the IMF to request a member to exercisecontrols to prevent the use of the IMF’s general re-sources to finance a large or sustained outflow ofcapital, and stated that failure to do so could result inthe member being declared ineligible to use theIMF’s general resources. Section 3 of the same arti-cle recognized the right of members to “exercisesuch controls as are necessary to regulate interna-tional capital movements,” as long as it were done ina manner that did not restrict current internationalpayments or transfers. This provision was reaffirmedin a subsequent decision of the IMF ExecutiveBoard, approving a report of its Committee on Inter-pretation, which provided that: “Subject to the provi-sions of Article VI, Section 3 concerning paymentsfor current transactions . . . members are free toadopt a policy of regulating capital movements forany reason, due regard being paid to the general pur-poses of the Fund. . . . They may, for that purpose,exercise such controls as are necessary . . . withoutapproval of the Fund.”6

Over the subsequent decades, however, two im-portant developments took place, which changed theenvironment envisaged in the Articles. First, startingin the late 1950s, an increasing number of countrieshave removed restrictions on the making of paymentsand transfers for current transactions and acceptedthe obligations under Article VIII, Sections 2, 3, and4 of the IMF Articles. The effectiveness of capitalcontrols depends to some extent on the ability to con-trol or at least monitor current account transactions,because (1) some current transactions can substitutefor capital account transactions that are otherwise re-stricted and (2) current transactions can create scopefor disguised capital transactions through leads andlags or under- and over-invoicing. The removal of re-strictions on current payments and transfers has to

17

3There are several reasons for the distinction that emerged be-tween current transactions and the making of payments and trans-fers for those transactions. Jurisdiction over the underlying cur-rent transactions was to have resided with a proposedInternational Trade Organization. However, opposition to this ap-proach from key constituencies led instead to reliance upon themuch less ambitious General Agreement on Tariffs and Trade(GATT). For a more in-depth discussion of the political contextunderlying decisions on current account convertibility and juris-diction, see James (1996).

4Helleiner (1994, pp. 33–38) discusses how the views ofKeynes and White influenced the provisions of the Articles ofAgreement regarding capital account issues. Keynes argued thatinternational capital movements should be allowed only “for le-gitimate purposes” and that there must be “a means . . . of con-trolling short-term speculative movements or flights of currency.”White, for his part, argued that “the task . . . is not to prohibit in-struments of control but to develop those measures of control . . .as will be the most effective in obtaining the objectives of world-wide sustained prosperity” (as quoted in Horsefield, 1969, pp. 32,64). See Boughton (2002) for a detailed analysis of the views ofKeynes and White, who differed in their assessments of why con-trols on capital movements were necessary.

5Such a view, shared by both Keynes and White, had been ini-tially advanced by League of Nations economists in the 1930s(see Nurkse, 1944).

6Executive Board Decision No. 541-(56/39), July 25, 1956.


CHAPTER 2 • GENERAL POLICY AND ANALYSIS

some extent diminished the effectiveness of any re-maining capital controls. More recently, the expan-sion of financial market innovation, including the de-velopment of new and more complex financialinstruments, has made it even more difficult to en-force capital controls effectively. These were impor-tant factors accelerating moves toward liberalizationamong industrial countries in the 1970s and 1980s.

Second, de facto capital account liberalization pro-ceeded in the context of such multilateral agreementsas the OECD Code of Liberalization of Capital Move-ments (1961)7 and the European Communities Direc-tives on Capital Account Liberalization (1986–88).The WTO’s General Agreement on Trade in Services(GATS) also served to ease restrictions on trade in fi-nancial services and, as a consequence, facilitated as-sociated capital movements among a wider group ofcountries. The increased freedom of capital move-ments, particularly among industrial countries, gener-ated large cross-border capital flows globally, withimplications for macroeconomic stability and ex-change rate management in many countries. Thesewere some of the developments that put into questionthe ability of the IMF to deal with these issues effec-tively and highlighted the potential role the IMF couldplay in ensuring orderly liberalization.

The role of the IMF in capital account issues

An important milestone in this process was theSecond Amendment of the IMF’s Articles of Agree-ment, which was put in place in the wake of the col-lapse of the Bretton Woods system of pegged ex-change rates. At that time, a new Article IV, interalia, provided for surveillance by the IMF over theexchange rate policies of members and establishedcertain obligations of members with respect to ex-change rate stability. The preamble to Article IV(Section 1) states that “the essential purpose of theinternational monetary system is to provide a frame-work that facilitates the exchange of goods, services,and capital among countries.”8 In enumerating spe-cific obligations of members, Article IV, Section 1required each member to “avoid manipulating ex-change rates or the international monetary system inorder to prevent effective balance of payments ad-justment or to gain an unfair competitive advantageover other members.”

Under the new framework, the Executive Boardadopted a decision setting out principles and proce-

dures for surveillance over members’ exchange ratepolicies (the 1977 Surveillance Decision).9 Theseprinciples noted the importance of restrictions on cap-ital movements and included, among the develop-ments that might indicate the need for discussion witha member, “the introduction or substantial modifica-tion for balance of payments purposes of restrictionson, or incentives for, the inflow or outflow of capital,”“the pursuit, for balance of payments purposes, ofmonetary and other domestic financial policies thatprovide abnormal encouragement or discouragementto capital flows,” and “unsustainable flows of privatecapital.”10 However, while unambiguously noting theimportance of the capital account for the purposes ofIMF surveillance, the 1977 Surveillance Decision im-plied neither the encouragement nor discouragementof capital account convertibility.

In the latter half of the 1990s, the IMF reassessedits mandate and jurisdiction over capital accounttransactions and considered the possibility of amend-ing the IMF’s Articles of Agreement (Box 2.1).11 Inthe event, the proposed amendment of the Articlesfailed to materialize. As it stands, Article IV (“Gen-eral Obligations of Members”) and the associated1977 Surveillance Decision of the IMF ExecutiveBoard, as amended in April 1995, define the role ofthe IMF in surveillance with respect to capital ac-count issues. It is now generally understood thatwhile the IMF does not have the authority to assessor enforce a standard of capital account convertibilityamong the general membership, it has a responsibil-ity to exercise surveillance over capital account poli-cies, albeit as part of its larger responsibility to exer-cise firm surveillance over exchange rate policies. Itis also understood that the IMF can use technical as-sistance for capital account issues. Ambiguity re-mains, however, because the Articles do not prescribea member’s specific obligation with respect to capitalaccount policies and technical assistance is not an ac-tivity explicitly mandated by the Articles.

General Operational Approach

Most of the intellectual and operational develop-ments of the 1990s related to capital account issues

18

7See Thiel (2003) for a discussion of the role of the OECDCode in encouraging capital account liberalization in recent ac-cession countries.

8The IMF’s Legal Department, however, has emphasized thatthis refers to a “purpose of the international monetary system,”and not necessarily of the IMF.

9Executive Board Decision No. 5392-(77/63), April 29, 1977.10The last item “unsustainable flows of private capital” was

added in the 1995 amendment.11As part of this debate, the distinction between restrictions on

the making of payments and transfers for capital transactions andthe underlying capital transactions did not figure as prominentlyas with the current account. This was largely because—for manycapital account transactions (e.g., long-term loans)—it was oftendifficult to distinguish operationally between the underlyingtransaction and its associated payments and transfers.



took place within the context of the management ini-tiatives to amend the Articles of Agreement to expandthe IMF’s mandate and jurisdiction. From 1995 to1999, staff prepared a number of policy papers forExecutive Board discussion, providing an analysis oflegal and other conceptual issues involved in amend-ing the Articles.12 As noted at the beginning of this

chapter, however, this section focuses on those devel-opments that shed light on the IMF’s actual opera-tional work. In fact, the IMF’s views and operationalwork on capital account issues evolved, respondingto new evidence or new developments (see Table2.1). We review this evolution in this section, whichforms an important part of the basis upon which theIMF’s country work will be assessed in the followingchapter.

19

Box 2.1.The Proposal to Amend the Articles of Agreement1

In the latter half of the 1990s, IMF managementproposed and actively promoted an amendment to theArticles of Agreement that would have transformed theIMF’s formal role in capital account liberalization andcapital account issues in general. The idea for anamendment had been raised within the IMF for sometime, at least since 1994, but it was in 1996 that theagenda to amend the Articles received priority in thework program of the IMF. During 1996 and 1997, theExecutive Board made intensive deliberations of theissues involved, to which IMF staff contributed signifi-cant intellectual inputs. At the level of the Board ofGovernors, the Interim Committee gave both encour-agement and specific directives from time to time. Thesupport of the Interim Committee reached its height inSeptember 1997, when, at its annual meeting in HongKong SAR, the Committee issued a communiqué out-lining the logic of its support for an amendment andrequesting the Executive Board to “accord high prior-ity” to submitting “a draft amendment to the Board ofGovernors.”

As the debate evolved, there emerged general agree-ment that the proposed amendment must involve twofundamental and distinct changes. First, the IMF was tobe endowed with a new purpose: to promote the liberal-ization of capital flows. Article I was to be amended toinclude the encouragement of the liberalization of capi-tal movements and the elimination of restrictions oncapital account transactions. Second, the IMF was toassume jurisdiction over restrictions in the capital ac-count. Jurisdiction would have established as a generalrule that member countries would be prohibited fromimposing restrictions on certain types of internationalcapital movements without the approval of the IMF.The amendment would also have resulted in a revisionof Article VI, which recognizes the right of members to“exercise such controls as are necessary to regulate in-ternational capital movements” as long as they do notrestrict current international payments or transfers.

In addition to the enthusiasm expressed by IMFmanagement, the vast majority—albeit not all—of in-dustrial countries favored the formalization of IMF au-thority on the regulation of international capital flows.Their authorities recognized that they lived in a world“very different from that faced by the Fund’s foundingfathers,” characterized by floating exchange rates andlarge capital movements among the major countries.Many developing countries, however, were moreguarded, considering that capital controls could be use-ful in some circumstances to deal with exchange ratepressure.2 Their apprehension was not fully erased bythe proposed provision for transitional arrangements(comparable to Article XIV for current transactions)and the language emphasizing the IMF’s role to ensure“orderly” liberalization. Consensus in favor of anamendment eluded the Executive Board, which contin-ued to discuss the possibilities for achieving morewidespread support for an amendment. In the event, inthe summer of 1997, most Executive Directors agreedthat inward direct investment, often sensitive politi-cally, would have to be excluded from the IMF’s ex-panded jurisdiction.

The East Asian crisis, contagion from which wasspreading through Asia and beyond at the time of the In-terim Committee’s Hong Kong SAR meeting in 1997,changed the dynamics of the debate in a fundamentalway. Because the crisis was unexpected and severe, therisks of capital account liberalization began to weigh onthe minds of policymakers who had previously empha-sized the benefits. In addition, opposition began toemerge from some influential members of the U.S.Congress, who felt reservations about giving more au-thority to the IMF when it was seeking an augmentationof its quota. Although IMF management never officiallyabandoned the idea, by the spring of 1999 it was clearthat sufficient support was not forthcoming to amendthe Articles, at least as it was drafted and proposed.

1This is based on a comprehensive analysis of this episode,as provided by Abdelal (2005).

2Concluding Remarks by the Acting Chairman, ExecutiveBoard Seminar on “Issues and Developments in the Interna-tional Exchange and Payments System,” November 16, 1994.

Two-Tiered Approach,” SM/99/220, September 3, 1999. Thepaper was never discussed by the Executive Board.

12The last of the series of formal Board papers discussing ap-proaches to amending the Articles was prepared in September1999 by the Legal Department, “The Role of the Fund in the Lib-eralization of Capital Movements—Further Considerations on a



Expanding role of the IMF

From the late 1980s, the IMF began to givegreater attention to capital account issues as part ofits surveillance work. Records show that the Execu-tive Board regularly, and with increasing frequency,held meetings to discuss international capitalflows.13 These early efforts, however, tended to be apositive analysis of what motivated internationalcapital flows and what the consequences would be,rather than an attempt to make a normative case for aparticular capital account policy. They also includeda review of measures taken by a number of develop-ing countries to access international capital markets,particularly in the aftermath of the debt-servicingdifficulties they had experienced in the 1980s, and

often emphasized the importance of sound macro-economic policies in attracting capital inflows.

In the early 1990s, in an environment in whichnearly all industrial countries had removed virtuallyall capital controls, staff prepared policy papers thatwere clearly advocating the benefits of capital ac-count liberalization, to which many Executive Direc-tors gave broad endorsement.14 While some of thesepapers raised questions of IMF jurisdiction and theneed to formalize the IMF’s role, it was not until themid-1990s that these jurisdictional issues receivedthe formal attention of the Executive Board.

Clearer support for capital account liberalizationemerged in the context of the so-called “Madrid Dec-laration on Cooperation to Strengthen Global Expan-sion,” adopted by the Interim Committee of the IMF’sBoard of Governors at its October 1994 meeting. Inthis meeting, Governors approved a statement wel-coming the “growing trend toward currency convert-

20

Table 2.1. Notable Events Affecting Capital Account Issues, 1991–2004

Date Events

June 1991 An unremunerated reserve requirement (URR) introduced in Chile.

September 1993 Chilean-style controls introduced in Colombia.

October 1994 “Madrid Declaration” issued by the Interim Committee, encouraging countries to removeimpediments to the free flow of capital.

December 1994 Mexican peso comes under pressure and is allowed to float.

July 1995 Executive Board’s first operational guidance on capital account liberalization issued to IMF staff.

December 1995 Staff operational note on capital account liberalization issued by the Policy Development andReview Department (PDR) and the Monetary and Exchange Affairs Department (MAE).

1995–96 OECD accession for the Czech Republic, Hungary, and Korea.

1996–97 Empirical evidence on the effectiveness of Chilean URR begins to appear.

1996–99 Executive Board deliberations on amending the Articles of Agreement.

March 1997 A supplement to the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictionspublished, with an expanded coverage of capital account regulations.

July 1997 Thai baht comes under pressure and is allowed to float.

September 1997 The Interim Committee meeting in Hong Kong SAR issued a communiqué supporting anamendment of the Articles.

November–December 1997 Executive Board approves Stand-By Arrangements for Indonesia and Korea.

August 1998 Russian default and devaluation.

September 1998 Capital outflow controls introduced in Malaysia.

January 1999 Brazilian devaluation.

May 1999 Financial Sector Assessment Program (FSAP) launched.

July 2001 New “integrated” approach to capital account liberalization discussed in an Executive Boardseminar.

May 2004 EU accession for eight transition economies, including the Czech Republic, Hungary, and Latvia.

13For example, the Board met during 1989–90 to discuss anumber of policy papers prepared by staff, including: “Policies toPromote Private Capital Inflows in Fund-Supported AdjustmentPrograms,” EBS/89/117; “Study on the Measurement of Interna-tional Capital Flows,” EBAP/89/269; “The Determinants andSystemic Consequences of International Capital Flows,”SM/90/128; and “Capital Market Financing for DevelopingCountries—Recent Developments,” SM/90/174.

14Some of these papers and the Executive Board’s reactions tothem are reviewed in Monetary and Exchange Affairs Depart-ment, “Issues and Developments in the International Exchangeand Payment System,” SM/94/202, August 1994.



ibility and [encouraging] member countries to removeimpediments to the free flow of capital.”15 Followingthe Madrid Declaration, in July 1995, the ExecutiveBoard gave its first operational guidance to the staff,when it met to review the recent experience of theIMF’s membership with capital account liberaliza-tion.16 Here, Directors gave general support to theidea that capital account issues should be coveredmore fully in Article IV consultations and that surveil-lance and technical assistance work be strengthenedto encourage and support capital account liberaliza-tion. These views were broadly endorsed by the In-terim Committee in its October 1995 communiqué.

It was around this time that IMF management andstaff began to give greater attention to capital accountissues in the actual operational work of the IMF. Aseries of notes were prepared during 1995, providingoperational guidelines for area department staff. Ofthese, a set of notes prepared by the Policy Develop-ment Review Department (PDR) and the ResearchDepartment (RES) in October 1995 discussed how toincorporate large unexpected capital inflows into pro-gram design, particularly when disinflation was animportant program objective; and how to identify thecauses of large capital inflows and determine appro-priate policy responses. For example, the PDR notestated that programs should include the quasi-fiscalcosts of sterilization within fiscal performance crite-ria and medium-term projections. The RES note sug-gested that the mix of instruments to deal with largecapital inflows would depend on the institutionalstructure of the country and the history of policies,adding that “temporary capital controls” might benecessary if the use of conventional macroeconomictools was restricted or their effectiveness limited.17

These were followed, in December 1995, by a“staff operational note” prepared by the Directors ofPDR and the Monetary and Exchange Affairs Depart-ment (MAE) and circulated to area departments, out-lining “the next steps to be followed by the staff inadapting Fund practices to elicit greater emphasis oncapital account issues, and to promote more activelycapital account liberalization.”18 The note requested

help and cooperation from area department staff intwo areas: (1) to give greater attention to capital ac-count developments in mission and technical assis-tance work; and (2) to assist in the collection of de-tailed information on capital account regulations fora pilot group of major emerging market economies.While the note’s guidance to the staff was clearly toencourage capital account liberalization, it also in-cluded a qualification: “Liberalization of capital ac-count transactions should generally be undertakenconsistent with progress in macroeconomic stabiliza-tion and structural reforms . . . in certain circum-stances, the use of capital controls to deter or slowsuch inflows may provide some temporary breathingroom for the authorities, while more fundamentalpolicy adjustments are being prepared.”

Over 1996–97, a significant improvement wasmade in the IMF’s capability to collect more detailedinformation on the regulatory framework of externalcapital account transactions. Initial work involvedadaptation of the well-established codes developedby the OECD; the expanded data on capital controlsclassified measures into 20 broad categories (10each for inflows and outflows). In December 1995, aquestionnaire was sent to a pilot group of 31 mem-ber countries, which was subsequently compiled as asupplement to the 1996 AREAER. With the success-ful completion of the pilot project, the coverage wasextended to all IMF member countries, and August1997 saw the publication of the 1997 AREAER withthe expanded coverage of capital account regulationsfor all countries. Subsequently, in 1998, the ex-panded data set allowed MAE to develop indices ofexchange and capital controls capable of providing aquantitative measure of the restrictiveness of a mem-ber’s exchange and capital control regime that iscomparable across countries.19

Pace and sequencing of capital account liberalization

The speed with which a controlled regime can (orshould) be liberalized depends on various factors, in-cluding risks, distortions, and institutional capac-ity.20 Developing effective regulatory frameworkstakes time, but a lengthy process may create wrongincentives and distortions. There are also political

21

15Similar support was evident in the Interim Committee com-muniqué of 1996 in which the members “encouraged the Fund, inpromoting liberalization in a global market setting, to pay in-creased attention to capital account issues and the soundness offinancial systems.”

16“Capital Account Convertibility—Review of Experience andImplications for Fund Policy,” SM/95/164. The Board discussionwas held on July 28, 1995. See EBM/95/73. The paper was subse-quently issued as an Occasional Paper (Quirk and others, 1995).

17These notes were circulated to the staff under a managementmemorandum dated October 25, 1995.

18“Strengthening Discussions and Information on Capital Ac-count Convertibility—Next Steps.” Cover memorandum, datedDecember 13, 1995.

19See, for example, Johnston and Tamirisa (1998), IMF (1999),and Miniane (2004).

20Lack of administrative capacity may argue either for oragainst faster reform, because there is no presumption that the re-source requirements of implementing a quick reform are eithersmaller or larger than those of managing a long transition processor administering capital controls. For a discussion of issues re-lated to the speed of reform, including the choice between a grad-ualist and a big-bang approach, see Nsouli and others (2002).



considerations. A big-bang approach may be appro-priate if a prolonged transition is likely to create re-sistance from vested interests or if different elementsof the existing system are so dependent upon eachother that a piecemeal reform is not possible withoutcreating significant distortions. A gradualist ap-proach, on the other hand, may be more appropriateif it takes time to build political consensus or if aslower process is more conducive to minimizing theadjustment costs.

Much of the scholarly literature in economics ad-vocated the big-bang approach in the context of tran-sition economies in the early 1990s, arguing that thelack of credibility in the reform made it more appro-priate to act quickly (Funke, 1993). In extending thebig-bang approach to nontransition contexts, manyexperts, including some at the IMF, argued that thebest route to an efficient financial sector was to liber-alize the capital account quickly, as it would allowmarket discipline to operate on the banking system(Guitián, 1996).21 Others in the IMF used the ineffec-tiveness of capital controls as the argument for fastercapital account liberalization, given their distortionaryeffects. For example, Mathieson and Rojas-Suarez(1993) stated: “Analyses of the sequencing of struc-tural reforms and stabilization policies for developingeconomies have traditionally argued that capital con-trols . . . are necessary to prevent capital flows thatwould undermine the reform program. These policyrecommendations, however, stand in sharp contrast toa growing body of empirical evidence that suggeststhat capital controls have often been evaded.”

Following the East Asian crisis, however, “se-quencing” emerged as an operational concept in theIMF’s approach to capital account liberalization. TheIMF staff emphasizes that “sequencing” as used inIMF terminology is an operational concept, involvingspecific measures of institutional building, which isdistinct from “order” as used in the literature on eco-nomic reform.22 A policy paper discussed by the Ex-ecutive Board in July 1998 stated: “The Asian coun-try experiences confirm that it is necessary toapproach capital account liberalization as an integralpart of more comprehensive programs of economic

reform, coordinated with appropriate macroeco-nomic and exchange rate policies, and including poli-cies to strengthen financial markets and institutions.The question is not so much one of the capital liberal-ization having been too fast, since some of the coun-tries in Asia have followed a very gradualist ap-proach. Rather, it is more to do with the appropriatesequencing of the reforms and, more specifically,what supporting measures need to be taken.”23

At the same time, the greater recognition of theneed for sound financial systems led, in May 1999,to the establishment of a Financial Sector Assess-ment Program (FSAP), which was to be adminis-tered jointly by the IMF and the World Bank. TheFSAP was meant to fill an identified gap in the inter-national financial architecture in support of crisisprevention, based on a judgment that existing ap-proaches at the IMF under Article IV consultationswere not sufficient for effective financial sector sur-veillance. Although participation is voluntary, over80 member countries have so far subjected their fi-nancial systems to assessment under the FSAP.24 Be-cause the FSAP’s key objective is an early detectionof financial sector vulnerabilities, the staff has usedits assessments as a basis for dialogue with the au-thorities of countries considering further capital ac-count liberalization; technical assistance on capitalaccount liberalization has also been given in con-junction with FSAP assessments.

Analytical work by staff on sequencing culmi-nated in a policy paper, which was discussed in anExecutive Board seminar in July 2001 and subse-quently issued as an Occasional Paper (Ishii and oth-ers, 2002). The paper stresses the importance of an“integrated” approach, which considers capital ac-count liberalization as part of a more comprehensiveprogram of economic reform and coordinates it withappropriate macroeconomic and exchange rate poli-cies as well as policies to strengthen the financialsystem. It analyzes different risks that might be posedto financial and macroeconomic stability by capitalaccount liberalization. In drawing operational princi-ples, it relies on the (both successful and not so suc-cessful) experiences of nine countries: Austria, Hun-gary, Korea, Mexico, Paraguay, South Africa,Sweden, Turkey, and the United Kingdom.25 The op-

22

21Guitián (1996) was initially presented in a conference held in1992. In a broader context, the “discipline effect” of internationalmarkets has been argued by some to operate on macroeconomicpolicymaking more generally. See, for example, Tytell and Wei(2004), who suggest that “financial globalization” may have en-couraged low-inflation monetary policies but not necessarily low-budget deficits.

22The early contributions in this literature were based on the“Southern Cone” experience of Argentina, Chile, and Uruguay inthe late 1970s, and emphasized the importance of achievingmacroeconomic stabilization, financial liberalization, and tradeliberalization before opening the capital account (McKinnon,1982; Edwards, 1984).

23Monetary and Exchange Affairs Department, “Developmentsand Issues in the International Exchange and Payments System—Background Studies,” SM/98/172, Supplement 2, July 8, 1998.

24A separate IEO evaluation of the FSAP process is under way.The terms of reference for this evaluation can be found atwww.imf.org/ieo.

25For example, the paper notes that capital controls in SouthAfrica and a cautious approach and early implementation ofstructural reforms in Hungary, respectively, may have limited thevulnerabilities of these countries to contagion from Russia.


www.imf.org/ieo


erational principles stress the importance of safe-guarding financial sector stability and maintainingconsistent exchange rate and macroeconomic poli-cies (see Box 4.3).

At the Executive Board seminar, Directors ex-pressed a range of views on the paper. While somethought that the approach was appropriate, othersexpressed the view that some of the suggested policymeasures could be implemented simultaneously andthat countries might want to “use windows of oppor-tunity” to move ahead quickly with capital accountliberalization. The Acting Chairman noted that speedwas not the issue—“what matters is the relationshipbetween capital account liberalization and otherpolicies, and that liberalization is sustainable.” Al-though there was a broad endorsement of the generalapproach proposed, the views expressed in a semi-nar—unlike those in a formal Board meeting—wereinformal, and no formal decision was taken. Conse-quently, the paper has not removed the ambiguitythat exists on the IMF’s formal policy advice on cap-ital account liberalization.

Multilateral Surveillance

Multilateral surveillance is an activity of the IMFthat, among other things, identifies major risks andvulnerabilities that may affect economic policymak-ing in member countries through global and regionallinkages. As the key outputs of multilateral surveil-lance exercises—the WEO and the ICMR/GFSR—are widely disseminated, they also serve as a channelby which the IMF communicates its views to thepublic.26 In this section, we review how capital ac-count issues were viewed by the IMF staff and theExecutive Board in the context of multilateral sur-veillance exercises.

Developments in advanced economies—“push” factors

In discussing the IMF’s policy advice on capitalaccount liberalization and management of capitalflows, one must first address the fundamental ques-tion of what determines the volume of global capitalflows to emerging and developing countries—the so-called “push” factors that largely originate in ad-vanced economies. In fact, it is well known that

global capital flows have been characterized by“boom-and-bust” cycles (Figure 2.1), and it is appro-priate to ask at the outset what the IMF’s multilateralsurveillance was saying about the causes of these cy-cles and the consequent policy implications for bothadvanced and emerging market countries.

Although a fuller discussion of this issue goeswell beyond the scope of this evaluation,27 it is use-ful to place the evaluation of the IMF’s multilateralsurveillance of “supply-side” factors within the fol-lowing stylized characterization of the two polarends of the debate:

• Some observers would argue that the fundamen-tal cause of “excessive” capital inflows toemerging markets followed by sudden outflowslies in weaknesses in both macroeconomic pol-icy (especially exchange rate policy) and in theframework governing financial institutions inemerging markets. According to this view, someof the solutions are to adjust macroeconomicpolicy settings and to strengthen institutionalframeworks (for example, through the various“standards and codes” initiatives). These ob-servers then note that solutions along these linesare under way (through more flexible exchange

23

26In 2002, the International Capital Markets Report (ICMR)changed its name to the Global Financial Stability Report (GFSR),but we refer to each report by the title used at the time it was pre-pared. The World Economic Outlook (WEO) and ICMR/GFSR re-ports, prepared by the staff in the context of multilateral surveil-lance, are released to the public, with the disclaimer that the viewsexpressed therein are those of the staff, and not necessarily thoseof the Executive Board.

Figure 2.1. Net Capital Inflows to Developing Countries1

(In percent of total developing country GDP)

Source: IMF database.1Changes in private foreign liabilities, including equity portfolio. Excludes

government borrowing.

–1.0

–0.5

0

0.5

1.0

1.5

2.0

2.5

0320019997959391898785831981

27A forthcoming IEO evaluation of the IMF’s multilateral sur-veillance is expected to address some of these issues in greaterdepth.



rate policies and the strengthening of financialsector and transparency frameworks in emerg-ing markets).

• An alternative and perhaps more pessimisticview would be that the unstable nature of theseflows has more to do with the fundamentals ofhow financial markets operate that are muchmore difficult to resolve (for example, environ-ments of suboptimal information that cause in-vestors to herd; and “informational cascades” orother market imperfections that lead to marketmyopia). Proponents of these views point to evi-dence suggesting that the incidence of crises hasnot declined, and may even have increased overthe long run (Eichengreen and Bordo, 2002;Persaud, 2001). These commentators usuallyemphasize the importance of “supply-side re-forms” in advanced economy financial marketsas part of the solution—although there are obvi-ous questions about the extent to which officialpolicies can, or should, seek to reduce volatilityin these markets.28

Most observers would agree that both strands ofthe debate are important. For the purposes of thisevaluation, the question is what the contribution ofthe IMF has been to the discussion on these issues,in terms of both analyses and policy advice.29

A review of multilateral surveillance documentssuggests that the IMF paid relatively little attention tothe push factors of global capital flows in the early1990s but has given this topic increased attention inmore recent years. This does not mean that the staffwas initially unaware of the importance of develop-ments in advanced economies in determining globalcapital flows. The early analysis, however, did notraise these issues in terms that stressed potential risksto emerging market economies. It appears that theIMF held a rather “fundamentalist” view of interna-tional capital flows. For example, the 1992 ICMRstated: “International capital flows continued to re-flect the current account imbalances of industrial and

developing countries and the international diversifi-cation of investment portfolios.” The staff tended toview any regulatory development that would increasedeveloping countries’ access to international capitalmarkets as beneficial.30

The prevailing idea that any measure that in-creases capital flows into emerging marketeconomies is good must be understood within thecontext of the period. During much of the 1980s, anumber of developing countries had lost access to in-ternational capital markets. Moreover, with the be-ginning of transition in former socialist economies, awidely held view stressed the need for greater globalsaving, and policy advice to advanced economiestended to be framed in these terms. For example, theOctober 1991 WEO noted the shortage of world sav-ing and called for industrial countries to consolidatetheir fiscal policies. The May 1995 WEO continuedto note the need to boost world saving: “World capi-tal flows and financial conditions are largely deter-mined by the industrial countries and the trend to-ward public dissaving is also heavily an industrialcountry problem. These are the countries where fiscalconsolidation can help boost world saving the most.”

As a result, multilateral surveillance at this time,while aware of the potential swings in capital flows,did not devote much attention to analyzing the po-tential risks to emerging market economies of capitalflow volatility. The 1994 ICMR, for example, notedhow hedge funds and other highly leveraged specu-lators had increased their exposure to emerging mar-kets, particularly in Latin America, but concludedthat these institutional investors were “often subjectto limitations—ranging from government regula-tions to self-imposed prudential restrictions—ontheir holdings of paper from developing country is-suers” though, given the still low level of exposure,the limits were not yet binding. The October 1994WEO noted a strong inverse relationship betweendevelopments in long-term interest rates in the majorcountries and stock prices in emerging markets,31

but included very little discussion of how these de-velopments in advanced economies posed policychallenges to emerging market economies.

The IMF staff expressed the view that the benefitsof greater integration brought about by portfolio di-versification outweighed the risks. The October

24

28See, for example, Ocampo and Chiappe (2003) for a proposalto introduce a countercyclical element into the regulation of fi-nancial intermediation and capital flows. See also Griffith-Jones(1998), Williamson (2002), and Griffith-Jones and Ocampo(2003). Metcalfe and Persaud (2003) emphasize the “fundamen-tal” causes of boom-bust cycles, but are skeptical about solutionsthat give a central role to improved information flows in crisisprevention. Such questions also apply to the debate over greaterdisclosure by hedge funds and other institutional investors.

29To the extent that crises are inevitable, more systemic crisisresolution procedures are also needed. In this context, the IMFhas been involved in discussions on various mechanisms, includ-ing collective action clauses and a Sovereign Debt RestructuringMechanism. See also Buiter and Sibert (1999) for the idea of anautomatic debt rollover mechanism. We will not discuss these cri-sis resolution issues here.

30For example, the 1992 ICMR mentioned the June 1991 deci-sion by the Japanese authorities to ease the credit rating standardsfor public placement of bonds in the Samurai market and the reg-ulatory changes in the United States to reduce the transactionscosts and liquidity problems facing developing country issues inU.S. capital markets as positive developments that “have facili-tated, or have the potential to promote, developing country accessto international bond markets.”

31The WEO then added a footnote: “There is no strong evidencethat recent capital flows are caused by speculative bubbles.”



1995 WEO, for example, remained optimistic: “Theincreased openness of developing countries and theirgreater integration into the world economy do notnecessarily mean greater vulnerability to externalconditions. Paradoxically, increased openness andgreater integration may reduce vulnerability becauseof stronger growth momentum in individual devel-oping countries and in the developing countries as agroup. . . . In addition, the impact on developingcountries of changes in the demand for their exportsmay be partially offset by countercyclical changes incapital flows, as has been the case recently, implyinga dampening of overall cyclical impulse from the in-dustrial countries.” Thus, the staff tended not to viewany developments in industrial countries that mightaffect capital flows to be an important risk factor foremerging markets.

In the latter part of the 1990s, and certainly fol-lowing the East Asian crisis, there was a fundamen-tal change in the way multilateral surveillanceviewed capital flow issues. It now began to paygreater attention to the linkage between industrialcountry developments and their capital flow and riskimplications for emerging market economies. In1998 and 1999, both the WEO and ICMR reports dis-cussed how an underestimation of risks by interna-tional investors and low interest rates in industrialcountries had contributed to large capital flows toemerging market economies. In this context, the Oc-tober 1998 WEO suggested that it would be wrong“to attribute financial crises exclusively to policyshortcomings in the crisis countries,” and called oninvestors and regulators in creditor countries to “rec-ognize the inherently fragile and volatile nature ofcapital flows by better pricing risks.” The ExecutiveBoard, in discussing the May 1999 WEO, “pointedto the need to improve the regulatory oversight, onthe supply side, of the highly leveraged activities offinancial institutions.”32

The staff’s analysis of risk factors inherent in fi-nancial integration has become increasingly sophis-ticated in more recent years. For example, the 2001ICMRs analyzed the relationship between financialmarket returns in mature markets and those in devel-oping countries, and its impact on global capitalflows to emerging markets; it also discussed thecross-border behavior of investors and how it might

affect aggregate private capital flows to emergingmarkets. The 2003 issues of GFSR included ananalysis of the “feast or famine” dynamics in emerg-ing debt markets (March issue) and the “boom-and-bust pattern and volatility” of capital flows (Septem-ber issue). However, the policy prescriptions drawnfrom this analysis have emphasized, not the actionsto be taken by creditor countries, but the actions tobe taken by emerging market economies, includingthe need for greater transparency in data and poli-cies, the need to develop local markets and, as ex-pressed by the Executive Board in March 2004, theneed to implement sound macroeconomic policiesconsistently.33

In making this observation, we are not implyingthat there was a set of policies for improving thefunctioning of the “supply-side” mechanism that theIMF should have been advocating or that there wereclear answers to what was an “appropriate” level ofvolatility. Clearly, there was no such consensus. Norare we suggesting that the IMF has done nothing toreduce the cyclicality of international capital move-ments on the supply side. In fact, the IMF has ad-dressed this issue from the standpoint of minimizingmoral hazard in investor behavior, by encouraginggreater exchange rate flexibility in recipient countriesand limiting access to IMF financing during a crisis.The proposal for a Sovereign Debt RestructuringMechanism (SDRM) and the encouragement of col-lective action clauses (CACs) can be considered notonly as the IMF’s search for a crisis resolution mea-sure but also as part of its efforts to minimize themoral hazard aspect of capital flow volatility on thesupply side. But it is worth noting that the IMF’s con-tributions to the broader debate on what, if anything,can be done by advanced countries to minimize thecyclicality of international capital flows (for exam-ple, through regulatory measures directed at institu-tional investors) have been much more limited.34

Gradual versus rapid liberalization

The documents prepared by the IMF staff in thecontext of multilateral surveillance during 1990–2003 consistently favored capital account liberaliza-

25

33In a recent note on the IMF’s medium-term strategy circu-lated to the Executive Board, the staff suggested that its researchprogram should include issues related to institutions and contrac-tual mechanisms that can help protect countries from externalvolatility. See Caballero (2003) for an example of a “hedging andinsurance instrument” to protect a country from volatility arisingfrom commodity price fluctuations. While these ideas are wel-come, the focus remains on the recipient countries.

34In this context, in July 2003, the IMF staff provided the BaselCommittee with comments on the proposed New Basel CapitalAccord (Basel II), noting that the use of ratings in setting capitalcharges could increase market volatility and procyclicality.

32From 1999 to 2000, the question of how to regulate highlyleveraged institutions received considerable attention in the offi-cial community, and a number of reports were prepared by vari-ous bodies, including the International Organization of SecuritiesCommissions (IOSCO) and the U.S. President’s Working Groupon Financial Markets. The IMF made its own contribution to thedebate through its work on the Financial Stability Forum’s Work-ing Group on Highly Leveraged Institutions. See Financial Stabil-ity Forum (2000).



tion. The staff in the early years emphasized effi-ciency gains and the need to attract foreign invest-ment as the predominant reasons for capital accountliberalization; the 1993 ICMR mentioned the stabiliz-ing role of capital account liberalization that wouldcome from a broader investor base; and in the discus-sion of the 1994 ICMR, Executive Directors stressedthe discipline effect of capital flow volatility onmacroeconomic and structural policies. In later years,the IMF staff also noted additional factors, such asthe need for better risk diversification, greater con-sumption smoothing, and an improvement in the do-mestic financial system.

In the early years, the staff generally favoredrapid liberalization on the grounds that capital con-trols were not effective. The staff was aware of theidea of sequencing, however. The October 1992WEO, for example, stated that countries with an “un-competitive” banking system should not liberalizethe capital account until domestic financial liberal-

ization was complete. Yet, the same report advocated“a comprehensive and rapid progress on all fronts.”Management and the Executive Board generallysupported these views. At the discussion of the May1994 WEO, the Managing Director observed that“[in] its surveillance under Article IV, the Fund wasmaking a great effort to convince countries of thebroader benefits of capital account convertibility.”Undoubtedly referring to a potential amendment ofthe Articles, he expressed hope that capital accountconvertibility would be part of the IMF’s “refreshedmandate.”

