Statistical Capacity Building - IMF -- International … capacity building : case studies and...

INTERNATIONAL MONETARY FUND

StatisticalCapacity Building

Case Studies andLessons Learned

Edited by

Thomas K. Morrison

With Case Studies by

Zia AbbasiNoel AtcherleyJaroslav KuceraGraham L. Slack

2005

© 2005 International Monetary Fund

Cataloging-in-Publication Data

Statistical capacity building : case studies and lessons learned / edited by Thomas K. Morrison with case studies by Zia Abbasi . . . [et al.] — [Washington, D.C.] : International Monetary Fund, 2005.

p. cm.

ISBN 1-58906-442-9

1. Economics — Statistical methods — Case studies. 2. Technical assistance — Case studies.3. International Monetary Fund. I. Morrison, Thomas K. (Thomas Keith), 1944 – II. Abbasi, Zia. III. International Monetary Fund. Statistics Dept.HB137.S73 2005

Price: US$23.00

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

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

PART I. OVERVIEW AND LESSONS LEARNED

1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Thomas K. Morrison

Purpose of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Scope and Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4Strategies and Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Lessons for Country Authorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Lessons for Technical Assistance Donors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Lessons Regarding Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Lessons for Transition and Postconflict Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

PART II. CASE STUDIES

2. Cambodia: Multisector Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Zia Abbasi

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Socioeconomic Situation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11The Evolution of Official Statistics in Cambodia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Strategies and Plans for Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Implementation and Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Statistical Organization Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Statistical legislation and regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Statistics coordination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Statistical planning frameworks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Statistics development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Statistical roles, functions, organization structure,

and staffing arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Government financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Staff training and development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Economic and Financial Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16National accounts statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Price statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Foreign trade statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Contents

iii

Balance of payments statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Government finance statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Monetary and financial statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Other economic statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Sociodemographic Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Population and demographic statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Poverty statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Dissemination and Use of Official Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Availability and quality of statistics and methods documentation . . . . . . . . . . . . . . . . 20Dissemination and public access to official statistics and methodology . . . . . . . . . . . . 20

Conclusions and Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Champions and leaders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Commitment and ownership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Financing for statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Awareness and understanding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22User consultation and involvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Phasing and mode of technical assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Selection and use of experts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Appropriate legislation and enforcement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Coordination mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Use of technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3. Bosnia and Herzegovina: Multisector Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Noel Atcherley

The Immediate Postwar Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25General Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Statistical Situation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Statistical organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Statistical output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Strategies and Plans, Implementation, and Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271996–97—Initial Postwar Years—A Limited Focus on Statistics . . . . . . . . . . . . . . . . . . 27

Strategies and plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Implementation and outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

1998–99—Support for Physical Infrastructure, State Agency, and Donor Planning, but Limited Technical Assistance . . . . . . . . . . . . . . . . . . . . . . 28


2000–03—Training, Coordination and Planning Attempts, Limited Collections, a Statistical Law, and Further Donor Meetings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30


2004 Onward—Implementation of Law, Priority Surveys, Declining Technical Assistance, and Possible Organizational Change . . . . . . . . . . 35

Strategies, plans, and implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Setting 2004, Conclusions, and Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

The Setting 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Economic and political . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Statistical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

iv

Contents

4. Bosnia and Herzegovina: Monetary Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39Graham L. Slack

Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39Institutional Setting (Capacity, Statistical Outputs, Shortcomings, and Challenges) . . . . 39

The Plan of Action and Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39Plans for Improving Statistical Capacity, Technical Aspects,

and Institutional and Systemic Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Selected Methodological Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42Reliability of Source Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42Data on Domestic Bank Financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Payment Bureau Accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Cashier’s Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Co-Circulation of Currencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Currency Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Tools of Technical Assistance in Bosnia and Herzegovina . . . . . . . . . . . . . . . . . . . . . . . . . 43Technical Assistance Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44IMF Statistical Advisor and Resident Representative . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Multisector Mission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Phases of Technical Assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Implementation Phase (1996–97) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Consolidation Phase (1997–98) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Facilitating Phase (1998 to Date) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Conclusions and Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5. Ukraine: Monetary Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Jaroslav Kucera

Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Statistical Capacity, Technical Assistance, and Institutional Issues . . . . . . . . . . . . . . . . . . . 49

Stage One: Fact Finding and Outlining the Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Stage Two . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Implementation of initial recommendations and further steps to improve monetary statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Substantial progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Stage Three . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

New plans of accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Improvements in the framework for the compilation

of monetary data based on new plans of accounts . . . . . . . . . . . . . . . . . . . . . . . . 51Conclusions and Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Tables4.1. Overview of Key Aspects of the Banking System in BiH in 1996 . . . . . . . . . . . . . . . . 414.2. Time Line of IMF Monetary Statistics Missions and Key Outcomes . . . . . . . . . . . . . . 42

Figures4.1. Institutional Arrangements for Data Collection and Compilation of Monetary

Statistics for BiH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464.2. Spreadsheet Architecture for Data Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Map4.1. Bosnia and Herzegovina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

v

Contents

Technical assistance is a key element in the role of the IMF Statistics Department. While in the earlyyears it was provided as a support to the statistical publications program and to assist member countriesin adopting internationally accepted methodologies, it evolved in the 1990s to respond to the broader de-mand that countries provide reliable data on a timely basis to policymakers and markets. In the wake ofthe financial crises of the 1990s, the Statistics Department spearheaded the drive to develop the data dis-semination standards, as well as the Data Quality Assessment Framework (DQAF) that has been appliedin many member countries in the context of Reports on the Observance of Standards and Codes (ROSCs).Technical assistance has been critical in assisting countries with their statistical capacity-building effortsto come into line with these international standards.

The effectiveness of technical assistance is evaluated in several different ways, including inspection vis-its and regular feedback from users. Statistical Capacity Building: Case Studies and Lessons Learnedcomplements these efforts by analyzing in greater depth the technical assistance experiences in four coun-try case studies. The studies focus on countries that were in transition and postconflict situations in the1990s and early 2000s. It is hoped that the conclusions and lessons learned from this report will informfuture technical assistance efforts, and will also be a helpful reference to the wider international com-munity, particularly countries undertaking statistical capacity-building programs and other technical as-sistance providers.

The report was coordinated and edited by Thomas K. Morrison, advisor in the IMF’s Statistics Depart-ment. The authors of the case studies are Zia Abbasi and Noel Atcherley (long-term resident experts)and Jaroslav Kucera and Graham L. Slack (senior economists in the Statistics Department).

Robert W. EdwardsDirector

Statistics DepartmentInternational Monetary Fund

Preface

vii

ADB Asian Development BankASYCUDA Automated System for Customs Data AdministrationBHAS Agency for Statistics of Bosnia and HerzegovinaBiH Bosnia and HerzegovinaBOP Balance of paymentsBOPSY Balance of Payments Statistics Yearbook, IMFBPM5 Fifth edition of the Balance of Payments Manual, IMFCARDS Community Assistance for Reconstruction, Development, and StabilizationCBBH Central Bank of Bosnia and HerzegovinaCDC Council for the Development of CambodiaCED Customs and Excise DepartmentCOM Council of MinistersCPI Consumer price indexCSES Socioeconomic Survey of CambodiaDFUD Department of International DevelopmentDHS Demographic and Health SurveyDQAF Data Quality Assessment FrameworkEC European CommissionESCAP Economic and Social Commission for Asia and the PacificEU European UnionEurostat Statistical Office of the European CommunitiesFederation Federation of Bosnia and HerzegovinaFIS Federation of Bosnia and Herzegovina Federal Institute of StatisticsFMoF Federation Ministry of FinanceGDDS General Data Dissemination SystemGDP Gross domestic productGFS Government finance statisticsGFSM Government Finance Statistics Manual, 1986 and 2001, IMFGosbank The State Bank of the Soviet UnionGTZ German Technical CommissionGVA Gross value addedHBS Household Budget SurveyIC International communityIFS International Financial Statistics, IMFIIP International Investment PositionILO International Labor OrganizationIMF International Monetary FundISI International Statistical InstituteISTAT Italian National Institute of StatisticsIT Information technologyITRS International Transactions Reporting System

Abbreviations

ix

JICA Japan International Cooperation AgencyJOCV Japan Overseas Cooperation VolunteersJSA Administered Account for Selected Fund Activities—JapanJVI Joint Vienna InstituteKM Convertible MarkaKOICA Korean International Cooperation AgencyLFS Labor Force SurveyLSMS Living Standards Measurement SurveyMAFF Ministry of Agriculture, Forestry, and FisheriesMDGs Millennium Development GoalsMEF Ministry of Economy and FinanceMFSM Monetary and Financial Statistics Manual, IMFMIME Ministry of Industry, Mines, and EnergyMoC Ministry of CommerceMOEYS Ministry of Education, Youth, and SportsMoH Ministry of HealthMoP Ministry of PlanningMoT Ministry of TourismMPT Ministry of Post and TelecommunicationsMSA Multisector AdvisorMTPW Ministry of Transport and Public WorksNBBH National Bank of Bosnia and HerzegovinaNBC National Bank of CambodiaNBFRY National Bank of the Federal Republic of YugoslaviaNBRS National Bank of Republika SrpskaNBU National Bank of UkraineNIS National Institute of StatisticsNOE Nonobserved economyNPRS National Poverty Reduction StrategyOHR Office of the High RepresentativePHARE Poland Hungary Action for Reconstruction of the EconomyPMAS Poverty Monitoring and Analysis SystemPMG Priority Mission GroupPMIS Poverty Monitoring Information SystemPPI Producer price indexRDD Research and Development DepartmentRGC Royal Government of CambodiaROSC Report on the Observance of Standards and CodesRS Republika SrpskaRSIS Republika Srpska Institute of StatisticsSCB Statistics capacity buildingSDDS Special Data Dissemination StandardSEDP Socioeconomic Development PlanSFOR Stabilization ForceSFRY Socialist Federal Republic of YugoslaviaSIAP Statistics Institute for Asia and the PacificSIDA Swedish International Development Agency1993 SNA System of National Accounts 1993STA Statistics Department technical assistance, IMFTA Technical assistanceTCAP Technical Cooperation Action Plan

x

Abbreviations

UK-DFID United Kingdom Department for International DevelopmentUNDP United Nations Development ProgramUNECE United Nations Economic Commission for EuropeUNFPA United Nations Population FundUNICEF United Nations Children’s FundUSAID United States Agency for International DevelopmentYUD Yugoslav dinars

xi

Abbreviations

PART I

Overview and Lessons Learned

Purpose of Studies

1.1 The importance of statistical information in thepublic arena is well established. Increasingly in thepast decade, events and developments have propelledefforts to improve the quality of statistics. The finan-cial crises of the 1990s increased attention to thetimeliness and reliability of statistics. In addition, aheightened emphasis worldwide on good governance,transparency, and accountability underpinned the driveto improve data quality.

1.2 In this context, statistical capacity building be-came a focus of governments of developing countriesand of donors that provide technical assistance (TA)in support of this effort. Growing out of crisis pre-vention and an emphasis on good governance andtransparency, demands arose from a broad range ofcountries for TA. Demands also came from particulargroups of countries and regions—for example, thetransition countries and a growing list of postconflictcountries. Both these groups required fundamental re-forms and rebuilding of the statistical infrastructure.

1.3 Responding to these needs, developing countrygovernments and TA providers alike focused scarceresources on increasing delivery of TA and imple-menting plans for improving statistical capacity. Overthe past decade, TA in statistics provided by the Inter-national Monetary Fund (IMF), for example, increasedmore than fourfold.

1. Overview

Thomas K. Morrison*

3

*The author would like to take this opportunity to thank Carol S.Carson, who was director of the IMF Statistics Department whenshe approved my one-year sabbatical in 2003, giving me the timerequired to organize this project. I would also like to express myappreciation to the four authors of the case studies, who did nothave the luxury of a sabbatical but somehow managed to carve outthe necessary time from their busy schedules. Joan Gibson pro-vided valuable editorial assistance. Thanks also to the variouscountry officials who provided valuable comments. The views ex-pressed in this paper are those of the authors and myself and do notnecessarily reflect those of the IMF.

1.4 Sufficient experience has now been gained to per-mit stepping back and examining the effectivenessand impact of capacity-building efforts and attempt-ing to learn lessons about how to improve these ef-forts. The country case studies in this report are basedon the experiences of the four authors providing IMFTA. Both Mr. Abbasi’s and Mr. Atcherley’s experi-ences have been as resident long-term multisector sta-tistics advisors, in Cambodia and Bosnia and Herze-govina (BiH), respectively. Both Mr. Kucera’s andMr. Slack’s experiences, on the other hand, have beenas monetary statistics experts involved in sustainedmedium-term programs of short-term missions to theUkraine and BiH, respectively.