The Mexican crisis had impact on the thinking ofsome staff (and Board) members (see Box 2.2). Theidea of sequencing became more prominent in staffanalysis, although the evolution of the institution’sstance on capital account liberalization would con-tinue for some time, at least until the East Asian cri-sis. The WEO began more systematically to call forgradualism and sequencing in capital account liber-

26

Box 2.2. Mexico: Capital Account Liberalization and the Crisis of 1994–95

From the late 1980s to the early 1990s, Mexico lib-eralized its capital account as part of a larger programof economic stabilization and reform. In 1989, the au-thorities eliminated major restrictions on FDI, allowedforeign investors to purchase nonvoting shares in theMexican equity market, and allowed Mexican firms toissue stocks in foreign markets; at the end of 1990,they allowed nonresidents to purchase domestic gov-ernment bonds; in 1993, they took additional measuresto internationalize the stock market and to liberalizeFDI. Although the stabilization and reform processtook place under an IMF-supported program, the ini-tiatives for capital account liberalization came fromthe Mexican authorities themselves within the contextof negotiations for the prospective North AmericanFree Trade Agreement (which started in 1990). By thetime Mexico began to negotiate for OECD accession,it had already achieved almost complete capital ac-count convertibility.

As early as 1989, Mexico’s efforts to liberalize theforeign investment regime received a strong endorse-ment of the IMF Executive Board. On the other hand,potential vulnerabilities created by financial liberaliza-tion (1988–89) and what turned out to be an ill-timedprivatization of banks (1991–92), when prudential reg-ulation was weak, did not receive adequate attention inthe IMF’s assessment of Mexico’s capital account lib-eralization strategy. On the contrary, at a Board meet-ing in 1991, “Directors expressed satisfaction with theprogress achieved so far in the reprivatization of thecommercial banks and the strengthening of the domes-tic financial system.” When Mexico received large cap-ital inflows between 1989 and 1993, relatively limiteddiscussion took place between the IMF and the Mexi-can authorities on how to manage the surge in inflows.

IMF staff, however, did communicate to the authoritiesits concern over the large current account deficit fi-nanced by short-term capital flows and the rapid in-crease in external borrowing.1 Executive Directors ex-pressed similar concern over this period. At thediscussion of the 1993 Article IV consultation, for ex-ample, Directors expressed the hope that there wouldbe a shift in external financing toward a greater share ofdirect investment.

The Mexican crisis of 1994–95 had only an incre-mental impact on the IMF’s thinking of capital accountliberalization, though it certainly influenced the viewsof some individuals within the institution. A number ofinternal and external experts interviewed explained thisas reflecting the predominant view held at the time thatthe crisis had largely resulted from inconsistency be-tween a pegged exchange rate and the pursuit of discre-tionary monetary policy, and not necessarily fromwrong sequencing in capital account liberalization.Even so, some in the IMF did become aware of thedanger of rapid liberalization when the prudential su-pervision of banks was weak. In fact, in discussing aninternal assessment of IMF surveillance in Mexico inApril 1995, Executive Directors suggested that “man-agement should also invite the Executive Board and thestaff to engage more decisively in capital account sur-veillance and discussion—a domain where the cultureof the Fund must no doubt still evolve.”

1In August 1991, the authorities imposed liquidity require-ments on short-term external borrowing by commercialbanks, which were extended in April 1992 to cover all foreigncurrency liabilities at a uniform rate of 15 percent.



alization, citing macroeconomic stability and finan-cial sector soundness as preconditions, a positionsupported by several Executive Directors, includingsome representing major industrial countries. TheOctober 1995 WEO recommended that countries inthe early phases of convertibility should liberalizeforeign direct investment (FDI) and trade-relatedflows before short-term flows. The 1997 issuesstressed the need for gradualism in the context ofsufficient exchange rate flexibility, sound macroeco-nomic policies, and a strong banking sector. The Di-rector of RES stated at one of the Executive Boardmeetings that a flexible exchange rate system wasnot a prerequisite for capital account liberalizationbut stressed the importance of having appropriateprudential supervision of the financial sector prior tomoving to capital account liberalization.

By 1999, the evolution in the institution’s think-ing was almost complete, and the voices of cautionhad become more dominant. In view of the weak ev-idence provided by the WEO to link capital accountliberalization with economic growth, an industrialcountry Executive Director made a statement in thesecond WEO discussion of 2001 questioning thewisdom of undertaking costly and risky reforms toliberalize the capital account in the expectation of“modest and uncertain” benefits, while a developingcountry Director added that the prerequisites forcapital account liberalization did not exist in most ofthe developing countries.

Although the staff position became more cautious,it remained in favor of capital account liberalizationas a long-term goal. The October 1998 WEO, for ex-ample, stated that countries should pursue a well-sequenced and prudent capital account liberalizationinstead of “turning the clock back,” in view of theirneed for external resources and the gains to be madefrom portfolio diversification. The WEO then sug-gested that countries should sustain structural re-forms to reduce information asymmetries or marketfailures in order to reduce resource misallocationsand excessive capital flow volatility. Likewise, in2001, the WEO suggested that those countries withsignificantly open capital markets should strengthenand improve their institutions while others that werenot fully involved in global capital markets shouldpursue capital account liberalization as the ultimategoal, though at the pace of their own choosing.


IMF staff in its multilateral surveillance workconsistently argued for tight fiscal policy and greaterexchange rate flexibility as the best tools to deal withlarge capital inflows. At the same time, the staff ex-pressed doubt at the effectiveness of sterilization,given its quasi-fiscal costs and its impact on the level

of interest rates. However, neither the staff nor Exec-utive Directors had much to say about conventionalmacroeconomic tools of capital flow management.At the 1995 ICMR discussion, an Executive Directorrepresenting a major industrial country expressedconcern over the lack of “clear policy guidelines” onmanaging capital inflows “from a monetary policystandpoint.”

Throughout the period, both the staff and Execu-tive Directors said much more about the temporaryuse of capital controls. The staff expressed a gener-ally negative position on the use of capital controlsas a tool to manage capital flows. In the early years,the staff was heavily focused on the long-term costsof such a policy. As empirical evidence on the effectof Chile’s inflow controls (particularly in lengthen-ing maturities) emerged, however, the staff began totake a less hostile attitude in its multilateral surveil-lance work toward the temporary use of market-based controls (Box 2.3). It remained opposed tothe use of administrative controls, particularly oncapital outflows.

During the first two years of the 1990s, theICMRs drew a lesson from developments in Europe(including the Exchange Rate Mechanism (ERM)crisis) that stressed the undesirability of capital con-trols and the need to strengthen the supervisory andregulatory systems. The staff in the April 1993ICMR observed that “countries faced with massiveand unrelenting capital market pressures often con-clude that it is less costly over the long run either torealign the parity or to resort to temporary floatingthan to impose capital controls.” The staff seemed torecognize the short-term benefits of capital controls,but a senior staff member stated at the May 1993Board discussion that capital controls were like“killing the goose that laid the golden egg,” espe-cially for those highly indebted countries trying toattract foreign investment. The staff’s preferred pol-icy, as expressed in the 1993 ICMR, was to “foster agradual expansion of the investor base and providesufficient resilience in the face of transient shifts inthe availability and terms of international financing”through macroeconomic stabilization and structuralreform and by improving transparency and informa-tion disclosure. Until about 1994, Executive Direc-tors generally expressed broad agreement with theseviews of the staff, although a few Directors were ofthe view that temporary controls could be useful inmanaging large short-term inflows.

A subtle change was in the making in 1994.While the WEO continued to advocate greater ex-change rate flexibility and tight fiscal policy as away of dealing with large capital inflows, it alsotook a more accommodating stance on the tempo-rary tightening of restrictions. The turnaround wasmore evident after 1995. The August 1995 ICMR

27



suggested that a mix of several tools (such as tightfiscal policy, foreign exchange market intervention,and temporary prudential controls) should be care-fully specified, depending on a country’s circum-stances; it also stressed the need to strengthen thebanking system against financial risks and to limitforeign exchange exposure. When the report wasdiscussed, some Executive Directors expressed sup-port for the temporary use of capital controls.

In the WEO discussion of May 1997, the Directorof RES clarified the staff position: “staff was not en-dorsing the use of capital controls; and some of thecountries had resorted to controls that were undesir-able.” During a Board discussion in October 1997,the RES Director stated that “staff encouraged coun-tries that faced capital inflow problems to considerliberalizing capital outflows as part of their overallstabilization strategy.” Even after the East Asian cri-sis, the WEO continued to express the view that,while controls to limit short-term inflows could be

helpful in specific circumstances (as in the case ofIndia), controls entailed longer-term costs and thatopen capital markets yielded substantial benefits; itfurther argued that use of capital controls duringcrises was not very effective and could “distract thegovernments from their prime task of strengtheningthe financial and macroeconomic environment.”

In subsequent years, there was little discussion oftemporary use of capital controls in multilateral sur-veillance work, reflecting the sharp decline in globalcapital flows to emerging market countries. Whenthe subject was discussed, the staff’s position re-mained nuanced. For example, when discussingMalaysia’s administrative controls on capital out-flows, the October 1999 WEO emphasized theprogress the country had made in financial and cor-porate sector restructuring, strengthening the regula-tion and supervision of financial markets, and imple-menting corporate governance reforms. It thenconcluded: “Any negative impact of the capital con-

28

Box 2.3. Chile:The IMF’s Views on the Use of Market-Based Controls on Inflows

Responding to a surge in capital inflows, in 1991, theChilean authorities introduced a 20 percent unremuner-ated reserve requirement (URR) on all foreign loans,except for trade credits, while relaxing the minimumstay requirements for foreign investment. The URR wasa noninterest-bearing deposit in foreign currency to belodged with the central bank for a specified period oftime (one year from May 1992), in an amount propor-tional to the value of the capital inflow.1 As experiencewas gained, the authorities introduced modifications totighten the URR over time.2 With capital flows toemerging market economies abating, they reduced therate of the URR to 10 percent in June 1998 and furtherto zero percent in September 1998. In 2001, the authori-ties removed all remaining restrictions on the capital ac-count, including the URR (see Box 1.2 for a review ofthe effectiveness of the Chilean URR).

The IMF staff’s position on the URR changed overtime. Initially, it did not oppose the use of the URR.The briefing paper for the 1992 Article IV consultation,for example, argued that the URR, along with the re-cent revaluation of the Chilean peso, could help coun-

teract the effect of capital inflows on spending. Thestaff report for the 1994 Article IV consultation, in July,noted a “significant reduction in short-term private cap-ital inflows” as a result of the URR. Later in the year,however, the staff’s position turned more negative. InNovember 1994, the staff strongly advised the authori-ties against any measure to tighten the URR. The brief-ing paper for the 1995 Article IV consultation statedthat capital controls increased the cost of capital, en-tailed allocative inefficiency, and would be increasinglyevaded over time, and argued for an orderly relaxationof these controls, beginning with an elimination of theone-year stay requirement and prior authorization re-quirement for capital inflows. During the 1996 consul-tation, the staff reiterated its view that the URR had notbeen effective in reducing capital inflows despite thetightening and that it should be eliminated.

As empirical evidence began to emerge, however, theIMF’s position was somewhat moderated. Internal doc-uments during 1997–98 indicate that the staff wasaware of some positive effects of the URR, particularlyin lengthening the maturity of inflows, but it continuedto argue against relying excessively on the URR as asubstitute for greater fiscal discipline. The views of Ex-ecutive Directors expressed at successive Board meet-ings remained mixed throughout the period. In general,more and more Directors over time saw the virtues ofthe URR in lengthening the maturity of capital inflows,increasing the share of non-debt-creating flows, andthereby reducing the volatility of capital inflows. How-ever, there were always some Directors who arguedthat the effectiveness of the URR in discouraging short-term capital inflows tended to diminish over time andcautioned against reliance on such a measure.

1The deposit requirement could be met by the payment of anup-front fee, equal to the financial cost of the URR (initiallycalculated at the London interbank offered rate (LIBOR)).

2For example, in 1992, the authorities extended the cover-age to foreign currency deposits in commercial banks in Janu-ary, raised the rate of the URR to 30 percent in May (exceptfor direct external borrowing by firms), and applied this ratefor all transactions in August. Subsequently, they further ex-tended the coverage to secondary American Depository Receipts (ADRs) in July 1995 and to FDI of a potentiallyspeculative nature in May 1996.



trols may have been offset by the increase in confi-dence from the acceleration of structural reformsand by generally sound macroeconomic manage-ment.” The staff—and Executive Directors—had amore favorable view of market-based controls, butstill justified them as a temporary measure and underspecial circumstances; they considered sustainablemacroeconomic policies to be preferred.

Assessment

Despite the ambiguity left by the Articles ofAgreements about its role in capital account issues,the IMF in the early 1990s responded to the changinginternational environment, characterized by largecross-border capital movements, by paying greaterattention to issues related to the capital account. Theearly efforts tended to be an analysis of factors influ-encing international capital movements. From themid-1990s, however, the IMF through its staff analy-ses began to advocate capital account liberalization.Concurrent with the initiatives to amend the Articlesto give the IMF an explicit mandate for capital ac-count liberalization and jurisdiction on members’capital account transactions, the IMF expanded thescope of its operational work in the area, encouragingthe staff to give greater emphasis to capital accountissues in Article IV consultations and technical assis-tance and to promote capital account liberalizationmore actively. At the same time, a significant im-provement was made in the capacity of the IMF tocollect information on members’ regulations of capi-tal account transactions.

In the early 1990s, the IMF’s multilateral surveil-lance work generally considered any measure thatwould promote capital flows to developing countriesto be a favorable development. This was understand-able in the context of the period when a number ofdeveloping countries were just beginning to regainaccess to international capital markets; this was alsoa period in which many held the view that, with thebeginning of transition in former socialist economies,there was a shortage of global saving. As a conse-quence, during this period, the IMF staff paid com-paratively less attention to potential risks of capitalflow volatility. In more recent years, particularly afterthe East Asian crisis, the staff has paid greater atten-tion to various risk factors, including the linkage be-

tween industrial country policies and internationalcapital flows as well as the fundamental causes andimplications of their boom-and-bust cycles. Thefocus of the analysis, however, remained on whatemerging market countries should do to cope withthe volatility of capital flows, rather than what, ifanything, advanced economies could do to reduce thecyclicality of such capital movements.

From the beginning of the 1990s, the IMF’s man-agement, staff, and Executive Board were all awareof the potential risks of capital account liberalizationand never promoted capital account liberalization in-discriminately. They also recognized the need for asound financial system in order to minimize the risksand maximize the benefits. Such awareness, however,largely remained at the conceptual level and did nottranslate into practical advice on preconditions, pace,and sequencing until later in the 1990s. Lacking theoperational content, such talk of risks failed to miti-gate the impact of the clear voice emanating from theIMF advocating capital account liberalization. Belowthe surface, however, a subtle change was takingplace within the institution. As preliminary evidenceemerged on the apparent effectiveness of Chile’s cap-ital control in the mid-1990s, some in the IMF, in-cluding within management and the ExecutiveBoard, began to take a more accommodating stanceon the use of capital controls at least as a temporarymeasure to deal with large capital inflows.

In the event, the proposed amendment of the Arti-cles failed to materialize, leaving ambiguity aboutthe role of the IMF in capital account issues. In themeantime, something of a consensus has emergedwithin the IMF that emphasizes the need to meetcertain preconditions for capital account liberaliza-tion and hence the importance of carefully determin-ing appropriate pace and sequencing in light ofcountries’ institutional capacity (the so-called “inte-grated” approach). The use of temporary capital con-trols also has become a more respected practicewithin the IMF, at least under certain conditions.These shared views, however, are unofficial as theyhave not been explicitly endorsed by the ExecutiveBoard, and ambiguity remains about the IMF’s workin capital account issues. In the following chapter,we turn to the question of how this ambiguity hasbeen dealt with in the context of a specific countryand see if the lack of a formal policy has affected thequality and consistency of the IMF’s advice.

29


This chapter reviews the IMF’s advice (or viewsexpressed) to member countries on capital ac-

count issues, particularly capital account liberaliza-tion and managing capital flows. It first discusses therole of the IMF in capital account liberalization dur-ing 1990–2002 in the full sample of emerging mar-ket economies, paying particular attention to how itsviews, as expressed in country work, evolved overtime. It then shifts to the IMF’s advice on managingcapital flows in the context of bilateral surveillance,covering macroeconomic and structural policies todeal with large capital inflows and, in a separate sec-tion, the temporary use of capital controls to dealwith both inflows and outflows.

Capital Account Liberalization

Out of the 27 countries examined, the evaluationidentified 18 countries for which the IMF staff pro-vided advice or otherwise expressed a view on capi-tal account liberalization in staff reports during1990–2002 (Table 3.1).1 The table summarizes theevaluation team’s best judgments on whether or notthe staff’s view or advice made a reference to theneed for proper sequencing. For each country, whenthe staff report for a particular year makes no men-tion of sequencing, it is indicated by 1; when men-tion is made of sequencing, it is indicated by 2.

The nine countries not listed in the table includesix countries that either had a relatively open capitalaccount or liberalized the capital account at the be-ginning of 1990 (or before becoming members of theIMF): Estonia, Latvia, Lithuania, Malaysia, Mexico,and Venezuela. In most of these countries, the IMFhad endorsed the authorities’ overall capital accountliberalization strategies. In the case of Mexico andVenezuela, this endorsement was given in the contextof IMF-supported programs;2 in the Baltic states ofEstonia, Latvia, and Lithuania, the IMF endorsed theliberal regimes already in place when they joined theIMF in 1992 or in subsequent policy dialogue. Theabsence of Colombia and Malaysia from the list maybe conspicuous, as they substantially opened the cap-ital account in the early to mid-1990s. Surprisingly,no discussion of capital account liberalization can befound in staff reports for these countries (Malaysia’scapital controls are discussed in the section “Tempo-rary Use of Capital Controls” of this chapter).3

Some countries had a very open capital accountbut are still included in the table, when remaining re-strictions were a subject of discussion. For example,Thailand, which had a very liberal capital accountregime in 1990 with respect to inflows, retainedsome restrictions on outflows, and the IMF staff ex-pressed its views on the liberalization of outflowcontrols (as well as on capital inflow promotionmeasures). Likewise, Chile abolished or eased mostadministrative controls on inflows from the late1980s to 1991, so the IMF staff’s subsequent viewson Chile mostly concerned the liberalization of out-flow controls (Chile’s capital inflow controls are dis-cussed in the section “Temporary Use of CapitalControls” of this chapter).

Advice to Member Countries,1990–2002

CHAPTER

3

30

1It is not always straightforward to make this judgment. For ex-ample, in some cases reference may be made to the “liberaliza-tion of the trade and exchange system,” in which case one cannotbe sure if capital account liberalization is included. In other cases,liberalization of capital transactions can mean either an overallopening of the previously closed capital account or a removal oftemporary capital controls (the latter case is treated separatelylater in this chapter). More fundamentally, the absence of an ex-plicit reference to capital account liberalization in staff reportsmay not mean that the staff expressed no view to the authorities inpolicy dialogue. In the case of program countries, we assumedthat the staff favored liberalization if a capital account openingmeasure was included in the authorities’ policy statements even ifno mention was made in the staff reports. The selection of coun-tries in Table 3.1 represents the IEO’s best judgments.

2Mexico launched a major program of economic liberalizationin the late 1980s, which the IMF subsequently supported with anarrangement under the Extended Fund Facility; and Venezuelalifted most capital account restrictions in January 1990 under anIMF-supported program.

3For Colombia, the staff report for the 1992 Article IV consul-tation endorsed the liberalization measures taken by the authori-ties during the previous years.


Chapter 3 • Advice to Member Countries, 1990–2002

We first discuss the role of the IMF in capital ac-count liberalization in terms of program conditional-ity and technical assistance. We then divide the pe-riod 1990–2002 into three subperiods, (1) the early1990s, (2) following the Mexican crisis; and (3) fol-lowing the East Asian crisis. For each period, we dis-cuss how the IMF viewed capital account liberaliza-tion within the context of a specific country and seehow its views might have changed over time.

Program conditionality and technical assistance

Of the 18 countries to which the IMF staff pro-vided advice on capital account liberalization, 13countries had an IMF-supported program at one timeor another. In none of these countries did the IMF re-quire capital account liberalization as formal condi-tionality for the use of its resources, where formalconditionality is understood to include prior actions,performance criteria, or structural benchmarks. In ad-dition to the country documents for these countries,the evaluation also examined PDR’s comprehensivedatabase on conditionalities for all programs, whichconfirms the almost complete absence of formal con-ditionality on capital account liberalization. This isconsistent with the established interpretation of theArticles of Agreement, as given by the IMF’s Legal

Department (LEG), which states that the IMF, as acondition for the use of its resources, cannot “requirea member to remove controls on capital movements.”4

The Articles, however, are interpreted to allow theIMF to require as conditionality certain actions re-lated to the capital account that are relevant to themandate of the IMF, notably elimination of paymentarrears (which typically arise from capital controls)and imposition of limits on external borrowing(which may implicitly require capital controls).Moreover, programs may also support capital ac-count liberalization as part of country authorities’overall package of economic policies. In fact, a num-ber of IMF-supported programs with some of thesample countries included references to aspects ofcapital account liberalization in the letters of intent(LOIs) or accompanying policy memorandums. LOIsare statements of the authorities’ policy intention anddo not constitute conditionality that links compliancewith disbursements of funds.

31

4See Legal Department, “Capital Movements—Legal Aspectsof Fund Jurisdiction Under the Articles,” SM/97/32, Supplement3, February 21, 1997. Records on structural conditionality indi-cate, however, that the 2001 program with Lesotho included com-pletion of “the first stage of capital account liberalization” as astructural benchmark. The IEO cannot ascertain why, given thestandard LEG interpretation, this particular measure was deemedappropriate to be included as formal conditionality.

Table 3.1.The Nature of the IMF’s Advice on Capital Account Liberalization in Selected Countries1

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

Bulgaria . . . . . . . . . . . . . . . . . . . . . 1 2 2 2 . . . . . .Chile2 . . . 1 1 1 1 . . . 1 . . . . . . . . . 2 . . . . . .China . . . . . . . . . 1 . . . . . . 2 2 . . . 2 . . . . . . . . .Croatia . . . . . . . . . . . . . . . . . . . . . 1 . . . . . . . . . 2 . . .Czech Republic . . . . . . . . . 1 2 2 . . . . . . . . . . . . . . . . . . . . .Hungary . . . 1 . . . 1 . . . . . . 2 . . . 2 . . . . . . . . . . . .India . . . 2 . . . . . . . . . 2 2 2 2 . . . . . . . . . . . .Israel 2 1 . . . . . . 1 . . . . . . . . . . . . . . . 2 1 . . .Peru 1 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Philippines . . . 1 1 . . . . . . . . . . . . . . . 1 . . . . . . . . . . . .Poland . . . 1 . . . . . . . . . . . . . . . . . . 2 . . . 1 1 . . .Romania . . . . . . . . . 1 . . . 1 . . . . . . 2 . . . . . . . . . . . .Russia . . . . . . 1 . . . . . . . . . . . . . . . . . . 2 . . . . . . 2Slovak Republic . . . . . . . . . . . . . . . . . . . . . 2 . . . . . . . . . . . . . . .Slovenia . . . . . . . . . 1 . . . . . . . . . 2 2 . . . . . . . . . . . .South Africa3 . . . . . . 2 . . . . . . 2 2 2 . . . . . . 2 2 . . .Thailand4 . . . 1 1 . . . . . . . . . 2 . . . . . . . . . . . . . . . . . .Tunisia . . . . . . . . . . . . 2 2 1 1 2 2 2 2 2

Source: IEO judgments based on IMF staff reports, supplemented by internal documents where necessary.1The IMF’s advice for a particular year is assessed when capital account liberalization was part of that year’s discussion with country authorities. “1” indicates that no

explicit mention is made of sequencing, and “2” indicates that mention is made. A shaded area corresponds to a period in which there was an IMF-supported program.2 The advice refers mostly to the liberalization of outflows after 1991.3In the case of South Africa, in addition to the financial sector, documents also explicitly spelled out country-specific risks whenever capital account liberalization

was raised.4The advice refers mostly to the liberalization of outflows.


CHAPTER 3 • ADVICE TO MEMBER COUNTRIES, 1990–2002

For example, Hungary’s 1991 LOI for an ex-tended arrangement expressed the authorities’ com-mitment to promote FDI by encouraging foreignparticipation in the banking sector and in the privati-zation process. Likewise, Russia’s 1992 LOI for thefirst credit tranche included the authorities’ intentionto issue necessary foreign exchange regulations cov-ering both capital and current transactions and to ex-tend the same convertibility to nonresidents as soonas the monetary arrangements in the ruble area hadbeen settled. The LOI for Croatia’s 1997 request foran extended arrangement was more explicit in stat-ing the authorities’ intention to “put the liberaliza-tion of capital account transactions on its policyagenda, including restrictions that relate to outwardportfolio and direct investment.”5

For the most part, available evidence suggests thatthe process of capital account liberalization was de-termined by country authorities’ economic and polit-ical agendas. In some cases, the process was drivenby prospective OECD or EU accession and, in lateryears, commitments under bilateral or regional tradeagreements. In other cases, such as Latvia (Box 3.1),it was driven by the countries’ desire to attract for-eign investment. The IMF’s deference to countryownership is particularly evident in countries thatchose a gradualist approach to capital account liber-alization, such as Hungary or India; in these cases,we found no evidence that the IMF pushed for afaster pace in program negotiations or other policydiscussions.

The IMF also provided technical assistance on is-sues that are broadly related to capital account liberal-ization. The documents we examined for 15 countriesincluded advice on such issues as the development ofindirect monetary policy instruments, establishmentor development of a foreign exchange market, anddrafting of a foreign exchange law. A Board paperprepared by the staff in 1995 noted that, while “theIMF’s treatment of the issue of capital account con-vertibility [had] been on a case-by-case basis in thecontext of its surveillance and use of IMF resourcesactivities, an effort to facilitate capital liberalization[had] been applied more generally through themedium of technical assistance to develop foreign ex-change markets” (Quirk and others, 1995).

In technical assistance discussions, the staff in theearly years tended to be more encouraging towardcapital account liberalization. The back-to-office re-

port for a 1995 technical assistance mission to theCzech Republic, for example, stated that the mission“argued strongly” in favor of a “more decisive pro-gram of liberalization.” A 1995 technical assistancereport for China concluded that an effective way toenhance the efficiency of the foreign exchange mar-ket was to eliminate restrictions that prohibited for-eign banks from operating in the domestic market.Even in India, where the staff supported the country’sgradualist approach in its surveillance work, a 1995technical assistance report on foreign exchange mar-ket development noted the need for a broad strategyfor capital account liberalization as a precondition fordeveloping a dynamic market. Little advice wasgiven, however, on the specifics of sequencing capitalaccount liberalization until later in the 1990s.

In the late 1990s, the nature of the IMF’s technicalassistance seemed to change in two important re-spects. First, it began to emphasize the notion of se-quencing. For example, in a December 1997 seminarheld in China, the staff stressed the need to coordinatecapital account liberalization measures with concur-rent measures to strengthen financial markets and in-stitutions and to sequence properly the liberalizationmeasures in FDI, portfolio inflows, and outward in-vestments. Second, technical assistance also becamemore accommodating of use of capital controls, espe-cially for prudential purposes. In January 1999, forexample, the technical assistance mission to Russiaargued that countries that had opened up their capitalaccount typically had in place regulations restrictingcertain transactions between residents and nonresi-dents, and recommended that the ability of nonresi-dents to access credit in the domestic market be re-stricted, especially where such borrowing could beleveraged to speculate against the ruble.

Surveillance in the early 1990s

In the early 1990s, the IMF viewed capital ac-count liberalization favorably in its country work.The countries that liberalized the capital accountmay have done so on their own, but the IMF approv-ingly accepted their liberalization plans. In Israel,the IMF even advised the authorities to acceleratethe pace of liberalization. For example, the staff re-port for the 1991 Article IV consultation with Israelstated: “The removal of foreign exchange controlsshould be speeded up to encourage capital inflowsand improve allocation.” In Thailand, in 1992 thestaff “urged” the authorities to remove the remainingrestrictions on capital outflows.

The IMF staff, in its discussions with country au-thorities, stressed the benefits of an open capital ac-count, including greater resource flows to supplementdomestic savings (as in Poland in 1991), better re-source allocations (Israel in 1991), lower interest

32

5Korea’s economic program supported under the 1997 Stand-By Arrangement included three measures to liberalize the capitalaccount: (1) greater foreign participation in the domestic financialsector; (2) opening of the equity, money, and corporate bond mar-kets; and (3) elimination of foreign borrowing by corporations.These measures were also included, not as formal conditionality,but in the memorandum attached to the LOI.



rates (Slovenia in 1993), and greater price stability(Russia in the early 1990s). The 1992 staff report onRussia noted that the anti-inflationary policies cru-cially depended on accelerating the pace of institu-tional and structural reforms, including opening theeconomy to foreign capital. For Israel (in 1994), thestaff listed Israel’s strong financial market and stablecurrency as reasons for moving forward with capitalaccount liberalization. The staff occasionally pointedout the problems posed by a weak banking system,lack of exchange rate flexibility, or weak fiscal policy,

but the awareness of these problems did not translateinto operational advice on the pace and sequencing ofcapital account liberalization (for example, Poland in1991; and the Czech Republic in 1993).

The staff’s views of the offshore Bangkok Inter-national Banking Facility (BIBF) were typical. TheBIBF was established in March 1993, with the goalof developing Thailand into a regional financial cen-ter. Various tax incentives were granted to BIBFbanks, including a lower corporate income tax rateand several outright exemptions. The IMF staff took

33

Box 3.1. Latvia

In 1991, independence for Latvia meant an immedi-ate dismantling of all the capital controls that existed inthe former Soviet Union. Independence also meant areorientation of trade away from the Soviet bloc, whichbrought about a deterioration in the terms of trade and ashortage of energy and raw materials. In the early1990s, coupled with the impact of a severe drought af-fecting agriculture, the economic outlook for Latviawas bleak. From 1992 to 1993, industrial output col-lapsed and unemployment soared, threatening to under-mine the broad political support for economic reforms.Under these circumstances, the Latvian authoritieschose not to replace the Soviet-era capital controls withnew ones and, in November 1991, enacted a liberal for-eign investment law in order to establish a legal frame-work for attracting foreign direct and portfolio invest-ment. Effectively, capital account liberalization wascompleted overnight and preceded domestic financialliberalization, the establishment of a central bank, andthe introduction of a national currency.1

Although the IMF had no formal role to play inLatvia’s decision to open the capital account,2 it sup-ported the authorities’ outward-looking, market-oriented reforms through financing and technical assis-tance.3 The IMF’s position on Latvia’s capital accountpolicy is evident in program documents. For example,in August 1992, the LOI for a Stand-By Arrangementstated that “[the] current stipulations in the foreign in-vestment law allowing liberal repatriation of capital andtransfers of shares in the event of liquidation [would] bemaintained.” Differences of view between the IMF staffand the Latvian authorities concerned the use of ex-change rate policy to deal with capital inflows. WhenLatvia received large capital inflows, the staff’s consis-

tent view was to encourage greater exchange rate flexi-bility. A briefing paper of January 1995 was typical, byarguing that it would be better for any real appreciationof the lats to be achieved through a nominal apprecia-tion than through higher monetary growth. The authori-ties on their part viewed nominal exchange rate stabilityas critical to price stability in a small, highly open econ-omy such as Latvia, and insisted on maintaining the defacto peg to the SDR, which they had introduced inearly 1994.

Latvia experienced a major banking crisis in the firsthalf of 1995, which involved a closure of several promi-nent banks, including the largest one. This was in factthe culmination of a crisis that had been brewing forsome time. In 1994, the licenses of several commercialbanks were suspended. When several major banks failedto submit audited reports in March 1995, this turned intoa major crisis of confidence in the banking sector. InMay, the announcement by the government to take overthe largest bank (accounting for 30 percent of total de-posits) precipitated capital flight amounting to 10 per-cent of total foreign exchange reserves, exerting down-ward pressure on the exchange rate. Through decisiveintervention in the banking sector and demonstratedcommitment to the fixed exchange rate policy, however,the crisis was resolved relatively quickly and capital in-flows resumed in the latter part of the year. A number ofLatvian experts consulted by the evaluation team haveexpressed the view that the banking crisis of 1995 hadlittle to do with capital account liberalization per se butresulted from the weak regulatory system that had al-lowed too many banks to be established without the nec-essary internal control procedures or sufficient capitaland allowed them to adopt unsound lending practices.4Fortunately, Latvia’s financial and foreign exchangemarkets lacked the depth and breadth to allow major in-ternational players to speculate against the currency.1In May 1992, the authorities introduced the Latvian ruble

initially as a supplement to the Russian ruble (to deal with acash shortage), before the formal introduction of the nationalcurrency lats in March 1993.

2Latvia signed the IMF Articles of Agreement in May1992.

3From 1992 to 1995, MAE provided extensive technical as-sistance on foreign exchange operations, central bank opera-tions, and banking supervision.

4Some experts have even suggested that banks were en-gaged in activities bordering on criminal behavior. When thelargest bank was suspended, most of the assets had beenstripped, leaving a negative net worth of around 8 percent ofGDP.



little note of this, however, and made no commentson its implications for capital inflows in documentsprepared around this time. The staff report for the1993 Article IV consultation saw the BIBF, not interms of its implications for capital inflows, butmore in terms of the benefits it would provide in im-proving competition and technical skills in the finan-cial sector. Even after the composition of capital in-flows shifted to shorter-term maturities in 1994, thestaff paid little attention to the vulnerabilities beingcreated by the BIBF. For example, the briefing paperfor the 1994 Article IV consultation noted the staff’sintention to discuss with the authorities the progressthey had made in developing the BIBF and the mea-sures to be taken to “complete the process of finan-cial market deregulation.”

This does not mean that the IMF staff was un-aware of the risks of rapid liberalization, and thestaff seems to have explicitly supported a cautiouspace in some cases. When the Indian government re-quested a Stand-By Arrangement (SBA) in 1991, forexample, the staff supported the authorities’ inten-tion, as spelled out in the LOI, to liberalize the capi-tal account slowly by first removing restrictions onFDI and equity inflows. In South Africa, the IMFagreed with the cautious attitude of the authoritiestoward capital account liberalization, albeit for acountry-specific reason involving the climate of po-litical uncertainty. Despite these specific instances,however, our review of country documents indicatesthat the IMF generally supported rapid capital ac-count liberalization in the early 1990s and offeredlittle practical advice on pace and sequencing.

Following the Mexican crisis

The Mexican crisis did not seem to have a signifi-cant impact on the broader institutional thinking ofthe IMF on capital account liberalization. As notedin Chapter 2, the “staff operational note,” instructingarea department staff to encourage countries to liber-alize the capital account, was issued in December1995, after the Mexican crisis. This may reflect thefact that this crisis was largely seen as one that hadresulted from the country’s choice of an inappropri-ate exchange rate regime under the circumstances;correctly or incorrectly, it was not necessarily seento have resulted from rapid capital account liberal-ization. Nevertheless, internal documents suggestthat the Mexican crisis did cause the staff’s policydialogue with some countries to become more cau-tious, questioning the wisdom of rapid capital ac-count liberalization, particularly on outflows andshort-term reversible inflows.

From 1995 to 1997, for example, some countryteams argued that the liberalization of outflows orshort-term flows could trigger capital flight (as in the

Czech Republic in 1995), exacerbate unstable capi-tal inflows (in Hungary in 1996), or make it morelikely to reintroduce capital controls (in India in1996). The idea of preconditions, familiar in the aca-demic literature (and which was appearing in thestaff’s analytical work), also became more evident inthe IMF’s country work during this period. In thestaff report for the 1996 Article IV consultation withthe Slovak Republic (issued in 1997), the staff cau-tioned against further liberalization because of thecountry’s large current account deficit. The staff wasalso aware of banking sector weakness (as in Chinain 1997), lack of exchange rate flexibility (in Hun-gary in 1996), and underdeveloped domestic finan-cial markets (in the Slovak Republic in 1997), andurged the authorities to undertake reforms in theseareas. In India, the staff in 1996 specifically advisedagainst lifting restrictions on purchases of govern-ment bonds by nonresidents until fiscal consolida-tion was achieved.6 In South Africa, the staff basedits support for gradualism on vulnerabilities arisingfrom the central bank’s large net open position in theforward exchange market.7

Even in those cases where the staff did not provideoperational advice to slow down the pace of liberal-ization, it clearly gave greater attention to the poten-tial vulnerabilities associated with large capital in-flows and the need for a strong financial system. InThailand, for example, the briefing paper for the1995 Article IV consultation indicated the staff’s in-tention to discuss with the authorities the policy im-plications of the recent increase in short-term capitalinflows, underscoring the importance of close bank-ing supervision and the need to maintain adequatecapital ratios for financial institutions. The staff re-port for the 1996 Article IV consultation paid atten-tion to the currency and maturity mismatches beingcreated through the BIBF. It expressed concerns overthe “rapid growth of foreign currency lending to localfirms, which is only partly hedged, suggesting thatforeign exchange risk needs to be closely monitored.”

Of course, these instances of cautious attitude andsupport for gradualism in the IMF’s country workmust be placed in context. As noted, the Mexicancrisis did not have a major institutional impact on theoverall philosophical orientation of the IMF. The

34

6As early as 1995, IMF staff had stressed that financial sectorreform, fiscal discipline, and development of indirect monetarypolicy instruments would be necessary before capital accountconvertibility could be achieved in India. In 1997, these ideas ledto a draft note on the pace and sequencing of capital account lib-eralization, which the staff submitted to the Indian authorities.The note explicitly listed preconditions for capital account liber-alization, including fiscal consolidation, banking sector reform,and exchange rate flexibility.