Scope and Structure

1.5 The case studies were selected to cover differentTA delivery mechanisms, sectors targeted, and coun-try situations. In the cases of Cambodia (see Chapter 2,Abbasi) and BiH (see Chapter 3, Atcherley), the me-chanism was IMF multisector TA, focusing on the sta-tistical system as a whole and including statistical or-ganization, the main macroeconomic statistics sectors(balance of payments, monetary, government finance,national accounts, and price statistics), and selectedsociodemographic statistics. In the cases of BiH (seeChapter 4, Slack) and Ukraine (see Chapter 5, Kucera),on the other hand, the mechanism was more narrowlyfocused single-sector (i.e., monetary statistics) TA,delivered by a series of short-term missions over amedium-term period.

1.6 All three countries are transition countries, hav-ing transformed over the past decade from planned tomarket economies, but in the different regional con-texts of Eastern Europe, the former Soviet Union, andAsia. The two case studies on BiH provide the oppor-tunity to examine two different approaches to TA inthe same country. In addition, while the authors’ ex-periences originate from their IMF assignments, the

Statistical Capacity Building

authors endeavor to recognize the capacity-buildingefforts in the countries in a broader context, includingnot only IMF activities but also those of the countryauthorities and other TA providers.

1.7 The four case studies are similarly structured,covering similar medium-term time frames in the late1990s and early 2000s. They generally begin by de-scribing the setting, giving a snapshot of the country’sstarting point in terms of general background, institu-tional setting, capacity, and statistical outputs. Theythen review the medium-term strategies and plans forimprovement that guided the capacity-building activ-ities. Following that is a description of the implemen-tation and outcomes, focusing on concrete outputs andquantitative indicators. Obstacles and shortcomings inimplementation are also noted and explained. The clos-ing sections present conclusions and lessons learned.They emphasize factors that contributed to achievingobjectives and, in cases where implementation fellshort, discuss the reasons and identify lessons learned.

Setting

1.8 The four studies describe the settings in the threecountries, first in terms of general sociopolitical andeconomic background—thereby providing the gen-eral context in which capacity-building efforts werelaunched—and second in terms of the status of statis-tical capacity building. In the first context, as notedearlier, all three countries were transition countries,although they were moving from different centrallyplanned systems and also from different types andlevels of sociopolitical and economic development.Ukraine was moving from the former Soviet Unionmodel, BiH from the Yugoslav model, and Cambodiafrom an Asian version of central planning. Ukrainehad the advantage that it was not postconflict, com-pared with both BiH and Cambodia, where basic in-frastructure had been destroyed and had to be rebuilt.As a result, poverty and a lack of resources were rel-atively greater in BiH and Cambodia than in Ukraine.All three countries benefited from large amounts offoreign assistance—which was most pervasive in BiH,where the Dayton Agreement set up a complex and di-vided political structure to rule the country. Ukraineand Cambodia were fortunate in experiencing theirtransitions with relative political stability and unity.

1.9 In terms of statistical capacity, all three countriesbegan their transition needing fundamental reforms

4

of their statistical systems, formerly based on cen-trally planned models. Interestingly, decentralizationitself created serious problems to be overcome in set-ting up new systems, as traditional sources of datacollection disappeared and the dispersion of respon-sibilities caused confusion and problems of coordi-nation. Staff in statistical agencies needed basic train-ing in the new methodologies. New institutions had tobe created or rebuilt in the two postconflict cases. Inall cases, the situations in the beginning were betterfor statistics produced by the central banks (e.g., mon-etary and balance of payments statistics), where rela-tively strong leadership and resources provided betterbases on which to build. Nevertheless, in all countrycases, statistical outputs at the beginning of the periodwere fairly limited. Statistical legislation was alsolacking in all three countries.

Strategies and Plans

1.10 The authors discuss strategies for statistical ca-pacity building and TA either for an entire statisticalsystem (Chapters 2 and 3, Abbasi and Atcherley) or fora sector or project (Chapters 4 and 5, Slack and Kucera).

1.11 Not surprisingly, the difficulties described in thecase studies dealing with systemwide strategies aregreater than those in the studies dealing with sector-specific strategies. In both Cambodia and BiH,Abbasi and Atcherley describe a lack of initial over-all strategies to guide the capacity-building efforts ofthe authorities and the donors. As a result, progress inthe early years in both cases was seriously hamperedby gaps in coordination and coverage, confused andoverlapping responsibilities, and a lack of effectiveleadership.

1.12 Coordinated strategic planning in Cambodia im-proved over time as authorities and donors gained ex-perience, but coordination in BiH has been more un-even. Cambodia made greater progress, because it wasnot hindered by the divided political structure experi-enced by BiH. A critical event in Cambodia was theconvening by the authorities in October 2002 of aworkshop, “Partnerships in Statistics Capacity Build-ing for Cambodia,” where a Statistics Master Plan forCambodia was agreed upon by the various statisticalagencies of the government and the donor community.This workshop probably would not have been possiblehad it not been for (i) the agreement of the majordonors in 2001 to finance jointly and work together on

1 • Overview

5

a medium-term Technical Cooperation Action Plan(TCAP) that paved the way for closer donor coopera-tion and (ii) Cambodia’s decision in 2001 to partici-pate in the IMF’s General Data Dissemination System(GDDS), which provided a comprehensive planningframework. Even in the more difficult situation of BiH,in Chapter 3 Atcherley describes four stages of theevolution of statistical strategy, leading finally to a sit-uation where statistical legislation is in place and anempowered state statistical office is ready to embarkon more prioritized strategic planning.

1.13 At the sector- and project-specific level in Chap-ters 4 and 5, Slack and Kucera point out that strategieswere agreed upon fairly early by the authorities and theIMF to improve monetary statistics over the mediumterm in Ukraine and BiH. However, it became morecomplicated than simply using a kind of generic blue-print or “boilerplate” approach. Slack describes howflexibility and a sensitivity to logistical and pragmaticconstraints were needed in deciding an appropriatephasing of TA delivery. The development of a unifiedstrategy in both cases was simplified because the IMFwas the major donor assisting with monetary statistics,and relatively little coordination with other donors wasnecessary. Although the IMF had developed an ana-lytical framework specifically for improving the mon-etary statistics of the former Soviet Union, Kucera ex-plains that this framework had to be adjusted to theparticular circumstances in Ukraine, given that it takestime to change the culture of institutions.

Implementation

1.14 The discussions of implementation in the four casestudies demonstrate clearly the importance of having awell-planned and coordinated strategy from the start.

1.15 In the two cases of TA focused only on monetarystatistics, clear initial strategies and definite objectivesled to a logical progression of outputs. While someunexpected setbacks and delays inevitably occurred,the existence of a “road map” served to keep the proj-ects broadly on track. In both cases, the principalmode of TA delivery was a sustained series (overseveral years) of short-term missions, supplementedby training courses. The missions, staffed by the sameexperts to the extent possible, could effectively evalu-ate progress since the last mission, advise about appro-priate next steps, and then leave the officials to man-age the implementation largely on their own during the

intervals between missions. This served to promote“ownership” of the capacity-building efforts and tobetter ensure their sustainability. Examples of somekey actions that facilitated implementation were (i) theformation of a working group in Ukraine to improveinternal coordination within the central bank and (ii) aneffort in BiH to bring the commercial banks more ef-fectively into the process through seminars explainingwhy the reforms were necessary and through actionsto reduce their reporting burden.

1.16 On the other hand, in the two cases where TAfocused on the statistical system as a whole, imple-mentation suffered from a lack of initial overall strat-egy. This problem was most serious in BiH where thefragmented political structure and a huge but uncoor-dinated foreign donor presence resulted generally (ex-cept for the financial sector) in a number of uncoordi-nated and sometimes inconsistent strategies, despitethe expenditure of substantial resources. In addition,the scarce staff resources of BiH were stretched thinby the demands of so many projects and were spreadover many activities, some less critical than others.Another serious problem was the phasing of inputs,exemplified by the lack of resources provided to de-velop source data.

1.17 These problems had the greatest impact on theBiH real sector statistics, whereas financial sector sta-tistics (i.e., balance of payments, monetary, and gov-ernment finance statistics) benefited from the fact thata single donor (i.e., the IMF) could take the lead in TAand agree with one national institution (i.e., the cen-tral bank) on unified and coordinated strategies. Onetelling example of the difference strong unified lead-ership can make was the personal intervention of theBiH Office of the High Representative (OHR), withstrong support from the governor of the central bank,in 2002 to impose the passage of statistical legislation,thereby strengthening the national statistical office.

1.18 Cambodia suffered as well from a lack of over-all strategy in the early years, but both the donors andthe authorities recognized and began to deal with theseproblems at a relatively early stage. Eventually, withthe critical support of “champions” for statistics (seeChapter 2, Abbasi), a coordinated and unified strat-egy was developed. A number of factors served tohighlight statistics in Cambodia and stimulate the de-mand for better statistics by policymakers. Statisticswere given prominence in the Millennium Develop-ment Goals (MDGs), the National Poverty Reduction


Strategy (NPRS), and the second Socioeconomic De-velopment Plan (SEDP).

1.19 Nevertheless, an important problem of phasingin Cambodia—the failure for so long to enact statisti-cal legislation despite the completion of many draftsand preliminary steps—hindered statistical capacity-building efforts. This lack of legislation adverselyaffected the implementation of TA, undermining theauthority of statistical agencies and leading to prob-lems in response rates and registration by businesses.

Lessons Learned

1.20 A main purpose of this project was to see whethera series of country case studies on statistical capacitybuilding, conducted within a single framework, wouldyield some common threads and patterns of conclu-sions and lessons learned. If so, the lessons involvedcould be more generally applicable than if they wereassociated only with an individual case study. This isnot to say that the findings of each case study are notrich enough by themselves to provide provocativelessons for TA practitioners. Indeed, much of the valueof the individual case studies is the flavor and the tex-ture of the lessons presented as real-world experience.

1.21 When the case studies are examined together,they do indeed provide common lessons. The mostcommon are listed below under four headings: (i) les-sons for country authorities, (ii) lessons for TA donors,(iii) lessons regarding prerequisites, and (iv) lessonsfor transition and postconflict countries. These cate-gories may be somewhat artificial (for example, thelessons for authorities and for donors are not mutuallyexclusive), but they serve to focus attention on wherethe lessons may be the most relevant.

Lessons for Country Authorities

• The importance of high-level and unified leader-ship in the government to push statistical capacitybuilding was emphasized especially in the multi-sector cases of BiH and Cambodia (see Abbasi’sfocus on “champions” in Chapter 2), but also in thesingle-sector cases (i.e., leadership within the cen-tral banks) of BiH and Ukraine. Unified leadershipin BiH and Cambodia was complicated by inade-quate attention to the impact of decentralization onthe statistical system. In BiH the fragmented politi-cal structure presented serious challenges, whereasin Cambodia high-level champions emerged to pro-

6

mote statistical capacity building. Some examplesof successful leadership in BiH are noted, however,such as the strong efforts by the governor of the cen-tral bank to take responsibility for and improve fi-nancial statistics, and the pushing through of statis-tical legislation in 2002.

• Taking ownership of statistical capacity-buildingactivities and reforms was key in all cases, particu-larly with respect to sustaining the reforms. Inter-related with the leadership lesson above, this lessonalso bespeaks the willingness of authorities to pro-vide the necessary staff and budgetary resources tocommit to the process. The Ukrainian authoritieswere relatively more successful in taking ownership,whereas in BiH the largely uncoordinated mannerin which TA was delivered by donors discouragedownership in the early years, and in Cambodia pro-gress has continued to be hindered by inadequatebudgetary resources.

• Taking charge of internal coordination of statis-tical activities in the government was critical. InUkraine, for example, internal coordination withinthe central bank was a serious obstacle until an inter-departmental working group coordinated the com-piling of monetary statistics within the bank. On theother hand, in BiH, internal coordination driven bythe governor was a positive feature in the centralbank’s efforts to improve statistics.

• Giving responsibility for donor coordination to asingle government institution was important in thetwo multisector cases. In BiH, donors improvedeffective coordination gradually but never fullyachieved it, whereas in Cambodia high-level sup-port of a forum for donor coordination could serveas a model for other countries. If the authorities donot take the lead in this area, experience shows thatdonors by themselves find it difficult to coordinatetheir respective activities. The identification of re-sponsibility for donor coordination and the relatedresponsibility for producing an overall strategy andwork plans should be part of statistical legislation.