7In 1995, South Africa lifted most controls on capital move-ments by nonresidents but retained restrictions for residents.



staff continued to encourage liberalization in suchcountries as Romania (1995) and Chile (1996). Anearlier IEO report (IEO, 2003) documents that, dur-ing this period, the IMF also encouraged the Koreanauthorities to remove remaining restrictions on long-term capital flows, with insufficient attention to se-quencing and supervision (see also Box 1.3). In thecase of Russia, the staff encouraged the opening ofthe public debt market to nonresidents, though ar-guably this was related more to the question of howto finance fiscal deficits than to that of capital ac-count liberalization per se. These sentiments wereshared by some on the Executive Board. When thestaff report for the 1997 Article IV consultation withIndia was discussed, some Executive Directors criti-cized the staff for its support of India’s gradualist ap-proach to capital account liberalization and arguedthat the benefits from further rapid liberalizationwould outweigh any risks.8

Following the East Asian crisis

The evolution of the IMF’s thinking on capitalaccount liberalization was a gradual process, and itis not correct to point out a single event as trigger-ing a fundamental shift. Even so, there is no deny-ing that the East Asian crisis had a major impact.From around 1998, one sees the IMF staff givinggreater attention than previously to the risk of rapidliberalization when preconditions were not met.During this period, expressions of caution towardcapital account liberalization or views accommo-dating of capital controls were no longer isolatedcases. In a wide range of cases, the staff now sawvirtue in limited capital account openness as ameans of protection from contagion (as in Sloveniain 1998, Romania in 1998, India in 1998, andChina in 1999) or as a way of providing breathingspace for necessary reforms (Russia in 2002). Thestaff report for the 1999 Article IV consultationwith China stated that the capital controls hadhelped the country reduce external vulnerability,while the staff report for the 1998 consultation withSlovenia noted that the country’s cautious approachto capital account liberalization had helped avoidserious contagion from the emerging market crisesof the recent years.

At the same time, the staff began to pay greaterattention to the sources of capital flow volatility, andthe need to address these as a precondition for fullcapital account liberalization. The staff in its countrywork suggested financial system weakness (in Bul-

garia in 1999, Croatia in 2001, and Russia in 2002)and the lack of market-determined interest rates(Slovenia in 1998) as among the factors contributingto capital flow volatility. The staff report for the2001 consultation with Russia (issued in 2002) ex-pressed support for cautious liberalization “in lightof the global uncertainties and potential risks frompremature liberalization with an underdeveloped do-mestic financial system.” For countries that re-quested advice on capital account liberalization, theIMF suggested that, as preconditions for liberaliza-tion, they should increase exchange rate flexibility(or introduce a floating exchange rate system withinflation targeting, if feasible), undertake bankingsector reform, and strengthen financial sector regula-tion possibly in the context of the FSAP. At the sametime, for those countries that had broadly met thepreconditions, the IMF staff at least implicitly en-dorsed their move toward full capital account liber-alization (as in the cases of Chile, Hungary, Israel,Poland, and South Africa).9

Macroeconomic and StructuralPolicies to Manage Capital Inflows

The 1990s saw a surge in capital flows to emerg-ing market economies. In part, this reflected global“push” factors. But at the same time, policies pursuedby these countries to liberalize the capital accountwere also an important “pull” factor contributing tothe pickup in capital inflows (see Calvo and others,1996). As countries experienced the large capital in-flows and associated macroeconomic challenges, thequestion of how to manage such inflows became aroutine subject of discussion between the IMF andcountry authorities. The evaluation found that thestaff expressed views on capital inflow issues in 16 ofthe 27 countries in the broader sample that experi-enced large capital inflows (see Table 3.2). This sec-tion reviews how, in response to the large capital in-flows, the IMF advised or otherwise expressed itsviews to the country authorities in these 16 countriesin terms of macroeconomic and structural policies.

Macroeconomic policies

There is a large literature that discusses policy op-tions available to countries desiring to manage largecapital inflows (for example, Goldstein, 1995). De-pending on the nature of the inflows, the optionsoften considered include sterilization through open

35

8All Directors recognized, however, that progress on domesticfinancial sector reform and fiscal consolidation would help re-duce these risks.

9The staff’s policy advice in South Africa throughout this periodwas to tailor the further capital account liberalization to thestrengthening of the balance of payments position, as reflected inthe elimination of the net open forward foreign exchange position.



market sales of domestic securities, increases in re-serve requirements, fiscal tightening, and greaternominal exchange rate flexibility. Further trade lib-eralization, removal of restrictions on capital out-flows, and tightening of controls on capital inflowsare also considered. The consensus, however, seemsto be that none of these policies is a panacea, as eachmay involve significant costs or otherwise bringabout other policy challenges.10 There is thus a diffi-

cult trade-off between the potential short-term costsof large capital inflows and the side effects of thepolicies to deal with them.

The 16 countries reviewed here used a combina-tion of various policy instruments at different timesto deal with the capital inflows, and the IMF pro-vided advice or expressed views on them. As earlyas 1993, the IMF staff (Schadler and others, 1993)stated that the conventional policy prescription in-cluded fiscal tightening and real appreciation(preferably through nominal appreciation) but ar-gued against sterilization (which would keep domes-tic interest rates high and therefore encourage more

36

Table 3.2.The IMF’s Advice on Managing Large Capital Inflows, 1990–20021

Policy Measures Advocated by the IMF__________________________________________________________________________________Greater Further Liberalization Tightening of Imposition

Tighten exchange rate trade of capital prudential of capitalYears2 fiscal policy flexibility Sterilization liberalization outflows regulation controls

Chile 1990–97 Yes Yes Yes3 Yes Yes

Colombia 1991–97 Yes Yes Yes

Czech Republic 1994–96 Yes Yes Yes1999–2001 Yes Yes Yes

Estonia 1996–97 Yes Yes4

2000–01 Yes Yes4 Yes

Hungary 1995–2000 Yes Yes Yes Yes

India 1994 Yes Yes Yes Yes Yes Yes2002 Yes Yes3 Yes

Latvia 1993–97 Yes Yes4

Malaysia 1991–96 Yes Yes Yes

Mexico 1990–93 Yes5 Yes3 Yes

Peru 1992–98 Yes Yes Yes3 Yes

Philippines 1994–97 Yes Yes

Poland 1995–98 Yes Yes

Russia6 1995, 1997 Yes Yes

Slovak Republic 1995–2001 Yes Yes

Slovenia 1995–97 Yes Yes Yes3 Yes

Thailand 1990–96 Yes Yes Yes Yes Yes Yes

Source: IEO judgments based on IMF staff reports.1 “Yes” indicates that IMF staff advised the policy measure concerned in one or more years.2 These years represent a period of substantial inflows.3No explicit advice was given, but the IMF staff did not oppose the use of sterilization by country authorities.4Tighter monetary policy.5Support given to the crawling peg.6In Russia, significant capital outflows were recorded in 1996.

10For example, sterilization has quasi-fiscal costs (as higher-yielding domestic securities are typically exchanged for lower-yielding industrial country securities) and may lose effectivenessas substitutability of assets increases; high reserve requirementsaffect the allocation of credit adversely by reducing financial inter-mediation; exchange rate flexibility may lead to a large real ex-change rate appreciation; elimination of restrictions on capital out-flows can send a positive signal to the markets, thus encouraging

further capital inflows; and fiscal policy lacks short-run flexibil-ity. Given the quasi-fiscal costs, moreover, there is a conflict be-tween use of sterilization and fiscal tightening.



inflows) and use of capital controls (which werejudged to be either ineffectual or distortionary).11

Broadly, a review of country experiences suggeststhat the IMF’s policy advice and views did not devi-ate much from this conventional wisdom. However,there were different emphases across time and acrosscountries.

Fiscal policy

The IMF staff’s most frequent advice was totighten fiscal policy, preferably through a cut in pub-lic expenditure. Fiscal tightening was expected to re-duce pressure on the exchange rate not only throughlower domestic absorption but also by limiting theincrease in the relative price of nontradables.12 Onthe other hand, fiscal tightening could promote capi-tal inflows by signaling the authorities’ commitmentto prudent macroeconomic management and therebycause the exchange rate to appreciate, especiallyover the medium term.

The advice to tighten fiscal policy was given con-stantly throughout the period of large capital inflowsin five countries: Chile, Colombia, Estonia, India,and Peru. The cases of Chile, Colombia, and Indiaare of particular interest because the IMF providedthis advice in connection with its advice on the useof capital controls. In Chile, the staff on some occa-sions discussed with the authorities the possibility ofrelaxing capital controls pari passu with fiscal re-straint. In Colombia, in 1995 and 1997, the staff sug-gested that fiscal tightening could create conditionsfor a relaxation of the controls. Likewise, in India,the staff argued that fiscal tightening, complementedby further trade and capital outflow liberalization,would make it unnecessary to resort to administra-tive controls.

In seven other cases (the Czech Republic, Hun-gary, Malaysia, Poland, the Slovak Republic, Slove-nia, and Thailand), the staff advised fiscal tighten-ing, but only occasionally or with apparently lessintensity. In Poland, for example, the advice wasgiven only in 1995.13 In Slovenia, it was offered as acontingency measure to be used in the event largecapital inflows reemerged. The fiscal policy adviceoffered in the Slovak Republic was similar: in 2001,the staff suggested that the authorities should “standready to rebalance the policy mix toward a tighter

fiscal stance if persistent upward pressure on the ex-change rate posed a challenge to macroeconomicmanagement.”

In most of these countries, when support for fiscaltightening was expressed, the policy was perceivedlargely as an auxiliary measure. For example, thestaff report for the 1994 Article IV consultation withMalaysia stated: “While fiscal policy can be ex-pected to make a contribution, attaining the neededdegree of restraint will ultimately fall largely on theside of monetary and exchange rate policy.” Like-wise, the staff report for the 1992 Article IV consul-tation with Thailand, after arguing that tighter mone-tary policy should be attempted through sterilization,suggested: “Sterilization would become increasinglydifficult to sustain over time, and fiscal policy wouldhave to be tightened.”

In the remaining countries (Latvia, Mexico, thePhilippines, and Russia), the IMF staff did not ad-vise the authorities to tighten fiscal policy in re-sponse to large capital inflows. This is not to say thatfiscal policy was not discussed in policy dialoguebetween the IMF staff and the authorities. To thecontrary, in all these countries, fiscal policy was amajor topic of discussion. However, discussion onfiscal policy took place as part of the authorities’overall macroeconomic strategy, and not necessarilyin terms of managing capital inflows. For example,the Philippines was already committed to achievingfiscal consolidation by 1997 through structural fiscalreforms under an IMF-supported program and, inthis context, there was probably little need to make aseparate case for further fiscal tightening.

Country-specific factors must have been an impor-tant part of the observed cross-country differences inthe intensity with which the IMF staff advised atightening of fiscal policy. The staff’s consistent ad-vice of fiscal tightening in Estonia, for example, maybe related to the fact that fiscal policy was the pri-mary macroeconomic policy instrument availableunder the currency board arrangement. On the otherhand, the less-intensive advice in Thailand may beexplained by the assessment, as expressed during the1994 Article IV consultation, that there was “littleroom for fiscal policy to contribute to restraining de-mand as the fiscal consolidation of the late 1980s in-volved sizable cuts in expenditure.” Fiscal tighteningwas apparently not advocated strongly in Latvia be-cause the staff viewed the capital inflows as resultingfrom increased confidence, and not from “high realinterest rates caused by the combination of lax fiscalpolicy and tight credit policy.”14

37

11Schadler and others (1993) reviewed experiences with surgesin capital inflows in Chile, Colombia, Egypt, Mexico, Spain, andThailand, focusing on what the policymakers did rather than howthe IMF provided advice or expressed views.

12This assumes that government consumption is more intensivein the use of nontradable goods.

13Fiscal tightening was advised in subsequent years, but not inthe context of managing capital inflows.

14Briefing paper for the first review under the SBA, January 16,1996.



38

Table 3.3. Fiscal and Other Macroeconomic Indicators in Selected Countries with Large Capital Inflows, 1990–2002(In percent of GDP; percent a year)

AverageAverage Initital Average Improvement Average Average

Net Private Fiscal Fiscal in Fiscal Primary Initial Average CurrentYears1 Capital Flows2 Balance Balance Position3 Balance Public Debt4 Inflation Balance

Countries for which fiscal tightening was advised with greater intensity

Chile 1990–94 7.4 3.2 –0.3 –3.5 1.3 42.8 17.5 –2.41995–97 7.4 2.1 –1.1 3.2 7.2 –3.5

Colombia5 1991–94 1.5 –0.1 –0.4 –0.3 1.4 n.a. 25.7 –0.41995–966 –1.3 –2.0 –1.8 –0.2 20.8 –4.91997 1.1 –3.2 –3.1 –1.1 18.5 –5.4

Estonia 1996–97 14.0 –1.8 0.1 1.9 0.6 7.5 17.1 –10.02000–01 7.8 –0.6 –0.1 0.5 0.2 2.8 4.9 –5.6

India 1994 3.2 –7.6 –7.6 n.a. –2.5 73.3 10.2 –0.52002 3.1 –9.7 –9.7 n.a. –3.4 80.8 4.3 1.4

Peru 1992–94 6.3 –3.9 –3.5 0.4 0.6 80.2 48.6 –6.11995–98 6.9 –1.3 2.6 1.1 9.6 –6.8

Countries for which fiscal tightening was advised with less intensity

Czech Republic 1994–96 10.1 –1.8 –1.3 0.5 –0.2 16.2 9.3 –3.61999–2001 6.9 –3.1 –3.1 0.1 –2.1 13.4 3.6 –4.2

Hungary 1995–976 2.8 –6.2 –4.6 1.5 4.7 84.3 23.4 –4.61998–2000 11.0 –3.8 2.3 3.2 60.6 11.4 –7.9

Malaysia7 1991–96 8.4 –0.9 1.3 2.1 n.a. 73.3 3.9 –6.4

Poland 1995–98 5.7 –2.8 –3.0 –0.2 0.5 50.8 18.6 –2.3

Slovak Republic 1995–98 10.1 0.3 –3.1 –3.4 –0.9 21.6 7.1 –6.71999–2001 5.8 –5.9 –6.3 –2.9 10.0 –5.6

Slovenia8 1995–97 2.3 0.0 –0.3 –0.3 0.9 17.8 10.6 0.1

Thailand 1990–93 10.6 4.4 3.4 –1.0 n.a. 18.4 4.8 –6.61994–96 10.5 2.3 –2.1 n.a. 5.6 –7.1

Countries for which noor little advice on fiscal tightening was given

Latvia 1993–97 8.6 0.6 –1.7 –2.3 –0.7 11.29 39.2 0.1

Mexico 1990–93 5.1 –2.8 –0.1 2.6 5.1 50.2 18.6 –5.0

Philippines 1994–97 9.8 –1.7 –1.1 0.6 n.a. 72.5 7.8 –4.2

Russia 1995, 1997 1.8 –6.5 –7.5 –1.0 –3.1 n.a. 106.4 0.4

Sources: IMF databases; and IEO estimates.1Unless noted otherwise, these years cover a period of large capital inflows.2Foreign direct investment, private portfolio investment flows, and other private investment flows.3Average balance less initial balance.4For most countries, gross debt of central or general governments; for Estonia, net debt; for Latvia, Peru, and Thailand, total debt of consolidated central govern-

ments (obtained from the IMF’s Government Finance Statistics database).5In calculating average net private capital flows, 1991 is excluded. Net capital inflows in 1993 amounted to 4.2 percent of GDP.6Net private capital outflows were recorded in 1996–97.7Net private capital outflows were recorded in 1994. For 1991–93, net private capital inflows amounted to 12.4 percent of GDP.8Net private capital outflows were recorded in 1996.9For 1994.



While we can try to explain the IMF’s policy ad-vice in some of these individual instances, a moresystematic assessment of its overall advice on fiscalpolicy is more difficult. When we assess fiscal policyadvice against the strength of capital inflows, infla-tion, and various fiscal indicators, we find little sys-tematic relationship: advice was given in some coun-tries but not in others; and the same advice wasgiven in countries with different fiscal positions(Table 3.3). Neither do we find any systematic rela-tionship between initial fiscal positions, changes inthe fiscal position, and the strength of IMF advice onfiscal tightening (Figure 3.1). One wonders, for ex-ample, why the IMF staff gave the same advice ofstrong fiscal adjustment to both Chile and Colombia,when Chile was already following a tight fiscal pol-icy and Colombia was not.15

Exchange rate policy

The IMF staff also frequently advised countriesreceiving large capital inflows to increase exchangerate flexibility. In an environment in which the insti-tution was moving (at the time) toward a “two-cor-ner” solution view of sustainable exchange rate pol-icy under high capital mobility, this advice appearsto be all tending toward one of the corners only. Itshould be remembered, however, that none of thecountries to which this advice was given had a dejure peg, let alone a currency board arrangement. Itwas in this context that the IMF advised the authori-ties to allow the exchange rate to appreciate in re-sponse to large capital inflows or to allow greaterscope for depreciation in order to discourage specu-lative inflows.

In 4 of the 16 countries (Hungary, India, thePhilippines, and Poland), this advice was given con-stantly throughout the period of large capital in-flows (see Table 3.2). In Poland, in 1997, the staffstressed that “allowing more flexibility within theexchange rate band would be an effective instru-ment to contain inflows, as increased uncertaintywould deter speculative movements.” A similar lineof argument was offered in the same year (as well asin other years) to Hungary: “From a medium-termperspective, the shift toward a more flexible ex-change rate regime would also help contain specula-tive capital inflows in the context of increased fi-nancial liberalization.”

In 9 countries (Chile, the Czech Republic, Latvia,Malaysia, Mexico, Peru, Russia, Slovenia, and Thai-land), the staff supported greater exchange rate flexi-bility, but with varying degrees of intensity overtime. In the Czech Republic, for example, the staff’sadvice was to increase the flexibility of the managedfloat when there was a surge in capital inflows in2000, but similar advice was not given during theearlier surge of 1994–96 when the exchange rate waspegged.16 Likewise, in Thailand, the staff stressedthe need for greater exchange rate flexibility in the

39

Figure 3.1. Fiscal Performance and the IMF’s Fiscal Policy Advice in Selected Countries1990–20021,2

Source: The IEO’s judgment based on Table 3.3.1Fiscal performance is classified into three broad categories: improvement,

little or no change, or deterioration. The abbreviations are as follows: CHL-2 = Chile (1995–97); COL-1 = Colombia (1991–94); COL-2 = Colombia (1995–96); CZE-2 = Czech Republic (1999–2001); EST-2 = Estonia (2000–01); HUN-2 = Hungary (1998–2000); LVA = Latvia (1993–97); MYS = Malaysia (1991–96); MEX = Mexico (1990–93); PER-1 = Peru (1992–94); PER-2 = (1995–98); PHL = Philippines (1994–97); POL = Poland (1995–98); SVK-2 = Slovak Republic (1999–2001); SVN = Slovenia(1995–97); THA-1 = Thailand (1990–93).

2“ ” indicates a country in which the advice of fiscal tightening was given with greater intensity; “ ” indicates a country in which the advice was given with less intensity; and “ ” indicates a country in which no or little advice of fiscal tightening was given.

–8 –6 –4 –2 0 2 4 6 8

Improvement

Little or no change

Cha

nge

in fi

scal

po

siti

on

Initial fiscal balance (In percent of GDP)

Deterioration

HUN-2 PER-2

MEX

PHLEST-2

MYS

COL-2 LVATHA-1

SVK-2CHL-2

PER-1

CZE-2

POL SVN

COL-1

15The difference in fiscal performance between Chile andColombia was much greater than suggested by Table 3.3. TheIMF staff in a recent paper characterizes Chile’s fiscal perfor-mance during the 1990s as by far the best among Latin Americancountries (Singh and others, 2005). Unlike most other LatinAmerican countries, Chile’s fiscal policy was not procyclical andits debt-to-GDP ratio steadily declined.

16In fact, the staff report for the 1995 Article IV consultationwith the Czech Republic stated that “the main aim of the policyresponse to the capital inflows should be to limit further real ap-preciation of the koruna, through avoidance of a nominal appreci-ation and a deceleration of price and wage increases.” A numberof Executive Directors, however, disagreed with the staff positionand called for a more flexible exchange arrangement and for anominal appreciation of the koruna.



mid-1990s (when the composition of inflows be-came more short term), but not in the early 1990s. InMexico, the staff did not explicitly offer advice, butsupported the country’s crawling peg regime (whileimplicitly expressing preference for a wide band orfaster crawl). In Latvia, the staff in principle sup-ported the country’s peg to the SDR as a nominal an-chor, but suggested greater flexibility if the inflowswere to become exceptionally large.17

In the rest of the countries (that is, Colombia, Esto-nia, and the Slovak Republic), the advice of greaterexchange rate flexibility was not offered. Clearly inthe case of Estonia, the country’s choice of a currencyboard arrangement argued against such advice. In theSlovak Republic, the staff during the 1997 Article IVconsultation discussed with the authorities the need to“consider making more active use of flexibility withinthe band” as a way of deterring contagion from theEast Asian crisis, but not as a way of managing largecapital inflows.

Several considerations seemed to influence theIMF’s policy advice on exchange flexibility in spe-cific instances. In some cases, competitiveness con-siderations were paramount. In Thailand, for exam-ple, in 1991–92 the staff believed that exchange rateappreciation was not advisable because of the largeand growing current account deficit. A similar reasonwas given for not recommending greater exchangerate flexibility in the Czech Republic during much ofthe 1990s. The staff report for the 1995 Article IVconsultation was typical in arguing that appreciationwas not warranted, “given the apparent slowdown inthe growth of exports in recent months, the sharp in-crease in imports, and the loss of most of the gains incompetitiveness from the initial depreciation.”

Another factor determining the staff’s advice wasits perception of the degree of actual exchange rateflexibility. In Peru, for example, the staff generallyconsidered the exchange rate arrangement to beflexible (though at times highly managed), so theadvice for greater flexibility, as offered more fre-quently after 1995, focused on how to manage theexisting policy. Likewise, in Chile, the staff in 1993apparently thought that substantial flexibility hadbeen introduced to the exchange rate in 1992 when

the Chilean authorities adopted a basket of curren-cies to determine the reference exchange rate andwidened the band. By 1994 or 1995, however, thestaff no longer considered Chile’s crawling peg tobe flexible enough. In 1995, it thus advised the au-thorities to abandon the exchange rate band in favorof a managed float and subsequently became muchmore emphatic about the need for greater exchangerate flexibility.

The staff’s advice for Malaysia changed from yearto year, as its perception of actual flexibility changed.In general, the staff considered Malaysia’s exchangerate policy to be flexible. Thus, it supported “the au-thorities’ intention to accept, if necessary, a moderateappreciation of the ringgit” (1991 Article IV consul-tation) or expressed caution against “substantial fur-ther upward movement” of the exchange rate “inview of the speed and extent of the recent apprecia-tion” (1992 Article IV consultation), depending onthe particular circumstance. But when the authoritiesbegan to intervene heavily to support the exchangerate in late 1993 despite the large capital inflows, thestaff noted during the 1994 Article IV consultation:“Movements in the exchange rate should reflect un-derlying market conditions. As developments in late1993 well illustrate, efforts to force an exchange ratepolicy that is not in line with underlying fundamen-tals are bound not to succeed for long and, ultimately,are more likely to complicate the conduct of mone-tary policy.” At this point, the staff strongly arguedfor greater exchange rate flexibility.

As in the case of fiscal policy, the review of thesecases does not support the criticism of a “one sizefits all” approach to the IMF’s policy advice. How-ever, although we can explain the IMF’s advice insome individual instances, a more systematic assess-ment of the factors underlying its overall advice onexchange rate flexibility is much more difficult. Fig-ure 3.2, which depicts the relationship between theintensity of policy advice on exchange rate flexibil-ity and the degree of actual flexibility, clearly showsthat actual flexibility was not a dominant determi-nant of IMF advice. At the same time, the figure alsosuggests that at least part of this diversity in policyadvice may be a reflection of the evolution of theIMF’s views on appropriate exchange rate policy foremerging market economies. It appears that the IMFstaff placed somewhat more emphasis on exchangerate flexibility in the latter part of the period thanhad been the case earlier.

Sterilization

The IMF staff expressed its views on sterilizationin all countries. In general, the staff took a stancethat accommodated the policy choices of the author-ities, but in two countries (Poland and the Slovak

40

17From early 1994, Latvia had a de facto peg to the SDR,which the authorities considered to be critical as a nominal an-chor. It appears that the IMF staff viewed the de facto nature ofthe SDR peg as providing room for flexibility. In reality, the au-thorities’ commitment to the peg was quite firm, although theywanted to maintain some uncertainty in the minds of market par-ticipants by not announcing it formally. A briefing paper of Octo-ber 1994 stated the mission’s intention to urge the authorities toreclassify the exchange rate regime from “independently float-ing” to “pegged to the SDR” or “other managed floating,” while amanagement comment in February 1995 asked why the Latvianauthorities did not want to “go to a currency board.”



Republic) the staff opposed the use of sterilization.In Estonia and Latvia, support was given to steriliza-tion in the broader sense of the word, in the form oftight monetary policy. For example, the staff reportfor Latvia’s Stand-By request in early 1995 stated:“The staff . . . argued for more active liquidity man-agement, preferably through the use of treasury bills,in order to encourage the development of a govern-ment securities market.” In many cases, the staff ad-vised tight fiscal policy or greater exchange rateflexibility to complement the use of sterilization. Inview of inflation risks, the staff advised sterilization(and tight monetary policy) in such countries asIndia (1994), the Philippines (1995), and Hungary(1996). In several countries, including Latvia, asnoted above, the staff encouraged the authorities todevelop secondary markets for government bonds inorder to increase the effectiveness of sterilization.

In the context of the policy advice offered, thestaff’s view differed from those of the Czech andThai authorities on the use of sterilization, but on

different sides of the argument. Both countries expe-rienced a surge in private capital inflows under apegged exchange rate system in 1995. In the CzechRepublic, the staff argued that there was room forsterilization,18 while the authorities were reluctant todo so aggressively because of its presumed costs andineffectiveness. In Thailand (where fiscal policy wastight), on the other hand, the staff favored greater ex-change rate flexibility, while the authorities viewedsterilization as superior.

Staff was aware of the limitations of sterilizationand accordingly qualified its advice by emphasizingits quasi-fiscal costs and the risk that such operationsmight increase the level of interest rates in a self-de-feating manner. In Thailand, for example, the staffnoted in 1994 that “maintaining the desired mone-tary stance may well require large-scale sterilization

41

Figure 3.2. IMF Support for Exchange Rate Flexibility in Selected Countries, 1990–2002

Source: The IEO’s judgment based on IMF documents.1As given in the context of managing large capital inflows.2Slovenia briefly adopted a managed float during the last two quarters of 1995.3According to MFD classification.

Mo

re in

tens

eL

ess

inte

nse

More flexible Actual exchangerate flexibility3

Intensity of IMF support for exchange rate flexibility1

Less flexible

Thailand, 1994–96 Slovenia, 1996–97 Chile, 1990–91, Peru, 1995–98 1994–97

Philippines, 1995–97 Hungary, 1995–2000

India, 1994 Poland, 1995–98

Latvia, 1994–97 Russia, 1995 Chile, 1992–93 Czech Republic, 2000–01

Philippines, 1994–95

Malaysia, 1991–96

Estonia, 1996–97, Thailand, 1990–93 Slovenia, 19952 Mexico, 1990–91 Mexico, 1991–93

Czech Republic, 2000–01 1999

Czech Republic, Russia, 1997 Colombia, 1991–97 Slovak Republic, 1994–96 1999–2001

Slovak Republic, Latvia, 1993 1995–98

Peru, 1992–94

18An accompanying selected background study showed that be-tween 35 percent and 65 percent of domestic monetary operationscould be offset by external capital inflows.



which could be very costly.” In Malaysia, in 1993,the staff argued: “Large-scale mopping-up opera-tions, in addition to their quasi-fiscal costs, may alsoinduce further inflows.” In the Slovak Republic, in2000, the staff noted that the authorities were “moresanguine” about the effectiveness of sterilized inter-vention to deal with short-term capital inflows butagreed that it would not be effective to deal with “thelarger, FDI-related inflows” being projected in con-nection with EU accession. Sterilization, however,remained a frequently used policy measure to dealwith large capital inflows well into the late 1990sand early 2000s, including in Slovenia (1997),19

Latvia (1998), the Czech Republic (1999), the Slo-vak Republic (2000), and India (2002).

Structural policies

Structural reform constitutes an important topicof policy dialogue between the IMF and membercountries. As such, the documents examined indicatethat a wide range of structural reform issues werediscussed between the IMF staff and country author-ities in all sample countries over the entire period.Of the many issues covered, some had obvious rele-vance to capital account liberalization, including pri-vatization (which may affect policies toward inwardforeign direct investment) and bank supervision(which has been increasingly recognized as a criticalprecondition for liberalization). However, the docu-ments identify only a limited number of structuralmeasures as discussed or proposed for the specificpurpose of managing large capital inflows. The moreimportant of these are noted below.

Further trade liberalization. Further trade liberal-ization, as a supporting policy to cope with the ef-fects of large capital inflows, was advised in at leastfour countries: Chile, Colombia, India, and Thai-land. In Chile, the advice offered in 1997 involved areduction in the uniform import tariff. In Colombia,the advice was to liberalize the trade regime to in-crease competitiveness, while the staff suggested inIndia that trade liberalization should be acceleratedthrough the elimination of import restrictions onconsumer goods.

Liberalization of capital outflows. Elimination ofrestrictions on capital outflows was suggested as acomplementary measure to offset the large capitalinflows in at least six countries: Chile, the Czech Re-public, Hungary, India, Slovenia, and Thailand. Inthe Czech Republic, the advice was given in 1994but, as noted earlier, was reversed in 1995 when the

staff recognized that such a measure was prematureas it would make the country more vulnerable tospeculative capital movements. The advice to Thai-land was given in 1992 and again in 1996 (this timeby suggesting that pension funds be allowed to in-vest in foreign assets). The advice was given toSlovenia in 1997 and to Hungary in 2000.

Tightening of prudential regulation. Prudentialregulation received much attention from the IMFstaff, but generally as part of broader surveillance ofthe financial sector. In four countries (India, Mexico,the Slovak Republic, and Thailand), however, theIMF staff advised a tightening of prudential regula-tions as a way of dealing with the large capital in-flows (the advice in two countries took the form ofendorsing what the authorities had already done). InMexico, in August 1991 the authorities imposed liq-uidity requirements on short-term external borrow-ing by banks, which a number of Executive Direc-tors welcomed; at the end of 1993, the staff exploredthe possibility of further tightening the rules on for-eign borrowing by commercial banks. In Thailand,the staff in 1992 recommended that the authoritiesrequire appropriate provisioning by commercialbanks for the guarantee they provided for foreignborrowing. In India, the authorities tightened pru-dential regulations on portfolio investment in 1994, ameasure the staff subsequently endorsed. In the Slo-vak Republic, on the other hand, the staff took theinitiative to suggest tightening foreign exposure reg-ulations for commercial banks during 1999–2000.

In some countries, additional measures were men-tioned. For example, in the Czech Republic, bankingsector reform (including privatization) was advised in1995 as a measure to reduce the high cost of interme-diation (hence the high level of interest rates thatwere attracting foreign capital inflows). In Colombia,accelerated labor market reform was suggested onseveral occasions as necessary to maintain competi-tiveness in the face of upward exchange rate pressurecreated by the large capital inflows. In Chile, struc-tural fiscal reform (in the form of a tax on nonrenew-able natural resources) was proposed in 1997 as away to moderate foreign investment flows in the min-ing sector. In India, the IMF staff in 1994 proposedthe elimination of a floor on bank lending rates inorder to reduce incentives for speculative capital in-flows by reducing the level of interest rates.

There may well be other instances where the IMFstaff or country authorities recognized structural re-forms as a component of the broader strategy to man-age large capital inflows (for example, through theireffect on efficiency and competitiveness). Financialmarket development, sometimes encouraged as part ofadvice on sterilization (see “Sterilization” above),may be one such instance. The difficulty in identify-ing these cases, however, is that, as structural reform

42

19At the time of the 1995 Article IV consultation, the Slovenianauthorities stated that the costs of sterilization had reached 1 per-cent of GDP in late 1994 and therefore suspended “proactive ster-ilization.” Large-scale intervention resumed in late 1996.



is usually set in a long-term, country-specific frame-work, it is often not mentioned as part of a response tolarge capital inflows even when a particular structuralmeasure could serve such a purpose. Nevertheless, itis probably fair to say that structural policies receivedmuch less attention than conventional macroeconomicpolicies, as measures to deal with large capital in-flows. Given the IMF’s focus and comparative advan-tage, this was probably appropriate, especially sincethe structural area that has received considerable, andincreasing, attention from the IMF—strengthening theregulatory framework for the financial sector—ismost closely linked to capital account sequencing.

Imposition of capital controls

Rather surprisingly, the IMF staff in the latter partof the 1990s advised the authorities of two countries(Estonia and Peru) to impose capital controls to dealwith speculative capital inflows.20 In neither of thesecountries, however, did the authorities accept the IMFadvice. In Estonia, the staff stated in 2000 that the au-thorities’ ability to offset the impact of capital inflowswas limited unless they were willing to introducetemporary capital controls or raise reserve require-ments, which were already very high. Previously, in1996 and 1997, the authorities had been moreamenable to introducing such a capital control mea-sure if necessary (and a system of reserve require-ments on foreign interbank liabilities was introducedin mid-1997). In 2000, however, they objected to sucha measure because it was precluded by an agreementwith the EU. In Peru, the staff was concerned with therapid credit expansion, which was increasinglyfunded by short-term foreign currency liabilities. In1997–98, the staff proposed to the Peruvian authori-ties that the coverage of marginal reserve require-ments on foreign currency deposits be extended to ex-ternal borrowing, while lowering the rate.21

Temporary Use of Capital Controls

The evaluation reviewed the relevant IMF docu-ments for 10 countries involving 12 cases of tempo-rary controls on either capital inflows or outflowsduring 1990–2002, with a view to trying to judge

whether the differences in approach, if any, reflecteda consistent set of underlying principles (Table3.4).22 Of the 12 cases reviewed here, 8 concernedcapital inflow controls, and the other 4 capital out-flow controls. We review below each of these cases.

Inflow controls

Of the eight cases of capital inflow controls, theIMF staff did not initially object to the introduction ofcontrols in five of them (Chile, Colombia, the Philip-pines, Slovenia, and Thailand); however, in three ofthese (Chile, Colombia, and the Philippines), the staffchanged its views over time. In the remaining threecountries (the Czech Republic, Hungary, andMalaysia), the staff initially opposed the introductionof controls, though it became more accepting of themover time in two cases (i.e., the Czech Republic andHungary). We explain below the context in which theIMF staff expressed its views on capital inflow con-trols by dividing the countries into those for whichthe staff’s initial reaction was positive and those forwhich it was negative.

Cases of positive initial staff response

The IMF staff initially supported the unremuner-ated reserve requirement (URR) introduced by Chilein 1991 and Colombia in 1993, as we discuss morefully elsewhere in the report (see Box 2.3 for Chile;Appendix 1, the section “Colombia: Market-BasedControls on Inflows”). Likewise, the staff was sup-portive of inflow controls introduced by the Philip-pines, Slovenia, and Thailand.

In January 2000, the Philippine authorities intro-duced a 90-day minimum holding period on nonresi-dent deposits. At the time of a program review inJuly 2000, the staff considered this to be a relativelyminor measure designed to limit speculative capitalflows and it supported the authorities’ intention notto introduce additional controls. In late 2000, whenthe authorities tightened the reporting and compli-ance rules for capital account transactions,23 the staffargued that the benefits of these measures needed tobe weighted against the costs, while acknowledging

43

20The control measures suggested by the IMF could also be in-terpreted as having a prudential element. However, these are “re-strictions” as the term is generally applied in the IMF, becausetheir impact discriminates against international (as opposed to do-mestic) transactions.

21In 1993—based on an earlier technical assistance (TA) reportby MAE—a briefing paper had indicated the staff’s intention toexplore ways of tightening credit policy through an increase inmarginal reserve requirements on U.S. dollar deposits (then at 50percent). In 1995, the staff took a position against capital controlsexcept as a last resort measure.

22The list of countries in Table 3.4 is not meant to be exhaus-tive. For example, according to a Board document, 29 countrieseither introduced new controls or tightened existing ones during1993–97 alone. See Monetary and Exchange Affairs Department,“Developments and Issues in the International Exchange and Pay-ments System,” SM/98/172, July 7, 1998.

23These measures included the prohibition of “engineeredswaps” (operations to exploit weaker regulation and lower taxa-tion for foreign exchange deposits relative to peso deposits),stricter reporting requirements for all foreign exchange transac-tions, and the requirement that bank-affiliated foreign exchangecorporations be subject to the same rules as the banks themselves.



that these measures would improve monitoring andhelp close the regulatory loopholes. At the sametime, the staff expressed regret that these measureshad resulted in less freedom of capital movements,and welcomed the authorities’ assurance that theywere not the start of a tighter regime of controls.

In Slovenia, in 1995, the authorities introduced anunremunerated reserve requirement on non-trade-related loans with a maturity of up to five years. Thismeasure was tightened during 1996 and remained ineffect until September 1999. The staff initially en-dorsed the measure as a prudent response to the un-certainties of capital flows, but very little was said inthe subsequent staff reports.24 Slovenia had a reason-ably sound fiscal policy (with virtual balance in1994–95 and a deficit of only 1 percent of GDP in1997) and, in 1994, active sterilization had been ter-minated because of its huge fiscal costs. The staff’stacit approval of the capital control measure can beunderstood in this context.

Thailand introduced market-based control mea-sures in 1995 and 1996.25 The briefing paper for the1996 Article IV consultation indicates that the staffviewed the use of temporary controls on short-term

instruments as ineffective, but did not “rule them outex ante provided they were used as a complement to,not a substitute for, the macroeconomic measuresproposed.” The staff report for the 1996 Article IVconsultation stated the mission’s view that thesemeasures would likely have some impact on foreignborrowing by banks, but they would not be expectedto have impact on nonbank sources of credit, orhence on the overall level of inflows. The ExecutiveBoard supported the staff’s view.

Cases of negative initial staff response

As noted in greater detail in Appendix 1, whencapital control issues became a topic of discussion inearly 1995 in the Czech Republic, the staff objected touse of capital controls by saying that experience inmany countries had proved them to be ineffective. Inmid-1995, however, the staff became more sympa-thetic to the authorities’ use of capital controls, if fi-nancial stability was threatened or if use of other pol-icy instruments was prevented by political constraints.This support for the capital controls was expressedwhile, as noted, the staff was not in favor of greaterexchange rate flexibility.