• Attention to user consultation was found to raisethe profile of statistics by fostering demand for sta-tistics and making products more relevant. The casestudy on Cambodia shows how the government usedseminars and other means to consult with users.

Lessons for Technical Assistance Donors

• Regarding donor coordination, donors should en-gage in projects in accordance with the strategy andpriorities of the host country, and not propose proj-

1 • Overview

7

ects that are tied to their own strategic interests andpriorities if these are at odds with the priorities of thehost country. If the host country does not have aclear overall strategy, then donors should do the bestthey can to coordinate among themselves. BiH (SeeChapter 3, Atcherley) has been an example of a hostcountry in which donors have at times not allocatedproject resources in accordance with the country’spriority needs.

• Donors should pay attention to the appropriate phas-ing of project inputs in view of the absorptivecapacity of the country and the most effective se-quence of inputs. The logical phasing for improvingmonetary statistics in BiH provided in Table 4.2 (seeChapter 4, Slack) is a good example of the effec-tive phasing of project inputs. BiH (see Chapter 3,Atcherley), on the other hand, showed how too muchattention to technical training initially at the ex-pense of developing source data can prevent signif-icant progress in improving national accounts.

• It is important to complement strategies and recom-mendations with hands-on TA and to work with theauthorities to show them how to implement recom-mendations in the day-to-day, nitty-gritty activities.The effectiveness of the hands-on approach is par-ticularly noted in Cambodia and BiH, which pro-vide examples of how important it is for experts tobe willing to work with the authorities on the mun-dane aspects of improving statistics, and not onlyon the big picture issues (see Chapter 2, Abbasi, andChapter 4, Slack). This hands-on approach is facil-itated where donors use resident long-term advisorsrather than only short-term missions. Similarly, it isalso critical that donors focus on direct data pro-duction activities and not solely on supporting ac-tivities such as seminars, study tours, and provisionof office space, important though they may be (seeChapter 3, Atcherley).

• Closely related to the last point is that donors shouldbe careful in their choice of experts. Several exam-ples are noted in Cambodia (Chapter 2, Abbasi) andBiH (Chapter 3, Atcherley) where the effectivenessof projects was compromised by experts who werenot adequately qualified.

• The choice of mode of delivery by donors (e.g., long-term advisor, short-term mission, series of short-termmissions, training, workshops) should match theparticular needs of the country’s situation. The twocases of TA in monetary statistics (Chapters 4 and 5,Slack and Kucera) are good examples of how a sus-tained series of short-term missions matched wellthe need to provide patient medium-term assistance

that allowed for the difficult process of culturechange to take place and learning/technology trans-fer to occur. Resident advisors can complementshort-term missions in countries requiring intensivesustained TA (which is especially the case in tran-sition and postconflict countries).

• The ultimate success or failure of a project will de-pend on its sustainability and on whether the donorshave designed an exit strategy. This is an importantcomplement to the ownership lesson listed aboveunder “Lessons for Country Authorities.” Again,the Slack and Kucera case studies (Chapters 4 and 5)are good examples of how the TA was geared con-sistently toward the objective of eventual total trans-fer to the authorities.

Lessons Regarding Prerequisites

• Many examples are provided, particularly in Chap-ter 2 (Abbasi) and Chapter 3 (Atcherley), of howimportant statistical legislation is for progress inimproving statistics in a variety of areas, such ascoordination, adequate authority of the statisticalagency, clear designation of responsibilities, andprofessional autonomy. Nevertheless, in BiH (Chap-ter 4, Slack) a culture of data reporting and goodwillamong data providers and compilers was an effec-tive short-term substitute for statistical legislation.

• Chapter 2 (Abbasi) and Chapter 3 (Atcherley) alsoprovide many examples of how critical adequateresources are, both in terms of staff and budgetaryresources, for achieving objectives in capacity build-ing. This lesson is closely related to the ownershiplesson for country authorities, as it is essential incases like Cambodia and BiH, where large donorassistance is provided, that the country authoritiesgradually take over more and more of the owner-ship and financing for statistical capacity building.Resource shortages, however, need not prevent theproduction of at least limited yet reliable statisticsin the near term. In BiH (see Chapter 4, Slack), care-ful staff selection, planning of workloads, and ded-icated local officials compensated for initial short-ages in staff numbers and technology.

Lessons for Transition and Postconflict Countries

These lessons are not exclusively for transition andpostconflict countries, but they are of particular im-portance for these cases (and also perhaps for a num-ber of very poor, resource-starved countries).


• To be effective, the scale of assistance has to be sig-nificantly larger than in typical cases of TA, becausefundamental changes in methodology and cultureare often involved, as well as the basic building orrebuilding of statistical infrastructure.

• Donors must be extremely flexible in their approachand must adapt their strategies and modes of deliv-ery to unusually difficult circumstances and lowabsorptive capacity.

• Both donors and country authorities should be pa-tient and prepared for sustained efforts over a

8

multiyear period, and should be focused consis-tently on eventual total ownership by the authoritiesand on donor exit strategies.

• Institutional and statistical organization issuesare critical for these countries. For example, it wasshown for both Cambodia and BiH that a decentral-ized statistical system seriously challenges post-conflict countries engaged in fundamental reform ofthe statistical system. A strong central statistical of-fice can be a great help in coordinating and imple-menting large statistical improvement programs.

PART II

Case Studies

Introduction

2.1 This case study broadly assesses the evolution ofCambodia’s official statistics and statistical capacitybuilding, emphasizing the partnership between theCambodian authorities and their TA partners, as wellas the lessons learned.

2.2 An assessment of Cambodia’s official statisticsand statistical capacity requires an appreciation of theCambodian context. Thus, the study begins with abroad overview of socioeconomic development inCambodia, the current legal and institutional environ-ment, and resources available to the government forofficial statistics.

2.3 The study then focuses on statistical capacityplanning and implementation in Cambodia between1993 and 2003, noting that much has been achievedin terms of improvements in the coverage, quality,and range of statistics, especially as donor coordina-tion and strategic planning improved over the period.The importance of the role of bilateral donors andinternational organizations in achieving these out-comes is also reviewed.

2.4 Critical success factors and constraints are iden-tified, including champions and leadership; ownershipand commitment by the authorities and local statisticscounterpart staff; awareness and understanding of theneed for sound official statistics by the authorities anddonors; phasing and mode of TA; effective user con-sultation and involvement; independence of officialstatistics and technical autonomy of the National Insti-tute of Statistics (NIS); and informal statistical coor-dination mechanisms.

2. Cambodia: Multisector Statistics

Zia Abbasi*

11

*The author is IMF multisector statistics advisor in Cambodia.The IMF is grateful to the Australian Bureau of Statistics for re-leasing Mr. Abbasi to take this assignment. His assignment in Cam-bodia is financed by the Administered Account for Selected FundActivities—Japan (JSA).

2.5 The case study concludes with a discussion of thelessons learned concerning how to improve the effec-tiveness of statistical capacity building.

Setting

2.6 An assessment of Cambodia’s statistical capacity,especially with respect to data quality, requires an ap-preciation of its current socioeconomic development,as well as an understanding of the legal and institu-tional environment and resources available to the gov-ernment for official statistics.

Socioeconomic Situation

2.7 Cambodia has come a long way since the signingof the Paris Peace Accords in October 1991. The pre-ceding two decades of war and the tragic loss of mil-lions of lives during the Khmer Rouge years had leftit nearly destitute in terms of both economic infra-structure and human capital. An estimated 39 percentof Cambodians lived below the poverty line in 1993.1

However, since 1993 Cambodia has made significantadvances in rebuilding its political, economic, and so-cial infrastructure. Three general elections have beenheld over the past decade, and a degree of political sta-bility has been achieved.2

2.8 Despite considerable effort by the Royal Gov-ernment of Cambodia (RGC), together with its bilat-eral and multilateral development partners, to improveliving standards and the quality of life of the Cambo-dian people, Cambodia still has one of the highestpoverty rates in Southeast Asia (estimated at about 36 percent in 1997).3 Its Human Development Index

1NIS, Socioeconomic Survey of Cambodia, 1993–94.2General elections were held in May 1993, July 1998, and July

2003.3NIS, Socioeconomic Survey of Cambodia, 1996.


score is currently ranked 130th out of 173 countriesand is the lowest in Southeast Asia.

2.9 Cambodia is a small economy, with gross domes-tic product (GDP) in 2002 of about US$3.9 billion anda per capita income of US$291 per annum.4 WhileGDP growth has averaged just over 6 percent perannum in recent years, GDP per capita growth has av-eraged about 3.5 percent. The economic base still de-pends largely on agriculture, fisheries, and forestry,with these sectors accounting for about 34 percent ofGDP. The industry sector accounts for about 26 per-cent of GDP, with the textiles, garments, and footwearindustries accounting for 50 percent of this sector. Theservices sector accounts for 24 percent of GDP and is largely driven by tourism. Most of the economicgrowth since 1995 has been due to the growth in thetextiles, garments, and footwear industries.

The Evolution of Official Statistics in Cambodia

2.10 The Cambodian national statistical system ishighly decentralized. In addition to the NIS of the Min-istry of Planning (MoP), all ministries (and in somecases, departments within ministries) and other insti-tutions have planning and statistics units responsiblefor producing statistics. While annual appropriationsfor official statistics have grown from an estimatedUS$0.2 million in 1993 to about US$1.5 million in2003, this is only about 0.2 percent of the total bud-get. The NIS budget appropriation’s share of the totalhas increased from about US$20,000 to US$427,300.However, these appropriations are still not commen-surate with the demand for official statistics, and theNIS and other RGC statistics units do not have in-dependent budgets. Over the past decade, the staffinglevel of the NIS has increased from 495 staff to morethan 650 staff. However, staff with statistical skills inother RGC statistics units have been significantly re-duced over the same period.

2.11 Compounding the need for adequate and auton-omous appropriations for official statistics and ap-propriate staffing levels is the legal and institutionalenvironment. The NIS and the few functioning RGCstatistics units do not have the legal and institutionalrequisites to effectively collect the range of quality sta-tistics required to monitor and measure development

12

results. No crosscutting statistics legislation is imple-mented to provide the legal basis for official statistics.Similarly, no clear demarcation exists between offi-cial national and sector statistics and other manage-ment information needed to monitor governmentoperations, services, and outcomes.

2.12 In the past decade, Cambodia has progressivelymoved from a centrally planned economy to a marketeconomy, resulting in a significant transition in allspheres of Cambodian society. This transition has re-sulted in the dismantling of central planning, the de-centralization of decision making in relation to gov-ernment operations, and a high level of privatization,as well as opportunities for the private sector to grow.The central role of the MoP in economic managementand central planning for the provision of governmentservices has increasingly devolved to the Ministry ofEconomy and Finance (MEF) and various other lineministries. This process unfortunately also led to aweakening of data coordination mechanisms and cen-tralized information management within the govern-ment, which was not the fault of decentralization butof inadequate attention to the maintenance of coordi-nated data systems during decentralization.

2.13 Similarly, coordination among donors assistingthe RGC in different sectors was not as effective as itcould have been in the early 1990s. Donors tended tofocus on only one government agency, with their TAprovision covering program development, imple-mentation, evaluation, and statistics for that agency.With a few exceptions, statistics were generally col-lected, compiled, and produced without taking intoaccount the needs of the wider user community out-side the sector in question. The government agenciesinvolved tended to monopolize the TA support, andthe NIS’s role in producing national statistics and itsstatistics collection and processing capacity were oftenoverlooked.

2.14 Another important factor that has affected the de-velopment of Cambodia’s official statistics is thatdonors tend to support the production of statistics ofspecific interest to them. As a result, the developmentof official statistics has varied across sectors and hasbeen limited for core national statistics, according tothe number of interested donors and the level of TAprovided. Agriculture, education, health, population,and poverty statistics have attracted relatively more TAsupport than macroeconomic, industry, services, envi-ronment, natural resource, and other social statistics.4 NIS, July 2003 National Accounts of Cambodia, 1993 to 2002.

2 • Cambodia: Multisector Statistics

13

2.15 To understand the environment in which Cam-bodia’s official statistics are being developed, donorsshould consider not only the issues constraining sup-ply but also the demand for these statistics. In fol-lowing up on the Millennium Declaration, Cambodiahas launched several initiatives, including several in-stitutional reforms, to meet its global and nationalcommitment to fight extreme poverty. The NationalPoverty Reduction Strategy (NPRS) builds upon apolicy stance of economic growth and poverty re-duction. The RGC’s reform agenda essentially focuseson deepening economic reforms and macroeconomicstability, including fiscal and monetary reforms, tradeand investment promotion, administrative reforms,military demobilization, and improved fishery andforestry management. The RGC has identified fourpriority sectors: education, health, rural development,and agriculture. In each sector, major programs ofeconomic and social infrastructure development areunderway. It has also prepared a Governance ActionPlan to enhance governance in key areas such as thejudicial and legal system, anticorruption programs,and public finance.