When the Hungarian authorities considered in-troducing capital controls in 1996, the staff ex-pressed the view that the controls as proposedwould easily be circumvented. When a reserve re-quirement on nonresident bank deposits was intro-duced in early 1999,26 however, the staff was sup-

44

Table 3.4.The IMF’s Position on Temporary Use of Capital Controls,1991–2001

Position When Controls Did thePeriod Were Introduced Position Change?

Controls on inflows

Chile 1991–98 Supportive YesColombia 1993–2000 Supportive YesMalaysia 1994 Not supportive NoCzech Republic 1995 Not supportive YesSlovenia 1995–99 Supportive NoThailand 1995–97 Supportive NoHungary 1996, 1999 Not supportive YesPhilippines 2000 Supportive Yes

Controls on outflows

Venezuela 1994–96 Not supportive NoThailand 1997–98 Partly supportive YesMalaysia 1997–2001 Not supportive YesRussia 1998–2000 Partly supportive No

Source: IEO judgments based on IMF documents.

24In 1998, the staff expressed views sympathetic to the tempo-rary use of capital controls on short-term banking flows as a wayof preserving domestic monetary policy autonomy in the transi-tion period to participation in the European Economic and Mone-tary Union (EMU), provided that they were market based.

25In 1995, a 7 percent reserve requirement was introduced onnonresident baht accounts with a maturity of less than one yearand on finance companies’ short-term foreign borrowing. In1996, the 7 percent requirement was extended to short-term bahtborrowing by nonresidents and short-term foreign borrowing bycommercial and BIBF banks.

26The reserve requirement was to be set below market rate. In theevent, the rate was set at zero percent, and was never activated.



portive, saying that it could help the country re-spond to a possible resumption of large capital in-flows. Unlike the case of the Czech Republic, how-ever, the staff throughout the period was strongly infavor of greater exchange rate flexibility.

In Malaysia, a set of administrative controls wasimposed in early 1994 on short-term capital in-flows. These included prohibiting residents fromselling short-term monetary instruments to nonresi-dents and a ban on forward transactions (on the bidside) and non-trade-related swaps by commercialbanks with foreign customers.27 The briefing paperfor the 1994 Article IV consultation stated that theauthorities should phase out these controls andallow the ringgit to appreciate. In 1995, the staffwelcomed the authorities’ decision to lift most ofthese controls.

Outflow controls

With respect to the use of outflow controls, theevaluation reviewed the related IMF documents forfour countries that introduced them during 1990–2002: Venezuela (1994), Thailand (1997), Malaysia(1997), and Russia (1998). The documents suggestthat the IMF’s position differed across countries andover time (Table 3.4, bottom panel).

As explained more fully in Appendix 1, in June1994, Venezuela introduced comprehensive priceand exchange controls, covering current transactionsand capital outflows. From the beginning, the staffconsistently argued for an elimination of all ex-change controls, saying that the “complete elimina-tion of controls need not lead to a renewal of capitaloutflows” if supported by tight macroeconomic poli-cies.28 In March 1996, in view of the weak bankingsystem, the staff proposed “a two-stage approach”involving the immediate elimination of exchangecontrols on current and nonportfolio capital transac-tions, accompanied by a public commitment to liber-alize portfolio investment gradually.29

When Thailand introduced capital controls onoutflows in May–June 1997 (see Box 3.2 for de-tails),30 the staff argued that it was essential to usethe capital controls as a breathing space in which toimplement a comprehensive macroeconomic policy

package but, following the outbreak of an all-out cri-sis in July 1997, it argued for their immediate elimi-nation. The staff’s view remained negative when theThai authorities attempted to tighten the existingcontrols in October 1999.31 When the authorities in-tensified the enforcement of the existing controls in2000,32 the staff argued that the controls had beenmore effective in reducing offshore baht liquiditythan in limiting exchange rate depreciation and thatany tightening of the controls would likely impedetrade and hedging activities.

The Malaysian authorities introduced capitalcontrol measures twice in connection with the EastAsian crisis of 1997–98, first in August 1997 andthen in September 1998 (see Box 3.2 for details onthe second set of measures). When the first set ofmeasures was introduced,33 the mission expressedits strong disagreement and recommended that theybe removed as soon as possible. A briefing paper ofJanuary 1998 expressed the staff’s intention to urgethe authorities to remove the swap limits imposed inthe previous summer. When the second set of mea-sures was introduced in September 1998, the IMFstaff initially advocated a more flexible exchangerate and the simultaneous removal of most of thecapital controls that had just been introduced. Whenthe administrative controls were replaced by a sys-tem of exit levy for portfolio investment in February1999, the staff recommended prudential risk-basedmanagement of cross-border transactions whilecontinuing to favor the immediate removal of theexit levy.

The Russian authorities introduced capital outflowcontrols in August 1998, while simultaneously de-faulting on domestic-currency-denominated govern-ment bonds. The controls included the suspension ofconversion operations through nonresident accountsfor those who had participated in the restructuring ofgovernment bonds,34 and repatriation of ruble bal-ances by other nonresidents—these measures effec-tively suspended capital transfers abroad by nonresi-dents (though some were current transfers as defined

45

27In August 1994, the authorities eliminated most of these con-trol measures.

28Briefing paper for the 1994 Article IV consultation, October28, 1994.

29In mid-April, the authorities abolished all exchange controls,on both current and capital transactions, in the context of an IMF-supported program.

30Most of the exchange restrictions and capital control mea-sures introduced were relaxed in January 1998.

31This involved a clarification that the existing nonresident bor-rowing limit of 50 million baht for each customer (with no under-lying real transaction) applied on a consolidated basis to sub-sidiaries as well.

32Stricter reporting requirements were applied to banks, andfines were imposed on banks found violating baht lending limitsto nonresidents without an underlying real transaction.

33Limits were imposed on ringgit offer-side swap transactions(those that are not related to trade) by banks with nonresidents,except for hedging for trade-related transactions and genuineportfolio and FDI flows.

34A noninterest-bearing transit account was created for theruble proceeds from government bond transactions.



46

Box 3.2. Outflow Controls in Thailand (1997) and Malaysia (1998)

ThailandIn May–June 1997, the Thai authorities introduced

temporary selective capital controls on outflows byprohibiting the lending and sale of baht to nonresidentsand requiring that any purchase before maturity ofbaht-denominated securities, or purchase of equities,from nonresidents be made in foreign currency. Themeasures were aimed at limiting the speculation ofnonresidents against the baht by delinking the onshoreand offshore markets.1

The IMF staff initially argued that these measurescould not substitute for more fundamental adjustmentmeasures and suggested that the breathing space pro-vided by the control measures should be used to imple-ment further fiscal consolidation, greater exchange rateflexibility, and financial sector reforms—a view en-dorsed by the Executive Board. Executive Directors,however, cautioned against use of permanent controlsbecause they would risk undermining the confidence oflong-term investors. A staff memorandum of June 1997noted that, although the new measures had been effec-tive in the short run, they had done “long-term damage”and led to a reduction in capital inflows.

Following the floating of the baht in July, the mis-sion took the position that the capital controls shouldbe eliminated as early as confidence was restored. TheAugust 1997 LOI for an SBA indicated the authorities’commitment to eliminate the restrictions on purchasesand sales of baht by nonresidents as well as sales ofdebt securities and equities for baht, once the currencywas stable. Over the subsequent months, the IMF staffargued for an early elimination of the restrictions forbaht sales to nonresidents (first review) and suggestedthat the unification of the onshore and offshore ex-change markets remained a priority (second review inJanuary 1998). In the event, at the end of January 1998,the authorities relaxed most of the control measures in-troduced in 1997.2

Malaysia In September 1998, the Malaysian authorities intro-

duced temporary capital outflow controls, while peg-ging the exchange rate to the U.S. dollar.3 The measureswere designed to restore a degree of monetary indepen-dence and included, among others, the introduction of aone-year holding period for the repatriation of portfolioinvestment. In February 1999, the one-year holding pe-

riod was replaced by a graduated system of exit levy (inwhich principal and profits were allowed to be repatri-ated by paying an exit tax, the amount of which was de-termined by the duration of the investment). In Septem-ber 1999, the exit levy was abolished, except for profitsfrom portfolio investment brought in after February1999. The system was entirely abolished in 2001.

According to internal documents, the IMF staff ini-tially believed that an orderly exit strategy should beconsidered as soon as possible because of ample evi-dence that outflow controls did not work. The staff ad-vocated a move to a nominal exchange rate band ac-companied by inflation targeting, but was not opposedto maintaining the controls on short-term inflows thathad been in place prior to September 1998 during theprocess of financial and corporate restructuring. It evensuggested considering a market-based “prudential”measure on capital inflows of the Chilean type. The fallof 1998 also coincided with the end of the active dia-logue of previous months between the IMF and theMalaysian authorities on policy issues, including IMFtechnical support. Internal documents, however, aresurprisingly silent on whether the disagreement overthe measures taken in September had any role to play inthe process.

In 1999, however, the staff became more accommo-dating as it recognized that the controls had been ad-ministered effectively and that Malaysia had usedwisely the breathing space created. The mission recom-mended a shift to the prudential risk-based manage-ment of cross-border capital transactions, while advo-cating an immediate removal of the exit levy on profitsfrom portfolio investment. The staff supported the au-thorities’ intention to use the capital controls as a tem-porary measure. Executive Directors commended theauthorities for wisely using the breathing space pro-vided by the capital controls. On the use of the controlsitself, however, the Executive Board was divided: anumber of Directors supported the maintenance of cap-ital controls in order to manage an orderly exit, whileothers urged an immediate removal of the exit levy onprofits. In 2001, the elimination of the system was wel-comed by both the staff and the Executive Board.

The assessment of the effectiveness of Malaysia’scapital outflow control is made difficult by the fact thatit was introduced in September 1998, after the conta-gion from elsewhere in Asia had worked its waythrough the region. The Malaysian experience has re-ceived both positive and negative academic assessment,depending on whether one thinks that, in the fall of1998, Malaysia still faced a significant risk of furthercapital flight (Kaplan and Rodrik, 2001; Johnson andMitton, 2003). If the capital controls indeed worked inMalaysia as intended, it may be due to the controls’strictly temporary nature, the supporting policies (in-cluding measures to strengthen the banking and corpo-rate sectors), Malaysia’s institutional capacity, and agenerally favorable external environment (see Abdelaland Alfaro, 2003).

1According to some studies, these measures had temporaryeffectiveness in delinking the two markets (Edison and Rein-hart, 2001).

2The prohibition on baht lending to nonresidents was re-placed by a 50 million baht limit per counterparty without anunderlying current or capital account transaction.

3The staff report for the 1999 Article consultation notedthat, in the view of LEG and MAE, the newly introduced re-strictions (including the recent modifications) were not in vio-lation of obligations under Article VIII.



by Article XXX).35 At the same time, the authoritiesrequested technical assistance from the IMF to helpdevelop a more effective and less intrusive mecha-nism for preventing illegitimate capital outflows. “Inview of the need to stem pressures on reserves,” thestaff did not oppose the maintenance of these restric-tions, but refrained from formally recommendingtheir approval in the absence of “a time-bound plan”for removing “all remaining restrictions.”

Assessment

In this concluding section, we attempt a brief as-sessment of how the IMF’s policy advice measuredup to some of the “tests” set out in Chapter 1,namely: (1) Was there any difference between theIMF’s general policy pronouncements and the ad-vice it gave to individual countries? (2) Was the IMFpolicy advice operational and based on solid evi-dence? (3) How did the IMF advice change overtime, and did this change keep pace with availableevidence? (4) Did the IMF give similar advice tocountries in similar situations? and (5) Was the pol-icy advice on the capital account set in a broader as-sessment of the authorities’ macroeconomic policiesand institutional framework? Given the absence ofconsensus in the academic and official policymakingcommunities, however, this assessment is not aboutwhether the IMF’s particular policy advice was rightor wrong; rather, it is meant only as a broad charac-terization of the IMF’s overall advice on capital ac-count issues.


In all the countries that liberalized the capital ac-count during the 1990s, whether partially or almostfully, the process was for the most part driven by thecountry authorities’ own economic and politicalagendas. When programs were involved, the IMFnever required capital account liberalization as for-mal conditionality for the use of its resources. TheIMF, however, did not hesitate to support capital ac-count liberalization as part of the authorities’ overallpolicy package as expressed in program documents.Most of the advice on capital account liberalizationwas given to member countries through policy dia-

logue and technical assistance (as broadly defined)in the context of surveillance or financial support.Against this broad background, the IMF’s approachtoward capital account liberalization evolved overtime.

In the early 1990s, the IMF encouraged membercountries to liberalize the capital account withoutnecessarily raising the issue of pace and sequencing.This attitude was particularly evident in those coun-tries that had program relationships with the IMF. Itwas not uncommon during this period for the LOIs toinclude the authorities’ intention either to further lib-eralize capital account transactions or to maintain theopening they had achieved. Occasionally, the staffexpressed concern over financial sector weakness ormacroeconomic instability, but this did not translateinto operational advice on pace and sequencing.From around 1994, and certainly after the Mexicancrisis, some within the IMF took a more cautious atti-tude toward capital account liberalization. Discussionof sequencing and the need to meet preconditions be-came more systematic and routine. Although thechange was gradual, there is no denying that the EastAsian crisis had a profound impact on this process.

This evolutionary process is reflected in the ap-parent inconsistency observed in the IMF’s adviceon capital account liberalization, particularly duringthe early years: from the very beginning, sequencingwas mentioned in some countries but not in others(see Table 3.1). This is not especially surprising be-cause there was no official policy. Consequently, thecontent of policy advice in country work was largelydetermined by individual staff members as they as-sessed the situation and as new evidence emerged.To the extent that country circumstances differ, uni-formity cannot be the only criterion to judge thequality of policy advice and our evidence suggestsno “one size fits all” approach. But in this case, it ap-pears that the IMF’s policy advice lacked not onlyuniformity but also consistency because, in the ab-sence of clear official guidelines, it was almost en-tirely left up to the discretion of individual countryteams and the resulting differences in approach be-tween countries are sometimes difficult to explain.

Managing capital flows

The staff’s policy advice to specific countries onmanaging capital flows was largely in line with thepolicy conclusions typically derived from the schol-arly literature on open economy macroeconomics.To deal with large capital inflows, this advice advo-cated tight fiscal policy and greater exchange rateflexibility, but discouraged the use of capital con-trols. The staff’s position on sterilization, consistentwith the conventional wisdom, emphasized its quasi-fiscal costs and its longer-term ineffectiveness but

47

35Article XXX (d) includes “payments of moderate amount foramortization of loans” (in addition to “payments due as intereston loans”) as current transactions. By appealing to this provision,in a memorandum to management dated August 18, 1998, theLegal Department took the position that these restrictions cover-ing nonresidents’ repatriation of proceeds from bond transactionsconstituted restrictions subject to approval under Article VIII,Section 2 (a).



was, to a surprising extent, supportive of the countryauthorities’ policy choices, whatever they may havebeen. In a few instances, the staff also recommendedfurther trade liberalization, liberalization of capitaloutflows, and tightening of prudential regulation asmeasures to deal with large capital inflows. Theseand other structural measures, however, received rel-atively little attention in the IMF’s policy advice.Given the IMF’s focus and comparative advantage,this was probably appropriate.

In order to make a full assessment of the IMF’soverall policy advice on managing capital inflows,one must understand the staff’s assessment of thecountries’ macroeconomic frameworks and institu-tional constraints under which advice is given. How-ever, internal country documents (let alone staff re-ports) generally do not provide sufficient analyticalbasis, much less a well-articulated analytical model,for understanding why a particular combination ofpolicies was advised in a particular case. Withoutsuch understanding, there appears to be inconsis-tency in the intensity with which advice for fiscaland exchange rate policies was given both acrosstime and across countries. It would be helpful forcountry documents to be more explicit in articulat-ing the rationale for advocating certain policy ac-tions, and not others.


Temporary use of capital controls has been a con-troversial subject, not only within the IMF but also inthe academic and official policymaking communi-

ties. It is possible here to make a broad generalizationthat the IMF staff was in principle opposed to the useof such instruments, either on inflows or outflows. Itsview was that they were not very effective, especiallyin the long run, and could not be a substitute for therequired adjustments in macroeconomic policies.Nevertheless, in some countries, the IMF staff dis-played a remarkable degree of sympathy with the useof capital controls even from the earliest days. Insome cases, it even suggested that market-based con-trols could be introduced as a prudential measure.

The key question is whether the difference re-flected the IMF’s assessment of the role of thesecontrols in the authorities’ overall macroeconomicframework. This may well be the case in some in-stances. For example, the staff’s tacit support forRussia’s outflow controls may be related to the factthat the Russian authorities had already agreed to acomprehensive program of fiscal consolidation andbanking reform, whereas it might have given lesssupport to Thailand’s use of outflow controls be-cause there was no agreed program in place. Simi-larly, in some cases there is evidence that differingjudgments on the appropriateness of temporary con-trols reflected differing assessments of underlyingexchange rate policies. Such an approach seems rea-sonable. More generally, however, documents arenot sufficiently explicit to allow us to make a defi-nite judgment about the consistency of the IMF’soverall advice on capital controls, except to say thatthe staff became much more accommodating of theuse of capital controls from the mid-1990s, and cer-tainly following the East Asian crisis.

48


This chapter reviews the IMF’s ongoing work ina sample of 14 countries for which outstanding

issues on capital account liberalization and the tem-porary use of capital controls were documented dur-ing 2003–04. The countries are divided into twogroups. The first group of nine countries were at dif-ferent stages in the ongoing process of capital ac-count liberalization during 2003–04, with some at anearly stage and others at a relatively advanced stage.The second group of six countries (with Russia over-lapping both groups) represent countries for whichthe IMF staff expressed a view on use of capital con-trols during 2003–04. The chapter first discusses theIMF’s advice on capital account liberalization andthen shifts to a discussion of how the IMF viewedthe temporary use of capital controls.

Capital Account Liberalization

The evaluation reviewed the documents for ninecountries that were moving toward greater capitalaccount convertibility during 2003–04: China, India,the Islamic Republic of Iran, Kazakhstan, Libya,Morocco, Russia, South Africa, and Tunisia (Table4.1). In addition, the evaluation looked at technicalassistance reports for two other countries in theprocess of capital account liberalization: Lesothoand Tanzania. Not only were these countries at dif-ferent stages of capital account openness, but theyalso differed in the way they interacted with the IMFduring this period. Some had a more formallyarranged dialogue on specific capital account liberal-ization issues with the IMF, while for others theIMF’s inputs were limited to what it provided as partof routine Article IV consultation discussions. Sevenof the nine countries reviewed here received an as-sessment of their financial systems under the FSAP.For each of these countries, we explain below thecontext in which advice on capital account liberal-ization was offered and the content of that advice.

In China, the IMF staff supported the country’sgradual approach to capital account liberalization.The briefing paper for the 2003 Article IV consulta-tion indicated that the staff’s support for gradual lib-

eralization was based on its view of “structural weak-nesses in the financial sector.” A staff note of March2004 argued that capital account liberalization shouldcome only after the establishment of a floating ex-change rate system, whose introduction should bephased in order not to pose risks to the weak financialsector. The gradual introduction of exchange rateflexibility was viewed as helpful to create strongerincentives for developing the foreign exchange mar-ket and for currency risk management, which in turncould facilitate further capital account liberalization.Capital controls were seen to support this process. Inthis context, the importance of pursuing consistentpolicies was stressed. Clearly, in China, the staff’sadvice was intimately connected with its assessmentof the financial sector.

Within the context of gradual liberalization (seeBox 4.1), the Indian authorities considerably liber-alized the capital account in 2002. The briefingpaper for the 2003 Article IV consultation indicatedthat the mission welcomed these measures but ex-pressed caution in proceeding with further signifi-cant opening, particularly in view of the large fiscaldeficits and the still weak financial system. Theback-to-office report suggested that, in this context,the mission reiterated the need for greater exchangerate flexibility, which would create incentives forrisk hedging. Additional measures were taken in2004 to liberalize the capital account, including eas-ing resident firms’ access to international capitalmarkets, raising the ceiling on the stock of govern-ment bonds that could be held by foreign investors,and relaxing the limits on capital outflows by resi-dents. During the 2004 Article IV consultation, thestaff again expressed support for the government’sgradual approach to capital account liberalizationand stressed the importance of addressing fiscal andfinancial sector weaknesses before proceeding fur-ther. It thus appears that the IMF’s advice on capitalaccount liberalization in recent years was framed inits assessment of India’s macroeconomic and struc-tural conditions.

Direct investment flows into the Islamic Republicof Iran picked up significantly during 2002–04, in partas a response to a revision of the foreign direct invest-

Ongoing Country Dialogue onCapital Account Issues

49

CHAPTER

4


CHAPTER 4 • ONGOING COUNTRY DIALOGUE ON CAPITAL ACCOUNT ISSUES

ment law. In supporting the authorities’ gradual ap-proach to capital account liberalization, the IMF staffemphasized the importance of financial sector reformand, in this context, encouraged the authorities toopen the banking system to foreign strategic investors.The staff consistently pressed the authorities totighten fiscal policy, increase exchange rate flexibility,and improve monetary policy instruments in order todeal better with the rising capital inflows. It did notsuggest an introduction of inflation targeting at thistime, however, because of the Islamic Republic ofIran’s weak institutions. The briefing papers for boththe 2003 and the 2004 Article IV consultations ex-pressed the view that the speed of liberalizationshould be linked to progress in addressing weaknessin the financial system and improving fiscal and mon-etary policy coordination. This clearly indicated thatthe staff’s view of capital account liberalization wasbased on its assessment of the Islamic Republic ofIran’s overall macroeconomic, financial sector, andinstitutional conditions.

In Kazakhstan, the authorities indicated their inten-tion gradually to liberalize the capital account, withtechnical assistance from the IMF, beginning in 2003(Box 4.2). The staff supported the authorities’ cau-tious approach to capital account liberalization, butstressed the need for supporting policies, including es-tablishing effective consolidated prudential supervi-sion and implementing sound risk management prac-tices. The briefing paper for the 2003 Article IVconsultation noted that the staff was more concernedwith the authorities’ ability to manage risks in the faceof large capital inflows than with the potential for cap-ital flight, and stated its intention to emphasize the im-portance of building a sound financial system “regard-

less of the pace of liberalization.” The staff’s discus-sions with the authorities during 2003–04 focused onthe need for greater exchange rate flexibility. On themonetary policy framework, however, the staff in re-viewing departments expressed doubt about the ap-propriateness or feasibility of inflation targeting in acountry highly dependent on oil revenues and atKazakhstan’s stage of institutional development. Thebriefing paper for the 2004 Article IV consultationnoted the same concerns but stated the staff’s inten-tion to discuss “plans to accelerate capital account lib-eralization in light of de facto openness of the capitalaccount and strong macroeconomic fundamentals.”Given Kazakhstan’s strong fiscal position, however,little concern was expressed about fiscal policy.

The Libyan authorities expressed in 2004 their in-tention to open up Libya’s economy to the globalmarket and to move forward with the implementationof structural reforms with IMF support. In the staff re-port for the 2004 Article IV consultation, the staff out-lined a preferred sequence for structural reform con-sisting of two phases. The first phase, which is to last1–12 months, involves adopting a medium-term bud-get framework, lifting sectoral credit restrictions,gradually liberalizing interest rates, and finalizing theplan to restructure public banks. The second phase,which is to last another 12–36 months, involves thebuilding of institutions for an efficient market econ-omy, including accelerating the process of creating asound investment climate, restructuring or privatizingpublic enterprises and banks, and developing moneymarkets and monetary policy instruments. The staffreport then suggested that, over the medium term,“capital account convertibility could be considered”once the first two phases have been sufficiently imple-

50

Table 4.1. Policy Environments for Capital Account Liberalization in Selected Countries1

1999 2000 2001 2002 2003 20042

China . . . . . . . . . . . . . . . . . .India3 . . . FSAP . . . . . . . . . . . .Iran, I.R. of . . . M, FSAP . . . . . . TA . . .Kazakhstan M FSAP . . . TA TA FSAPuLibya . . . . . . . . . . . . . . . . . .Morocco . . . . . . . . . FSAP . . . . . .Russia3 . . . . . . . . . FSAP . . . . . .South Africa3 . . . IT, FSAP FSAPu . . . . . . . . .Tunisia . . . . . . FSAP TA . . . . . .

Source: IMF database.1A shaded area corresponds to a period in which the country had a floating exchange rate (including a managed float). M and

IT indicate the introduction of a monetary policy anchor and inflation targeting, respectively. FSAP (FSAPu) indicates the year inwhich an assessment of the country’s financial sector was made (or updated) under the Financial Sector Assessment Program.TA indicates the year in which the country received IMF technical assistance with regard to capital account liberalization is-sues.

2As of the second quarter.3There was a monetary policy anchor in these countries during the whole period.


Chapter 4 • Ongoing Country Dialogue on Capital Account Issues

mented. The IMF staff strongly urged the authoritiesto take advantage of the window of opportunity pro-vided by favorable macroeconomic and external con-ditions (with a budget surplus, a current account sur-

plus, and low inflation) to pursue the needed reformsrigorously.

In Morocco, the Article IV consultation discus-sions in both 2003 and 2004 focused on the need

51

Box 4.1. India

In the early 1990s, India made a historic departurefrom the inward-looking and interventionist policies ofthe past to embark on outward-looking reforms(Ahluwalia, 2002). Trade was considerably liberalizedover time, and liberalization of FDI received consider-able attention. The reforms of 1991 removed many ofthe restrictions on FDI inflows. In 1993, foreign institu-tional investors were also allowed to purchase shares oflisted Indian companies, opening a window for portfo-lio investment. Recognizing the link between currentand capital transactions, however, the authorities in-cluded certain safeguards in foreign exchange regula-tions when current account convertibility was estab-lished (Jadhav, 2003). These were: (1) surrenderrequirements; (2) need to present documentary evi-dence; and (3) indicative limits for representative cur-rent account transactions.1

Along with current account convertibility, capital ac-count transactions expanded substantially, and Indiaexperienced a surge in capital inflows from the end of1993 through 1994. In view of the expansion in capitalflows, it became an urgent priority for the authorities todetermine the systematic approach they should take to-ward the capital account. It was against the backdrop ofthese developments that, in March 1997, the ReserveBank set up a Committee on Capital Account Convert-ibility in order to “chalk out the road map and timeframe for achieving capital account convertibility”(Tarapore, 1998; see also Reserve Bank of India, 1997).The Committee’s report, released in early June 1997,concluded that India was “ready for a cautious andphased move” toward capital account convertibility andoutlined a three-year program to meet a set of certainpreconditions covering fiscal discipline, price stability,and banking system soundness.

The Indian authorities for the most part acceptedthese preconditions and embarked on a gradual ap-proach to greater capital account openness. In this ap-proach, controls have been applied more strictly foroutflows than for inflows; for residents than for nonres-idents; for banks than for institutional investors; and forindividuals than for corporations (Jadhav, 2003). FDIhas been encouraged by progressively expanding boththe automatic and the case-by-case routes. Access toexternal long-term borrowing has generally been lim-ited to corporations and development financial institu-tions, and subject to annual ceilings, but the thresholdfor automatic approval has been raised over time.Short-term borrowing is strictly controlled, except for

trade-related purposes. Liberalization of capital out-flows has just begun.

Various internal documents suggest that the IMFstaff consistently supported India’s gradual approach tocapital account liberalization. The staff during the 1995Article IV consultation, for example, endorsed themaintenance of an annual limit on external commercialborrowing, while suggesting that sufficient flexibilitybe retained to allow adequate financing for high prior-ity infrastructure projects. The 1996 mission stressedthe importance of sequencing, arguing that priorityshould be given to FDI and equity investment and thatliberalization of restrictions on short-term debt capitaland outward investment should be gradual.2 At thesame time, it cautioned the authorities against the risksof rapid capital account liberalization in the absence ofa strong fiscal position and a more robust financial sec-tor. In April 1997, the staff prepared a comprehensivenote on the pace and sequencing of capital account lib-eralization in support of the just-formed Committee onCapital Account Convertibility, stressing the need to se-quence capital account liberalization with other struc-tural and macroeconomic policy measures to containpotential risks. It also emphasized the importance ofexchange rate flexibility and the need to have market-based instruments for monetary control. The briefingpaper for the 1998 Article IV consultation expressedthe staff view that contagion to India from the EastAsian crisis had been limited, in part because of its rel-atively closed capital account.

The Executive Board broadly endorsed these staffviews. There were, however, always dissenting voices,particularly before the East Asian crisis. At the Boardmeeting to discuss the 1996 Article IV consultation,some Directors remarked that a more rapid pace of cap-ital account liberalization could be beneficial in en-couraging fiscal consolidation and financial reforms. Atthe 1997 meeting, while welcoming the proposal putforward by the Committee on Capital Account Convert-ibility, some Directors argued that there was consider-able scope to move forward at an early stage to liberal-ize further equity inflows and FDI. Such views becamerarer following the East Asian crisis. The summing upof the 1998 Board meeting simply noted: “Directorsencouraged the authorities to persevere with the phasedopening of the capital account.”

1These limits were raised over time.

2From the point of view of promoting fiscal discipline, themission further advised against liberalizing restrictions onforeign investment in government securities, even if such ameasure could foster the development of the government se-curities market.



for fiscal consolidation, “a new monetary policyframework,” and financial sector restructuring asthe essential prerequisites for moving ahead withcapital account liberalization. The 2003 FSAP re-port noted that, while the banking system was in areasonably good shape, removal of capital accountrestrictions could put adverse pressure on banks.The staff consistently argued that a pegged ex-change rate could not be sustainable in the absenceof restrictions on capital account transactions. Itadvised the authorities to consider exiting from thepeg over the medium term when the preconditions

for capital account liberalization have beenachieved. Under the existing macroeconomic con-ditions, the staff advised against attempting agreater integration of the Moroccan economy withworld financial markets.

In Russia, a new law to liberalize foreign ex-change transactions was submitted to the Duma inearly 2003. This included the immediate reductionof surrender requirements to 30 percent from 50 per-cent and the replacement of the approval require-ments by reporting requirements for many capitalaccount transactions. The proposals also included

52

Box 4.2.Technical Assistance

During the 1990s, the IMF staff provided a numberof countries with technical assistance (TA) on issuesbroadly related to capital account liberalization. For ex-ample, an Executive Board paper prepared by MAE inJuly 1998 stated that “about one quarter” of “the 54countries that received technical assistance on ex-change rate systems” during 1994–97 “received assis-tance on capital account liberalization.”1 However, anexamination of TA documents and staff’s explanationin Board papers indicate that the TA advice was quitebroad and general. It was often given to help build sup-porting institutions for an open capital account, includ-ing the reform of the foreign exchange system, im-provements in monetary control, the strengthening ofdomestic financial systems, and the preparation of newforeign exchange legislation. It was in the context ofthese more general discussions that the staff in somecases encouraged the authorities to make progress incapital account liberalization.

More specific TA on sequencing capital account lib-eralization—including on specific measures of institu-tion building and coordination with other reforms—was not given until the very end of the 1990s. The firstcountry to request such technical assistance on se-quencing was Tanzania (1999), followed by Kaza-khstan (2002), Tunisia (2002), the Islamic Republic ofIran (2003), and Lesotho (2003). To date, only thesefive countries have formally requested IMF technicalassistance on sequencing capital account liberalization.Of these, technical assistance to Tanzania and Lesothowas given in the context of an IMF-supported program(Lesotho also underwent an assessment of the financialsystem under the FSAP, which included advice on cap-ital account liberalization).

In providing advice to these countries, the staff advo-cated a gradual approach to capital account liberaliza-tion. While the staff argued that the benefits of capitalaccount liberalization were large, it also highlighted therisks that would arise in the absence of supporting poli-

cies and institutions, particularly a sound financial sec-tor and effective prudential regulation. In terms oforder of liberalization, the staff’s consistent positionwas that long-term flows should be liberalized beforeshort-term flows, and that the internationalization ofdomestic currency should be limited at an early stage ofthe liberalization process in order to retain greatermonetary autonomy.

Among the five countries, only two—Kazakhstanand Tunisia—have completed the initial objectives ofthe TA received. In the case of Tanzania and the IslamicRepublic of Iran, on the other hand, data limitationsprevented the staff from providing constructive advice,except to suggest strengthened efforts at better data col-lection and monitoring capacity. As to Lesotho, its spe-cial monetary relationship with South Africa compli-cated the domestic strategy for capital accountliberalization; the staff’s recommendation was there-fore limited to stressing the need to establish an appro-priate environment for liberalization, including largerforeign exchange reserves, a more sound financial sec-tor, and more prudent fiscal policy.

In Kazakhstan and Tunisia, the staff devised a multi-stage strategy of starting with measures that could beimplemented immediately before proceeding to thosethat required prior institution building. At the same time,it advised the authorities to introduce an early warningsystem to monitor speculative capital flows as the firstline of defense, and to enforce such a mechanism bypursuing appropriate exchange rate and monetary poli-cies. In particular, the staff urged the central banks to in-crease exchange rate flexibility and to expand the menuof monetary policy instruments (e.g., repo facilities andgovernment securities). In the case of Kazakhstan, thestaff even listed use of capital controls to complementthese measures, by stating that capital controls on short-term flows could be reinstated “under very exceptionalcircumstances” in which monetary or financial stabilitywas threatened by large capital flows.2

1Monetary and Exchange Affairs Department, “Develop-ments and Issues in the International Exchange and PaymentsSystem,” SM/98/172, July 7, 1998.

2Monetary and Exchange Affairs Department, “Kaza-khstan: Capital Account Liberalization and Financial SectorSupervision,” March 2002.



the possibility of introducing unremunerated reserverequirements to discourage short-term capital flows(see the section “Temporary Use of Capital Con-trols”). The briefing paper for the 2003 Article IVconsultation indicated that the staff welcomed thesesteps, but expressed concern over the weak state ofthe financial system if the capital account was to beopened quickly. The back-to-office report stated thatthe mission had advised the authorities that capitalaccount liberalization should be carefully phasedand had recommended, under the circumstances,only an early removal of controls on outflows. Fur-ther, the briefing paper for the 2004 Article IV con-sultation recorded the staff’s view that a sound bank-ing system was all the more important because of“the relatively ambitious schedule for liberalizationof capital controls.”

In South Africa, the IMF staff for a long timetook a very cautious attitude toward capital accountliberalization, particularly in view of the centralbank’s net open position in the forward exchangemarket. During the 2003 Article IV consultation,the mission reiterated its support for the gradual ap-proach but suggested that, in view of the possibleimpact on exchange rate volatility, completion ofthe process should wait until international reserveshad been built up to more comfortable levels.1The open forward position of the central bank waseliminated in February 2004, after which the staff’sposition clearly changed. The 2004 Article IV con-sultation report indicated that, given the relativelyhealthy banking system, the strength of the SouthAfrican rand presented “an opportune time to moveahead” in relaxing controls, particularly on residentcompanies. The mission continued to support theauthorities’ gradual approach to capital account liberalization, but expressed the view that the current strength of the rand could be exploited tomove more quickly with the removal of remainingcontrols.

As noted in greater detail in Appendix 1, theTunisian authorities have been pursuing gradual cap-ital account liberalization since the mid-1990s. Thebriefing paper for the 2004 Article IV consultationnoted that the staff would advise the authorities toestablish a new monetary framework as a nominalanchor and to increase exchange rate flexibilitygradually in their phased movement toward capitalaccount convertibility. At the same time, the need fora sound banking system was repeatedly stressed.During the discussions, the staff urged the authori-ties to address banking sector weaknesses to supportthe gradual liberalization of the capital account.

Temporary Use of Capital Controls

During 2003–04, use or possible use of capitalcontrols was on the agenda for discussion betweenthe IMF staff and the authorities of at least sixcountries: Bulgaria, Colombia, Croatia, Romania,Russia, and Venezuela. In four of these countries(Bulgaria, Croatia, Romania, and Russia), the sub-ject of policy discussion was market-based controlson inflows. In Colombia, a minimum stay require-ment was introduced to limit short-term capitalflows, while the controls in Venezuela were tar-geted at outflows and initially introduced as part ofa system of comprehensive exchange controls cov-ering both current and capital account transactions.We first discuss the four cases of market-basedcontrols, followed by a discussion of the othercases of administrative controls.

Market-based controls

Bulgaria experienced a rapid credit growth fi-nanced by external borrowing within the context of acurrency board arrangement. In the briefing paperfor the 2004 Article IV consultation in March, thestaff expressed its intention to advise the authoritiesto implement additional measures to contain creditgrowth if necessary, including market-based controlson capital inflows as a last resort measure. A similarview was expressed in the briefing paper for a staffvisit in September as well as in the subsequent brief-ing paper for a Stand-By review.

Financed by external borrowing, private credit ex-panded rapidly in Croatia. In view of the authorities’commitment to the use of the exchange rate as anominal anchor,2 the IMF staff in its November 2003briefing paper for a Stand-By review argued for atight monetary policy while suggesting that the au-thorities consider raising the marginal deposit re-quirements on banks’ liabilities or introducing mar-ket-based controls on capital inflows, if necessary.3In the briefing paper for the 2004 Article IV consul-tation, the staff repeated the same point but empha-sized that capital controls should remain a last-resortand “stop-gap” option to deal with large and unex-pected capital inflows “until other policies adjust orinflows abate.” In the event, in July 2004, the author-ities introduced a marginal reserve requirement,which required banks to deposit 24 percent of net in-creases in their foreign liabilities into a special inter-

53

1A similar view had been expressed by the 2000 FSAP report.

2Croatia had a managed float, but the authorities were reluctantto allow greater exchange rate flexibility because of their successwith exchange rate-based stabilization in the previous years.