2.16 Related to the Millennium Development Goals(MDG), the NPRS, and various economic, fiscal, andmonetary reform initiatives, the demand for officialstatistics has increased substantially. An integratedand relevant national statistical system has, therefore,emerged as an important component of overall re-forms. High-standard and quality statistics are neededto capture economic growth and poverty reduction,as well as other dimensions of social development(e.g., gender equality, education and literacy, healthand nutrition) and related demographic, economic, en-vironment and natural resources, and social statistics.

Strategies and Plans for Improvement

2.17 The RGC authorities recognized at an early stagethe need for good-quality national statistics on the Cam-bodian economy, society, and population, and reflectedthat accordingly in the first five-year SocioeconomicDevelopment Plan (SEDP). However, at that time, noseparate strategic plan focused solely on statistical de-velopment. The production of official statistics at thesector level was viewed as a planning function of eachof the relevant ministries. Given the limited financial resources available to the government, it focused ini-tially on moderately redeploying resources and staff tostrengthen the planning function within line ministries.

2.18 The authorities’ initial position was that the exist-ing administrative processes and regulations wouldcontinue to provide the data needed to compile officialstatistics. What was not appreciated at the time was thatthe administrative data collections set up to provide sta-tistics for a centrally planned economy were neithersuitable nor of sufficient quality to meet the informationneeds of policymakers in a transition economy. Further-more, the central government’s beginning to decentral-ize decision making and devolve responsibility was alsoweakening the MoP’s capacity to ensure effective datacoordination. Another compounding factor was uncleardefinition of responsibilities for specific statistics. Thisresulted in duplication of effort and uncertainty aboutwhich sets of statistics were official (e.g., consumerprice indexes (CPIs) were produced by the MEF and theNational Bank of Cambodia (NBC) during 1994 and1995, and national accounts were produced by the MEFand MoP from 1993 to 1995).

2.19 By early 1994, the authorities realized that whiledevolution was important in improving governmentoperations and providing services, responsibility forbasic official statistics needed to be centralized, andstronger data coordination mechanisms were needed.This led to the establishment of the NIS, as the na-tional statistical institute, within the MoP in that year.The NIS was given sole responsibility for key officialnational statistics, such as the CPI, national accounts,and population statistics. The implementation of theSubdecree on Statistical Obligations in 1997 elevatedthe NIS from a department to a directorate general andstrengthened its coordination role.

2.20 The authorities also realized that the administra-tive data collections were not providing suitable sourcedata to compile key economic and sociodemographicstatistics and that statistics staff had little or no experi-ence in compiling the statistics required. As resourceswere scarce, the authorities recognized the need to en-list the donor community in financing and providingTA in the following areas: developing statistical ca-pacity, skills, and infrastructure; conducting surveys;and compiling key economic and sociodemographicstatistics.

2.21 While the authorities would have preferred toengage the donor community in supporting an inte-grated approach to developing the national statisticalsystem, the government was constrained by the dif-fering priorities of the various multilateral and bi-lateral donors. As a result, the authorities developed


separate statistical plans and partnerships with keymultilateral donors for each area of statistics from1993 to 1997. The main partners in statistical devel-opment over the past decade have been the Asian De-velopment Bank (ADB) and IMF, for economic andfinancial statistics; and the United Nations Develop-ment Program (UNDP) and the United Nations Popu-lation Fund (UNFPA), for sociodemographic statis-tics. While these multilateral donors provided themain financing and/or TA in developing Cambodia’sofficial statistics, the authorities also engaged bilateraldonors and other multilateral donors to supplementmajor development activities, conduct one-off surveys,and support sector statistics.

2.22 By 1997, both the authorities and donors recog-nized that separate TA agreements had hampered ef-fective data coordination and made it difficult to re-duce duplication, especially in collecting data. Thisled to improvements in the coordination of TA amongdonors and among line ministries. It also resulted in astrengthening of the NIS’s role in facilitating moreeffective data coordination and integration between1998 and 2002.

2.23 The second SEDP also helped focus on statisti-cal capacity development and the need for improveddata coordination and integration. The SEDP includedinitiatives to expand data collection and compilationfor a number of sectors, especially poverty statistics.However, statistical development was still seen as partof the overall development objectives for each eco-nomic and social concern and as the responsibility ofthe relevant line ministry.

2.24 By 2002, statistical development planning hadreached a degree of maturity and become more inte-grated, reflecting the authorities’ two main objectives,economic growth and poverty reduction. The statisti-cal objectives set out in the second SEDP and the firstNPRS reflect this integrated approach. In addition, thefirst NIS Strategic Plan (2002–2006) and the five-yearNIS Forward Work Program (2003–2007) were re-leased in July 2002. These documents outline long-term statistical development strategies and annualdevelopment priorities for the NIS and the nationalstatistical system. The authorities also convened a sta-tistics workshop in October 2002, “Partnerships inStatistics Capacity Building (SCB) for Cambodia.”Participants included all line ministries with responsi-bility for official statistics and the donor communityin Cambodia. In addition to the NIS plan and work

14

program, many other statistical capacity developmentproposals were presented at the workshop. In terms ofcontent, the NIS plan and work program and the pro-posals discussed at the workshop constitute the mainelements of a statistics master plan for Cambodia. TheNIS authorities intend to develop the master plan in2005 for release in early 2006.

2.25 Cambodia’s adoption of the IMF’s General DataDissemination System (GDDS) framework has alsofurther improved statistics planning and monitoring.The NIS prepared the GDDS metadata templates,with the assistance of an IMF GDDS mission and themultisector advisor (MSA), in December 2001. TheRGC officially adopted the GDDS in February 2002.Short-term and medium-term plans to improve theproduction and dissemination of official statistics havebeen developed and are being implemented as re-sources permit. The NIS has monitored progress andcompleted its first GDDS Review in February 2003.Good progress has been made in implementing short-term plans, especially for balance of payments, na-tional accounts, consumer prices, money and banking,education, and poverty statistics.

Implementation and Outcomes

Statistical Organization Prerequisites

Statistical legislation and regulations

2.26 Considerable attention has been given to devel-oping statistical legislation in recent years. The primeministerial Subdecree on Statistics Obligations wasdrafted and promulgated in 1997. The drafting of thestatistics law commenced in earnest in early 2000,with ADB providing support. The draft law was sub-stantially revised between November 2001 and May2002, with IMF MSA assistance, to incorporate thefundamental principles of official statistics and to rec-ognize the need for stronger coordination mechanisms,given the highly decentralized nature of the nationalstatistical system. The draft law was approved by theCouncil of Ministers (COM) in June 2002 and sub-mitted to the National Assembly in August 2002. Thelaw was not yet enacted by the beginning of 2005. Inaddition, a draft Subdecree on Statistics Obligations,Roles and Responsibilities has been prepared and isbeing discussed by an interministerial technical work-ing group. The Subdecree on Designated Official Sta-tistics and related statistical regulations are currentlybeing prepared. Statistical regulations have also been


15

prepared and approved by the NBC Governor to im-plement the International Transactions Reporting Sys-tem (ITRS) collection and the release of the new quar-terly Balance of Payments, Cambodia publication.

Statistics coordination

2.27 Since 1997, a significant improvement has oc-curred in statistical TA coordination arrangementsamong most line ministries and the NIS, as well asamong the donor community. Good examples of effec-tive partnerships between the authorities and the donorcommunity are the 1998 General Population Census ofCambodia, the 1999 Socioeconomic Survey of Cambo-dia, and the 2000 Demographic and Health Survey.

2.28 The draft statistics law and subdecrees specifi-cally provide for establishing the Statistics AdvisoryCouncil and Statistics Coordination Committee (includ-ing provisions relating to the composition and func-tions of these bodies), as well as for implementingother formal coordination mechanisms and institu-tional arrangements. The NIS has established technicalworking groups, with ADB assistance, for statisticallegislation, agriculture statistics, environment andnatural resources statistics, and foreign trade statis-tics. The authorities have also expanded and im-proved statistics cooperation, with IMF MSA assis-tance, among the NIS, ministries, and other institutionsin relation to data-sharing arrangements.

2.29 The IMF MSA has assisted the NIS in improvingthe coordination of donor financing and TA for statisti-cal capacity building for the MEF, MoP, NBC, and NIS.The MSA has worked closely with staff of the ADB, theEconomic and Social Commission for Asia and thePacific (ESCAP), the German Technical Commission(GTZ), the IMF, the Japan International CooperationAgency (JICA), the Japan Overseas Cooperation Vol-unteers (JOCV), the UNDP, the United Nations Chil-dren’s Fund (UNICEF), the UNFPA, and other expertsand volunteers. Assistance has been provided to the au-thorities in coordinating ADB experts and IMF missionsand in preparing ADB TA and IMF mission proposals.

Statistical planning frameworks

2.30 With IMF MSA assistance, the NIS imple-mented a Strategic and Forward Work Program Plan-ning and Monitoring Framework in February 2002.MoP/NIS senior officials meet monthly and NIS de-partment heads meet weekly to monitor progress in

implementing capacity development and the statisticswork program. The NIS Forward Work Programs(2002–2006) and (2003–2007) and the NIS StrategicPlan (2002–2006) were approved and released in Apriland July 2002, respectively. Work program planningand monitoring has also been introduced for statisticsunits in the MEF; the NBC; the Ministry of Educa-tion, Youth and Sports (MOEYS); and the Ministry ofHealth (MoH) through the GDDS framework. Simi-larly, awareness of the need for work plans and mon-itoring has been significantly raised for most other sta-tistics units of the RGC by the MSA through thedevelopment of SCB project proposals, as well as theconduct of the SCB workshop in October 2002.

Statistics development

2.31 A broad review of current statistical capacitywas conducted with IMF MSA assistance from Janu-ary to May 2002, and the findings were released in theReport on Statistical Capacity Building for Cambo-dia’s Official Statistics in June 2002. The report, con-sidered a draft statistics master plan, includes broadprograms, financing, and TA requirements for devel-oping and improving official statistics. As mentionedabove, the SCB workshop was conducted in October2002 and discussed 58 project proposals. The work-shop was successful in improving donor and govern-ment understanding of Cambodia’s current statisticalcapacity, as well as in raising donor and governmentawareness of the need to:• Implement appropriate statistical legislation and

regulations and ensure that the legislation and reg-ulations are enforced;

• Increase RGC appropriations and donor financingfor official statistics;

• Establish a central body to advise the RGC and NISon statistical development priorities and emergingdemands for statistics;

• Improve coordination within the donor communityand the government on TA issues;

• Establish a central statistics coordination committeeand implement formal mechanisms and other insti-tutional arrangements to ensure effective coordina-tion of statistical capacity building and informationsharing;

• Strengthen staffing, information technology (IT) re-sources, and facilities commensurate with the de-mand for official statistics, and implement moreeffective resource monitoring;

• Improve remuneration and career paths for statisticsstaff;


• Implement an integrated skills development strat-egy for statistics staff;

• Introduce performance management and merit-basedrecruitment and promotion processes for statisticsstaff;

• Ensure efficient use of resources by eliminatingduplication and leveraging the survey and dataprocessing capacity of the NIS;

• Ensure official statistics are produced on an impar-tial and independent basis, and introduce transparentstatistical policies and practices;

• Ensure that the methodological basis for officialstatistics is in line with international standards andgood practice;

• Ensure the quality of statistical methods and theaccuracy and reliability of statistics; and

• Ensure that official statistics and related metadata arereadily accessible, comparable, and timely, and thatperiodicity of release is aligned to data user needs.

Statistical roles, functions, organization structure,and staffing arrangements

2.32 The IMF MSA assisted the director general ofthe NIS in developing a detailed organizational re-structuring proposal in early 2002. A detailed imple-mentation plan, Structures, Roles, and Functions ofthe National Institute of Statistics (July 2003), hasbeen released and is in the process of being imple-mented. The plan includes:• Role and function statements for NIS statistical

units, as well as advisory, coordination, and man-agement bodies;

• Staff competency and duty statements;• Organizational charts for central, provincial, and

district offices;• Recruitment, selection, appointment, probation, per-

formance management, and promotion procedures;• Secrecy agreements and security clearances;• Staff streaming and remuneration; and• Job rotation, training, and development.

2.33 A similar implementation plan on the estab-lishment of an NBC statistics department has beenprepared and approved for implementation. Similarrestructuring plans are being developed and/or imple-mented for some other RGC statistics units, for ex-ample the Ministry of Veterans and Women’s Affairs.

Government financing

2.34 Although not significant, moderate increaseshave been made in government appropriations for

16

NIS statistical activities over the past decade. The2003 budget allocation was US$427,300, an increaseof approximately 20 percent over the 2002 allocation.Similarly, remuneration for senior managers (i.e., di-rectors and above) has increased considerably, and afurther 26 NIS senior statistician positions were iden-tified as part of the donor-assisted Priority MissionGroup (PMG) initiative. Staff in these positions wereto be paid salary supplements from 2004 onward.However, overall financing is still not commensuratewith the demand for statistics.

Staff training and development

2.35 In 2001, ADB financed the establishment of anin-house statistics training program and the NIS Train-ing Center. Findings from a review of NIS trainingand development needs, conducted in January 2002,were documented in the NIS Training Program Re-view Report. The NIS Training and DevelopmentStrategy (March 2002) outlines future directions andtraining courses to be implemented.

2.36 On-the-job training has been provided in macro-economic statistics, especially for balance of pay-ments, labor force, monetary, national accounts, andprice statistics. Data collection and compilation bylocal officials leading to the establishment of self-sustaining statistical systems have been emphasized.In addition, a number of MEF, NBC, and NIS officialshave been sponsored for courses, tertiary studies, andworkshops by the ADB, ESCAP, the IMF, and theStatistics Institute for Asia and the Pacific (SIAP) overthis period. An estimated 10 percent of statistical staffhave gained postgraduate tertiary qualifications in ap-plied statistics through this TA. While overall skillshave significantly developed from a very low skillsbase, skills development has not been integrated acrossgovernment, and some agencies and ministries havebenefited far more than others.

Economic and Financial Statistics

National accounts statistics

2.37 The development of national accounts statisticshas accelerated significantly in recent years, comparedwith the 1990s. This has been partly due to substantialsupport by the ADB (the only donor supporting na-tional accounts development between 1994 and 2000),which provided financing for the significant surveyprogram required to build the national accounts data-base. The development of national accounts statistics


17

in the past two years is a good example of successfulpartnerships between the authorities and TA partners(the ADB and IMF). Significant ADB expert assis-tance and IMF MSA assistance have been provided tothe NIS in improving compilation methods for bothproduction and expenditure accounts aggregates in thepast two years, resulting in significant revisions in1993 to 2000 estimates, including an average increasein real GDP levels of 10 percent.

2.38 Further improvements in the source data andcompilation methods were made during the 2003 esti-mation round, resulting in improved estimates, as wellas an average increase in real GDP levels of 4 percent.The 2003 release of National Accounts of Cambodia,1993–2002 is approximately three times the size ofthe first bulletin (1993–96) released in 1997. Theanalytical content and number of statistical tables re-leased over the years have also significantly increased.Additional detailed tables were included for the in-dustry sector, services sector, household final con-sumption expenditure, and the experimental nationalincome account. The technical notes section of thebulletin has also increased substantially.

2.39 The national accounts are now broadly consis-tent with the System of National Accounts 1993 (1993SNA), although full implementation is expected totake several years. New compilation methods havebeen implemented and revised estimates have beenproduced for many sectors. Experimental national ac-counts using the income approach were developed forthe first time for 1993–2002. The MSA has also as-sisted in improving consistency of the 2001 and 2002estimates with other macroeconomic statistics.

2.40 The NIS has significantly improved access tothe administrative data required to compile the ac-counts and, with ADB experts, improved the qualityof NIS survey data. It has improved coordination offinal administrative and survey data for the relevantreference year for use in compiling national accountsestimates. It is more effectively using existing andnew data sources. Access to the Ministry of Agricul-ture, Forestry, and Fisheries (MAFF) data on cropsand livestock, fisheries, forestry, and rubber produc-tion has been improved, and basic supply/use tableshave been developed for rice, fish, and some otheragricultural products.

2.41 Improved data access has resulted in improvedestimates for international tourism and new estimates

for domestic tourism; new estimates for casinos; im-proved estimates for the finance industry and financialservices charges; improved estimates for public ad-ministration; and new estimates for nonprofit organi-zations serving households. New NIS population pro-jections data have been used to improve compilation ofvarious series, especially household consumption ex-penditure. Access to the Council for the Developmentof Cambodia’s (CDC) investment and nongovernmentorganizations expenditure data, to the Customs and Ex-cise Department (CED) data, and to NBC unrecordedexports and imports estimates have resulted in moreconsistent national accounts estimates. Improved ac-cess to MAFF prices and other price data has led to re-visions in price deflators.

2.42 The AsDB TA has underpinned the develop-ment of the national accounts database. It has assistedthe NIS in conducting three establishment surveys, in1993, 1995, and 2000; several Phnom Penh surveysand two national labor force surveys, in 2000 and2001; the quarterly Survey of Hotels and Manufac-turing Establishments; and the quarterly Retail TradeSurvey. More recently, the ADB has financed a num-ber of one-off agriculture and informal sector surveysfor agriculture, marine and inland fisheries, forestry,rubber plantations, traditional textile manufacturers,small retailers and wholesalers, guesthouses, and landand water transport operators. The results from all ofthese surveys have helped produce new and revisedformal and informal sector gross value added (GVA)ratios and production estimates for a number of in-dustries from 1994 to 2002.

2.43 Work has commenced on compiling quarterlynational accounts. The NIS intends to produce exper-imental quarterly national accounts from March quar-ter 2000 to June quarter 2005. The estimates will beconsistent with the annual accounts on both expendi-ture and production bases. Source data are reasonablytimely and comparable in quality to the annual sourcedata currently used. An ESCAP expert has workedwith the IMF MSA and NIS staff on developing thebasic methodology. The IMF MSA worked with NIScounterparts on compilation methods and the devel-opment of the Excel spreadsheets.

2.44 The ADB provided significant support for sta-tistical skills development in conducting economicsurveys and compiling national accounts statisticsover the past decade. In March 2002, the ADB expertand the IMF MSA conducted a three-week national


accounts statistics course for 40 data compilers andusers from the NIS and other ministries. In addition,two staff have attended the IMF national accounts sta-tistics course at IMF headquarters. IMF MSA andADB experts have provided ongoing assistance andon-the-job training to improve the quality of datacollection, data quality assurance, compilation, andanalysis and dissemination.

Price statistics

2.45 The NIS has significantly expanded its price sta-tistics program since the first release of the monthlyPhnom Penh CPI in 1995, extending consumer pricecollection to five provincial cities in addition to PhnomPenh, with ADB TA. The NIS has produced a linkedCPI series, combining the 2000 series with the 1994series for the Phnom Penh CPI, with MSA assistance.Assisted by an ADB expert, it has also released twoprovincial CPI publications. The MSA assisted theNIS in developing a methodology to link the variouscities’ CPI series to produce an Other Urban CPI andan Urban CPI. A new quarterly Urban CPI publicationwas released in January 2003.

2.46 Work has begun on producing a quarterly pro-ducer price index (PPI) for Cambodia, with ADB andIMF statistics department technical assistance (STA).A producer price survey commenced in August 2003,and the PPI for March quarter 2000 to Septemberquarter 2003 was also released soon thereafter.

2.47 The MSA conducted an abridged IMF price sta-tistics course in March 2003, for 40 data collectors,compilers, and users from the NIS and other min-istries. In addition, two staff have attended the IMF’sJoint Vienna Institute (JVI) price statistics course. On-going assistance and on-the-job training have beenprovided by ADB experts, the IMF MSA, and IMFmissions to improve the quality of data collection,data quality assurance, compilation, and analysis anddissemination. The explanatory and technical notes inquarterly prices publications have been expanded toprovide more information on concepts, sources, andmethods.

Foreign trade statistics

2.48 The development of foreign trade statistics hasbeen less successful, mainly owing to the lack of fi-nancing. However, the NIS has established a ForeignTrade Statistics Technical Working Group, which

18

meets quarterly. The MSA has assisted the NBC andNIS to improve access to existing CED data and hasprepared new report specifications for volume, value,and price data (by commodity and country). Despitesignificant resource constraints, CED has continued tocooperate with the NBC and the NIS and has com-pleted the programming to generate the reports re-quired, with the IMF IT advisor’s assistance. Annualreports have been provided for 2001 and 2002, andmonthly and quarterly reports for 2003.

Balance of payments statistics

2.49 The IMF posted a resident balance of paymentsstatistics advisor to assist the NBC on balance of pay-ments, foreign investment, and foreign trade statisticsfrom May 1995 to May 1997. The NBC has signifi-cantly progressed in improving balance of payments(BOP) and international investment position (IIP) sta-tistics data sources and compilation methods, as wellas consistency with the standards of the IMF Balanceof Payments Manual, fifth edition (BPM5). Compil-ing annual and quarterly BOP and IIP statistics, it pub-lishes annual BOP statistics and quarterly BOP and IIPstatistics. In addition, a Summary Concepts, Sourcesand Methods Manual was released in October 2003.

2.50 The NBC, assisted by the IMF MSA and mis-sions in 2001–2003, has reviewed and improved com-pilation methods and access to existing administrativedata sources for unrecorded exports and imports, re-exports, transportation, travel, other services, income,current and capital transfers, foreign investment, andCambodian investment abroad. More effective use isbeing made of partner country data. BOP and IIP es-timates are now comparable. Revised quarterly esti-mates have been prepared for all quarters from 1998to 2002. More detailed data have been compiled for anumber of aggregates, especially travel credits andunrecorded trade.

2.51 In January 2003, the NBC financed and imple-mented a monthly ITRS to collect much-needed dataof international transactions via the banking system,improving balance of payments estimates. The ITRSdata have been assessed as being of good quality andcovering approximately 90 percent of the total valueof transactions going through the banking system. Arevised questionnaire for foreign direct investmententerprises has also been developed and pilot tested.Unfortunately, the lack of financing has meant that aregular survey has yet to commence.


19

Government finance statistics

2.52 The MEF is progressively implementing therecommendations of the March 2001 IMF mission.MEF reporting of monthly data for inclusion in theIMF’s monthly International Financial Statistics (IFS)has continued, provided on a basis consistent with the IMF’s Government Finance Statistics Manual,1986 (GFSM 1986). The MEF has begun producinggovernment finance statistics (GFS) broadly based onthe IMF’s Government Finance Statistics Manual,2001 (GFSM 2001), submitting preliminary 2001GFS to the IMF Statistics Department. Progress inimproving GFS has generally been slow, owing tolimited financing and the lack of full-time counterpartstaff.

Monetary and financial statistics

2.53 The NBC has significantly progressed in im-proving monetary statistics over the past decade. Ithas introduced a new chart of accounts that conformswith international standards for commercial and spe-cialized banks and has revised the statistical compila-tion procedures to reflect the classification scheme ofthe new chart of accounts. It has reviewed mappingprocedures used for moving from the old NBC chartof accounts to the new chart of accounts and revisedprocedures used for compiling the NBC’s accountstatement. It has also implemented a new methodol-ogy for compiling monetary and financial statistics, asrecommended in the IMF’s Monetary and FinancialStatistics Manual (MFSM). The IMF MSA and the2002 mission assisted the NBC in improving the cov-erage and quality of monetary statistics.

Other economic statistics

2.54 Various line ministries and their departmentsproduce summary annual industry statistics based onadministrative or statistical collections. However, thereliance on ad hoc donor funding and TA, as well asthe lack of sufficient coordination of statistical activi-ties, have resulted in inadequate coverage of some in-dustries. For example, MAFF, the Ministry of Com-merce (MoC), and the Ministry of Tourism (MoT)appear to be better resourced to produce some of thecore data required, but the CDC; the Ministry of In-dustry, Mines and Energy (MIME); the Ministry ofPost and Telecommunications (MPT); and the Min-istry of Transport and Public Works (MTPW) are un-derresourced. In addition, the current portfolio struc-

ture for the economic ministries does not favor the col-lection, coordination, and dissemination of industryand sector statistics.

Sociodemographic Statistics

Population and demographic statistics

2.55 Significant financing and TA by the UNFPAhave underpinned the development of population sta-tistics for Cambodia. The NIS successfully conductedthe 1996 Cambodian Demographic Survey and the1998 General Population Census of Cambodia andhas continued to compile and disseminate demo-graphic and population statistics over the years. It hasreleased 16 publications and analytical reports. Hav-ing conducted in 1998 the first population census inCambodia since 1962, the NIS intends to continue tohold decennial censuses.

2.56 The 1999 Cambodia Socioeconomic Survey, con-ducted with the assistance of the UNDP, the SwedishInternational Development Agency (SIDA), and theWorld Bank, provides estimates of the total popu-lation, household composition, age structure, maritalstatus, economic activities, and housing characteris-tics and amenities.