3A new foreign exchange law, in May 2003, gave the centralbank additional authority to restrict capital inflows.



est-free account held with the central bank, for anunlimited time period.4

Romania has also experienced a rapid creditgrowth in recent years. Romania, however, main-tains restrictions on short-term capital inflows. Inthis context, in the November 2003 briefing paperfor a request for a Stand-By Arrangement,5 thestaff’s position was that the authorities should con-tinue to restrict the access of nonresidents to short-term domestic currency instruments until price sta-bility was firmly achieved.6 As things turned out,while the growth of domestic currency credit slowedin 2004, foreign currency credit steadily grew. In thebriefing paper for the first Stand-By review, the staffexpressed its intention to discuss the possibility ofintroducing greater exchange rate flexibility.7 Effec-tive August 2004, the authorities raised the reserverequirements on foreign exchange liabilities to 30percent from 25 percent. A briefing paper of January2005 indicated the staff’s intention to advise the au-thorities to extend the reserve requirements on for-eign currency liabilities to those with a residual ma-turity of more than two years as a way of creatingscope for limited monetary easing.8

In Russia, the law establishing an unremuneratedreserve requirement (URR) on both inflows and out-flows took effect on July 1, 2004. At the time, theRussian authorities abolished most of the existing ad-ministrative controls on capital movements (see alsothe section “Capital Account Liberalization”). TheURR thus served as a device to liberalize the capitalaccount while retaining safeguards against excessiveflows. The IMF staff and management views weresomewhat mixed. During 2003, when the provisionsof the law were being discussed against the back-ground of steady capital inflows, the staff had writtenin a briefing paper that it intended to recommend as apolicy response to this situation fiscal tightening andpossibly introduction of additional temporary con-trols on inflows. To this, management suggested thatthe staff should focus on the measures currently pro-posed by the authorities without advocating any newcontingent capital control measures.

When the law was enacted, the staff consideredthe introduction of the URR as a significant im-provement over the previous system of administra-tive controls. At the same time, it viewed the largecapital inflows as being caused largely by Russia’sexchange rate policy9 and considered that fiscal pol-icy was too loose, which burdened monetary policywith the task of controlling inflationary pressure.Under these circumstances, the staff thought thatthere was no need to activate the URR if appropriateadjustment was made in exchange rate and macro-economic policies. According to the back-to-officereport for the 2004 Article IV consultation, the mis-sion argued that while the URR was appropriate as atransitional measure from the system of permits andcontrols, its effect on capital inflows would be onlytemporary and its main effect would be to raise thecost of capital for small and medium-sized domesticborrowers. It then noted that controls should not be asubstitute for tight fiscal policy and the adoption of amore flexible exchange rate policy. In the case ofRussia, the staff’s assessment of capital controls wasclearly framed within its assessment of the authori-ties’ macroeconomic strategy.10

Administrative controls

In Colombia, faced with a sharp and sustained ap-preciation of the Colombian peso, in December 2004the authorities introduced a one-year stay require-ment for inward portfolio investment by nonresi-dents, while allowing the profits from these invest-ments to be repatriated freely. Interviews suggestthat the IMF staff, while skeptical of the effective-ness of the capital control in arresting aggregate cap-ital inflows (given its limited coverage), understoodthe political economy context in which it was intro-duced. The issue was not raised in negotiations for aStand-By Arrangement in early 2005.

As explained in greater detail in Appendix 1, theVenezuelan authorities introduced “temporary”comprehensive foreign exchange controls in Febru-ary 2003, covering both current and capital accounttransactions. While the staff recommended that allrestrictions on current transactions be eliminated, itexpressed its understanding that capital controls

54

4In February 2005, the authorities raised the deposit require-ment to 30 percent from 24 percent; in March 2005, they an-nounced that they would liberalize controls on capital outflows.

5The Stand-by Arrangement was to be treated as precautionary.6Under EU accession agreements, the authorities had made a

commitment to allow nonresidents to hold domestic currency de-posits, tentatively from April 2005, and to liberalize the treasurybill market in 2007. In early 2004, however, the authorities madea decision to delay by 12 months the implementation of theircommitment to this schedule of capital account liberalizationunder EU accession agreements. This decision was supported bythe IMF staff.

7Romania abandoned the de facto (unannounced) crawlingband in November 2004.

8This measure was taken in February 2005.

9From early 2003, the authorities allowed the ruble to appreciateagainst the U.S. dollar but temporarily discontinued this policy inearly 2004. The ruble’s appreciation resumed later in the year.

10There was a contradiction in the staff assessment at a differ-ent level. In 2003, the staff’s position was that Russia should onlygradually dismantle administrative controls on capital flows, inview of its weak banking system and weak enforcement of pru-dential norms. In 2004, however, the staff took a position againstthe URR which had been introduced to replace the administrativecontrols, while the macroeconomic situation had not materiallychanged from 2003.



might be necessary in the short run, mainly on capi-tal outflows. In 2004, the staff advised a gradual ap-proach to liberalize capital account restrictions,which had to be well sequenced and supported bymacroeconomic and institutional reforms.

AssessmentIn ongoing country work, the IMF staff has in gen-

eral been accommodating of the authorities’ policy

choices when they have involved a gradual approachto capital account liberalization or temporary use ofcontrols. In terms of capital account liberalization,the IMF staff was sometimes more cautious than theauthorities (for example, in Russia in 2003) whentheir preferred policy was rapidly to liberalize thecapital account. In most cases, it appears that the stafftook a medium-term perspective and emphasized theimportance of meeting certain preconditions, themost important of which were fiscal consolidation, a

55

Box 4.3. The IMF’s “Integrated” Approach to Capital Account Liberalization

The IMF’s “integrated” approach, developed fromthe late 1990s to the early 2000s and as outlined in Ishiiand others (2002), was discussed in an Executive Boardseminar in July 2001.1 Although the Board has neverendorsed it as official policy, it nonetheless appears toenjoy wide acceptance among the IMF staff as repre-senting the institution’s current thinking on capital ac-count liberalization.

The approach is called “integrated” because it consid-ers capital account liberalization as part of a comprehen-sive program of economic reforms in the macroeco-nomic policy framework, the domestic financial system,and prudential regulations. In terms of sequencing, thepolicy paper sets forth ten general principles: (1) capitalaccount liberalization is best undertaken with soundmacroeconomic policies; (2) those reforms that supportmacroeconomic stabilization should be given priority;(3) reforms that are operationally linked should be im-plemented together; (4) domestic financial reformsshould be complemented by prudential regulation and fi-nancial restructuring; (5) sequencing should take ac-count of the concomitant risks of various types of instru-ments; (6) pace should take account of conditions in thenonfinancial sector; (7) reforms that take time should bestarted early; (8) reforms need to take account of the ef-fectiveness of the controls in place; (9) the pace, timing,and sequencing of liberalization need to take account ofpolitical and regional considerations; and (10) thearrangements for policy transparency and data disclosureshould be adapted to support capital account opening(see Ishii and others, 2002, pp. 16–18).

The paper argues that these general principles on se-quencing are to be applied in a specific instance in sucha way as to reduce or better manage various types ofcredit, market, and liquidity risks that are typically mag-nified by cross-border transactions. In particular, thepaper identifies a number of risks associated with differ-ent types of capital flows and suggests key policy mea-sures to manage those risks. The operational “methodol-ogy” would first make a diagnosis of the existingregulations and institutions and, on the basis of this di-

agnosis, develop a three-stage plan for sequencing andcoordinating capital account liberalization with otherpolicies: (1) the first stage involves achieving a high de-gree of macroeconomic stability and developing mar-kets and institutions, fostering good risk managementby banks and other economic entities, and remedyingthe most important shortcomings in prudential regula-tion, with low-risk capital flows (such as FDI) being al-lowed to take place first; (2) the second stage entails aconsolidation and deepening of the progress made in thefirst stage, with considerable further capital account lib-eralization taking place; and (3) in the third and finalstage, all remaining capital controls are lifted, as macro-economic and financial sector conditions have improvedto the point where risks are effectively managed.

As a statement of general principles, few would dis-agree with the prudence and judiciousness of the inte-grated approach, and its emphasis on the need to tailorthe pace and sequencing to particular country circum-stances is welcome. However, while the paper statesthat “a gradual approach would not by itself guaranteean orderly liberalization,” there is an unmistakable biastoward gradualism in this approach. More importantly,by emphasizing all of the potential interlinkages with-out any clear approach to identifying a hierarchy ofrisks, this approach may inadvertently create a false no-tion that a country can achieve full capital account con-vertibility only when it has fully developed all relevantmarkets, institutions, and regulatory frameworks. Inthis connection, a group of Asian experts have de-scribed the IMF’s integrated approach as including“virtually every conceivable aspect of microeconomic,institutional, and macroeconomic policy possible,”“unnecessarily complex,” and “unoperational, as itlacks a clear hierarchy of priorities” (Asian PolicyForum, 2002). While this assessment is perhaps too se-vere, the principles, while giving some—albeit incom-plete—indications on prioritization, do not provide cri-teria for judging what reforms are more critical thanothers overall. It is certainly true that such judgmentswould have to be made in the context of country-spe-cific circumstances. However, because the approachidentifies a complex set of risks and the requisite mea-sures without clear criteria for balancing those risks, ithas proven to be difficult to apply in practice.

1A further elaboration of this approach is given by Kara-cadag and others (2003).



sound financial system, and the adoption of a floatingexchange rate (usually with inflation targeting). Thestaff’s generally cautious attitude toward capital ac-count liberalization is in part a reflection of the new“integrated” approach (Box 4.3), which has beenmore fully applied in countries that have requestedtechnical assistance from the IMF.

In terms of advice on temporary use of capital con-trols, the IMF staff seldom challenged the authorities’decision and even had a positive attitude toward mar-ket-based controls, if only as a temporary measure.There was, however, a slight difference in emphasisacross countries. In some cases (as in Russia in 2004),the staff expressed a view quite forcefully that capitalcontrols, no matter how useful they might be in theshort run, could not be expected to be effective overtime and should not be used as a substitute for appro-priate adjustment in macroeconomic and exchangerate policies. In others (as in Colombia), the use ofcontrols introduced by the authorities did not figureprominently in policy discussions. In still other cases(as in Bulgaria and Croatia), the staff recommended amarket-based control as a last resort measure.

Clearly in some countries (for example, China andIndia), the IMF’s advice on the sequencing of capitalaccount liberalization was framed in its assessmentof the countries’ macroeconomic or financial sectorconditions. In other cases (as in the Islamic Republicof Iran, Kazakhstan, and Libya), its cautious ap-proach was conditioned by its assessment that therequisite institutional infrastructure for an open capi-tal account was not yet in place. In one case (SouthAfrica), on the other hand, a favorable assessment ofthe overall macroeconomic situation prompted thestaff to suggest that the remaining controls could beremoved quickly. In terms of temporary use of capitalcontrols, the staff proposed market-based instrumentsin the form of marginal reserve requirements onbanks’ foreign exchange liabilities when it saw alimit to the effectiveness of conventional macroeco-nomic policy tools to deal with large capital inflows.These are indications that the staff is giving greaterattention than previously to the overall economic pol-icy environment and institutional constraints of thecountries concerned in providing advice on capitalaccount issues.

56


This report has evaluated the IMF’s approach tocapital account liberalization and other related

capital account issues. It first reviewed the IMF’sgeneral operational approach and analysis as theyevolved from the early 1990s into the early 2000s.The report then assessed the IMF’s country work for 1990–2004 in terms of (1) its role in capital ac-count liberalization, (2) advice to member countrieson managing capital flows, including the temporaryuse of capital controls, and (3) ongoing work oncapital account issues. Most of the analysis oncountry work was based on IMF documents for asample of over 30 emerging market and developingeconomies.

This concluding chapter summarizes the majorfindings of the evaluation and presents two broadrecommendations designed to help improve theIMF’s operational work on capital account issues.

Major Findings

Major findings are summarized below under (1) the IMF’s general operational approach andanalysis, (2) the IMF’s country work, and (3) over-all assessment of the IMF’s approach to capital ac-count liberalization and related issues.

The IMF’s general operational approach and analysis

The Articles of Agreement left considerable ambi-guity about the role of the IMF in capital account is-sues. Even so, in the early 1990s the IMF respondedto the changing international environment, character-ized by large cross-border capital movements, bypaying greater attention to issues related to the capi-tal account. From the mid-1990s, staff analysesbegan clearly to advocate capital account liberaliza-tion. Concurrent with the initiatives to amend the Ar-ticles to give the IMF an explicit mandate for capitalaccount liberalization and jurisdiction on members’capital account policies, management and staff ex-panded the scope of the IMF’s operational work oncapital account issues in Article IV consultations and

technical assistance in an effort to promote capital ac-count liberalization more actively.

The IMF’s analysis prior to the mid-1990s tendedto emphasize the benefits to developing countries ofgreater access to international capital flows and topay comparatively less attention to the potential risksof capital flow volatility. More recently, however, theIMF has paid greater attention to various risk factors,including the linkage between industrial countrypolicies and international capital flows as well as themore fundamental causes and implications of theirboom-and-bust cycles. Still, the focus of the analysisremains on what emerging market countries shoulddo to cope with the volatility of capital flows (for ex-ample, through macroeconomic and exchange ratepolicy, strengthened financial sectors, and greatertransparency). The IMF has addressed the moral haz-ard aspect of boom-and-bust cycles, for example, byencouraging greater exchange rate flexibility in re-cipient countries and attempting to limit access toIMF resources during a crisis, but has not been at theforefront of the debate on what, if anything, can bedone to reduce the cyclicality of capital movementsthrough regulatory measures targeted at institutionalinvestors in the source countries.

From the beginning of the 1990s, the IMF’s man-agement, staff, and Executive Board were aware ofthe potential risks of premature capital account liber-alization and there is no evidence to suggest thatthey promoted capital account liberalization indis-criminately. They also acknowledged the need for asound financial system in order to minimize the risksof liberalization and maximize its benefits. Suchawareness, however, largely remained at the concep-tual level and did not lead to operational advice onpreconditions, pace, and sequencing until later in the1990s. At the same time, a subtle change was takingplace within the institution. As preliminary evidenceemerged on the apparent effectiveness of Chile’scapital controls in the mid-1990s, some in the IMFbegan to take a favorable view of the use of capitalcontrols as a temporary measure to deal with largecapital inflows.

In the event, the proposed amendment of the Ar-ticles put forward in the late 1990s failed to garner

Major Findings andRecommendations

CHAPTER

5

57


CHAPTER 5 • MAJOR FINDINGS AND RECOMMENDATIONS

sufficient support, leaving ambiguity about the roleof the IMF. In the meantime, something of a con-sensus—the so-called “integrated” approach—hasemerged within the IMF that places capital accountliberalization as part of a comprehensive program ofeconomic reforms in the macroeconomic policyframework, the domestic financial system, and pru-dential regulations. While few would disagree withthe prudence and judiciousness of the new ap-proach, it has proved to be difficult to apply as anoperational guide to sequencing because it empha-sizes all of the potential interlinkages but does notprovide clear criteria for identifying a hierarchy ofrisks. Moreover, these views remain unofficial, asthey have not been explicitly endorsed by theBoard. The IMF still does not have a formal state-ment from the Board providing guidance on what itspolicy advice is, and a number of senior staff ex-pressed unease about this lack of clarity on theIMF’s formal position.

The IMF’s country work

The evaluation reviewed the IMF’s country workin a sample of emerging market economies and as-sessed its approach in terms of the following criteria:(1) Was there any difference between the IMF’s gen-eral policy pronouncements and the advice it gave toindividual countries? (2) Was the IMF policy adviceoperational and based on solid evidence? (3) Howdid the IMF’s advice change over time, and did thischange keep pace with available evidence? (4) Didthe IMF give similar advice to countries in similarsituations? and (5) Was the policy advice on the cap-ital account set in a broader assessment of the au-thorities’ macroeconomic policies and institutionalframework?

The role of the IMF in capital accountliberalization

During the 1990s, the IMF encouraged capital ac-count liberalization in its country work, but the eval-uation suggests that in all the countries that liberal-ized the capital account, partially or almost fully, theprocess was for the most part driven by the countryauthorities’ own economic and political agendas. Innone of the program cases examined did the IMF re-quire capital account liberalization as formal condi-tionality, although aspects of it were often includedin the authorities’ overall policy package presentedto the IMF. In the early years, the issues of pace andsequencing were seldom raised. Occasionally, staffexpressed concern over financial sector weakness ormacroeconomic instability, but this did not lead tooperational advice to individual countries on paceand sequencing. From around 1994, and certainly

after the Mexican crisis, some within the IMF be-came more cautious in their policy advice on capitalaccount issues in country work, but it was only afterthe East Asian crisis that the whole institution’s ap-proach clearly changed. In later years, the IMF tooka more cautious attitude toward capital account lib-eralization in country work, emphasizing pace andsequencing and the need to satisfy certain precondi-tions. Through much of the 1990s, there was appar-ent inconsistency in the IMF’s advice on capital ac-count liberalization (for example, sequencing wasemphasized in some countries but not in others),suggesting that the staff took a pragmatic approachin country work in the absence of clear officialguidelines. On the other hand, there was little sys-tematic difference in terms of policy advice betweenprogram and nonprogram countries.

Advice on managing capital flows

As countries experienced large capital inflows andassociated macroeconomic challenges in the 1990s,the question of how to manage large capital inflowsbecame a routine subject of discussion between theIMF and the country authorities. The staff’s policyadvice on managing capital flows did not deviatemuch from the policy conclusions typically derivedfrom the scholarly literature on open economymacroeconomics. To deal with large capital inflows,it advocated tightening fiscal policy and greater ex-change rate flexibility. Advice on sterilization, in linewith the conventional wisdom, emphasized its quasi-fiscal costs and its longer-term ineffectiveness but, toa surprising extent, was supportive of country author-ities’ policy choices, whatever they may have been.In a few instances, the staff also recommended fur-ther trade liberalization, liberalization of capital out-flows, and tightening of prudential regulation as mea-sures to deal with large capital inflows, but these andother structural measures received relatively little at-tention. In terms of detail and emphasis, the staff’sviews differed both across time and across countries,but country documents do not provide a clear analyti-cal basis to make a definite judgment about the con-sistency of the IMF’s overall advice on managingcapital inflows.


Use of capital controls has been a controversialsubject, not only within the IMF but also in the aca-demic and official policymaking communities. It ispossible here to make a broad characterization thatthe IMF staff was in principle opposed to the use ofsuch instruments, either on inflows or outflows. Itsview was that they were not very effective, espe-cially in the long run, and could not be a substitute

58


Chapter 5 • Major Findings and Recommendations

for the required adjustments in macroeconomic poli-cies. Even so, from the earliest days, the IMF staffdisplayed a remarkable degree of sympathy withsome countries in the use of capital controls. In afew cases, both before and after the crises of1997–98, it even suggested that market-based con-trols could be introduced as a prudential measure. Asa general rule, the IMF staff in its country work, inline with the evolution of the institution’s view, be-came much more accommodating of the use of capi-tal controls over time, albeit as a temporary, second-best instrument.

Ongoing country dialogue on capital account issues

In ongoing country work (as documented for2003–04), IMF staff has been quite accommodatingof authorities’ policy choices when they have in-volved a gradual approach to capital account liberal-ization or temporary use of controls. In terms of cap-ital account liberalization, the staff has sometimesbeen more cautious than the authorities (for exam-ple, Russia in 2003) when their preferred policy wasto liberalize the capital account quickly. In mostcases, the staff has taken a medium-term perspectiveand emphasized the importance of meeting certainpreconditions, the most important of which are fiscalconsolidation, a sound financial system, and theadoption of a floating exchange rate (usually withinflation targeting). There has been some variationacross countries, however. For example, at least inone country (South Africa), the IMF staff has urgedthe authorities to move more quickly in removingthe remaining restrictions in view of favorable exter-nal conditions.

In terms of advice on temporary use of capitalcontrols, IMF staff seldom challenged the authori-ties’ decision and even supported market-based con-trols in some cases. There was a slight difference inemphasis across countries. In a few countries (as inRussia in 2004), the staff expressed forcefully theview that capital controls, no matter how useful theymight be in the short run, could not be expected to beeffective over time and should not be used as a sub-stitute for appropriate adjustment in macroeconomicpolicies. In others (as in Colombia), the use of con-trols introduced by the authorities did not figureprominently in policy discussions. In still other cases(as in Croatia), the staff recommended a market-based control, albeit as a last resort measure.

Overall assessment

These findings allow us to provide answers to thefollowing fundamental and related questions: (1)Did the IMF pressure member countries to liberalize

the capital account in the 1990s? and (2) Did theIMF encourage capital account liberalization prema-turely, without ensuring that necessary institutionsand regulations were in place to maximize its bene-fits and minimize its risks?

The answer to the first question is a definitive no,at least for the emerging market countries examined.In none of these countries did the IMF use the mostbinding tool of influence at its disposal: conditional-ity in the use of its resources. This is consistent withthe interpretation of the Articles of Agreement,which states that the IMF, as a condition for the useof its resources, cannot require a member to removecontrols on capital movements. In several programcountries, however, aspects of capital account liber-alization were included in the authorities’ package ofeconomic policies presented to the IMF. These maywell reflect different degrees of “pressure” in spe-cific instances, but the evaluation has not uncoveredany such cases. In summary, the IMF undoubtedlyencouraged countries that wanted to move aheadwith capital account liberalization, and even acted asa cheerleader when it wished to do so, especially be-fore the East Asian crisis, but there is no evidencethat it exerted significant leverage to push countriesto move faster than they were willing to go. Theprocess of liberalization was often driven by the au-thorities’ own considerations, including OECD orEU accession and commitments under bilateral orregional trade agreements.

The answer to the second question is less clear-cut. Yes, the IMF did encourage capital account lib-eralization and this encouragement probably gotahead of the prevailing evidence in the early 1990s.At the same time, in so doing, it made the point ofhighlighting the risks inherent in an open capital ac-count as well as the need for a sound financial sys-tem, even from the beginning. The problem was thatthese risks were insufficiently highlighted and therecognition of these risks and preconditions did nottranslate into operational advice on pace and se-quencing until later in the 1990s (and even thereafterthe policy advice has often been of limited practicalapplicability).

In this connection, it should also be noted thatthe IMF’s analysis in the earlier period emphasizedthe benefits to developing countries of greater ac-cess to international capital flows, while payingcomparatively less attention to the risks inherent intheir volatility. As a consequence, its policy advicewas directed more toward emerging market recipi-ents of capital flows, focusing on how to managelarge capital inflows and their boom-and-bust cy-cles. Little policy advice was offered, in the contextof multilateral surveillance, on how source coun-tries might help reduce the volatility of capitalflows through regulatory measures on the supply

59



side. In more recent years, the IMF’s analysis ofsupply-side factors has become more sophisticated,and the institution has also addressed the moralhazard aspect of investor behavior. Even so, thefocus of policy advice remains on the recipientcountries. Admittedly, this reflects the lack of a the-oretical and empirical consensus on what practicalsteps could be taken in this area, but the IMF has played a relatively limited role in exploring options.

The evaluation suggests that the IMF has learnedover time on capital account issues. This seems tohave affected the work of the IMF through twochannels. First, the IMF’s general approach did re-spond—albeit gradually—to new developments ornew evidence. Second, independent of how the gen-eral approach changed, some of the learning be-came more quickly reflected in the IMF’s countrywork through its impact on individual staff mem-bers. As a result, in the case of capital account is-sues, the IMF’s general approach lagged behind developments in some of the country-specific ap-proaches taken in the field.

The lack of a formal IMF position on capital ac-count liberalization and the associated partial dis-connect between general operational guidelines andcountry work had different consequences. On theone hand, it gave individual staff members freedomto use their own professional and intellectual judg-ment in dealing with specific country issues. On theother hand, the disconnect reflected the inherentambiguity of this aspect of the IMF’s work on capi-tal account issues and led to some lack of consis-tency in country work. Country work must of neces-sity be tailored to country-specific circumstances,so uniformity cannot be the only criterion for judg-ing the quality of the IMF’s policy advice. Even so,it appears that the apparent inconsistency to a largeextent reflected reliance on the discretion of indi-vidual staff members, and not necessarily the con-sistent application of the same principles to differ-ent circumstances.

In more recent years, somewhat greater consis-tency and clarity have been brought to bear on theIMF’s approach to capital account issues. For themost part, the new paradigm upholds the role ofcountry ownership in determining pace and sequenc-ing; takes a more consistently cautious and nuancedapproach to encouraging capital account convertibil-ity; and acknowledges the usefulness of capital con-trols under certain conditions, particularly controlson inflows. But these are still unofficial views,though they may well be widely shared within theinstitution. While the majority of staff members nowappear to accept this new paradigm, some continueto feel uneasiness with the lack of a clear formal po-sition by the institution.

Recommendations

The evaluation suggests two main areas in whichthe IMF can improve its work on capital account issues.

Recommendation 1. There is a need for moreclarity on the IMF’s approach to capital account is-sues. The evaluation is not focused on the argumentsfor and against amending the Articles of Agreement,but it does suggest that the ambiguity about the roleof the IMF with regard to capital account issues hasled to some lack of consistency in the work of theIMF across countries. This may reflect the lack ofclarity in the Articles, but with or without a changein the Articles it should be possible to improve theconsistency of the IMF’s country work in otherways. For example:

• The place of capital account issues in IMF sur-veillance could be clarified. It is generally un-derstood that while under current arrangementsthe IMF has neither explicit mandate nor juris-diction on capital account issues, it has a re-sponsibility to exercise surveillance over certainaspects of members’ capital account policies.However, much ambiguity remains on the scopeof IMF surveillance in this area. The cleareststatement of the basis for surveillance of capitalaccount issues is embodied in the 1977 Execu-tive Board decision calling for surveillance toconsider certain capital account restrictions in-troduced for balance of payments purposes, butthe qualification limiting the scope to balance ofpayments reasons is too restrictive to cover therange of capital account issues that surface inthe IMF’s country work. On the other hand, thebroader statement of the IMF’s surveillance re-sponsibility, found in the preamble to Article IV,is too wide to serve as an operational guide tosurveillance on capital account issues. Therewould be value if the Executive Board were for-mally to clarify the scope of IMF surveillanceon capital account issues. Such a clarificationwould recognize that capital account policy isintimately connected with exchange rate policy,as part of an overall macroeconomic policypackage, and that in many countries capitalflows are more important in this respect thancurrent flows; capital controls can be used tomanipulate exchange rates or to delay neededexternal adjustment; and a country’s capital ac-count policy creates externalities for other coun-tries. Capital account policy is therefore of cen-tral importance to surveillance.

• The IMF could sharpen its advice on capital ac-count issues, based on solid analysis of the par-

60


Chapter 5 • Major Findings and Recommendations

ticular situation and risks facing specific coun-tries. Given the limited evidence that exists inthe literature on the benefits or costs of capitalaccount liberalization in the abstract, the IMF’sapproach to any capital account issue must nec-essarily be based on an analysis of each case.For example, if a capital control is involved, theIMF must ask in the context of a specific coun-try what objectives the control is designed toachieve; if it is accomplishing them; andwhether there are more effective or less distor-tionary ways of achieving the same objectives.Such assessments need to be set in an overallconsideration of the macroeconomic policyframework and whether controls are being usedas a substitute for, or to seek to delay necessarychanges in, such policies. The evaluation indi-cates that this is already done in some but not allcases. If a capital control measure is judged use-ful to stem capital flight under certain circum-stances, the IMF should ask what supportingpolicies are needed to make it more effective orless distortionary (for example, setting up a sys-tem of monitoring external transactions). Interms of providing advice on capital account lib-eralization, just to spell out all the risks inherentin opening the capital account is of limited use-fulness to countries seeking IMF advice. To as-sist the authorities decide when and how to openthe capital account, the IMF should providesome quantitative gauge of the benefits, costs,and risks (and, indeed, practicality) of moving atdifferent speeds. Admittedly, this is not an easytask. Drawing on the well-established literatureon welfare economics, the IMF must ask suchquestions as: What distortions are being createdwhen one market is liberalized but not another?What is the nature of risks being borne by resi-dents when capital account transactions are lib-eralized only for nonresidents? And what are thecosts to the economy (in terms of investmentflows) of allowing equity inflows but not debtinflows?

• The Executive Board could issue a statementclarifying the common elements of agreementon capital account liberalization. At present,there remains considerable uncertainty amongmany staff members on what policy advice toprovide to individual countries. This has led tohesitancy on the part of some within the staff toraise capital account issues with country au-thorities. The Executive Board could provideclear guidance to staff on what the IMF’s offi-cial position is. This is not to say that the Exec-utive Board must come up with a definitivestatement on all aspects of pace and sequenc-

ing. Given the lack of full consensus, oneshould not expect such a definitive view fromthe Board. However, Board guidance on whatare the minimum common elements on whichthere is broad, if not universal, agreementwould be useful to the staff and member coun-tries. Although the details are for the Board to decide, such a statement might include someor all of the following elements: (1) that in afirst-best world there would be no need for con-trols over capital movements (though financialmarkets may not always operate accordingly);(2) that controls should not be used as a substi-tute for adjusting macroeconomic or structuralpolicies; (3) a broad (as opposed to unnecessar-ily complex) framework of sequencing basedon the consensus in the literature on the orderof economic reforms; (4) the importance of tak-ing country-specific circumstances into ac-count; and (5) that risks can never be totallyeliminated, so they should not be used as a rea-son for permanently delaying liberalization.

Recommendation 2. The IMF’s analysis and sur-veillance should give greater attention to the supply-side factors of international capital flows and whatcan be done to minimize the volatility of capitalmovements. The IMF’s policy advice on managingcapital flows has so far focused to a considerable ex-tent on what recipient countries should do. While thisis important, it is not the whole story. As discussed inthe evaluation report, the IMF’s recent analyses havegiven greater attention to supply-side factors, includ-ing the dynamics of boom-and-bust cycles in emerg-ing market financing. The IMF has also establishedan International Capital Markets Department (ICM)as part of an effort to better understand global finan-cial markets; it participates actively in the work of theFinancial Stability Forum, which was established tomonitor potential vulnerabilities in global financialmarkets; and it has proposed a Sovereign Debt Re-structuring Mechanism (SDRM), encouraged the useof collective action clauses (CACs), and has at-tempted to place limitations on countries’ access toIMF resources in a crisis, in an effort to reduce theperceived moral hazard that may have led capitalmarkets to pay insufficient attention to the risks of in-vesting in developing countries and contributed to theboom-and-bust cycles of capital movements. Theseare important and welcome initiatives, but the IMFhas not yet fully addressed issues of what, if any-thing, can be done to minimize the volatility of capi-tal flows by operating on the supply side— as yet, lit-tle attention seems to be paid to supply-side risks andpotential mitigating actions in the industrial countriesthat are home to the major global financial markets.The IMF could usefully provide more input into ad-

61



vanced country financial supervision and other finan-cial market policy issues globally. Are current globalsupervision guidelines designed to help create stabil-ity?1 What if any action could be taken on the supplyside to reduce cyclicality and herd behavior? Admit-tedly, this is a difficult topic on which little profes-

sional consensus exists. Yet, this is an area where asignificant debate has taken place in the academicand policymaking communities and to which theIMF could contribute further. Indeed, one of thebroad themes identified as potential priorities for theIMF’s research program over the medium term—oninstitutions and contractual mechanisms that can helpprotect countries from external volatility—goes someway in this direction, but should not focus only onpolicies in countries that are recipients of capital inflows.

62

1To give one example where the IMF did provide such inputs,in July 2003, the staff commented on the proposed New BaselCapital Accord (Basel II), pointing to its potentially procyclicaleffects.


This appendix presents four specific contexts inwhich the IMF interacted with member countries

in terms of their capital account policies: (1) theCzech Republic, which liberalized the capital accountrelatively quickly in the mid-1990s; (2) Colombia,which introduced market-based controls to deal withlarge capital inflows during much of the 1990s; (3) Venezuela, which introduced controls on capitaloutflows in 1994 and 2003; and (4) Tunisia, whichhas pursued a gradual approach to capital account lib-eralization. These four cases are chosen for diversityof experience and are designed to illustrate what therole of the IMF was in each case and how the IMFviewed different policy issues regarding the capitalaccount. Although we briefly explain the policy mea-sures taken by the authorities to give context, thefocus remains on the IMF.

The Czech Republic: Liberalization in a Transition Economy


At the beginning of the transition process,Czechoslovakia was the most centrally plannedeconomy in Central Europe (Bakker and Chapple,2003). This in part explains the authorities’ generalpropensity for rapid reforms, aided by recourse toforeign capital (Blejer and Coricelli, 1995). Follow-ing the division of the country into two republics, in1993, the Czech authorities began to map the stepstoward capital account liberalization. According to anumber of former officials interviewed, the processwas largely driven by the country’s prospective ac-cession to the OECD and the EU.1 The liberalizationof capital transactions, which initially proceededthrough a progressively liberal application of exist-ing controls (first on banks and then on firms), wasvirtually completed in October 1995 with the enact-ment of a new Foreign Exchange Act.

Some restrictions did remain (mostly on the out-flow side but some on inflows), including restric-tions on the issuance of debt and money market se-curities abroad by residents and on foreign securitiestransactions executed through domestic agents.Moreover, the new Foreign Exchange Act enacted in1995, while codifying the framework for a liberalforeign exchange regime, included a provision underwhich the authorities could introduce an unremuner-ated reserve requirement on nonresident deposits ifnecessary.2

The remaining restrictions were eliminated withinthe framework of a plan agreed with the EU. In Feb-ruary 1996, the Czech Republic, in formally apply-ing for EU membership, adopted a tentative five-year plan for full capital account liberalization,which was revised in mid-1998 in the light of the1997 currency crisis (see below). The authoritieslifted the restrictions on foreign security transactionsin 1999, those on the issuance of debt securitiesabroad by residents in 2001, and the ban on pur-chases of agricultural land by nonresidents in 2002.In this manner, the Czech Republic virtually com-pleted the full liberalization of the capital accountbefore joining the EU in May 2004.

Policies to deal with capital flows

With the progress of capital account liberaliza-tion, a large amount of foreign capital flowed intothe country (Figure A1.1). The authorities respondedto the surge in inflows by taking various measures,including sterilization and increases in reserve re-quirements. Monetary policy assumed most of theburden of adjustment, however, and fiscal policy re-mained loose. In late 1994, the nature of capital in-flows evidently turned more speculative, with a shiftfrom long-term external borrowing by Czech com-panies to inflows of nonresident bank deposits. Astaff memorandum of November 1994 stated that theauthorities at this time considered introducing capi-tal controls, though the idea was overruled by the

A More Detailed Assessment ofSome Country Cases

APPENDIX

1

63

1The Czech Republic signed an association agreement with theEU in October 1993 and became the first transition country tojoin the OECD at the end of 1995. 2This provision has never been activated.


APPENDIX 1

prime minister. In 1995, the authorities introducedmeasures targeting short-term speculative inflows:3

a foreign exchange transaction fee in April and alimit on banks’ net short-term liability positions tononresidents in August.4 These measures, however,remained largely ineffective, with net private capitalinflows in 1995 amounting to as much as 15 percentof GDP.

In early 1996, capital inflows began to slow, coin-ciding with the decision by the authorities to widenthe fluctuation band of the exchange rate peg in Feb-ruary (from 0.5 percent on either side of parity to 7.5percent). Capital inflows continued to declinethroughout 1996, and a sharp reversal in 1997 culmi-nated in a currency crisis. Responding to the situa-tion, in April 1997 the authorities announced a pack-age of tight macroeconomic policies designed torestore internal and external balance. After a briefrespite, the currency soon began to fall again, forc-ing the authorities to abandon the exchange rate peg5

and to strengthen the macroeconomic policy pack-age in May 1997. In the event, these measures al-

lowed the Czech Republic to pull itself out of thecrisis rather quickly despite the turbulence in EastAsia, and the exchange rate stabilized after a modestdepreciation of 10 percent from the original parity.The country was relatively unaffected by contagionfrom the subsequent East Asian and Russian crises.

When the economy recovered in 1998, the authori-ties revised their strategy toward foreign capital flowsin favor of an explicit promotion of FDI. Recognizingthe weak corporate governance and low productivityof firms that had resulted from a mass (voucher) pri-vatization scheme, the authorities introduced mea-sures to encourage more concentrated shareholding,followed by attempts to sell the entities to foreignstrategic investors (OECD, 1996 and 1998). In addi-tion, they introduced a foreign investment incentivescheme and offered subsidies to qualified greenfieldprojects. Thanks to the new FDI incentive scheme, theCzech Republic again began to receive large capitalinflows, this time mainly driven by FDI, while thereremained a net outflow of short-term capital.

The IMF’s views of capital account measures

The process of capital account liberalization wasinitiated by the Czech authorities. As noted, it was inpart motivated by the country’s prospective acces-sion to the OECD and the EU, and the explicit roleof the IMF was consequently rather limited, exceptto endorse the decisions taken by the authorities. Thecountry’s first and only Stand-By Arrangement withthe IMF, agreed in 1993 and treated as precaution-ary, made no reference to capital account liberaliza-tion.6 The IMF, however, expressed a range of viewson capital account measures taken by the authoritiesthroughout the period.

In the early 1990s, in keeping with the predomi-nant thinking of the time, the IMF was clearly en-couraging the authorities to liberalize the capitalaccount rapidly. By 1994, however, the attitude ofthe area department staff had become more cau-tious, particularly as the banking sector weaknessescame to the surface. A briefing paper of May 1994supported the authorities’ decision to liberalize thecapital account “in a phased manner,” given theproblems in the banking system and the volatilityof capital flows.7 The area department’s views be-came even more cautious following the Mexican

64

–20

–15

–10

-5

0

5

10

15

20

Net private capital flows

(right scale;in percent of GDP)

022000989694921990

Figure A1.1. Czech Republic: Capital Account Openness and Net Private Capital Inflows1

Sources: IMF database; and Appendix 5.1Net foreign direct investment, net private portfolio investment, and net

private other investment.

Less

ope

nM

ore

open

Capital account openness(left scale, in percent)

0

50

100

Less

ope

nM

ore

open

3In early 1995, there were signs of informal tightening of capi-tal inflow controls. A staff memorandum, for example, noted asharp decline in approved capital transactions.

4The first measure was abolished in May 1997 and the secondmeasure in November of the same year.

5Inflation targeting was formally adopted in December 1997.

6The LOI included the intention of the authorities to “extend”current account convertibility by dismantling restrictions on cur-rent account transactions. In the context where the country hadnot yet accepted the obligations under Article VIII, the focus ofIMF support was necessarily placed on the liberalization of thecurrent account.

7Interestingly, the paper also noted that the authorities hadadopted a gradual approach to capital account liberalization, “fol-lowing a well-publicized debate.”


Appendix 1

crisis. A briefing paper of April 1995 expressed theview that “unduly quick liberalization [of capitaloutflows], for short-term easing of the capital in-flow pressure” would create vulnerability to crisisand proposed that the mission not press for moreambitious liberalization of capital outflows thanwas then envisaged.