2.57 The 2000 Demographic and Health Survey(DHS), conducted with the assistance of the UNFPA,UNICEF, and the United States Agency for Inter-national Development (USAID), provides new esti-mates of fertility and mortality rates. In addition, theDHS provides information on household populationby age, sex, and residence and housing characteristicsand amenities.

Poverty statistics

2.58 The UNDP has provided the majority of the fi-nancing for compiling official poverty indicators andfor conducting the Socioeconomic Survey of Cambo-dia (CSES) 1993–94, 1996, 1997, and 1999 throughthe Capacity Development for Socioeconomic Surveysand Planning Project. As mentioned earlier, the ADBhas also contributed extensively to the previous twosurveys. Other TA partners have included the Inter-national Labor Organization (ILO), SIDA, UNICEF,and the World Bank. The UNDP is also financing thePoverty Monitoring and Analysis System (PMAS) andthe development of the Poverty Monitoring Informa-tion System (PMIS).


2.59 Estimates of the distribution of income and con-sumption by household in Cambodia are based on theCSES conducted in 1993–94, 1996, 1997, and 1999.Consumption poverty indicators derived from the CSES(such as the headcount ratio and poverty gap) are basedon (i) a food poverty line, which refers to a minimalstandard of living just sufficient to meet food con-sumption needs; and (ii) a higher poverty line, which in-cludes a minimal allowance for nonfood consumption.

Dissemination and Use of Official Statistics

Availability and quality of statistics and methods documentation

2.60 The availability and quality of data sources fornational accounts, BOP, CPI, labor force, and mone-tary and financial statistics, as well as for the socio-economic statistics, have gradually improved. Doc-umentation of BOP, national accounts, labor force,prices, government finance, and monetary statistics hasalso improved (for example, in STA mission reports,GDDS templates, and statistical publications and insummary concepts, sources, and methods documents).The Directory of Statistical Sources, begun in January2003 and completed in 2004, includes broad concepts,methods, and sources metadata in relation to statisticsproduced by all RGC statistics units.

Dissemination and public access to officialstatistics and methodology

2.61 The NIS first launched the census website inDecember 2000 and then the expanded NIS website,www.nis.gov.kh, in July 2002, both with UNFPA as-sistance. It will expand and improve the content of thewebsite, including statistics and related metadata, asresources permit. Metadata for priority socioeconomicstatistics included in the GDDS templates are also nowavailable via the IMF public website. The NIS has ex-panded and improved the statistical content and tech-nical notes of its publications (e.g., the content of theStatistics Yearbook increased 33 percent). It has re-leased new publications on child labor, provincial andurban consumer price indices, and new electronicproducts. As a result of the establishment of the DataUsers Service Center in mid-2002, the NIS has alsoimproved access to official statistics, also with UNFPAassistance. In addition, the MoC and MoT have statis-tics pages on their websites, and the NBC has begundeveloping its website (including statistics web pages).Official statistics are also released in the standardprinted form by a number of agencies. In addition, the

20

annual reports of various other ministries (includingdepartment-level annual reports in some instances) andinstitutions generally provide summary statistical data.

2.62 Under the framework of the project, CapacityDevelopment for Poverty Monitoring and Analysis—led by the Council for Social Development in partner-ship with SIDA and the UNDP—the RGC has estab-lished a PMAS. The PMAS aims at timeliness andreliability of data, including the monitoring of process,inputs, outputs, and outcomes. The PMAS also aims atachieving both openness and result orientation forinformed policy dialogue and decision making, bystrongly emphasizing the measurement of poverty inrelation to key macroeconomic, sector, and budgetaryreforms. The system is expected to be the main toolfor regular reporting on Cambodia’s progress towardachieving its national development goals.

Conclusions and Lessons Learned

Sustainability

2.63 The implementation of TA over the past decadehas contributed to sustainable capacity building in thefollowing areas:• Improved statistical legislation and coordination

arrangements;• Improved strategic planning, work program devel-

opment, and work program monitoring by the NBCand NIS;

• Effective leadership and management of statisticalprojects in the NBC and NIS;

• Improved staff development and management inthe NBC and NIS;

• Improved statistical skills of local counterparts in-volved in producing macroeconomic and financialstatistics;

• GDDS short-term and medium-term planning andframework implementation for relevant line ministries;

• Expanded and improved data and metadata dissem-ination by the NBC and NIS via publications, elec-tronic products, and websites;

• Expanded and improved documentation and dis-semination of concepts, sources, and methods usedin producing macroeconomic, financial, and prior-ity sociodemographic statistics;

• Improved quality of data collected by the NBC and NIS;

• Improved consistency with international standardsand improved compilation methods; and

• Improved access to administrative data sources.


21

2.64 While TA has had a measurable impact, it is lesslikely to be sustained without further follow-up assis-tance, in relation to:• Enforcement of statistics legislation;• Improved formal statistical coordination arrange-

ments and statistics governance;• Establishment of central RGC statistics units and an

integrated national statistical system;• Improved planning and work program development

and monitoring by all RGC statistics units;• Effective leadership and management of statistical

programs, as well as improved recruitment and re-muneration of statistics staff for RGC statistics units;

• Appropriate financing for statistics capacity build-ing and ongoing activities;

• Improved data and metadata dissemination by otherRGC statistics units via publications, electronicproducts, and the NIS website;

• Improved quality of administrative data sources;• Expanded and improved data collection activities;

and• Improved dissemination of official statistics and re-

lated metadata by RGC statistics units other than theNIS and the NBC.

2.65 Sustainability in these areas would be ensuredby implementing the UNDP Capacity Developmentof the National Statistical System Program and the re-lated TCAP proposal, as well as a continuation of TAprovided by the ADB, the IMF, JICA, the UNFPA,and other donors over the medium term.

Lessons Learned

2.66 Several factors have contributed to developingCambodia’s official statistics and statistical capacityover the past decade, including champions and lead-ership; ownership and commitment by the authoritiesand local statistics staff; awareness and understandingof the need for sound official statistics by the author-ities and donors; phasing and mode of STA; effectiveuser consultation and involvement; independence ofofficial statistics and technical autonomy of the NIS;informal statistical coordination mechanisms; and useof technology.

Champions and leaders

2.67 A critical success factor contributing to devel-oping Cambodia’s official statistics has been its cham-pions. This is evident in the case of the developmentof official statistics produced by the NIS and NBC.

The governor of the NBC, both as governor and in hisprevious position as minister of planning, has cham-pioned the development of Cambodian statistics overthe past decade and has played a leading role in theirdevelopment, as well as in increasing appropriationsfor statistics for these two agencies. The minister ofplanning, secretary of state in charge of statistics, anddirector general of the NIS have also championed thedevelopment of Cambodia’s official statistics and theimportant expansion of the NIS and its statistics pro-gram since 1998. However, champions have not comeforward in the areas of international trade and gov-ernment finance statistics; agriculture, industry, andservices statistics; social sector statistics; environmentand natural resources statistics; and regional or sub-national statistics.

2.68 The success of any project depends greatly oneffective leadership and management. The people andproject management and technical skills of senior of-ficials at the MoP, NBC, and NIS responsible for sta-tistics have been developed significantly in recentyears through donor sponsorship of overseas studies.Effective leadership and management have contrib-uted vastly to the improvements in MoP, NBC, andNIS economic and poverty statistics and to the suc-cessful conduct of the population census and severalmajor establishment and household surveys by theNIS. While the MoP, NBC, and NIS have been suc-cessful, this success is to the credit of a dozen or soeffective leaders and managers, and more needs to bedone in developing the depth of leadership and man-agement in these organizations.

Commitment and ownership

2.69 While strong commitment and ownership of sta-tistics development are evident at the MoP, NBC, andNIS, they are less evident in other ministries andagencies responsible for producing official statistics.The success of any project depends on ownership ofproject objectives by the authorities and local coun-terparts. The successful development of BOP and na-tional accounts statistics, the introduction of the ITRScollection, and successful conduct of NIS householdsurveys are all examples of the strong commitment ofthe authorities. Future TA initiatives should continueemphasizing the need for commitment and ownershipby both the authorities and local counterparts.

2.70 The commitment of local counterparts at theMoC, MOEYS, MoP, MoT, NBC, and NIS is partly


responsible for the major improvements over the pastseveral years in balance of payments, commerce, mon-etary and financial, national accounts, prices, primaryeducation, population, poverty, and tourism statistics.However, the level of commitment of local counter-parts is limited as a result of poor pay and conditions.Inadequate remuneration continues to have a detri-mental impact on capacity building and the quality ofofficial statistics, despite relatively large increases inthe past two years. The overall impact of low remuner-ation means that a number of counterparts do not worknormal office hours and knowledge transfer andprogress in implementing objectives have been slowerthan they could have been. For example, TA is oftenhampered by the lack of full-time counterparts. MoPand NIS authorities are attempting to address thischallenge by including local counterparts in the PMGinitiative. A need exists for the relevant authorities re-sponsible for other RGC statistics units to include keystatistical staff in the PMG initiative.

Financing for statistics

2.71 Possibly the most significant constraint in de-veloping statistics is appropriate financing. Govern-ment financing as a percentage of the total budget forCambodia’s official statistics is among the lowest ofall Southeast Asian countries. Almost all survey activ-ity is donor financed. With the exception of financingfor the consumer prices and ITRS collections, no bud-get allocations exist for regular surveys. To producegood-quality statistics, a significant increase in gov-ernment budget appropriations for statistics is essen-tial. The additional funding needs to be directed toconducting surveys:• annual agriculture, establishment, financial sector,

and government-owned enterprise surveys.• quarterly surveys of domestic and international in-

vestment; international trade in services; selectedproducer, trade, and services prices; business prof-its; and the labor force.

• expanded monthly consumer prices surveys.

2.72 The NIS has yet to conduct agriculture sectorand establishment censuses, and the absence of thesebenchmark data reduces the quality of annual nationalaccounts estimates. The establishment census is alsoneeded to develop an integrated business register.

2.73 Additional financing is also required to strengthenadministrative collections. Lack of sufficient fundingfor the CDC and CED to redevelop systems and to

22

collect foreign trade prices and volumes data and in-ternational investment data has delayed progress indeveloping merchandise trade price indices and im-proving IIP statistics. Considerable scope remains forfurther improving the range and quality of Cambo-dia’s economic and financial statistics, if appropriatefinancing is made available.

Awareness and understanding

2.74 Increased awareness and understanding of theneed for good-quality national statistics have alsocontributed to the development of official statistics inCambodia, especially since 1998. Examples includethe commitment of significant donor and governmentfinancing for developing poverty statistics, the con-ducting of the 1998 General Population Census ofCambodia, and the CSES. More recently, the SCBWorkshop conducted in October 2002 significantlyimproved donor and government understanding ofCambodia’s statistics capacity and raised donor andgovernment awareness of the need to increase financ-ing for capacity development.

User consultation and involvement

2.75 The MoP, NBC, and NIS have a good track rec-ord of ensuring that data users are actively involved inplanning, developing, conducting, and evaluating theestablishment and household surveys. The agenciesalso conduct regular user seminars whenever the na-tional accounts or the results of major surveys are re-leased. In addition to attracting additional financing,this contact results in more relevant statistics for datausers. While other RGC statistics units consult to a de-gree, consultation is far less frequent, thus contribut-ing to poorer quality and less relevant statistics. Theresulting lower profile of these statistics has also ledto lower levels of TA.

Phasing and mode of technical assistance

2.76 Effective phasing and the mode of TA are alsocritical success factors. For example, IMF TA com-menced with a multisector statistics mission in 1992and followed through with a combination of residentadvisors and short-term missions. The combination ofa resident ADB senior statistics advisor in the early1990s and a resident IMF MSA since November 2001,working with ADB short-term experts, has also beeneffective. This has resulted in statistics on balance ofpayments, consumer prices, labor force, money and


23

banking, and national accounts that are more devel-oped than other economic statistics. Similarly, thecombination of long- and short-term experts providedby the UNFPA was also a major factor in successfullydeveloping NIS demographic and population statistics.