MAE and PDR, however, remained sanguine.From late 1994 to early 1995, MAE, through itstechnical assistance work, continued to emphasizethe benefits of removing all remaining controls, cit-ing their ineffectiveness. In February 1995, a techni-cal assistance report stated that both administrativeand market-based capital controls had proved to beineffective except in the very short run, by appealingto the experience of other countries.8 Commentingon the April 1995 briefing paper, PDR opposed thecautious approach of the area department and arguedfor a continued liberalization of remaining outflowrestrictions. Management’s comments on the paperemphasized the need to be pragmatic in order toavoid an early reversal of policies.

A similar evolution can be seen in the attitude ofthe area department staff toward use of capital con-trols. By 1995, when capital controls reemerged as atopic for policy discussion, the staff’s earlier am-bivalence was gone. The staff report for the 1995 Ar-ticle IV consultation stated that “[given] the politicalconstraints on the early use of other instruments ofpolicy, the introduction of capital controls had be-come unavoidable.” It added, however, that capitalcontrols should be market based and nondiscrimina-tory across borrowers, and should be limited toshort-term inflows; and they should be seen as atemporary measure intended to buy time for morefundamental correction in policy, “given the likelyprogressive leakages over time” and the associatedallocative inefficiency.

The staff report for the 1996 Article IV consulta-tion took the view that capital controls on short-terminflows had “helped lengthen the maturity of banks’foreign liabilities,” though they had been “less suc-cessful in limiting the total volume of capital in-flows.” Judging from the comments on briefing pa-pers and minutes of Board discussions, it appearsthat management was more skeptical of capital con-trols. The Executive Board, on the other hand, wasgenerally more sympathetic, though several Direc-tors questioned the effectiveness of capital controlsand argued that use must be temporary.

Throughout much of this period, capital inflowswere an important topic of discussion between thestaff and the authorities.9 In late 1994, the staffthought that the inflows were being mainly driven bylarge interest rate differentials and demand by Czechcompanies for long-term credit. On the grounds thatthese factors reflected the banking sector’s limitedability to intermediate, it argued that the authoritiesshould not only further liberalize outflow controls butalso improve the banking sector. In the 1995 ArticleIV consultation discussions, the staff argued that therewas room for sterilization, because capital inflowswere no longer primarily driven by interest rate differ-entials. Throughout the period, the staff’s consistentposition was that tight fiscal policy was desirable.

As to exchange rate policy, the staff supported theauthorities’ policy of resisting any appreciation of theexchange rate.10 In fact, the staff did not advise theauthorities to increase exchange rate flexibility until2000. As noted earlier (Chapter 3, the section “Tem-porary Use of Capital Controls”), part of this re-flected the staff’s assessment that most of the gains incompetitiveness from the initial depreciation hadbeen lost by 1995, and that any nominal appreciationshould be avoided in order to maintain competitive-ness. The staff did mention that an exchange rateband could be a useful exit mechanism from the peg,but did not consider that a period of large capital in-flows should be the time to attempt an exit as itwould surely result in further appreciation. The staffreport for the 1995 Article IV consultation noted:“While the introduction of a band might providesome help in stemming the capital flows, it would re-sult in an immediate stepped-up real appreciation ofthe currency that would further worsen the externalbalance.”11 As late as 1999, the staff continued toargue that “any significant upward pressures on theexchange rate (potentially arising from substantialforeign direct investment inflows related in part tothe planned privatization) should be resisted.”

65

8Monetary and Exchange Affairs Department, “Issues Relatedto External Liberalization,” February 1995. The earlier back-to-office report noted that the mission had argued for a “more deci-sive” program of liberalization and helped “strengthen the handof those in favor of more rapid liberalization.”

9At least 12 staff memoranda were prepared on the subject be-tween 1993 and 1997.

10The staff report for the 1995 Article IV consultation offeredcommentary on the reluctance of the Czech authorities to allowthe exchange rate to appreciate: “They are very conscious of theexperience in the 1920s when revaluation of the currency in re-sponse to heavy capital inflows induced by successful stabiliza-tion was followed by an export slump and banking crisis that re-quired the Government to step in.”

11In discussing this report, however, some Executive Directorsdisagreed with the staff assessment and called for greater ex-change rate flexibility and an appreciation of the koruna. Thestaff report for the 1996 Article IV consultation notes that, fol-lowing the Board discussion, the staff adopted “that position dur-ing follow-up discussions with the authorities in November 1995,after taking into consideration revised data that mitigated con-cerns about export performance,” but it appears that the staff’ssupport for exchange rate flexibility disappeared quickly.


APPENDIX 1

Assessment

The Czech Republic relatively quickly lifted mostrestrictions on capital inflows while its fiscal policyremained expansionary, its banking system wasweak, and it maintained a fixed exchange rateregime. This experience can be contrasted to that ofHungary, which followed a more gradual and se-quential approach to capital account liberalization(see Box A1.1). A number of experts consulted bythe evaluation team have expressed the view thatthese factors, and not the mere fact that the capitalaccount was liberalized, explain why a currency cri-sis took place in the Czech Republic in the spring of

1997. This is not to minimize the risk of opening thecapital account with a weak banking system. As itturned out in the Czech Republic, the costs of bailingout the banking system amounted to more than 10percent of GDP.

The IMF staff initially encouraged the Czech au-thorities to liberalize the capital account rapidly butit soon recognized that there were weaknesses in thebanking system and pressed for banking reformfrom an early stage. The problem was that the staffdid not translate this recognition of the banking sec-tor weakness into operational advice on the pace andsequencing of capital account liberalization, for ex-ample, by suggesting that the authorities slow down

66

Box A1.1. Hungary

Hungary took a gradual approach to capital accountliberalization. It first liberalized FDI. Liberalization ofportfolio investment began toward the middle of the1990s in the context of OECD accession, which culmi-nated in the Foreign Exchange Act of 1995. It contin-ued, however, to maintain controls on short-term capi-tal inflows, notably restrictions on purchases ofshort-term domestic instruments by nonresidents andrestrictions on external lending to nonresidents by resi-dents in the domestic currency. The authorities clearlystated at that time that the final step to full capital ac-count convertibility should be taken only after the sys-tem of regulation and prudential supervision was firmlyin place for banking and securities market activities.This was achieved in June 2001, when Hungary elimi-nated the remaining restrictions and at the same timewidened the exchange rate band substantially (from2.25 percent to 15 percent).

The process was largely defined by the country’s po-litical agenda and institutional capacity. In line with theestablished practice, none of the three successive IMF-supported programs with Hungary in the 1990s in-cluded conditionality related to capital account liberal-ization. However, the LOI for the 1991 extendedarrangement included the authorities’ intention to en-courage foreign participation in the banking sector andin the privatization process. Likewise, the LOI for the1993 SBA referred to the prospective Foreign Ex-change Act, which would provide the basis for continu-ing “the process of liberalization of the trade and ex-change system”; the LOI for the 1996 SBA indicatedtheir intention to take measures to “facilitate a furtherliberalization of capital flows.” In this manner, the IMFsupported the country’s overall strategy of capital ac-count liberalization.

Some experts have argued that prolonged use of cap-ital controls explains the resilience Hungary demon-strated through the turbulent years of the late 1990s(Nord, 2003). The country was little affected by theEast Asian crisis, and it had little difficulty in manag-ing the contagion from the Russian crisis. Others, how-

ever, have emphasized the importance of sequencing(Ishii and others, 2002). Hungary began banking re-form from the late 1980s and allowed a significantpresence of foreign banks from the beginning (Szaka-dat, 1998). Moreover, the negative legacy of the social-ist era (characterized by large fiscal and externaldeficits) was much greater in Hungary than in othercentral European countries, causing the country to ex-perience an economic crisis relatively early in the tran-sition process, in late 1994. As a result, it was able toundertake many of the necessary but painful macro-economic and microeconomic reforms before embark-ing on capital account liberalization.

The Hungarian experience is instructive not only forthe sequencing the country took but also for illustratinghow the IMF viewed capital account issues during thisperiod. When there was a surge in short-term capital in-flows in early 1996, the IMF staff considered the maincause to be the large interest rate differential under anarrow crawling peg. When the authorities began toconsider introducing capital controls, the staff was di-vided on the issue. A staff memorandum of March1996 suggests that the area department, along withPDR and RES, took an accommodating view of mar-ket-based controls, while MAE was adamantly op-posed. When in early 1999 a law was enacted to allowthe introduction of a reserve requirement on nonresi-dent bank deposits, the IMF had already unified its po-sition. The staff unanimously supported the right of theauthorities to activate the measure in the event of asurge in short-term capital flows.1 After the economiccrisis of 1994, Hungary maintained reasonable fiscaldiscipline. This is why the IMF staff’s advice of fiscaltightening was not as consistent as in many other coun-tries. Rather, the staff consistently argued for a move togreater exchange rate flexibility.

1The reserve requirement was stipulated to be at below-market yields. In the event, this has never been activated, withthe rate maintained at zero.


Appendix 1

the pace of liberalization. The staff generally did notfavor greater exchange rate flexibility before the cur-rency crisis of 1997, in part owing to its view that thegains in competitiveness from the initial gains hadbeen eliminated and that any movement would be inthe direction of appreciation once the exchange ratewas allowed to float.

The staff’s country work in the Czech Republic re-flected the changing views of the risk of capital ac-count liberalization within the profession, much morequickly than the general policy guidelines emanatingfrom the IMF’s headquarters. Soon after the Mexicancrisis, even before its full implications were widelydiscussed, the staff began to take a cautious approachto further capital account liberalization, particularlyon the outflow side, and became more accepting ofmeasures to control short-term capital inflows.

In terms of policies to deal with large capital in-flows, the staff consistently advised the authorities totighten fiscal policy, but to no avail. The staff shouldhave given a greater warning about rapid capital ac-count liberalization and begun to advise alternativepolicy measures, such as greater exchange rate flexi-bility, when it was evident that its preferred policywas unlikely to be followed. In retrospect, the staffunderestimated the contribution of the fixed exchangerate to the capital inflows. The IMF staff should haveargued for greater exchange rate flexibility in the mid-1990s when the pace of speculative inflows pickedup, regardless of its fiscal policy advice.

Colombia: Market-Based Controls on Inflows

Colombia’s unremunerated reserverequirement

In September 1993, the Colombian authorities in-troduced an unremunerated reserve requirement(URR) on external borrowing. At the time, the author-ities were concerned with the large capital inflows,the pace of which had picked up as the authorities ac-celerated the opening of the economy (Figure A1.2).In 1991, for example, regulations on inward foreigninvestment were relaxed; in 1992, restrictions onlong-term external borrowing as well as purchases ofdomestic assets by foreign investment funds weresubstantially eased. Responding to the surge in capitalinflows, the authorities took various measures, includ-ing sterilization, further import liberalization, and eas-ing of capital outflow restrictions. The authorities alsoallowed the Colombian peso to appreciate somewhatand, in 1991, introduced a 3 percent withholding taxon transfers and nonfinancial private services (the ratewas increased to 10 percent in 1992), but the inflowsincreased sharply in 1993. It was under these condi-

tions that the authorities introduced the URR, whichwas modeled after a similar measure introduced ear-lier by Chile.

The URR was a system requiring that a desig-nated percentage of foreign loans with a maturity ofless than a designated maximum be kept as a depositat the central bank, at zero interest for a designatedholding period (Ocampo and Tovar, 1999). Alterna-tively, the deposit could be redeemed immediatelyby paying the equivalent cost calculated at apreestablished discount rate. Initially, exemptionswere made for certain loans and credit transactions,including loans to finance the import of capitalgoods, trade credits with a maturity of up to sixmonths, credits to finance investment abroad, andcapital inflows related to privatization and conces-sions. FDI was never subject to the URR.12

The URR was a market-based measure and, inthis respect, was very much in line with the coun-try’s overall strategy of economic liberalization. Infact, as the authorities introduced the URR, theyconcurrently eliminated most remaining administra-tive controls on capital movements. For example,surrender requirements on services and transfers

67

Figure A1.2. Colombia: Capital Account Openness and Net Private Capital Inflows1

Sources: IMF database; and Appendix 5.1Net foreign direct investment, net private portfolio investment, and private

other investment.

–2

–1

0

1

2

3

4

5

6

Net private capital flows(right scale; in percent of GDP)

022000989694921990Le

ss o

pen

Mor

e op

en

0

20

40

60

80

100Capital account openness

(left scale; in percent)

12This exemption created an incentive to substitute FDI fordebt, for example, by establishing a holding company in a tax-haven country. External borrowing by such holding companiesand their transfer of funds to Colombian firms were regarded asFDI.


APPENDIX 1

(except for interest and profits) were eliminated. Atthe same time, surtaxes on remittances of earningsfrom foreign investments were reduced, and ap-proval was granted to lending denominated in for-eign currency and hedging operations by foreign ex-change intermediaries.

Initially, the URR was set at 47 percent for oneyear, applicable to external borrowing with a matu-rity of up to 18 months. Depending on the strengthof capital inflows, the URR was tightened from timeto time (see Table A1.1 for details). During 1994, forexample, the applicable maturity was lengthened to60 months, and the share of a loan subject to theURR was raised to 43–140 percent.13 After someeasing in early 1996, in March 1997 the authoritiesraised the maturity of external borrowing subject tothe URR back to 60 months. In May 1997, the sys-tem was switched to a Chilean-style URR (in whichno maximum maturity is specified), with the loanshare of 30 percent for 18 months (except that, un-like the Chilean system, the deposit needed to bemade in domestic currency).14 Following the EastAsian crisis, in 1998, the authorities eased the URRand, in May 2000, reduced the applicable loan shareto zero.15

The IMF’s stance on the URR

Initially, IMF staff was not opposed to the URRitself. During a staff visit of early 1994, however, themission cautioned the authorities against introducingany additional control measures in the absence offiscal tightening. During the 1994 Article IV consul-tation, the mission took the view that the effects ofthe URR were likely to be limited in reducing infla-tion and preserving competitiveness. Thus, it arguedfor a tighter fiscal policy, further liberalizing tradeand reducing labor market rigidities. The staff alsosuggested that restrictions on external borrowingwere increasingly being circumvented, and thatthese could not only inhibit productive investmentbut also send a wrong signal to the markets. Accord-ing to the Summing Up of the Executive Boardmeeting, “A few speakers encouraged the authoritiesto remove the recent restrictions on external borrow-ing, but others considered that capital controls—de-spite their shortcomings— would be an acceptabletemporary response to capital inflows.”

At the time of a staff visit in mid-1995, the mis-sion argued that contagion from the Mexican crisishad been limited because of Colombia’s tight mone-tary policy stance and restrictions on external bor-rowing introduced in August 1994. In view of theemerging pressure on the foreign exchange and stockmarkets associated with a political crisis, the staffrecommended a tightening of both fiscal and mone-tary policies. During the 1995 Article IV consulta-tion, the staff argued that, given the slowdown in cap-

68

Table A1.1. Colombia:The Unremunerated Reserve Requirement, 1993–2000

Deposits________________________________________Maximum Applicable Holding period Applicable share

Effective Date Maturity (In months) (Percent of loan)

September 1993 18 12 47

March 1994 36 12 9318 6424 50

August 1994 60 Variable1 43–1402

February 1996 48 Variable1 10–853

March 1996 36 18 50

March 1997 60 18 50

May 1997 Eliminated4 18 30

January 1998 . . . 12 25

September 1998 . . . 6 10

May 2000 . . . . . . 0

Source: Banco de la República.1Corresponded to the maturity of the loan.2The maximum rate applied to loans with a maturity of 30 days or less.3The maximum rate applied to loans with a maturity of 180 days or less.4The URR was made applicable to all foreign loans irrespective of maturity.

13At the same time, all foreign investments unrelated to tourismor plant and infrastructure were prohibited.

14A senior official interviewed by the evaluation team gavesimplicity as the reason for the change.

15Although the share of a loan subject to the URR is currentlyset to zero, the URR as a control system still exists.


Appendix 1

ital inflows, the authorities could ease restrictions onexternal borrowing, while noting that if such a mea-sure were reinforced by fiscal tightening, it wouldlikely ease pressure on domestic interest rates. At theExecutive Board meeting, some Directors argued fora phased relaxation of the restrictions on externalborrowing, accompanied by a tighter fiscal policy.16

An intensification of political difficulties furtherdiminished the prospect for fiscal adjustment in1996, and the central bank came under intense pres-sure to ease liquidity conditions. The easing of theURR in early 1996 took place in this context. Thesituation, however, quickly changed as capital in-flows seemed to pick up again toward the end of theyear. In January 1997, the government announced apackage of “economic emergency” measures, in-cluding a tax on capital inflows (which would belevied in addition to the URR). When the tax wasruled unconstitutional by the Constitutional Court inMarch 1997, the authorities immediately respondedby tightening the URR.

In the briefing paper for the 1997 Article IV con-sultation, the staff expressed the view that the recur-rent use of capital controls in Colombia had servedmainly to crowd out the private sector and had onlybought time for the policymakers to strengthen thefundamentals; but that tightening of restrictions onexternal borrowing could temporarily help ease up-ward pressure on the real exchange rate and shift thecomposition of borrowing in favor of longer-termmaturities. During the consultation discussions, thestaff noted that the effectiveness of capital controlswas likely to be eroded over time and continued toargue for fiscal tightening, which would help createconditions for a gradual relaxation of the restrictionson external borrowing. At the Executive Board meet-ing, Directors emphasized that fiscal consolidationwas critical to avoiding a further real appreciation ofthe currency, while some Directors expressed con-cern over the recent broadening of restrictions onforeign borrowing.

The East Asian crisis of 1997–98 drasticallychanged the external environment faced by Colom-bia. Owing to pressure on the peso, the country lost asubstantial amount of foreign exchange reserves.During the 1998 Article IV consultation, the IMFstaff encouraged an elimination of the tax on profitremittances and other restrictions in order to pro-mote capital inflows. The staff and the Board wel-comed the easing of the URR and the subsequentfloating of the currency in September 1999.

Assessment

It is difficult to assess the effectiveness of theColombian URR because its intensity changed overtime and, while it was in place, fiscal policy was ex-pansionary and some administrative controls on ex-ternal borrowing remained.17 The IMF staff’s viewof the URR seemed to moderate over time. It wasnot initially opposed to the URR. This may have re-flected the staff’s understanding that the URR inColombia was a tool to manage the transition froman administrative control regime to a liberal one: theURR was introduced as a number of administrativecontrols were lifted.

The view then turned negative, and it was sometime before the staff recognized the potential use-fulness of the URR as a temporary measure, in linewith the evolution of the institution’s general think-ing on the temporary use of capital controls. TheExecutive Board, on the other hand, was more con-sistent, although the composition of views maywell have shifted: there were always some Direc-tors who recommended a relaxation of the URR,while there were also others who were more sup-portive of the measure. The gradual moderation ofthe staff’s negative view toward Colombia’s URR,however, failed to highlight the fact that it was infact used as a substitute for the needed correction infiscal policy.

In fact, IMF staff consistently advised fiscaltightening as the most effective measure to mitigatethe pressure on real appreciation created by thelarge capital inflows. Political constraints provedformidable, however, and did not allow the Colom-bian authorities to pursue successfully the IMF’spreferred strategy. When this became evident, IMFstaff could have suggested other policy options tocomplement fiscal tightening. Admittedly, thatwould have been a difficult task. To advise a tight-ening of the URR could have been counterproduc-tive and highly distortionary, given the loose fiscalpolicy. On the other hand, to advise an eliminationof the URR would not have served the purpose ofcontrolling the capital inflows. Greater exchangerate flexibility—going beyond the crawling band of7 percent introduced in January 1994—may wellhave been the only sensible policy alternative avail-able at the time, and the staff could have pushedharder for this policy. In the event, it only contin-ued to insist on fiscal tightening.

69

16The Colombian representative stated at the January 1995meeting that the restrictions on external borrowing would remainfor some time because, as in the case of Chile, they could helpdeter short-term speculative inflows and thereby moderate an ap-preciation of the currency.

17If one could isolate the impact of the URR from those ofother factors, it may well be that the URR was effective in limit-ing capital inflows and lengthening their composition when it wasintensely applied. Some recent research seems to support such aconjecture (e.g., Ocampo and Tovar, 1999 and 2003; Villar andRincon, 2003).


APPENDIX 1

Venezuela: Use of Capital Controls on Outflows

Controls on capital outflows in 2003

In February 2003, the Venezuelan authorities intro-duced temporary comprehensive foreign exchangecontrols on both current and capital transactions. Thedecision was made in an environment of great politi-cal uncertainty and economic difficulties, which hadresulted in large capital outflows (Figure A1.3), andfollowed an earlier decision to suspend foreign ex-change trading temporarily in late January. At thesame time, the exchange rate was fixed (to be deval-ued by 20 percent a year later), and price controlswere also introduced. This was the second timeVenezuela had introduced capital outflow controls inrecent years, the previous occasion being in June 1994(Quirk and others, 1995; Ariyoshi and others, 2000).18

The control regime worked in the following way.First, all foreign exchange proceeds needed to be sur-rendered to the central bank. Second, foreign ex-change was allocated, with priorities given to the im-port of basic goods, foreign debt service, andrepatriation of profits and dividends. In view of con-cerns over possible misuse of discretionary powers,

the authorities made it clear that the system of foreignexchange allocation would be managed in a transpar-ent manner. In addition, they stressed from the begin-ning that the imposition of foreign exchange controlswas a temporary emergency measure and that the con-trols would be gradually relaxed and eventually elimi-nated as foreign exchange earnings from state oil ex-ports were restored. The system was considerablyeased in 2004, when a wider coverage of transactionswas made eligible for foreign exchange allocation.

The 2003 system of foreign exchange controlswas designed to minimize the problems encounteredunder an earlier system. In the system introduced inJune 1994, the exchange controls also covered bothcurrent and capital account transactions; capital out-flows unrelated to the amortization of external debtand the repatriation of capital by nonresidents wereprohibited; foreign exchange earnings were to be sur-rendered to the central bank; and limits were set onthe allocation of foreign exchange for education,travel abroad, and family remittances, and for trans-fers of profits from certain investments. However,there was considerable evasion of capital controlsthrough permitted current account transactions. Thus,the authorities came to the conclusion that an effec-tive system of controls over capital account transac-tions required a more comprehensive arrangementfor monitoring all foreign exchange transactions.

How the IMF viewed the capital controls

According to a memorandum of February 2003,the staff thought that the new control regime wouldgive rise to a parallel foreign exchange market andan arbitrary system of foreign exchange rationingthat could be subject to political manipulation.19 Thebriefing paper for a subsequent fact-finding missionnoted the staff view that a flexible exchange rate sys-tem should be introduced and foreign exchange re-strictions on current transactions eliminated. Thestaff, however, believed that capital controls mightbe necessary in the short run in order to reduce pres-sure on the exchange rate, the balance of payments,and the banking system.

In evaluating the IMF’s views of the capital con-trols in Venezuela, it is important to keep in mindthat the control regime also covered exchange con-trols on current account transactions, which are sub-ject to IMF approval under Article VIII. In this re-spect, the position of the IMF staff in 2003–04 isstrikingly different from the position taken earlier.During 1994–96, the staff took the position that allexchange controls—both for current and capital

70

18The first occasion was also in the context of political uncer-tainty, accented by an evolving banking crisis and declining oilprices.

Figure A1.3. Venezuela: Capital AccountOpenness and Net Private Capital Inflows1



–20

–15

–10

–5

0

5

10

15

20

Net privatecapital flows

(right scale; in percent of GDP)

022000989694921990

Capital accountopenness

(left scale; in percent)

Less

ope

nM

ore

open

0

20

40

60

80

100

19A parallel foreign exchange market did emerge, in part sup-ported by residents’ access to the ADR market in New York.


Appendix 1

transactions—should be eliminated, along with therestoration of a flexible exchange rate regime.20

While it was willing to accept a gradual eliminationof capital controls as oil revenues were restored, itsview on the desirability of restoring full capital ac-count convertibility was unambiguous. In responseto the concern of the authorities that an eliminationof capital controls might lead to capital flight, thestaff stated that capital outflows would not result iftight fiscal policy was maintained21—a view broadlyendorsed by the Executive Board in its discussion ofthe 1994 Article IV consultation. In the context of anegotiation for a Stand-by Arrangement from 1995to 1996, the staff proposed a “two-stage approach in-volving the immediate liberalization of most currentaccount transactions accompanied by the gradualliberalization of capital transactions.”22

In contrast, in 2003, the IMF staff was much moreaccommodating of capital controls, though it firmlyopposed the maintenance of exchange controls forcurrent transactions. As part of an overall strategy toeliminate foreign exchange controls, the missionsupported a floating exchange rate mechanism forall current transactions, an export surrender require-ment to the domestic interbank market (as opposedto the central bank), and, if necessary, explicit con-trols on capital outflows. The staff argued that, giventhe aim of capital controls to reduce capital flight,controls should focus on capital outflows, whileother controls should be eliminated to reduce the ad-verse impact on domestic real activity.

During the 2004 Article IV consultation, the IMFmission urged the authorities to reinstate a floatingexchange rate regime and to remove all restrictionssubject to Article VIII or indicate a timetable fordoing so. The mission also argued for a gradualelimination of most capital controls, stressing thatthe process must be well sequenced and supportedby reforms in fiscal, monetary, and exchange ratepolicies, including the adoption of inflation target-ing. Given the comfortable international reserve po-sition, the mission urged the authorities rapidly toeliminate foreign exchange controls on all currentaccount transactions and to develop a strategy foreliminating “the majority” of capital controls. The

authorities in principle agreed with the staff position,though they preferred a somewhat slower pace ofliberalization and stressed the importance of politi-cal developments in determining the precise timing.The elimination of capital controls would come later,except perhaps for controls on short-term flows.

Assessment

The IMF’s position on the 2003 system differedfrom that on the 1994 system in an important re-spect. On both occasions, the IMF opposed the im-position of foreign exchange controls on currenttransactions, which are subject to IMF approvalunder Article VIII. In terms of controls on capitaloutflows, the IMF’s position in 2003 was much moreaccommodating. The IMF staff was willing to see atleast some of the control measures, particularly onoutflows, as part of Venezuela’s foreign exchangeregime.23 In 1994, on the other hand, the IMF hadargued for a full elimination of capital controls, al-though it was pragmatic enough to recognize thevirtue of gradual liberalization.

The positions taken by the staff on two occasionscertainly reveal how the views of the IMF on the useof capital controls has changed over the years. At thesame time, they also highlight the reluctance IMFstaff now seems to feel about expressing its positionforcefully on capital control issues. Not all capitalcontrols can be appropriate tools of economic policyunder all circumstances. The appropriateness of aparticular capital control measure must be judged onthe basis of an assessment of the overall macroeco-nomic policy and institutional framework underwhich it is introduced. In 1994, the staff had judgedthe capital controls to be inappropriate, given the un-sustainable macroeconomic policies, and argued thattheir elimination would not lead to capital flight ifsupported by tight macroeconomic policies. In 2003,no such assessment of the place of capital controls inthe overall macroeconomic policy framework wasoffered.

Tunisia: Gradual Liberalization


Tunisia has pursued a gradual approach to capitalaccount liberalization since the mid-1990s. Consid-erable de facto liberalization has taken place but, as

71

20The back-to-office report of July 1994 expressed the initialreaction of the staff that the price and exchange controls would“introduce serious distortions without really restraining exchangerate and price pressures.”

21From 1993 to 1994, Venezuela’s fiscal deficit deterioratedsharply to 7.3 percent of GDP from 2.9 percent of GDP; inflationincreased from 38 percent to 61 percent.

22In the event, in April 1996 the authorities eliminated all ex-change controls on current and capital account transactions aspart of an agreed economic program. In approving the program,Executive Directors commended the Venezuelan authorities onthe elimination of all exchange controls.

23In commenting on the draft report, the IMF staff stated thatits accommodative stance, as expressed in the 2004 staff report,referred only to the need to retain foreign exchange surrender re-quirements for the state oil company, and should not be taken as ageneral endorsement of permanent capital controls.


APPENDIX 1

indicated by the constant value of the index of dejure openness (Figure A1.4), restrictions remain onalmost all categories of capital account transactions,particularly for residents. In terms of speed and se-quencing, the experience of Tunisia is similar to thatof India (see Box 4.1 in the main text). Unlike thecase of India, however, capital account liberalizationin Tunisia has taken place in the context of extensiveIMF support, which included an assessment of its fi-nancial sector under the FSAP and technical adviceon sequencing.

The Tunisian authorities have been engaged in aprogram of broad economic liberalization since1986, aimed at reducing the role of government in thecountry’s economic activities (Nsouli and others,1993). The new development strategy has been deci-sively more outward looking and market oriented,and involved a gradual dismantling of restrictions ondomestic financial and international transactions.Macroeconomic stability was achieved, and prudentmacroeconomic policies have been maintained. Theprocess of financial liberalization and financial mar-ket development began in the late 1980s, followed bythe launch of banking sector restructuring in the early1990s. The trade regime was liberalized from the out-set of the reform process, with a focus on graduallyreducing quantitative restrictions on imports.

Along with trade liberalization, progress wasmade toward currency convertibility. In terms of cur-

rent account convertibility, the authorities graduallyeased foreign exchange controls by decentralizingthe allocation of foreign exchange, increasing ex-change allocations for invisible transactions, andeasing regulations on opening foreign currency ac-counts and using them for making payments abroad.Following these measures, in January 1993 Tunisiaaccepted the obligations under Article VIII of theIMF Articles of Agreement (Souayah, 1996).

A step-by-step approach was taken to the liberal-ization of capital account transactions. Among thefirst to be liberalized were transactions related to theresident export sector and nonresident FDI in certainexport-oriented sectors. However, many restrictionsremained on inward portfolio investment by nonresi-dents and outward capital transactions (except forexport-oriented activities and enterprises). It wasonly in 1995 that the Tunisian authorities initiated aconcerted effort to liberalize the capital account.24

The pace of capital account liberalization, how-ever, has been slow. In part, this reflected Tunisia’sconsensus-building culture. The authorities adopted aparticular liberalization measure only after it had re-ceived broad sociopolitical support and they were cer-tain that it would not be reversed.25 The back-to-officereport of November 2000 stated that the authoritieswere concerned about the risks of unstable portfolioflows. Yet the authorities have publicly stated on anumber of occasions that achieving full capital ac-count convertibility remains the official policy of thegovernment.26 This is still an ongoing process.

The IMF’s early views on Tunisia’s capitalaccount liberalization

The IMF supported Tunisia’s economic reformsthrough a series of financing arrangements, and hasmaintained a highly collaborative relationship withthe Tunisian authorities. Exchange of views on capitalaccount liberalization over the years has taken placewithin the context of this broad policy dialogue.

Responding to the authorities’ expressed intentionto pursue capital account liberalization, in June1995, the IMF staff suggested the following se-quence for removing restrictions on external capital

72

Figure A1.4. Tunisia: Capital Account Openness and Net Private Capital Inflows1



0

2

4

6

8

10

12

14

16

18

20

Net private capital flows(right scale; in percent of GDP)

022000989694921990

Less

ope

nM

ore

open

0

20

40

60

80

100Capital account openness(left scale, in percent)

24In the same year, Tunisia signed an Association Agreementwith the EU, which implied the eventual goal of full trade liberal-ization and capital account convertibility.

25A point emphasized in a March 1997 back-to-office report byan IMF mission.

26For example, see the statement of the central bank governor,“La Convertibilité Totale du Dinar,” May 24, 2001. In this state-ment, the governor listed five preconditions to be met before fullconvertibility was established: (1) strengthening of the country’sproductive capacity; (2) continuation of sound and sustainablemacroeconomic policies; (3) maintenance of a sustainable bal-ance of payments position; (4) a sound banking and financial sys-tem; and (5) sufficient international reserves.


Appendix 1

transactions: (1) FDI, (2) portfolio inflows, (3) for-eign borrowing, and finally (4) outward portfolio in-vestment. The staff also stressed the need to pursueprudent fiscal and monetary policies, encourage fi-nancial sector reform, and improve the domestic fi-nancial markets. Executive Directors agreed with thestaff and encouraged the authorities to take furthersteps toward full currency convertibility, while not-ing the preconditions to be satisfied. Both the staffand Executive Directors welcomed the associationagreement with the EU as a signal of the authorities’commitment to full integration with the global econ-omy. From 1996 through 1998, the staff clearly fa-vored a faster pace, with an immediate and completeliberalization of inward FDI, while acknowledgingthe importance of further financial sector reform. Abriefing paper of September 1998 added to the list ofpreconditions the privatization of the provision offorward foreign exchange cover.

The staff’s position turned clearly more cautiousin 1999. The briefing paper for the 1999 Article IVconsultation expressed the view that the restrictionson the capital account had helped Tunisia weatherthe East Asian crisis by limiting its exposure toshort-term capital flows, and hinted that the staffmight reconsider the whole liberalization strategy.Moving further in this direction, the briefing paperfor the 2000 Article IV consultation stated that lim-ited capital mobility had allowed monetary policyautonomy and “had served Tunisia well.” During theconsultation discussions, the staff cautioned againstsignificant further capital account liberalization untiladditional progress was made toward strengtheningthe banking system. While the staff continued tosupport the lifting of restrictions on outward FDI, itraised the fragile banking sector, undeveloped do-mestic financial markets, and uncompleted trade lib-eralization as reasons to argue against broader capi-tal account liberalization. Executive Directorsbroadly supported the staff’s position.

The IMF’s inputs in the ongoing process

Since 2001, the IMF has played a more explicitrole in Tunisia’s capital account liberalization. Inearly 2001, the Tunisian authorities requested an as-sessment of its financial sector by the IMF and theWorld Bank under the jointly administered FSAP.The resulting report, issued in November 2001,noted that the series of measures taken since theearly 1990s had significantly liberalized the regula-tory framework for capital account transactions infavor of export-oriented activities and corporations,but that other transactions, including inward directand portfolio investments, were still subject to alarge number of restrictions. The report then con-cluded that a wide-ranging capital account liberal-

ization would be premature before a “financial mar-ket economy” emerged. The 2001 FSAP assessmentsubsequently formed the basis for the IMF’s adviceon sequencing capital account liberalization.

In September 2001, the staff and the authoritiesagreed that the focus of the 2002 Article IV consul-tation discussions would be on (1) sequencing andpace of capital account liberalization and (2) ex-change rate policy in the context of capital accountliberalization.27 As agreed, the 2002 Article IV con-sultation involved intense discussions on capital ac-count liberalization, in which the authorities ex-pressed their preference for a gradual approach. Thestaff agreed, stressing the importance of proceedingwith caution, and encouraged the authorities to pre-pare ground work by (1) making further progress inestablishing a monetary policy framework that couldprovide a nominal anchor to inflation and exchangerate expectations, (2) allowing market conditions toplay a greater role in determining the exchange rate,and (3) reducing the existing vulnerabilities of thebanking system.28

The background paper prepared by the staff,drawing on the 2001 FSAP assessment, spelled outachievements and remaining challenges in the areasof (1) macroeconomic stabilization, including ex-change rate policy; (2) financial sector liberalization,including financial sector supervision; (3) systemicliquidity framework (meaning availability of market-based monetary policy instruments); (4) governmentsecurities market; (5) corporate sector restructuring;(6) legal framework; and (7) gaining access to inter-national capital markets (emphasizing the need to di-versify sources of balance of payments financing bygiving the private sector greater access to external fi-nancing). The paper then drew on previous work byMAE staff to argue that long-term capital accounttransactions needed to be liberalized before short-term movements; FDI inflows should be among thefirst categories of transfers to be liberalized; and thetransfer and use of domestic currency abroad shouldbe limited in the early phases of liberalization.29

In terms of specific sequencing, the backgroundpaper advocated a three-phase approach to capitalaccount liberalization. The first phase included stepsthat could be taken immediately, such as removingrestrictions on FDI by nonresidents and long-termloans to listed firms, and allowing limited nonresi-dent investment in local currency government secu-rities. The second phase involved steps that pre-

73

27Back-to-office report for a staff visit, September 20, 2001.28Briefing paper for the 2002 Article IV consultation.29The paper, entitled “Liberalization of the Capital Account in

Tunisia—Progress Achieved and Prospects for Full Convertibil-ity,” was included as Chapter 2 of the Selected Issues paper forthe 2002 Article IV consultation. SM/02/155, May 2, 2002.


APPENDIX 1

sumed the attainment of a solid banking system, aflexible exchange rate regime, and a market-basedmonetary policy framework, such as liberalizing out-ward FDI, portfolio investment by institutional in-vestors, and inward portfolio investment in debt in-struments. The third and final phase entailed stepsrequiring a robust financial sector and a resilient bal-ance of payment position, such as lending by resi-dents to nonresidents. Some Tunisian officials inter-viewed by the evaluation team indicated that theyfound the paper’s exposition of various country ex-periences to be particularly useful, although theoverall approach lacked operational specificity. Thebroad conceptual scheme of the staff advice, how-ever, received wide support when it was discussedby the Executive Board.

During the 2003 Article IV consultation, the staffreviewed progress under the capital account liberal-ization plan. The staff stressed the need for a floatingexchange rate regime to preserve independent mone-tary policy, while noting the importance of adoptinga “monetary reference target” to provide a nominalanchor as steps were taken to further liberalize thecapital account.30 Executive Directors welcomed theplanned move to a floating exchange rate regime butstressed that full capital account liberalizationshould be delayed until the proposed monetary pol-icy framework was established and the transition to afloating exchange rate complete.

Assessment

The case of Tunisia illustrates how the IMF hasapplied the new conceptual framework of sequenc-ing to a specific country context. Indeed, it is one of

only a few examples of the IMF’s new “integrated”approach in operation. The IMF staff has warned theauthorities of all potential risks involved in movingtoward full capital account convertibility; it has insome cases discouraged the authorities from movingfurther until banking sector problems were better ad-dressed; and it has spelled out how institutional andregulatory constraints could condition the pace andsequencing of removing restrictions. This new ap-proach has received wide support from the ExecutiveBoard and also appears to be accepted by most IMFstaff members.