Selection and use of experts

2.77 The appropriate selection and use of experts is acritical success factor for TA. Occasionally, donorsselect experts on the basis of their seniority in nationalstatistics agencies at the time they retired or on the as-sumption that the higher the tertiary qualification thebetter. The general view of the authorities is that thesecriteria are not necessarily effective and that the keyto a successful expert is applied skills and results ori-entation. The most effective experts are hands-on, arewilling to learn as much as to teach—irrespective oftheir seniority or tertiary qualifications—and are will-ing to lead by example. TA is less effective when anexpert assumes that giving broad direction and pro-viding conceptual training are all that is required. Ef-fective TA in Cambodia requires an expert to under-stand local culture and to appreciate, for example,that counterparts may be reluctant to ask questionsor challenge advice, even when they suspect that theproposed methodology is not appropriate in theCambodian context. To ensure data quality, expertsmust be willing to supervise directly the developmentof the establishment register or population frame andto review the data and coding at each stage of pro-cessing, from completed survey questionnairesthrough output validation. To paraphrase the cur-rent senior minister of finance, “At times the expertteaches the local counterparts to cook and at othertimes the expert needs to cook so that the meal is pre-pared on time.”

2.78 TA is also less effective when the workloads andterms of reference of the experts are too ambitious.Workloads and expectations of experts are sometimestoo high, given that local counterparts may not havethe level of commitment or the skills required to com-plete the work program within the deadlines specified.

Appropriate legislation and enforcement

2.79 The delay in recent years in enacting the statis-tics law has adversely affected the overall statisticalsystem. While the NIS is technically independent inproducing and disseminating statistics, other RGCstatistics units are not necessarily so. Technical inde-

pendence is a fundamental prerequisite for objective,good-quality, relevant, and timely official statistics.The NIS requires the authority under law to establishthe Statistics Advisory Council, the Statistics Coor-dination Committee, and other formal coordinationarrangements. The NIS and RGC statistics units in-volved in business registration and establishment sur-vey activities need the backing of law to improve re-porting and registration by businesses. Response ratesfor business and establishment collections can be aslow as 10 to 20 percent, although NIS establishmentsurveys have generally higher response rates (about70 percent).

2.80 Statistical legislation should ensure:• Autonomous budget allocations for statistics, thereby

encouraging debate on better financing;• Reporting by the NIS and the Statistics Advisory

Council to the government on statistical capacitybuilding and outputs, thereby increasing account-ability and transparency;

• Data provider understanding that noncompliancewill result in possible fines and imprisonment,thereby improving response rates and providingmore accurate data reporting; and

• Penalization of statistics officers who profit from orrelease confidential individual information, therebybuilding the trust of data providers.

2.81 However, even the best statistics law will fail todeliver the desired outcomes without proper law en-forcement. Effective implementation of the current anti-corruption and judicial reforms is therefore essential.

Coordination mechanisms

2.82 Data coordination has improved significantly asa result of establishing informal arrangements. Asmentioned above, the NIS has established technicalworking groups, with ADB assistance, for statisticslegislation, agriculture statistics, environment and nat-ural resources statistics, and foreign trade statistics.Statistics cooperation has also been expanded and im-proved among the NBC, NIS, ministries, and institu-tions in relation to data-sharing arrangements. TheIMF MSA has assisted the NIS in improving the coor-dination of donor financing and TA for statistics ca-pacity building for the MEF, MoP, NBC, and NIS. TheMSA has also worked closely with the NIS/MoP onimproving the coordination of TA provided by expertsand volunteers from the various ADB, ESCAP, GTZ,IMF, JICA, JOCV, UNDP, UNFPT, UNICEF, and


other organizations. Assistance has been provided tothe authorities in coordinating ADB and IMF STAmissions and in preparing ADB and IMF STA missionproposals.

2.83 However, the current formal coordination mech-anisms between RGC institutions and ministries forstatistical activities and capacity building have provento be severely constrained. Communication amongministries about information sharing generally requiresa written request by one minister to the other minis-ter—a process that can typically take about a month toget the desired result. More formal arrangements, suchas the Statistics Coordination Committee and specificprovisions in the subdecrees requiring direct informa-tion exchange and cooperation between the NIS andRGC statistics units, will help reduce turnaround times.

2.84 The level of decentralization is also a significantconstraint. For example, in addition to the MAFF sta-tistics office, six other departments and offices withinthe ministry are involved in statistical collection anddissemination. Similar levels of decentralization in

24

the MEF and some other ministries make coordina-tion more difficult. It also means much duplication instatistical infrastructure and statistics produced, andthat scarce resources are being spread too thin. Thelaw includes an article specifying the need for min-istries to establish centralized statistics units. Thisshould improve capacity within each ministry andmake TA and skills transfer more cost effective andeasier to implement.

Use of technology

2.85 A final success factor contributing to more rapidstatistical development is the use of technology. Com-puterization of statistical activities in the NIS CentralOffice and major provinces and statistics units in theCDC, MAFF, MEF, MoC, MOEYS, MoP, MoT, andNBC has vastly improved the quality and timelinessof official statistics. In addition, establishing the Na-tional Information Development Authority in 2001,with TA from the Korean International CooperationAgency (KOICA), contributed to building IT capac-ity for other line ministries.

The Immediate Postwar Setting

General Background

3.1 Bosnia and Herzegovina (BiH) in 1996 emergedfrom a four-year war that had commenced only a monthafter the country had declared independence from theSocialist Federal Republic of Yugoslavia. The conflictleft more than 200,000 people (about 5 percent of thepopulation) dead, half the population displaced abroadand locally, and the civil infrastructure, including hous-ing, severely damaged.

3.2 The Dayton Agreement that led to the end of thewar in December 1995 divided the country into twoentities, the Federation of BiH (Federation) and Re-publika Srpska (RS).1 Under the Dayton Agreement,the main constitutional powers were assigned to eachentity, leaving the State (the term generally used todescribe the central government) with foreign affairs,monetary policy, foreign debt service, trade and cus-toms policies, and some more limited functions. TheAgreement provided for the Stabilization Force (SFOR)to assume military control to ensure peace and for theOffice of the High Representative (OHR), appointedby the 50-member Peace Implementation Council, to be responsible for civilian implementation of theAgreement.

3.3 The economic situation of BiH in the immediatepostwar period is reflected in an estimated GDP ofUS$2.8 billion (per capita US$665) in 1996 com-pared with US$10.6 billion prewar (1990) (per capita

3. Bosnia and Herzegovina:Multisector Statistics

Noel Atcherley*

25

*The author was IMF multisector statistics advisor in Bosniaand Herzegovina (BiH) in 2002–04. He has previously served in a similar capacity in Ukraine and Russia, and as balance ofpayments statistics advisor in Namibia, and prior to those assign-ments served in the Australian Bureau of Statistics. His project inBiH was financed by the Administered Account for Selected FundActivities—Japan (JSA).

1The map at the beginning of Chapter 4 (Map 4.1) may be a help-ful reference.

US$2,446). In addition, there were high unemploy-ment, low output from the mainly government-ownedenterprises, and circulation of several currencies.

3.4 With military control and civilian guidance fromthe international community (IC) accompanied by largeofficial aid flows and assistance from international fi-nancial organizations, reconstruction began quickly.Governments were established, administrative agen-cies created, and a new central bank established witha currency board arrangement that provided full cur-rency convertibility and contributed to a progressivelowering of inflation.

3.5 Substantial international donor assistance wascommitted at the end of the war (some US$5 billionfor civilian aid over a three-to-four year period). In ad-dition to the SFOR and the OHR, many internationaland regional organizations, together with foreign gov-ernment agencies and hundreds of nongovernment or-ganizations, established a significant presence in BiH.BiH joined the IMF one week after the Dayton Agree-ment was signed. In the first few postwar years, hu-manitarian aid and reconstruction were the main focusof donor assistance. Substantial military and civil as-sistance was undertaken directly by donors and notchanneled through government budgets.

Statistical Situation

Statistical organization

3.6 The structure of the postwar statistical servicesreflected the constitutional powers and governmentstructures established by the Dayton Agreement. Pre-war, the main headquarters statistical office for BiHwas the regional office of the Yugoslav Statistical In-stitute in the capital, Sarajevo, with small offices else-where. That office began to operate independently in1992 and evolved in 1997 into the Federation ofBosnia and Herzegovina Federal Institute of Statistics(FIS), operating under the Statistics Law of 1995. In


the Federation, the government has a decentralizedstructure of 10 cantons (which spend about half of thetotal Federation budget), each with municipalities.Thus, the FIS has statistical staff in the cantons fordata collection, while a few large cantons have theirown statistical offices.

3.7 In RS, the government has a more centralizedstructure with only the central government and mu-nicipalities. Thus, the small former statistical office ofthe Sarajevo branch became the nucleus for the RS In-stitute of Statistics (RSIS) during the war and oper-ated under a new statistics law from 1996. Hence, inthe immediate postwar period, RSIS started with amuch smaller staff (less than 50) than the FIS (whichhad about 100) and also was substantially isolatedfrom international statistical TA for the initial postwaryears. Because numerous experienced staff had leftthe country, many postwar staff were new and inex-perienced. In 2001, following the creation of the smallDistrict of Brcko, with powers similar to those of theentities, a Brcko Bureau of Statistics was establishedwith three staff within the Revenue Agency, and a sta-tistics law was provided.

3.8 Each entity statistical office operated largely in-dependently and sought to compile and publish basicstatistical data, mostly based on limited administrativeby-product data. Staff had generally been trained underthe statistical system of the former Yugoslavia, andboth statistical institutes were using Yugoslavianmethodologies, which differed significantly from in-ternational statistical standards. Until the establish-ment of the Agency for Statistics of Bosnia andHerzegovina (BHAS) in Sarajevo in August 1998with encouragement of the IC, no coordination of en-tity statistics took place, and no official data werecompiled at the country level. Having only ninestaff, the fledgling BHAS operated under the StateCouncil of Ministers (COM) without any statisticallegislation.

3.9 In 1997, BiH established an independent centralbank—the Central Bank of Bosnia and Herzegovina(CBBH)—under the Law on the Central Bank of BiH.The Bank initially assumed central responsibility formonetary statistics and later assumed responsibility forbalance of payments and government finance statistics,in accordance with Article 68 of the law, which em-powered it to collect statistics on “economic and finan-cial matters.” The statistical offices were responsiblefor other statistics, notably real sector statistics.

26

Statistical output

3.10 For the real sector, the FIS had produced, since1995, national accounts in current prices based on theIMF’s System of National Accounts 1993 (1993 SNA)and confined largely to GDP on a production basis.2

On the other hand, RSIS continued to publish nationalaccounts according to the Material Balances System.Both entities used accounting reports and other ad-ministrative data collected by the payment bureaus asthe main data source for the national accounts. Retailprice indices, published by both entities, were proxiesfor consumer price indices and were based on gener-ally sound methodological principles but with seri-ously outdated weights. A producer price index (PPI)was available for the Federation but not for RS. Laborstatistics were weak for both entities. No populationcensus had been done since 1991, and, owing to themassive movements of people in and out of the coun-try during the war, the country had no reliable popu-lation statistics.

3.11 Balance of payments statistics were not com-piled. The CBBH commenced this task in 1998. Nogovernment finance statistics (GFS) were compiledby the authorities for the consolidated general govern-ment in BiH. Budget data available from entity min-istries of finance and the state treasury used nonstan-dard classifications, excluded substantial military andcivilian expenditures externally financed “off-budget,”excluded significant expenditure arrears, and sufferedfrom problems with the time of recording.

3.12 No countrywide monetary statistics were com-piled at the beginning of 1996. However, by year end,monetary data began to be reported to the IMF. Soonafterward there was a transfer of responsibility forthis task from the National Bank of BiH (NBBH) tothe CBBH.

3.13 The statistical outputs in the immediate postwarsituation, therefore, were very limited indeed. Exceptfor monetary data compiled in 1996, country-levelmacroeconomic statistics did not exist, owing to threefactors: the socialist origin of the statistical offices ina prewar centrally planned economy, a governmentstructure where entities had the main responsibilitiesfor compiling statistics, and the lack of a state statis-tical office to ensure consistency and coordination.

2 International Monetary Fund, 1993, System of National Ac-counts 1993 (Washington).

3 • Bosnia and Herzegovina: Multisector Statistics

27

3.14 Real sector statistics were especially limited,were based on differing standards in the two entities,and suffered from the lack of directly collected databased on 1993 SNA requirements. The main sourceof data for the national accounts, for example, wasaccounting reports. These reports were not lodged byall registered incorporated businesses, there was agrowing number of unregistered and unincorporatedbusinesses, and reported data were affected by taxevasion. More generally, the substantial growth in thenonobserved economy was largely omitted from na-tional accounts estimates.

3.15 The capabilities of the statistical offices wereconstrained by a lack of training in international sta-tistical standards, inadequate staff (especially in RS),very inadequate office space in RS, the lack of even thebare minimum of computing and other office equip-ment, lack of coordination, and serious deficiencies insource data. There was no statistics legislation at theState level, and both the entity statistical offices oper-ated under new legislation that was not consistentwith international best practices. There appeared to belittle government priority for improving statistics.