The IMF staff has assessed Tunisia’s macroeco-nomic policies as being broadly prudent, and viewedits recent structural reforms as generally successful.At the same time, the staff considered the state of thebanking sector to be still insufficient to support afully open capital account. Given the government’spublicly stated commitment to achieving full con-vertibility of the dinar, however, one cannot avoidthe impression that the recent involvement of theIMF has not had much impact on the pace of capitalaccount liberalization, which began almost 10 yearsago. How quickly to liberalize the capital account(as well as how much risk to tolerate in the process)must remain the decision of a sovereign governmentand, given Tunisia’s consensus-building culture,there may not be much the IMF could have done toalter the pace of liberalization, in line with the com-mitments Tunisia made in its association agreementwith the EU. However, part of the problem may alsobe the lack of clear priorities in sequencing in theIMF’s new approach (see Box 4.3). By emphasizingall the potential interlinkages without identifying ahierarchy of risks, the integrated approach may havecreated an inevitable tendency to err on the side ofcaution. Despite the staff’s encouragement, and de-spite the government’s stated intention and interna-tional commitments, the process of capital accountliberalization has been painstakingly slow.

74

30In this monetary framework, base money was to serve as anoperating target. The system would be replaced by inflation tar-geting when conditions were met.


The IMF staff prepared more than 100 researchpapers on capital account issues between the

early 1990s and the early 2000s. The volume of re-search output in this area increased significantly inthe second half of the 1990s. The findings of staffresearch in this area broadly corresponded to theviews expressed in multilateral surveillance (seeChapter 2, the section “Multilateral Surveillance”),indicating that there was considerable synergy be-tween these two areas of activity. Research consis-tently found that permanent capital controls wereineffective, while staff research began to see the temporary use of capital controls in a more favorable light over time, at least as a short-term measure. The review of staff research pro-vided below is not meant to be comprehensive,but to cover only those studies that either reflectedor influenced the evolution of ideas within theIMF.

Early Work on Capital Controls andCapital Flow Management

Mathieson and Rojas-Suarez (1990), Mendoza(1990), and Calvo and others (1992) were amongthe first to analyze capital controls. Mathieson andRojas-Suarez (1990) showed that exchange ratepolicy would be affected by the removal of capitalcontrols as the economy would become more vul-nerable to foreign shocks, but that there was no sin-gle optimal exchange rate regime consistent with aparticular process of liberalization. Mendoza’s the-oretical study (1990) showed that the use of capitalcontrols had little, if any, impact on the output,consumption, and welfare of a small open economyfacing balance of payments problems. Calvo andothers (1992) argued that a case could be made for the policy mix of a tax on short-term inflows,exchange rate flexibility, and an increase in mar-ginal reserve requirements, and noted that capitalcontrols could be effective only in the short run because investors could find a way to evade themover time.

Two significant policy-oriented papers were is-sued as Occasional Papers during 1993.1 First,Mathieson and Rojas-Suarez (1993) advanced theidea that capital controls had lost effectiveness in the1980s with the liberalization of exchange and tradecontrols. They identified channels of evasion such asunder- and over-invoicing, transfer pricing policies,and leads and lags. This does not mean that capitalcontrols cannot affect certain types of capital trans-actions and market participants, but the authors ar-gued that, given the distortionary effects, adjustmentof macroeconomic policies was generally more ap-propriate than imposition of capital controls whenfaced with large capital movements. They then con-cluded that, in order to support capital account con-vertibility, efforts should be made to strengthen theprudential supervision of financial institutions, es-tablish more flexible interest rates, and restructureand recapitalize domestic financial institutions. The“consistency of macroeconomic, financial, and ex-change rate policies is more important for sustainingan open capital account than is the sequencing of theremoval of capital controls.”

The other Occasional Paper, by Schadler and oth-ers (1993), was an analysis of how countries had re-sponded to surges in capital inflows. In particular, itused the recent experiences of Chile, Colombia,Egypt, Mexico, Spain, and Thailand to document thepolicies adopted and the effectiveness of these mea-sures. It argued that tight fiscal policy was the onlymeans to prevent overheating and avoid a real appre-ciation “regardless of [the] cause” of the inflows. Itsassessment of sterilization, the most common policytool, was generally negative because its quasi-fiscalcost and its effect on the level of interest rates madeit infeasible on a sustained basis. The authors werecautious toward exchange rate flexibility because a

Staff Research

APPENDIX

2

75

1Compared with Working Papers, Occasional Papers tend to bemore department driven and less individually motivated, and havegreater internal status and outside visibility. Some Occasional Pa-pers are initially written as Board papers and are discussed by theExecutive Board in a formal meeting or an informal seminar be-fore they are published.


APPENDIX 2

change in the equilibrium real exchange rate mightnot be warranted. They recognized a case for capitalcontrols when “bandwagon effects are important orthere are doubts about the capacity of the economyto absorb inflows efficiently,” but found little evi-dence to argue for their effectiveness. Instead, theyargued that the easing of the external constraint pro-vided an ideal opportunity to address structuralweaknesses by liberalizing trade, moving towardcapital account convertibility, and reforming the fi-nancial sector.

Later Work on Capital Controls

Studies that appeared in 1994 and later reinforcedthe argument that capital controls were ineffective.For example, Johnston and Ryan (1994) argued thatcapital controls were not effective in developingcountries, and caused problems in macroeconomicmanagement with little effect on the balance of pay-ments. The authors then advocated rapid capital ac-count liberalization, given its positive impact on cap-ital inflows and domestic financial development. Areview of theoretical and empirical literature byDooley (1996) concluded that controls were some-what effective in creating a wedge between domesticand international interest rates, but there was littleevidence to show that they were effective in signifi-cantly affecting the volume of capital flows. At thesame time, the study noted that capital controls pre-viously employed by many industrial countries hadbeen effective (relative to developing country experi-ence), and concluded that administrative capacitywas a critical factor in determining the effectivenessof controls. Once the apparatus of control was re-moved, however, reintroducing controls in a liberal-ized regime would be unlikely to be effective.

As the experience of Chile with market-basedcontrols became widely known (see Boxes 1.2 and2.2), some on the IMF staff began to see temporaryuse of controls in a more favorable light. In 1996,Galbis (1996) argued that there were grounds for thetemporary use of a tax on capital inflows, while not-ing that quantitative controls on capital flows wereinefficient and discriminatory and should be the firstto be removed. Laurens and Cardoso (1998), how-ever, stressed that Chilean-style controls could be apolicy option only for a limited number of develop-ing countries because of the high level of enforce-ment capacity required for its implementation. Onthe other hand, Lopez-Mejia (1999) argued that thecapital controls in Chile, Colombia, and Malaysiahad proved useful in lengthening the maturity ofcapital inflows.

Determinants of capital controls received some at-tention in IMF research. The seminal work of Grilli

and Milesi-Ferretti (1995) used a large sample of over60 countries to find that capital controls were morelikely to be present in a country if it was less open, itsincome lower, its public sector larger, its central bankless independent, its exchange rate less flexible, andits current account deficit larger. The authors foundlittle evidence that capital controls were associatedwith higher economic growth, but controls tended tobe associated with higher inflation and lower real in-terest rates. Likewise, Johnston and Tamirisa (1998)identified additional factors to explain the impositionof capital controls by governments, including balanceof payments reasons, macroeconomic management,weak domestic regulatory systems, and the stage ofeconomic development.

Work on Sequencing

As early as 1994, staff research, while supportingcapital account liberalization, was already aware ofthe need for sequencing, which was well knownfrom the literature on the order of economic liberal-ization. For example, Quirk (1994) argued that capi-tal account liberalization should be implementedwith credible fiscal policy. Galbis (1994) argued that“a pragmatic approach to the sequencing issue [was]necessary as there [were] only a few general princi-ples valid for all countries.” He added that a casecould also be made from the literature that an earlyintroduction of capital account liberalization in thereform process could promote acceleration of do-mestic financial reforms. The conventional wisdomfrom the literature was reiterated by the previouslycited work of Galbis (1996), who listed fiscal con-solidation, noninflationary finance of public deficits,macroeconomic stability, an appropriate monetary-fiscal policy mix, and a strong domestic financialsector as preconditions for capital account liberaliza-tion. Surprisingly, however, exchange rate flexibilitywas not accorded the same emphasis it receivestoday as desirable for an open capital account.2

An Occasional Paper by Quirk and others (1995)was much more explicit on sequencing. The paperincluded the idea that one must consider a set ofpreconditions and the sequencing of liberalizationin moving toward capital account convertibility,and highlighted the danger of opening the capitalaccount too rapidly without supporting policies. Itthen noted that the most important preconditionwas domestic financial market reforms, including

76

2An earlier expression of the view intimating the need for ex-change rate flexibility under high capital mobility is found inGoldstein and Mussa (1993), who argued that greater capitalflows have “made the conditions more demanding for operatingdurably and successfully a fixed exchange rate arrangement.”


Appendix 2

strengthened prudential regulations. In terms of se-quencing, it suggested that (1) with a strong bal-ance of payments position, exchange rate pressurecould be minimized by liberalizing capital outflowsbefore inflows; and (2) one might also want to limitpotentially more destabilizing short-term inflowsby first liberalizing long-term inflows, such as di-rect investment. The authors, however, added that“such fine-tuning” might be difficult in practice as“liberalization of one component of the capital ac-count” would create pressure to liberalize all capitaltransactions.

Toward the end of the 1990s, even before the EastAsian crisis, staff research began to focus on thepace and sequence of capital account liberalizationin a more explicitly operational way. Johnston andothers (1997) documented the sequence of financialsector reforms and capital account liberalization fol-lowed by Chile, Indonesia, Korea, and Thailand, andsuggested that the speed should depend on macro-economic and exchange rate policies. Likewise,Johnston (1998) argued that prudential measuresshould not be considered to be equivalent to capitalcontrols because they were not meant to restrict cap-ital flows directly, but were designed to support thegains achieved in moving toward capital accountconvertibility by providing safeguards. These andother contributions were later compiled as a book,which was published by the IMF (Johnston and Sun-dararajan, 1999). An influential Occasional Paper byEichengreen and others (1998) discussed the role ofsequencing in a broader context of discussion on therisks of capital account liberalization and the needfor sound macroeconomic and prudential policies tominimize those risks.

In the early 2000s, there was a proliferation ofwork on pace and sequencing. For example, Kara-cadag and others (2003) considered hierarchy andinterlinkages among financial markets, and made aproposal on the modality of sequencing. In particu-lar, the authors emphasized the importance of under-taking central banking reforms and other measuresthat would allow a more effective conduct of mone-tary and exchange rate policies, and the need to im-plement technically and operationally connectedmeasures simultaneously. Kaminsky and Schmukler(2003) were skeptical of the need to follow a particu-lar order of liberalization, but nevertheless acknowl-edged the importance of doing institutional reformsbefore opening the capital account. Duttagupta andothers (2004) used country experience to argue thatattaining exchange rate flexibility before capital ac-count liberalization had the advantage of enabling

the economy to absorb capital account shocks at alower cost to the real economy. The authors also ar-gued that a transition to exchange rate flexibilityshould involve a gradual elimination of existingasymmetries (if any) in capital account openness be-tween outflows and inflows in order to facilitate anorderly correction of any potential misalignment inthe exchange rate.

More Recent Work

The areas of research on capital account issuesalso expanded in the early 2000s. We review here twostrands of research covering (1) the impact of capitalaccount liberalization and (2) analyses of market dy-namics. First, among recent studies to quantify theeffect of capital account liberalization on economicgrowth or policy discipline, Edison and Warnock(2003) supported the view that removal of restric-tions provided developing countries with increasedaccess to international capital markets, but found noevidence that capital controls created a bias in favorof domestic capital. An Occasional Paper by Prasadand others (2003) found no strong relationship be-tween capital account openness and growth (but sug-gested the importance of the quality of domestic in-stitutions in defining that link), while Tytell and Wei(2004) suggested no robust or causal relationship be-tween liberalization and fiscal discipline (althoughthere was a weak discipline effect on inflation).

A number of recent studies have investigated theworking of financial markets, particularly as it re-lates to international linkages through capital flows.For example, Arora and Cerisola (2001) provided aquantitative indication of how U.S. monetary policyinfluenced sovereign bond spreads in emerging mar-ket economies, and concluded that the spreads wereinfluenced not only by country-specific fundamen-tals but also by the stance and predictability of U.S.policy. Herding among international institutional in-vestors was the topic of empirical studies by Boren-sztein and Gelos (2000) and Gelos and Wei (2002); aliterature review on herd behavior was provided byBikhchandani and Sharma (2001). More recently,Chan-Lau (2004) analyzed, among other things, themain determinants of the emerging market asset al-location of pension funds in industrial countries,while Ong and Sy (2004) showed the importance offoreign investor presence in securities markets inemerging market economies and how asset alloca-tion decisions by mature market funds could possi-bly affect emerging market countries.

77


This appendix reviews public speeches and state-ments of IMF management during 1990–2004,

in order to see what messages were communicatedto the public on capital account issues. Much of theinformation in this appendix relies on various issuesof the IMF Survey.

In the early 1990s, IMF management viewedcapital account liberalization, along with macroeco-nomic discipline and IMF financial support, as essential ingredients of sustained growth for devel-oping countries. Management, however, was ex-plicit in spelling out the potential risks of capital ac-count liberalization. In 1994, for example, theManaging Director stated: “The Fund encouragescountries to liberalize their capital account restric-tions, while adopting policies that ensure that therisks involved are avoided and the potential benefitsfully realized.”1

Following the Mexican crisis, management fo-cused on the need for strong financial institutions, acompetitive domestic financial system, and effectivesupervision and regulation; it opposed use of capitalcontrols, including market-based ones. In 1995, theManaging Director stated that the IMF’s response tothe challenges of globalization was to strengthensurveillance and to secure appropriate resources toassist countries. Surveillance needed to be strength-ened, particularly in terms of attention to capital ac-count developments and financial flows. At a semi-nar held in April 1995, the First Deputy ManagingDirector said that the pace of capital account liberal-ization depended on the liberalization process of thedomestic financial sector and that a strong financialsystem was a prerequisite.

In September 1995, in responding to criticismsthat the IMF was an impediment to capital accountliberalization, the Managing Director wrote an arti-cle for the Wall Street Journal emphasizing thatfreedom of capital movements is “an objective thatthe IMF seeks to promote.” At the same time, hestated that, in the absence of certain prerequisites,

“open capital accounts may impose considerablecosts in terms of financial and economic instability,and risk costly reversal” and listed as the necessaryprerequisites a strong financial system and macro-economic stability. He then noted that, in the cir-cumstances of some developing countries, “certainkinds of measures to discourage capital inflows orinfluence their character might be appropriate”(Wall Street Journal, September 27, 1995).

In 1997, there was a marked change in manage-ment’s view of capital controls. While fiscal disci-pline and greater exchange rate flexibility remainedthe preferred policies, the First Deputy ManagingDirector stated that market-based controls were lessharmful than administrative ones, which were inef-fective and costly. He continued to advocate liberal-ization of outflows as a tool to manage capital in-flows. In 1998, he again reiterated the same views,namely, that controls on outflows should be removedas the country circumstances became appropriate,but market-based controls could be retained to dis-courage short-term inflows.

At the same time, management began to paymore attention to sequencing and gradualism. TheManaging Director emphasized the importance ofsound macroeconomic policies, a strong domesticfinancial system, phased capital account liberaliza-tion, properly sequenced reforms, and timely andaccurate dissemination of information. At a meetingof the Pacific Basin Economic Council held in May1999, the Managing Director stated that controlswere more effective on inflows than on outflows,and that they worked best when they were market-based and temporary. He then added that strongermacroeconomic policies and banking sectors—notthe controls per se—were the key factors behind thesuccess of the countries that imposed controls afterthe crisis.

The Managing Director, at the January 2001Asia-Europe Meeting of finance ministers fromAsia and Europe, conceded that there had been ex-cessively rapid capital account liberalization insome emerging market countries, and emphasizedthe need for preconditions to be met before pro-ceeding with full liberalization. At the same time,

Public Communications

APPENDIX

3

78

1Transcript of remarks made at a meeting of financial and busi-ness leaders in Korea in October 1994.


Appendix 3

he advised countries with open capital accounts notto reimpose controls, but rather to strengthen insti-tutions. He then noted the mixed experience withthe use of capital controls and called for “furtherresearch and analysis to assess the costs and bene-fits of controls in particular circumstances.” In

2003, at the January Asia-Pacific Economic Coop-eration meetings, the Managing Director stressedthe need for sequencing, saying: “Ensuring carefulsequencing, particularly in relation to the develop-ment of well-regulated and well-managed financialsectors, is a critical ingredient to success.”

79


Cross-Border CapitalTransactions and CapitalAccount Openness in SelectedCountries, 1989–2002

APPENDIX

4

80

Private Capital Flows1 (Annual Averages)____________________________________________________________________________________________________________As a percent of GDP In billions of U.S. dollars_____________________________________________________ _____________________________________________________

1989–93 1994–97 1998–2002 1989–93 1994–97 1998–2002_________________ ________________ _________________ ________________ ________________ _________________Liabilities Assets Liabilities Assets Liabilities Assets Liabilities Assets Liabilities Assets Liabilities Assets

Average 2.9 –0.3 6.1 –0.5 5.4 –1.4 2.4 –0.4 5.9 –1.7 4.9 –2.7

Bulgaria 0.40 –2.34 1.39 –4.76 7.62 0.71 0.76 –0.04 0.15 –0.42 1.05 0.11Chile 6.50 0.16 8.86 –0.76 7.15 –5.03 2.60 0.03 6.30 –0.57 5.19 –3.68China 2.38 –1.24 6.55 –1.53 4.21 –3.05 12.09 –5.80 48.43 –12.47 45.85 –31.74Colombia 1.36 –0.58 0.85 –1.56 2.61 –1.72 0.83 –0.31 0.91 –1.53 2.23 –1.48Croatia 1.66 –0.93 6.01 2.23 11.37 0.22 0.10 –0.11 1.18 0.43 2.30 0.09Czech Republic 1.44 0.71 7.54 –3.53 7.42 0.66 0.61 0.22 4.10 –2.06 4.67 0.55Estonia 1.55 0.00 16.50 –6.41 12.95 –5.21 0.02 0.00 0.69 –0.26 0.77 –0.31Hungary 4.44 –0.28 7.36 –1.09 9.38 –1.01 1.66 –0.08 3.21 –0.51 4.65 –0.52India 1.81 0.50 2.40 –0.04 2.20 –0.15 5.24 1.41 8.58 –0.14 10.16 –0.74Israel 2.49 –2.14 5.02 –1.09 5.33 –4.36 1.41 –1.27 3.97 –0.77 5.77 –4.72Latvia 3.18 0.00 14.79 –5.27 12.98 –4.15 0.21 0.00 0.78 –0.31 1.03 –0.34Lebanon 1.20 4.81 –0.46 32.30 –1.93 15.51 0.00 0.36 –0.22 3.98 –0.33 2.58Lithuania –2.15 0.00 7.77 –1.22 9.33 –0.89 –0.06 0.00 0.65 –0.11 1.09 –0.10Malaysia 11.12 –0.60 6.73 –2.57 –5.32 –1.51 6.44 –0.52 6.38 –2.43 –4.49 –1.30Mexico 5.52 –1.17 4.64 –0.87 2.95 0.26 18.24 –3.20 16.47 –2.83 16.13 1.80Peru 1.61 0.45 8.35 –0.50 1.55 0.42 0.55 0.16 4.41 –0.24 0.86 0.22Philippines 2.73 0.58 8.20 1.82 5.20 –2.35 1.31 0.26 6.21 1.46 3.83 –1.73Poland –2.29 –1.88 –1.17 0.92 6.30 –1.13 –2.21 –1.24 –0.24 1.69 10.83 –1.93Romania 1.34 –0.27 4.52 –0.52 4.95 0.10 0.16 –0.08 1.54 –0.16 2.05 0.04Russia 0.17 0.54 4.18 –5.45 2.57 –6.25 –1.38 1.00 16.08 –19.92 6.44 –16.33Slovak Republic 0.71 0.59 11.82 –2.72 11.51 –1.11 0.12 0.09 2.40 –0.60 2.53 –0.21Slovenia –0.21 –0.71 3.78 –0.96 6.75 –2.76 0.02 –0.10 0.73 –0.18 1.40 –0.58South Africa 0.02 –0.70 5.65 –3.05 4.69 –3.35 0.04 –0.84 8.31 –4.48 6.22 –4.43Thailand 10.47 –0.03 5.69 0.24 –4.53 –0.84 9.89 0.13 9.92 0.28 –5.30 –0.95Tunisia 6.91 –0.24 9.73 –1.63 12.85 –3.03 0.99 –0.03 1.76 –0.30 2.60 –0.61Ukraine 4.31 0.00 3.14 0.13 2.48 –2.10 1.60 0.00 1.12 0.06 0.88 –0.75Venezuela 8.39 –3.51 4.62 –4.62 2.52 –6.68 4.35 –1.81 4.19 –3.43 2.55 –6.94

Sources: IMF, WEO and other IMF databases.1Portfolio investment, other private investment, and foreign direct investment. Excludes government borrowing.2Taken from IMF, Annual Report on Exchange Arrangements and Exchange Restrictions, various issues. The index indicates the number of restricted categories divided by

the total number of capital control categories. A smaller number means a smaller number of existing restrictions on capital account transactions.


Appendix 4

81

Memorandums_________________________________Capital Account Openness2

Technical____________________________________

Global IMF-supported assistance in banking ranking Openness index programs and external sectors__________________________in 2002 1995 2002 Change (1990–2002) (1990–2002)

81 0.7 0.6 –0.1

97 0.9 0.8 –0.1 Yes Yes58 1.0 0.4 –0.6 Yes Yes

132 0.9 1.0 0.1 No Yes114 0.9 0.9 0.0 Yes Yes114 0.9 0.9 0.1 Yes Yes44 0.6 0.3 –0.4 Yes Yes44 ... 0.3 ... Yes Yes1 0.8 0.0 –0.8 Yes Yes

132 1.0 1.0 0.0 Yes Yes1 0.8 0.0 –0.8 No No

31 0.2 0.2 0.0 Yes Yes83 0.3 0.7 0.5 No Yes31 0.4 0.2 –0.2 Yes Yes

114 0.8 0.9 0.1 No Yes83 0.9 0.7 –0.2 Yes Yes1 0.2 0.0 –0.2 Yes Yes

114 0.9 0.9 0.0 Yes Yes83 0.9 0.7 –0.2 Yes Yes73 0.8 0.6 –0.2 Yes Yes97 0.8 0.8 0.0 Yes Yes

132 0.7 1.0 0.3 Yes Yes73 0.8 0.6 –0.3 Yes Yes

114 0.9 0.9 0.0 No Yes97 0.6 0.8 0.2 Yes Yes

114 0.9 0.9 0.0 Yes Yes114 1.0 0.9 –0.1 Yes Yes83 0.3 0.7 0.4 Yes Yes


Capital Account Openness in 12Sample Countries, 1990–2002

APPENDIX

5

82

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

India

Czech Republic

Chile

Latvia

Hungary

Colombia


Appendix 5

83

Sources: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER); Miniane (2004); and IEO estimates.

1The index shows in percentage terms how many of the 10 types of capital account transactions are subject to restrictions; a lower value indicates greater capital account openness.

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

0

20

40

60

80

100

120

022000989694921990

Tunisia

Mexico

Lithuania

Venezuela

Thailand

Malaysia


The IEO team has spoken to more than 30 cur-rent and former members of IMF staff and the

Executive Board. In addition, the following individ-uals have provided their views to the IEO, mostlythrough personal interviews but also through semi-nars and workshops. We express our gratitude fortheir time and apologize for any errors or omissions.They assume no responsibility for any errors of factor judgment that may remain in the report.

Former and current officials of internationaland regional organizations

Bank for International Settlements

David Archer Benjamin CohenMadhu Mohanty Ramon MorenoPhilip Turner Augustin Villar

European Commission

Maria-Rosario Areizaga Maurice-Pierre GuyaderKen Lennan Oliver SchmalzriedtHeliodoro Temprano

Organization for Economic Cooperation andDevelopment

Robert Ley Paul O’BrianPierre Poret

United Nations Conference on Trade andDevelopment

Yilmaz Akyuz Andrew CornfordHeiner Flassbeck Benu Schneider

Current and former officials of membercountries

Chile

Jorge Cauas Lama Luis Eduardo EscobarNicolas Eyzaguirre Ricardo Ffrench-DavisLeonardo Hernandez Luis Oscar Herrera

Esteban Jadresic Manuel MarfanCarlos Massad Felipe Morande Bernardita Piedrabuena Klaus Schmidt-HebbelClaudio Soto Kathleen UribeRodrigo Valdes

Colombia

Gerardo Hernandez Salomon KalmanovitzGuillermo Perry Jose Antonio OcampoHerman Rincon Jose Dario UribeMiguel Urrutia Leonardo Villar

Czech Republic

Oldrich Dedek Vladimir DlouhyTomas Holub Pavel MertlikJan Miladek Petr ProchazkaPetr Sedlacek Pavel StepanekJoseph Tosovsky Jiri Vetrocsky

Hungary

Laszlo Akar Lajos BokrosAkos Peter Bod Sandor DavidTibor Erhart Klara KamarasAgota Repa Gorgy SuranyiGyorgy Szapary

India

Montek Singh Ahluwalia R. BannerjiSurjit Bhalla Himadri BhattacharyaShyamala Gopinath Sujan HajraNarendra Jadhav F.R. JosephAshok Lahiri Rakesh Mohan Kirit S. Parikh Rajiv RanjanY.V. Reddy Ajay ShahV.K. Sharma S.S. Tarapore

Latvia

Helmuts Ancans Uldis CerpsJuris Kravalis Andris LiepinsZoja Medvedevskiha Einars RepseUldis Osis Guntis Valujevs

List of Interviewees

APPENDIX

6

84


Appendix 6

Mexico

Pedro Aspe Samuel AlfaroAlfonso Guerra Javier GuzmanDavid Madero Suarez Javier MaldonadoMiguel Mancera Roberto MarinoManuel Ramos Francia Julio SantaellaAlberto Torres Jesus Marcos YacamanAlejandro Werner

Tunisia

Badreddine Barkia Negib BouselmiSamir Brahimi Habib El Montacer SfarHabib Essafi Chedly EzzaouiaSamira Ghribi Brahim HajjiBelhadj Jameleddine Golsom JaziriAloui Messaoud Mohamed RekikAli Ridha Ben Achour Monia SaadaouiAbdelhamid Triki

Other countries

Bryan Chapple Lorenzo Bini SmaghiMarco Committeri Giorgio GomelIgnazio Visco Vicenzo Zezza

Academics and other individuals

Jacques Ardant Suman BeryTahar Ben Marzouka Slah E. BouguerraMauricio Cardenas J.B. ChandradharaWilliam Cline Michael DooleyChristian Gardeweg Morris GoldsteinLaszlo Halpern Joseph JoyceFaycal Lakhoua Rohini MalkaniAshwini Mehra Abdelhamid MiladiChristian Moreno Michael MussaAditya Narain Indranil PanBandi Ram Prasad Shubhada RaoRicardo Rocha Girts RungainisMohan Shenoi Kanhaiya SinghEdwin Truman Fabio Villegas RamirezJohn Williamson

85


Abdelal, Rawi, 2005, “Freedom and Its Risks: The IMFand the Capital Account,” Working Paper Series No.05-056 (Cambridge, Massachusetts: Harvard Busi-ness School).

———, and Laura Alfaro, 2003, “Capital and Control:Lessons from Malaysia,” Challenge, Vol. 46, No. 4(July–August), pp. 36–53.

Ahluwalia, Montek S., 2002, “Economic Reform in IndiaSince 1991: Has Gradualism Worked?” Journal ofEconomic Perspectives, Vol. 16 (Summer), pp. 67–88.

Ariyoshi, Akira, Karl Habermeier, Bernard Laurens, InciOkter-Robe, Jorge Ivan Canales-Kriljenko, and An-drei Kirilenko, 2000, Capital Controls: Country Ex-periences with Their Use and Liberalization, IMFOccasional Paper No. 190 (Washington: InternationalMonetary Fund).

Arora, Vivek, and Martin Cerisola, 2001, “How Does U.S.Monetary Policy Influence Sovereign Spreads inEmerging Markets?” IMF Staff Papers, Vol. 48,pp. 474–98.

Arteta, Carlos, Barry Eichengreen, and Charles Wyplosz,2001, “When Does Capital Account LiberalizationHelp More Than It Hurts?” NBER Working PaperNo. 8414 (Cambridge, Massachusetts: National Bu-reau of Economic Research).

Asian Policy Forum (APF), 2002, “Policy Proposal for Sequencing the PRC’s Domestic and External Finan-cial Liberalization” (Tokyo: Asian DevelopmentBank Institute).

Bakker, Age F.P., and Bryan Chapple, eds., 2003, CapitalLiberalization in Transition Countries: Lessons fromthe Past and for the Future (Cheltenham, UnitedKingdom, and Northampton, Massachusetts: EdwardElgar).

Bikhchandani, Sushil, and Sunil Sharma, 2001, “Herd Be-havior in Financial Markets,” IMF Staff Papers, Vol.47, pp. 217–310.

Blejer, Mario I., and Fabrizio Coricelli, 1995, The Makingof Economic Reform in Eastern Europe: Conversa-tions with Leading Reformers in Poland, Hungary,and the Czech Republic (Aldershot, Hants, UnitedKingdom, and Brookfield, Vermont: Edward Elgar).

Bloomfield, Arthur I., 1946, “Postwar Control of Interna-tional Capital Movements,” American Economic Re-view, Vol. 36, No. 2, pp. 687–709.

Borensztein, Eduardo R., and R. Gaston Gelos, 2000, “APanic-Prone Pack? The Behavior of Emerging MarketMutual Funds,” IMF Working Paper No. 00/198(Washington: International Monetary Fund).

Boughton, James M., 2002, “Why White, Not Keynes? In-venting the Post-War International Monetary Sys-tem,” in The Open Economy Macromodel: Past, Pre-sent and Future, ed. by Arie Arnon and Warren Young(Boston: Kluwer Academic Publishers).

Buiter, Willem H., and Anne C. Sibert, 1999, “UDROP: AContribution to the New International Financial Archi-tecture,” International Finance, Vol. 2, pp. 227–48.

Caballero, Ricardo J., 2003, “The Future of the IMF,”American Economic Review, Papers and Proceedings,Vol. 93, May, pp. 31–38.

Calvo, Guillermo, Leonardo Leiderman, and Carmen Rein-hart, 1992, “Capital Inflows and Real Exchange RateAppreciation in Latin America: The Role of ExternalFactors,” IMF Working Paper No. 92/62 (Washington:International Monetary Fund).

———, 1996, “Inflows of Capital to Developing Coun-tries in the 1990s,” Journal of Economic Perspectives,Vol. 10, No. 2, pp. 123–29.

Chan-Lau, Jorge A., 2004, “Pension Funds and EmergingMarkets,” IMF Working Paper No. 04/181 (Washing-ton: International Monetary Fund).

Cho, Yoon Je, 2001, “The Role of Poorly Phased Liberal-ization in Korea’s Financial Crisis,” in FinancialLiberalization: How Far, How Fast? ed. by GerardCaprio, Patrick Honohan, and Joseph E. Stiglitz(Cambridge and New York: Cambridge UniversityPress).

Desai, Padma, 2003, Financial Crisis, Contagion, andContainment (Princeton, New Jersey: PrincetonUniversity Press).

de Vries, Margaret G., 1969, “Exchange Restrictions: TheSetting,” in The International Monetary Fund,1945–1965: Twenty Years of International MonetaryCooperation, Vol. II, ed. by J. Keith Horsefield(Washington: International Monetary Fund).

Dooley, Michael P, 1996, “A Survey of Literature on Con-trols over International Capital Transactions,” IMFStaff Papers, Vol. 43 (December), pp. 639–87.

Duttagupta, Rupa, Gilda Fernandez, and Cem Karacadag,2004, “From Fixed to Float: Operational Aspects ofMoving Toward Exchange Rate Flexibility,” IMFWorking Paper No. 04/126 (Washington: Interna-tional Monetary Fund).

Edison, Hali J., Michael Klein, Luca Ricci, and TorstenSlok, 2002, “Capital Account Liberalization and Eco-nomic Performance: Survey and Synthesis,” IMFWorking Paper No. 02/120 (Washington: Interna-tional Monetary Fund).

References

86


References

Edison, Hali J., and Carmen M. Reinhart, 2001, “StoppingHot Money,” Journal of Development Economics,Vol. 66, pp. 533–53.

Edison, Hali J., and Francis E. Warnock, 2003, “A SimpleMeasure of the Intensity of Capital Controls,” Journalof Empirical Finance, Vol. 10, pp. 81–103.

Edwards, Sebastian, 1984, “The Order of Liberalization ofthe External Sector in Developing Countries,” Essaysin International Finance, No. 156 (Princeton, New Jer-sey: Department of Economics, Princeton University).

Eichengreen, Barry, 2001, “Capital Account Liberalization:What Do Cross-Country Studies Tell Us?” World BankEconomic Review, Vol. 15, No. 3, pp. 341–65.

———, Michael Mussa, Giovanni Dell’Ariccia, EnricaDetragiache, Gian Maria Milesi-Ferretti, and AndrewTweedie, 1998, Capital Account Liberalization: The-oretical and Practical Aspects, IMF Occasional PaperNo. 172 (Washington: International Monetary Fund).

Eichengreen, Barry, and Michael D. Bordo, 2002, “CrisesNow and Then: What Lessons from the Last Era ofFinancial Globalization?” NBER Working Paper No. 8716 (Cambridge, Massachusetts: National Bu-reau of Economic Research).

Eichengreen, Barry, and David Leblang, 2002, “CapitalAccount Liberalization and Growth: Was Mr. Ma-hathir Right?” NBER Working Paper No. 9427(Cambridge, Massachusetts: National Bureau ofEconomic Research).

Financial Stability Forum, 2000, Report of the WorkingGroup on Highly Leveraged Institutions (Basel).

Fischer, Stanley, 1998, “Capital Account Liberalizationand the Role of the IMF,” in Should the IMF PursueCapital Account Convertibility? Essays in Interna-tional Finance, No. 207 (Princeton, New Jersey: De-partment of Economics, Princeton University).

Forbes, Kristin J., 2003, “One Cost of the Chilean CapitalControls: Increased Financial Constraints for SmallerTraded Firms,” NBER Working Paper No. 9777(Cambridge, Massachusetts: National Bureau of Eco-nomic Research).

Ffrench-Davis, Ricardo, and Heriberto Tapia, 2004, TheChilean-Style of Capital Controls: An Empirical As-sessment (Santiago: United Nations Economic Com-mission for Latin America and the Caribbean).

Funke, Norbert, 1993, “Timing and Sequencing of Re-forms: Competing Views and Role of Credibility,”Kyklos, Vol. 46.

Galbis, Vicente, 1994, “Sequencing of Financial SectorReforms: A Review,” IMF Working Paper No. 94/101(Washington: International Monetary Fund).

———, 1996, “Currency Convertibility and the Fund: Re-view and Prognosis,” IMF Working Paper No. 96/39(Washington: International Monetary Fund).

Gallego, Francisco, and Leonardo Hernandez, 2003, “Mi-croeconomic Effects of Capital Controls: The ChileanExperience During the 1990s,” International Journalof Finance and Economics, Vol. 8, pp. 225–54.

Gallego, Francisco, Leonardo Hernandez, and KlausSchmidt-Hebbel, 2002, “Capital Controls in Chile:Were They Effective?” in Banking, Financial Integra-tion, and International Crises, ed. by Leonardo Her-nandez and Klaus Schmidt-Hebbel (Santiago: CentralBank of Chile).

Goldstein, Morris, 1995, “Coping with Too Much of aGood Thing: Policy Responses for Large Capital In-flows in Developing Countries,” Policy ResearchWorking Paper No. 1507 (Washington: World Bank).

———, and Michael Mussa, 1993, “The Integration ofWorld Capital Markets,” IMF Working Paper No.93/95 (Washington: International Monetary Fund).

Gelos, Gaston, and Shang-Jin Wei, 2002, “Transparencyand International Investor Behavior,” IMF WorkingPaper No. 02/174 (Washington: International Mone-tary Fund).

Gourinchas, Pierre-Olivier, and Olivier Jeanne, 2004,“The Elusive Gains from International Financial Inte-gration,” IMF Working Paper No. 04/74 (Washington:International Monetary Fund).

Griffith-Jones, Stephany, 1998, Global Capital Flows:Should They Be Regulated? (London: Macmillan).

———, and Jose A. Ocampo, 2003, What Progress on In-ternational Financial Reform? Why So Limited?(Stockholm: Almqvist & Wiksell International).

Grilli, V., and G.M. Milesi-Ferretti, 1995, “Economic Ef-fects and Structural Determinants of Capital Con-trols,” IMF Staff Papers, Vol. 42, pp. 517–51.

Guitián, Manuel, 1996, “The Issue of Capital AccountConvertibility: A Gap Between Norms and Reality,”in Currency Convertibility in the Middle East andNorth Africa, ed. by Saleh M. Nsouli and ManuelGuitián (Washington: International Monetary Fund).

Henry, Peter Blair, 2003, “Capital-Account Liberalization,the Cost of Capital, and Economic Growth,” Ameri-can Economic Review, Papers and Proceedings, Vol.93 (May), pp. 91–96.

Helleiner, Eric, 1994, States and the Reemergence ofGlobal Finance: From Bretton Woods to the 1990s(Ithaca, New York: Cornell University Press).

Horsefield, J. Keith, ed., 1969, International MonetaryFund, 1945–1965: Twenty Years of InternationalMonetary Cooperation, Vol. III (Washington: Interna-tional Monetary Fund).

Independent Evaluation Office (IEO), 2003, The IMF andRecent Capital Account Crises: Indonesia, Korea,Brazil (Washington: International Monetary Fund).

———, 2004, The IMF and Argentina, 1991–2001 (Wash-ington: International Monetary Fund).

International Monetary Fund (IMF), 1999, Exchange RateArrangements and Currency Convertibility: Develop-ments and Issues, World Economic and FinancialSurveys (Washington: International Monetary Fund).

Ishii, Shogo, Karl Habermeier, Jorge Ivan Canales-Kril-jenko, Bernard Laurens, John Leimone, and JuditVadasz, 2002, Capital Account Liberalization and Financial Sector Stability, IMF Occasional Paper No. 211 (Washington: International Monetary Fund).

Jadhav, Narendra, 2003, “Capital Account Liberalization:The Indian Experience,” Reserve Bank of India(Mumbai).

James, Harold, 1996, International Monetary CooperationSince Bretton Woods (Washington: InternationalMonetary Fund; and New York and Oxford: OxfordUniversity Press).

Johnson, Simon, and Todd Mitton, 2003, “Cronyism and Capital Controls: Evidence from Malaysia,”

87


REFERENCES

Journal of Financial Economics, Vol. 67 (February),pp. 351–82.

Johnston, R. Barry, 1998, “Sequencing Capital AccountLiberalizations and Financial Sector Reform,” IMFPaper on Policy Analysis and Assessment No. 98/8(Washington: International Monetary Fund).

———, and Chris Ryan, 1994, “The Impact of Controls onCapital Movements on the Private Capital Accounts ofCountries’ Balance of Payments: Empirical Estimatesand Policy Implications,” IMF Working Paper No.94/78 (Washington: International Monetary Fund).

Johnston, R. Barry, and Natalia T. Tamirisa, 1998, “WhyDo Countries Use Capital Controls?” IMF WorkingPaper No. 98/181 (Washington: International Mone-tary Fund).

Johnston, R. Barry, Salim M. Darbar, and Claudia Eche-verria, 1997, “Sequencing Capital Account Liberal-ization: Lessons from the Experiences in Chile, In-donesia, Korea, and Thailand.” IMF Working PaperNo. 97/157 (Washington: International MonetaryFund).

Johnston, R. Barry, and V. Sundararajan, eds., 1999, Se-quencing Financial Sector Reforms: Country Experi-ences and Issues (Washington: International Mone-tary Fund).

Joyce, Joseph P., and Ilan Noy, 2005, “The IMF and theLiberalization of Capital Flows,” Department of Eco-nomics Working Paper No. 05-01, Wellesley College.

Kaminsky, Graciela Laura, and Sergio L. Schmukler,2003, “Short-Run Pain, Long-Run Gain: The Effectsof Financial Liberalization,” IMF Working Paper No.03/34 (Washington: International Monetary Fund).

Kaplan, Ethan, and Dani Rodrik, 2001, “Did theMalaysian Capital Controls Work?” NBER WorkingPaper No. 8142 (Cambridge, Massachusetts: NationalBureau of Economic Research).

Karacadag, Cem, V. Sundararajan, and Jennifer Elliot, 2003,“Managing Risks in Financial Market Development:The Role of Sequencing,” IMF Working Paper No.03/116 (Washington: International Monetary Fund).

Kim, Soyoung, Sunghyun H. Kim, Yunjong Wang, 2001,“Capital Account Liberalization and MacroeconomicPerformance: The Case of Korea,” Policy Analysis01-01, (Seoul: Korea Institute for International Eco-nomic Policy).

Laurens, Bernard, and Jaime Cardoso, 1998, “ManagingCapital Flows: Lessons from the Experience ofChile,” IMF Working Paper No. 98/168 (Washington:International Monetary Fund).

Le Fort, Guillermo, and Sergio Lehmann, 2003, “The Un-remunerated Reserve Requirement and Net CapitalFlows: Chile in the 1990s,” CEPAL Review No. 81(Santiago: United Nations Economic Commission forLatin America and the Caribbean).

Lopez-Mejia, Alejandro, 1999, “Large Capital Flows: ASurvey of the Causes, Consequences, and Policy Re-sponses,” IMF Working Paper No. 99/17 (Washing-ton: International Monetary Fund).

Mathieson, Donald J., and Lilliana Rojas-Suarez, 1990,“Financial Market Integration and Exchange RatePolicy,” IMF Working Paper No. 90/2 (Washington:International Monetary Fund).

———, 1993, Liberalization of the Capital Account: Ex-periences and Issues, IMF Occasional Paper No. 103(Washington: International Monetary Fund).

McKinnon, Ronald I., 1982, “The Order of Economic Lib-eralization: Lessons from Chile and Argentina,”Carnegie-Rochester Conference Series on Public Pol-icy , Vol. 17 (Autumn), pp. 159–86.

Mendoza, Enrique G., 1990, “The Macroeconomic Effectsof Capital Controls and the Stabilization of the Bal-ance of Trade,” IMF Working Paper No. 90/109(Washington: International Monetary Fund).

Metcalfe, Michael, and Avinash Persaud, 2003, “Do WeNeed to Go Beyond Disclosure?” in The Road to In-ternational Financial Stability: Are Key FinancialStandards the Answer? ed. by Benu Schneider (Basingstoke, Hampshire, United Kingdom, and NewYork: Palgrave Macmillan).

Miniane, Jacques, 2004, “A New Set of Measures on Cap-ital Account Restrictions,” IMF Staff Papers, Vol. 51,No. 2, pp. 276–308.

Nadal-De Simone, Francisco, and Piritta Sorsa, 1999, “AReview of Capital Account Restrictions in Chile inthe 1990s,” IMF Working Paper No. 99/52 (Washing-ton: International Monetary Fund).

Nord, Roger, 2003, “The Liberalization of the Capital Ac-count in Hungary: Experiences and Lessons,” in Cap-ital Liberalization in Transition Countries: Lessonsfrom the Past and for the Future, ed. by Age F.P.Bakker and Bryan Chapple (Cheltenham, UnitedKingdom, and Northampton, Massachusetts: EdwardElgar).

Nsouli, Saleh M., Sena Eken, Paul Duran, Gerwin Bell,and Zuhtu Yucelik, 1993, The Path to Convertibilityand Growth: The Tunisian Experience, IMF Occa-sional Paper No. 109 (Washington: InternationalMonetary Fund).

Nsouli, Saleh M., Mounir Rached, and Norbert Funke,2002, “The Speed of Adjustment and the Sequencingof Economic Reforms: Issues and Guidelines for Pol-icymakers,” IMF Working Paper No. 02/132 (Wash-ington: International Monetary Fund).

Nurkse, Ragnar, 1944, International Currency Experi-ence: Lessons from the Inter-War Period (Geneva:League of Nations).

Ocampo, Jose A., and Camilo E. Tovar, 1999, “Price-BasedCapital Account Regulations: The Colombian Experi-ence,” Serie Financiamiento del Desarrollo No. 87,Development Finance Unit (Santiago: United NationsEconomic Commission for Latin America and theCaribbean).

———, 2003, “Colombia’s Experience with Reserve Re-quirements on Capital Inflows,” CEPAL Review No. 81 (Santiago: United Nations Economic Com-mission for Latin America and the Caribbean).

Ocampo, Jose A., and Maria L. Chiappe, 2003, Counter-Cyclical Prudential and Capital Account Regulationsin Developing Countries (Stockholm: Almqvist &Wiksell International).

Ong, Li Lian, and Amadou Sy, 2004, “The Role of MatureMarket Mutual Funds in Emerging Markets: Myth orMayhem?” IMF Working Paper No. 04/133 (Wash-ington: International Monetary Fund).

88


References

Organization for Economic Cooperation and Development(OECD), 1996 and 1998, OECD Economic Surveys:The Czech Republic (Paris).

Persaud, Avinash, 2001, “Fads and Fashions in the PolicyResponse to Financial Market Crises,” in FinancialInnovations and the Welfare of Nations, ed. by Laurent L. Jacque and Paul M. Vaaler (Norwell,Massachusetts: Kluwer Academic Publishers).

Prasad, Eswar S., Kenneth Rogoff, Shang-Jin Wei, and M.Ayhan Kose, 2003, Effects of Financial Globalizationon Developing Countries: Some Empirical Evidence,IMF Occasional Paper No. 220 (Washington: Interna-tional Monetary Fund).

Quinn, Dennis, 1997, “The Correlates of Change in Interna-tional Financial Regulation,” American Political Sci-ence Review, Vol. 91, No. 3, September, pp. 531–51.

Quirk, Peter J., 1994, “Capital Account Convertibility: ANew Model for Developing Countries,” IMF WorkingPaper No. 94/81 (Washington: International Mone-tary Fund).

———, Owen Evans, and others, 1995, Capital AccountConvertibility: Review of Experience and Implica-tions for IMF Policies, IMF Occasional Paper No.131 (Washington: International Monetary Fund).

Reserve Bank of India, 1997, “Report of the Committeeon Capital Account Convertibility,” Mumbai.

Rodrik, Dani, 1998, “Who Needs Capital Account Convert-ibility?” in Should the IMF Pursue Capital AccountConvertibility? Essays in International Finance, No.207 (Princeton, New Jersey: Department of Econom-ics, Princeton University).

Schadler, Susan, Maria Carkovic, Adam Bennett, andRobert Kahn, 1993, Recent Experiences with Surgesin Capital Inflows, IMF Occasional Paper No. 108(Washington: International Monetary Fund).

Singh, Anoop, Agnes Belaisch, Charles Collyns, Paula DeMasi, Reva Krieger, Guy Meredith, and RobertRennhack, 2005, Stabilization and Reform in LatinAmerica: A Macroeconomic Perspective on the Ex-perience Since the Early 1990s, IMF OccasionalPaper No. 238 (Washington: International MonetaryFund).

Souayah, Abdelmoumen, 1996, “Currency Convertibilityin Tunisia,” in Currency Convertibility in the MiddleEast and North Africa, ed. by Manuel Guitián andSaleh M. Nsouli (Washington: International Mone-tary Fund).

Stiglitz, Joseph, 2000, “Capital Market Liberalization,Economic Growth and Instability,” World Develop-ment, Vol. 28, No. 6, pp. 1075–86.

———, 2002, Globalization and Its Discontents (NewYork: W.W. Norton).

———, 2004, “Capital-Market Liberalization, Globaliza-tion, and the IMF,” Oxford Review of Economic Pol-icy, Vol. 20, No. 1, pp. 57–71.

Szakadat, Laszlo, 1998, “Hungary’s Ponzi Game,” in Hun-gary: Towards a Market Economy, ed. by LaszloHalpern and Charles Wyplosz (Cambridge, England,and New York: Cambridge University Press).

Tarapore, Savak S., 1998, “An Indian Approach to Capital-Account Convertibility,” Should the IMF Pursue Cap-ital Account Convertibility? Essays in InternationalFinance, No. 207 (Princeton, New Jersey: Depart-ment of Economics, Princeton University).

Thiel, Eva, 2003, “Recent Codes-Based Liberalization inthe OECD,” in Capital Liberalization in TransitionCountries: Lessons from the Past and for the Future,ed. by Age F.P. Bakker and Bryan Chapple (Chel-tenham, United Kingdom, and Northampton, Massa-chusetts: Edward Elgar).

Tytell, Irina, and Shang-Jin Wei, 2004, “Does FinancialGlobalization Induce Better Macroeconomic Poli-cies?” IMF Working Paper No. 04/84 (Washington:International Monetary Fund).

Villar, Leonardo, and Hernan Rincon, 2003, “The Colom-bian Economy in the Nineties: Capital Flows andForeign Exchange Regimes,” in Critical Issues in Fi-nancial Reform: A View from the South, ed. by A.Berry (New Brunswick, New Jersey: TransactionPublishers).

Wade, Robert, 1998–99, “The Coming Fight over CapitalFlows,” Foreign Policy (Winter), pp. 41–54.

———, and Frank Veneroso, 1998, “The Gathering WorldSlump and the Battle over Capital Controls,” New LeftReview, NLR I/237, September/October, pp. 13–42.

Williamson, John, 1990, “What Washington Means byPolicy Reform?” in Latin American Adjustment: HowMuch has Happened? ed. by John Williamson (Wash-ington: Institute for International Economics).

———, 2002, “Proposals for Curbing the Boom-and-BustCycle in the Supply of Capital to Emerging Markets,”WIDER Discussion Paper No. 2002/3 (Helsinski:United Nations University, World Institute for Devel-opment Economics Research).

89


Statement by the Managing Director

IMF Staff ResponseIEO Comments on Management/Staff Response

Summing Up of IMF Executive BoardDiscussion by the Acting Chair


The Independent Evaluation Office’s study on theIMF’s approach to capital account liberalizationshould help in disseminating the lessons for Fundpractice and enhancing the learning culture of theFund. The staff has prepared a statement indicatingits reactions to the report’s recommendations in theperiod ahead.

The Board discussion of the IEO report will pro-vide an opportunity to draw out its implications forthe Fund’s policies and procedures. The questions oncapital account issues raised in this report will be ad-dressed in the ongoing strategic priorities review, theresults of which will help shape our agenda goingforward.

STATEMENT BY THE MANAGING DIRECTOR ON THE

EVALUATION BY THE INDEPENDENT EVALUATION OFFICE OF THE

IMF’S APPROACH TO CAPITAL ACCOUNT LIBERALIZATION

Executive Board MeetingMay 11, 2005

93


1. The IEO report (SM/05/142) highlights the dif-ficulties and complexities the Fund faces in provid-ing advice on capital account issues. The two mainrecommendations put forward in the report are: theneed for (i) greater clarity in the scope of the Fund’spolicy advice on capital account issues to its mem-bership, and (ii) greater attention to supply-side fac-tors of international capital flows with a view tominimizing their volatility. This statement elaborateson some of the report’s analytical underpinnings.The report is timely in view of the ongoing work onthe Fund’s strategic priorities.

2. The staff considers the sample underlying thereport’s evaluation of our advice on capital accountissues a fair representation of the diverse member-ship, although it believes that a finer distinctionamong these countries would have been useful. Thestaff appreciates the considerable work that was in-volved in evaluating 27 countries, and finds the de-scription of the cases broadly accurate. However, indiscussing the staff’s policy advice, the report couldhave made a useful distinction between three cases:countries with an actual or potential balance of pay-ments or banking crisis; countries facing a majorcapital inflow; and those under “normal” conditionsbut with some remaining capital controls. This dis-tinction would help to clarify and nuance the staff’sadvice on capital account liberalization and imposi-tion of temporary capital controls.

3. Relatedly, the report’s finding of some apparentinconsistencies in the Fund’s advice on capital ac-count liberalization across countries needs to bemore nuanced. The Fund is sometimes criticized asbeing a monolithic institution following a “one-size-fits-all” approach. In contrast, the IEO concludes thatits evidence suggests no “one-size-fits-all” approachby the Fund staff. Indeed, the staff has used its ownprofessional judgment in approaching capital accountliberalization in individual countries, and there wasactive debate on the issue within the staff. In its as-

sessment, however, the report seems critical of thisdiscretion, and calls for greater consistency in coun-try work and institutional approach. The apparentlack of consistency in the Fund’s advice to its mem-bers may have a number of causes, as acknowledgedin the report, including the tailoring of policy adviceto country-specific circumstances or the absence ofan official position in the Fund on capital account is-sues. While we do not disagree with the report’s find-ing that the latter could have contributed to somevariation in policy advice across the membership, thereport could have gone further into explaining thedifferent country circumstances that would warrantdifferentiated advice. Indeed, a detailed case studyapproach of the sampled countries would have per-haps made a more convincing analysis.

4. The staff largely concurs with the evaluation’stwo key findings on the Fund’s policy advice to itsmember countries on capital account issues (Chap-ters III and IV). The IEO report finds that the Fundstaff has been quite accommodating of the authori-ties’ policy choices when they involved a gradual ap-proach to capital account liberalization or temporaryuse of capital controls. Moreover, in no case did theFund require capital account liberalization as formalconditionality for use of its resources. In particular,the Fund has not pressured member countries—cer-tainly not the emerging market countries sampled—to liberalize their capital accounts or to move fasterthan they wanted to go. In terms of advice on thetemporary use of capital controls, the IEO concludesthat the Fund staff seldom challenged the authori-ties’ decisions and even supported market-basedcontrols in some cases as a second-best option. Itnotes that the staff pointed out the risks inherent inan open capital account as well as the need for asound financial system, even in the early 1990s.However, the IEO considers that these risks were in-sufficiently highlighted and did not translate into op-erational advice until later in the 1990s. We would

STAFF RESPONSE TO THE EVALUATION BY THE

INDEPENDENT EVALUATION OFFICE OF THE



94


Staff Response

stress that the IEO sees recent improvements in thisarea. Indeed, substantial analytical work has beencarried out by the Monetary and Financial SystemsDepartment (MFD), the International Capital Mar-kets Department (ICM), Research Department(RES), and area departments.

5. The report does not do justice to the roleplayed by external forces in promoting capital ac-count liberalization. While it acknowledges the pres-ence of exogenous factors, the country cases do notfully reflect their role. In particular, free trade agree-ments (FTAs) and multilateral trade liberalizationthrough the World Trade Organization have coveredsignificant components of the capital account relatedto financial services and investment chapters. In ad-dition, recent bilateral investment treaties (BITs) fre-quently cover a broad range of instruments (includ-ing portfolio investments, inter-bank transactions,and sovereign debt) with no balance of paymentssafeguards; that is, signatories, to a point, cannot im-pose capital controls even during times of macroeco-nomic or financial stress. Moreover, a major sourceof pressure for liberalization is often a country’s owncommercial sector, which may be looking forsources of competitive financing. All in all, the pol-icy choices of emerging markets may have reflectedthese external factors at least as much as Fund policyadvice.

6. The report proposes two main recommenda-tions in light of the Fund’s experience with capitalaccount issues. As discussed below, the Fund is al-ready implementing some aspects and plans to ad-dress these issues further in its future work programand in the strategic priorities review.

7. Recommendation 1 is that more clarity isneeded on the Fund’s approach to capital accountissues. There are three aspects to this proposed rec-ommendation, which are discussed in turn.

• The report suggests that the place of capital ac-count issues in Fund surveillance could be clari-fied, and there would be value if the ExecutiveBoard were to clarify formally the scope of Fundsurveillance on capital account issues. TheBoard and the Fund staff have recognized thatcapital account developments and vulnerabilitiesconstitute an increasingly important focus of theFund’s work on promoting stability, and theprocess of clarifying the scope of Fund surveil-lance to include capital account issues is alreadyunderway. The Executive Board noted in thecontext of recent Biennial Surveillance Reviews(2002, 2004) that Fund surveillance needed toadapt to “a changing global environment, mostnotably to the rapid expansion of internationalcapital flows.” This adaptation would imply abroadening of the coverage of surveillance “from

a relatively narrow focus on fiscal, monetary andexchange rate policies, to a broader purview en-compassing external vulnerability assessments,external debt sustainability analyses, financialsector vulnerabilities, and structural and institu-tional policies that have an impact on macroeco-nomic conditions” (SUR/02/81). The Board hasalso highlighted the risks of opening the capitalaccounts before floating the exchange rate, andespecially the risks of sudden outflows (PIN No.04/141). More recently, in the context of the dis-cussion of the “Fund’s Medium-Term Strategy—Framework and Initial Reflections” (BUFF/05/ 60), the Board has called for additional work oncapital account issues.

• Relatedly, the report proposes that the ExecutiveBoard could issue a statement clarifying thecommon elements of agreement on capital ac-count liberalization. While this is an issue forthe Executive Board to decide, the staff agreesthat it would be useful to have some clear opera-tional guidance that lays out the broad principlesthat Fund staff needs to follow in its policy ad-vice across countries and that the outline ofthese principles in the IEO report is a usefulstarting point for such guidance. Indeed, staff(especially MFD) has already undertaken policyresearch and operational work on various as-pects of capital account issues (including capitalaccount liberalization and financial sector re-form, country experiences with use and subse-quent liberalization of capital controls, and sequencing), in the context of FSAPs and tech-nical assistance. We would, however, cautionthat there is no single “right” approach. Our ad-vice needs to take into account the different situ-ations confronting our members. It also mustrecognize that for many members full capital ac-count liberalization is an aspiration that willlikely take substantial time to achieve.

• Against this backdrop, the Fund’s future workon capital account issues should seek to but-tress efforts to promote financial stability,while helping ensure that controls are not usedto impede adjustment. An approach would beto build on the existing Fund expertise in thisarea, and to ensure that policy advice on capitalaccount issues is fully incorporated into themainstream of bilateral and multilateral sur-veillance, with analytic work being used tostrengthen the basis for policy advice and tech-nical assistance. This strategy would imply ameasured approach to liberalization to facili-tate countries’ integration into global economywhile maintaining stability, rather than promot-ing liberalization, per se.

95


STAFF RESPONSE

• The report also suggests that the Fund couldsharpen its advice on capital account issues,based on solid analysis of the particular situa-tion and risks facing specific countries. Thestaff endorses this recommendation, which is inline with best practice and longstanding generalguidance on Fund surveillance. In its policy ad-vice to countries the staff will continue to drawon all existing analytical work as well as any fu-ture findings that may arise from staff studiesthat have been requested by the Board, or fromother sources. Further, technical assistance pro-vided by the Fund on operational issues pro-vides hands-on advice to countries on how toproceed with capital account issues. We alsonote that MFD is reviewing, for the upcomingBiennial Review of Exchange Arrangement, Re-strictions and Markets, the role of capital con-trols during financial stress and issues relatingto financial liberalization and capital accountregulations. However, the IEO’s suggestion thatthe Fund should provide some quantitativegauge of the benefits, costs, and risks of movingat different speeds is likely to prove difficult toput into practice, given the conflicting theoreti-cal and empirical evidence on the subject andthe political and economic complexities thatcapital account issues typically involve.

8. Recommendation 2 is that the Fund’s analysisand surveillance should give greater attention to thesupply-side factors of international capital flowsand what can be done to minimize the volatility ofcapital movements. Staff agrees with the crux of thisrecommendation, although given the large numberof staff studies already completed (and more underway), we are not sure what other specific actions, ifany, the IEO may have in mind. We agree with thereport’s assessment that this is a difficult topic onwhich little professional consensus exists. That said,past work and current initiatives demonstrate that thestaff recognizes that reducing volatility and cyclical-ity in the supply of capital are critical to achievingcapital markets stability.

• To strengthen Fund surveillance in this regard,several research initiatives have been underway. ICM has aimed to focus on global financialmarket linkages, financial and risk-relatedflows, and global asset allocation decisions ofinstitutional investors. Recent Global FinancialStability Reports (GFSRs) have focused on thevolatility of private capital flows to emergingmarkets (September 2003), and provided analy-sis of the institutional investor base in emergingmarkets (March 2004). The next GFSR will in-clude studies of home bias and global diversifi-cation, and financial stability considerations of

regulatory and accounting policies. Further, aconsiderable number of studies by IMF staff(appearing in both Fund documents and outsidejournals) have in recent years examined supply-side aspects, including: the potential impact ofinterest rates and risk appetite in advancedcountries on emerging market spreads; herdingbehavior, contagion, and the possible role of in-stitutional investors and hedge funds in this re-gard. The Fund has also delved into deepeningdomestic capital markets by examining ways inwhich local securities and derivative marketscan be developed to tap on a more stable seg-ment of the investor base.

• The Fund is undertaking a number of initiativesat the country level. Through the Financial Sec-tor Assessment Program (FSAP) initiative, theFund is also contributing to identifying poten-tial vulnerabilities and risks in the financial sys-tems, particularly of systemically importantcountries, whose vulnerabilities may spill overto other countries owing to strong ties in the fi-nancial sector. Further, by way of the Reportson Observance of Standards and Codes(ROSC), Special Data Dissemination Standards(SDDS), General Data Dissemination System(GDDS), and technical assistance, the Fund ishelping to improve information flow to in-vestors, which in turn contributes to greater sta-bility. In addition, the Fund has developed oper-ational guidance to assist in maintaining thesoundness of relatively large and complexbanking organizations, with a view to creatingstability in domestic and international financialsectors.

• Several broader initiatives have also been under-taken by the Fund. The Capital Markets Consul-tative Group (CMCG), established in July 2000by the IMF’s Managing Director, provides aforum for informal dialogue between partici-pants in international capital markets and theIMF. More generally, the IMF is working on im-proving its understanding of recent trends in andfuture prospects of foreign direct investment(FDI) in emerging market countries (EMCs), theinvestment strategies of large multinational com-panies, and the determinants of FDI in EMCs.The Fund has also played a leading role in dis-cussions about the possibility of a statutory sov-ereign debt restructuring mechanism (SDRM)for orderly debt workouts, and has encouragedthe use of collective action clauses (CACs) tostrengthen crisis resolution and reduce uncer-tainty associated with debt restructuring. As indi-cated in the report, the staff comments on theproposed New Basel Capital Accord (Basel II)

96


Staff Response

have highlighted that using credit ratings to setcapital charges could increase volatility and pro-cyclicality.

• With so many initiatives under way at theFund, we are puzzled by the report’s findingthat the Fund pays too little attention to sup-ply-side risks. The past and current work,noted above, shows that the Fund staff hasbeen at the forefront of the analysis and policydebate in this area. However, staff recognizesthat it cannot rest here, and it will strive to en-hance further its understanding of supply-sidefactors and their implications for our member-ship. However, it would have been more help-ful if the IEO had proposed specific actionsthat could be taken on the supply-side to mini-mize volatility and cyclicality.

• Beyond the present list of activities within theFund to deal with the supply-side risks of capitalflows, the staff emphasizes that additional inter-nationally coordinated efforts could help givethese issues higher priority among policymakersin advanced economies. In this regard, the Finan-cial Stability Forum (FSF), established in 1999and, in whose work the Fund participates, worksas a conduit for promoting international financialstability, improving the functioning of financialmarkets, and reducing systemic risks. Further, thePrinciples for Stable Capital Flows and Fair DebtRestructuring in Emerging Markets, a private-sector-led initiative with the support of a numberof emerging market country issuers, are alsoaimed at improving the engagement between sov-ereign debtors and their creditors, with a view topromoting global financial stability.

97


We would like to offer a few points of clarifica-tion, in response to the comments made by manage-ment and staff on the IEO report.1

The staff suggests that the report is critical of thediscretion exercised in tailoring policy advice oncapital account liberalization to country-specific cir-cumstances (paragraph 3). This is not what the re-port says. The report tries to explain how the stafftailored its advice on capital account issues to coun-try-specific circumstances but concludes that a fullassessment is not possible because country docu-ments generally do not provide sufficient analyticalbasis for understanding why a particular combina-tion of policies was advised in a particular case. TheIEO report does not take issue with discretion assuch and we certainly agree that the IMF did notadopt a “one size fits all” approach in its approach tocapital account liberalization (and other related is-sues) in individual countries. Rather, the message ofthe report is that this potentially admirable discretionneeds to be guided by general principles and there-fore would have benefited from a clearer official po-sition in the IMF on capital account issues. Indeed,we read the staff response as agreeing that it wouldbe useful to have some further operational guidancethat lays out the broad principles that staff shouldfollow in its policy advice across countries.

The staff notes that the IEO report does not dojustice to the role played by external forces in pro-moting capital account liberalization (paragraph 5).We agree that the exogenous factors noted by thestaff were an important influence on capital accountliberalization in many countries. Indeed, the reportdoes refer to these external forces; some of the de-tailed country cases in Appendix 1 and some boxeson country experiences are quite explicit in spellingout the external forces that influenced the decisions

of these countries to open their capital accounts. Thefocus of the report, however, remains on the IMF’sapproach. If the forces driving capital account liber-alization were indeed beyond the IMF’s control, thisdoes not alter the fact that the IMF had a potentiallycritical role to analyze the risks involved and to offerappropriate policy advice to minimize those risks.An assessment of the IMF’s approach to capital ac-count liberalization does not depend on where theimpetus came from.

Regarding Recommendation 1, the broad ap-proach to the IMF’s future work on capital accountissues set out in the third bullet of paragraph 7 of thestaff response is consistent with what we had inmind. The staff also notes the practical difficulty ofproviding a quantitative gauge of the benefits, costs,and risks of liberalizing the capital account at differ-ent speeds. We concur that it is not an easy task, butassessing the trade-offs involved in different ap-proaches is critical if the IMF’s advice on sequencingcapital account liberalization is to be useful. Se-quencing necessarily involves a piecemeal process inwhich some markets are liberalized while others re-main closed. Then, sequencing is the right approachonly if the risks avoided exceed the risks created bypartial liberalization. While it is unlikely to be possi-ble to quantify precisely the trade-offs involved, pol-icy advice would be of limited usefulness unlesssome guidance on the nature and magnitude of thetrade-offs were provided. Welfare economics—inwhich the consequence of opening a market whendomestic distortions exist has been an importanttopic—may provide insight for thinking about how tomake policy advice on sequencing more operational.

Finally, regarding Recommendation 2, the staffnotes a number of initiatives that are already underway, to analyze supply-side factors influencing thevolatility of international capital flows. We welcomethese initiatives, some of which, but not all, werenoted in the report. The point we make in the report

INDEPENDENT EVALUATION OFFICE COMMENTS ON

MANAGEMENT/STAFF RESPONSE TO THE EVALUATION OF THE



98

1Paragraph references are to the staff response.


Independent Evaluation Office Comments

is that the IMF’s analysis of supply-side factors hasbeen largely descriptive and that the policy implica-tions drawn are mainly targeted at recipients of capi-tal flows. There are, of course, good reasons forthis—including the evolving nature of the literaturewhich has generated limited consensus on specificactions that could be taken on the supply side tominimize volatility and cyclicality. Therefore, ourrecommendation was not a call for the IMF to backany specific policy measure but rather to strive to en-

hance further its understanding of supply-side fac-tors and their operational or policy implications,with the focus on how such factors influence inter-linkages between countries. But we agree that con-siderable progress has already been made in this di-rection, including through the various forthcomingactivities noted by the staff in its response. Not all ofthese were known to us at the time the IEO reportwas prepared. For example, the planned thrust of thenext GFSR is the type of activity we had in mind.

99


Executive Directors welcomed the report by theIndependent Evaluation Office (IEO) on the IMF’sapproach to capital account liberalization. Theynoted that financial integration can confer benefits tothe global economy by promoting an efficient allo-cation of savings and a diversification of risks, withsome Directors emphasizing the importance of or-derly and well-sequenced liberalization. Directorsstressed the increasing significance of capital ac-count issues in Fund surveillance, and of fully un-derstanding and addressing the difficulties and com-plexities faced by the Fund in providing advice inthis area. They thus welcomed the opportunity thatthe report provides to explore how the Fund’s effec-tiveness in this area can be further advanced.

Directors appreciated the IEO’s efforts in evaluat-ing the Fund’s experience since the early 1990s witha large sample of representative countries. Theynoted that the report offers a broadly accurate ac-count of the evolution of Fund thinking and practiceon the issues surrounding capital account liberaliza-tion and capital flow management, including the tem-porary use of capital controls. Such a history pro-vides a crucial background for a discussion of theseissues. Directors welcomed the IEO’s confirmationthat the Fund did not apply an inappropriate “one-size-fits-all” approach to capital account liberaliza-tion in individual countries. Some Directors notedthat, while it was important to apply discretion in in-dividual cases, it would have been helpful for policyadvice to have been guided by general principles. Di-rectors concurred with the report’s finding that theFund did not pressure members to liberalize theircapital account sooner than desired by the authorities,and generally did not challenge the use of temporarycapital controls. They considered that the Fundshould continue to adopt a flexible approach to capi-tal account liberalization that takes due account ofcountries’ specific circumstances and preferences. At

the same time, Directors recognized that in the earlierperiod the risks of an open capital account had not al-ways been sufficiently highlighted in the Fund’s op-erational policy advice, and that little policy advicehad been offered in the context of multilateral sur-veillance. Directors were encouraged, however, thatin recent years substantial strides have been made,based on the lessons of experience and supported bythe Fund staff’s extensive analytical work on capitalaccount issues and financial system stability. In thisconnection, some Directors suggested that the IEOreport could have offered specific suggestions onhow to ensure greater consistency and coordinationbetween ongoing analyses on capital account liberal-ization issues at a conceptual level and their practicalapplication in country operational work.

Directors expressed a variety of views on the im-portance of factors motivating capital account liber-alization, such as free trade agreements and bilateralinvestment treaties. It was acknowledged that suchagreements are negotiated voluntarily by country au-thorities when they are considered to be in the na-tional interest. Some Directors felt that the role ofsuch agreements in capital account liberalizationshould not be underestimated. At the same time,many Directors saw a key role for Fund involvementin policy advice on capital account issues as a meansof promoting orderly and nondiscriminatory capitalaccount liberalization.

Directors also commented on the two main rec-ommendations of the IEO report.

Recommendation 1. There is a need for moreclarity on the IMF’s approach to capital account issues.

Directors stressed that the Fund has long attachedimportance to capital account issues and vulnerabili-ties, and that the process of clarifying their role insurveillance is well under way. They noted that the

THE ACTING CHAIR’S SUMMING UP

IEO REPORT ON THE

EVALUATION OF THE IMF’S APPROACH TO

CAPITAL ACCOUNT LIBERALIZATION


100


Summing Up by the Acting Chair

Executive Board, in its various discussions includingin the context of the Biennial Surveillance Reviews,has called for Fund surveillance to adjust to thechanging global environment, notably the expansionin capital flows. The Fund has provided country-spe-cific guidance to member countries on strengtheningdomestic policies and practices to manage risks re-lated to capital account liberalization, including inthe context of FSAPs and ROSCs. Furthermore, re-gional and global surveillance has increasingly fo-cused on global financial market linkages, looking atdemand- and supply-side factors, and the impliedcosts and benefits of capital account liberalization.Some Directors, however, saw merit in further clari-fying the scope of Fund surveillance to recognize ex-plicitly the central importance of capital accountpolicies. Such clarification will also serve to im-prove the consistency of the Fund’s approach in thisarea across countries. Directors agreed that the Fundhas an inherent responsibility to its members to ana-lyze the benefits and risks involved in a world ofopen capital markets, and to provide practical,sound, and appropriate policy advice to its memberson those issues. On the broader aspects of the Fund’srole in capital account liberalization, most Directorsdid not wish to explore further at present the possi-bility of giving the Fund jurisdiction over capitalmovements. However, a number of Directors feltthat the Fund should be prepared to return to thisissue at an appropriate time. Directors also notedthat additional work on capital account issues is con-templated in the context of the Fund’s ongoingstrategic review.

Directors expressed a variety of views on themerit of an Executive Board statement clarifying theelements of agreement on capital account issues. Anumber of Directors supported such a statement,which could build on the “integrated” approach thathas gradually evolved in the Fund’s operationalwork, as outlined in the IEO report. However, manyDirectors underlined the challenge that would befaced in developing such a statement in view of theinherent difficulty in developing common guide-lines that adequately take into account country-spe-cific circumstances. Further, these Directors notedthat, at present, firm theoretical and empirical con-clusions are lacking, and accordingly, such a state-ment might need to be crafted in rather generalterms, thereby not providing significant additionalguidance. Directors noted that they would have anopportunity to come back to this issue in the contextof the Fund’s ongoing strategic review. More gener-ally, Directors stressed that staff will need to con-tinue to exercise their informed professional judg-ment and discretion.

Directors saw scope for sharpening the Fund’s ad-vice on capital account issues. They emphasized that

Fund staff should continue to draw upon all availableresearch in its policy advice to members, and thatfurther research and study are needed to fully under-stand how best to obtain the benefits and manage therisks of capital account liberalization, including se-quencing issues. Directors urged the staff to base itspolicy advice on solid analysis of individual countrysituations. To this end, a number of Directors recom-mended that staff reports should include a clearerand more systematic and analytical rationale forFund advice. Directors also encouraged the staff tofurther strengthen its technical expertise on capitalaccount issues. With regard to the IEO’s suggestionthat the Fund staff should aim to provide more quan-titative assessments of the benefits, costs, and risksof liberalizing the capital account at different speeds,a few Directors saw merit in the proposal, while oth-ers considered it to be very difficult to implementbecause of the technical challenges and economiccomplexities involved.

Recommendation 2. The IMF’s analysis and sur-veillance should give greater attention to the supply-side factors of international capital flows and towhat can be done to minimize the volatility of capitalmovements.

Directors welcomed the various initiatives underway in the Fund to strengthen research, analysis, andsurveillance of the supply side of capital flows, andagreed with the IEO’s view that considerableprogress has already been made in this area. A num-ber of recent analytical staff studies have examinedsupply-side features of capital flows, and Directorsnoted that the recent Global Financial Stability Re-ports have examined the determinants and volatilityof capital flows to emerging market countries in-cluding their institutional investor base. Directorsfurther pointed to the Capital Markets ConsultativeGroup, which serves as an informal forum for dia-logue between participants in international capitalmarkets and Fund management, as well as to the vis-ibility given to supply-side issues by staff at the Fi-nancial Stability Forum and the Basel Committee ofBank Supervisors. A number of Directors accord-ingly felt that the Fund is already paying due regardto supply-side factors that influence the volatility ofcapital flows.

Directors encouraged the staff to continue tobuild on the work already being undertaken at theFund in order to further its understanding of supply-side factors and their operational and policy implica-tions, including how such factors influence interlink-ages and imbalances among countries in the contextof the Fund’s multilateral and bilateral surveillance.In particular, they suggested that more attention bedevoted to the spillover effects from regional devel-opments and from policies in systemically important

101


SUMMING UP BY THE ACTING CHAIR

advanced and developing countries. Directors cau-tioned that any expanded work on supply-side issuesshould not entail Fund involvement in the regulationof the sources of capital, noting that the Fund shouldinstead coordinate with the FSF and other fora thathave the necessary expertise and mandate in the set-ting of standards. Some Directors would have wel-comed suggestions by the IEO for additional spe-cific actions that the Fund could take to address risksrelated to the volatility of capital flows.

In concluding, Directors agreed that the Fund’sfuture work on capital account issues should seek tobuttress efforts to promote financial stability, whilehelping ensure that controls are not used as a substi-

tute for adjustment. The aim would be to build onthe existing Fund expertise in this area, and to en-sure that policy advice on capital account issues isfully incorporated into the mainstream of bilateraland multilateral surveillance, with analytical workbeing used to strengthen the basis for policy adviceand technical assistance. This strategy would implyorderly and nondiscriminatory liberalization aimedat facilitating countries’ integration into the globaleconomy while maintaining stability. As a follow-up to the findings of the IEO report, Directorslooked forward to capital account issues being ad-dressed in the context of the Fund’s ongoing strate-gic review.

102


asean.elibrary.imf.org€¦ · © 2005 international monetary fund production: imf multimedia...

Documents