Strategies and Plans, Implementation,and Outcomes

Introduction

3.16 A notable feature of postwar TA in statistics toBiH is that there were no agreed overall strategies orplans for improvement on the part of governments ordonors. In other words, no broad, let alone detailed,plan was devised by donors or BiH governments todeal with matters such as the following: (i) whichBiH agency would be responsible for which statis-tics, (ii) how country-level statistics would be coor-dinated and produced, (iii) which international stan-dards should be used, (iv) which statistical outputswere priorities and by when they should be finished,(v) how a reasonable balance between economic andsocial/demographic statistics was to be struck, and(vi) how those statistics were to be produced. The rea-sons for this lack of overall strategy and plans can beascribed mainly to the following factors: the highlydecentralized structure of BiH government, particu-larly the lack of an effective State body to coordinatestatistical developments; the focus of governmentson the many pressing problem areas of a postconflictsociety; and the lack of coordination among foreigndonors. Donors did indeed coordinate somewhat on a

few specific statistical activities but not in terms ofoverall strategies or agreed plans.

3.17 Most TA in statistics in postwar BiH can bedivided into four periods: the initial postwar years1996–97, with little TA in statistics; the next coupleof years 1998–99, when donors mainly developedtheir longer-term programs for assisting BiH; the pe-riod 2000–2003, when most donors extensively trainedBiH statisticians or initiated some statistical collec-tions and when a new State statistics law was estab-lished. The fourth period is just beginning, with seri-ous though limited donor support for priority datacollections and at least the opportunity for a muchmore fruitful phase in donor TA, especially becauseof developing improvements in the government orga-nization of statistics in BiH.

3.18 IMF TA in statistics followed a different pat-tern than that described above. Focused TA began in1996 and continued thereafter, aiming at measurableresults in compiling needed macroeconomic statisticsand at the same time establishing adequate legisla-tion, institutional capabilities, and statistical infra-structure. Nevertheless, it is convenient to consider alldonor statistical strategies under these four periods.

1996–97—Initial Postwar Years—A Limited Focus on Statistics

Strategies and plans

3.19 As stated above, no overall strategies and planswere developed during this period. The two entities’statistical offices were largely left on their own toacquire knowledge of international statistical method-ologies while continuing with limited statistical out-put. However, late in the period some European Com-mission (EC) programs were available to providesome study visits to European Union (EU) countries,some training abroad, and attendance at some techni-cal working groups and meetings in the StatisticalOffice of the European Communities (Eurostat) andelsewhere.

Implementation and outcomes

3.20 Significant initiatives were taken in developingmonetary statistics. With IMF assistance, includingtwo missions, the NBBH developed monthly mone-tary statistics during 1996 for policymaking purposesand for reporting to the IMF and other users. During1997, with the establishment of the CBBH, an IMF


mission worked with the CBBH to transfer the capa-bility to new statistical staff and to resolve new prob-lem areas. This served as the basis for regular compi-lation of the monthly monetary statistics from January1997 and for dissemination in the IMF’s InternationalFinancial Statistics (IFS) and by the CBBH.3 Mone-tary statistics were the first sector of macroeconomicstatistics based on international standards that wereproduced by BiH.

1998–99—Support for Physical Infrastructure,State Agency, and Donor Planning, but LimitedTechnical Assistance


3.21 During this period, the major goal of donors wasthe establishment of a State agency with the role of co-ordinating the development of country-level statisticsand negotiating with donors. Without such a body,donors were concerned that little country-level statis-tical development would occur.

3.22 The EC and Swiss Federal Statistical Office alsosought to meet some immediate office and equipmentneeds of the entity and new State statistical offices.The EC also planned some expert support to the BHASto assist it in undertaking its new coordination role.Also, the EC planned continued funding for BiH stat-isticians to attend expert meetings in Eurostat and forstudy visits to some EU countries. Some EC consul-tancies for countries in the program, Poland HungaryAction for Reconstruction of the Economy (PHARE),proposed that assessments be made of social, regional,and education statistics, and of household surveys andclassifications.

3.23 In these years, donors developed their longer-term strategies and plans for TA to BiH statistics. Fol-lowing the creation of the BHAS as the prerequisite forfurther donor assistance, BHAS sponsored the first sig-nificant international donors’ meeting for TA in statis-tics, held in Sarajevo in November 1998 and chairedby the OHR (a preliminary meeting of selected donorshad been held in Geneva in January 1998). Donorswere now prepared to consider technical support forBiH statistics. The Italian National Institute of Statis-tics (ISTAT) had assisted in arranging the meeting bycoordinating among donors and BiH statistical offices.The IC was widely represented at the meeting.

28

3.24 The BHAS and the entity institutes had eachthought in broad terms about their needs, and theyoutlined their requests for TA. BHAS proposals in-cluded in particular (in order of priority) English train-ing for all staff; a population census in 2001; demo-graphic statistics; statistics for national accounts, trade,household consumption, labor force, employment/unemployment and wages; computer training; sam-pling techniques; and communications facilities. TheFIS sought support for training in international stan-dards, national accounts, internal and foreign tradestatistics, and household budget surveys.

3.25 The RSIS requested support for national ac-counts based on the 1993 SNA, a population census,development of business registers, surveys on house-hold consumption and labor force, the introduction ofEC industrial products classification and sample sur-veys of industry, foreign trade statistics based on EUstandards, environmental statistics, migration statis-tics, a new office for the RSIS, and staff training. Butthere was no coordinated strategy or plan agreed at thecountry level, nor did the individual statistical officesindicate how those statistics were to be developed, es-pecially with respect to the balance needed betweentraining and source data collection.

3.26 Donors also had not developed any overall strat-egy and plans for assisting with BiH statistics butrather indicated what they each were planning to con-tribute. The EC had sought funds for a DM 5 millionnational PHARE program for BiH (content not spec-ified), and it was also funding a foreign coordinator inthe BHAS appointed by the State COM. The UnitedNations Development Program (UNDP) proposed aJapanese-funded project probably focusing on the na-tional accounts, social and demographic statistics, andlabor force and household surveys. The World Bankenvisaged assistance in external debt statistics, na-tional accounts, institution building, and a labor forcesurvey; it also envisaged possible assistance in infor-mation technology, household budget surveys, and, if needed, environmental statistics. Bilateral donors,notably Italy, Sweden, Switzerland, and Turkey, indi-cated similarly wide-ranging assistance.

3.27 The IMF indicated its objective of building thecapacity of institutions to compile macroeconomicstatistics in the area of GFS and BOP statistics; fur-ther improving monetary statistics; and disseminatingthe data internationally via IMF publications. Mis-sions would also involve developing a framework for

3 International Monetary Fund, International Financial Statistics(Washington). A monthly and annual publication.


29

consolidating data for all governments, which wouldoccur after a State agency has been identified to co-ordinate this work. Finally, the IMF advised that itwould provide a resident statistical advisor to adviseon institutional issues and the four main areas of macro-economic statistics.

3.28 Donors also urged the need for a three-yearwork program for priority tasks for the BHAS and theentity institutes and the need for the substantialstrengthening of cooperation and working relation-ships between the State and entity statistical offices.

3.29 The OHR suggested there was no further needfor coordination meetings of donors because the pro-cess was moving to implementation. However, an ad-ditional limited donor meeting occurred in 1999 toprovide a more specific plan for three household sur-veys: The Living Standards Measurement Survey(LSMS) to be funded by the World Bank, the UNDP,and the Department for International Development(DFID); the Household Budget Survey (HBS) to befunded by Italy; and the Labor Force Survey (LFS) tobe funded by the World Bank and UNDP.

3.30 As the basis for the IMF’s planned TA, the IMFhad assessed the TA needs in detail for macroeco-nomic statistics by sending a multisector statisticsmission in January 1998. The mission reviewed thekey macroeconomic statistical systems to assess theirsuitability for economic analysis and policy formula-tion, commenced work on developing unified systemsfor compiling macroeconomic statistics, and identi-fied TA needs. Initial detailed recommendations andaction plans were provided to the CBBH and the sta-tistical offices.

3.31 In summary, during this period, donors attemptedto provide effective assistance, but there was no singlestrategy or agreed set of plans for improvement. Basi-cally, coordination was left to bilateral communica-tions between the BHAS, entity statistical offices, theEC coordinator in the BHAS, and other donors. Nofurther donor coordination meetings were foreseen asnecessary. In this process, the seeds were sown of sub-sequent problems.


3.32 A major focus of donors, especially the EC andIMF, during this period was the creation of a Stateagency to coordinate the development of country-level

statistics, negotiate with donors, foster cooperationamong entity statistical offices, and undertake BiH’sinternational responsibilities in statistics. This wasachieved when the COM decided in August 1998 toestablish the BHAS. It had no authority to collect sta-tistics directly nor any authority over entity statisticaloffices. Its management was to be undertaken by amanagement board of three members, one from eachethnic group. All three members had to agree beforethe board could take any action. Six staff, in additionto the board, were provided to the BHAS, which had no legislative powers. The COM agreed that anEC-funded semiresident coordinator would participatein managing the management board for the first year.The BHAS began work operations in 1998. It pub-lished a very limited range of country-level statisticscommencing in 1999. The EC and the Swiss FederalStatistical Office quickly funded some initial rentedoffices and later fitted out BHAS’s government officesand provided some computers and other equipment.

3.33 With this institutional obstacle to funding oftechnical programs by most donors overcome, it wasexpected that the general commitments made at theNovember 1998 donor meeting would be firmed up inconsultation with the BHAS and entity statistical of-fices and then would be implemented. The actual donorTA and statistical outputs during this period were verylimited, however, especially in real sector statistics.The EC funded attendance of statisticians to sometechnical meetings of Eurostat and some study visitsto EU countries. Several EC-funded assessmentswere done on social, regional, and education statistics.ISTAT provided some coordination among donors.The EC was developing its substantial PHARE pro-gram, but the program was implemented only fromSeptember 2000 onward. However, the EC continuedto fund a semiresident coordinator in the BHAS to li-aise between the EC and BiH statistical offices. UNDP/World Bank/DFID/ISTAT plans for household sur-veys remained to be finalized. The World Bank projectwas not undertaken in this period. Sweden providedsome short-term consultancies and missions, includingsome on national accounts and, like the EC, providedtraining in English and IT, which assisted some statis-ticians acquiring some knowledge in these fields.

3.34 Support proposed by various donors for a 2001population census was not pursued. The reason wasthat the OHR advised in 1999 that it would be in-appropriate to conduct a census until Annex VII ofthe Dayton Agreement (which deals with return of


displaced people) was implemented, which could takea number of years. Thus, a serious gap in BiH statis-tics remained; the last census was in 1991.

3.35 Regarding the IMF’s plans, further work con-tinued to improve monetary statistics and respond tochanging circumstances in the banking sector, withassistance from annual missions and the statistical ad-visor from May 1999. Details of the work undertakenare given in Chapter 4 (Slack).

3.36 The IMF provided an expert mission in 1999 towork with newly appointed CBBH statistical staff andthe IMF statistical advisor to develop annual BOP sta-tistics. The statistics were first produced in 1999 start-ing from 1998 and were based on the fifth edition ofthe IMF Balance of Payments Manual (BPM5). BOPstatistics became the second field of macroeconomicstatistics based on international standards to be pro-duced and disseminated. This work involved assess-ing and determining the best available data sourcesand estimation methods and training staff in BOP com-pilation both on the job and at seminars. The CBBHbegan publication of BOP statistics in its 1999 quar-terly bulletins.

3.37 Concerning GFS, no State agency had beenfound during this period to coordinate this work. Thus,the role of implementing functional classifications wasundertaken as part of a U.S. Treasury project to assistthe entities and the State to develop modern treasuryfunctions. The project commenced in this period butwas not completed until 2002 for entities and the Stateand continues at subentity level with assistance fromthe United States Agency for International Develop-ment (USAID).

3.38 IMF efforts to develop annual country-level na-tional accounts estimates based on international stan-dards commenced in 1999 with entity and State statis-tical offices. Each office had allocated staff for nationalaccounts work—though only one each in the RSIS andBHAS—and some training was provided for thesestaff on the job by the IMF statistical advisor and at theIMF’s Joint Vienna Institute (JVI).

3.39 In summary, during 1998 and 1999 the main ac-complishments of donor support were the initiation ofthe BHAS; the participation of some statisticians in ex-pert meetings and some study visits; the first compila-tion of BOP statistics; further improvement of mone-tary statistics; and the commencement of development

30

of country-level national accounts (though with con-tinued poor source data). Donors did not plan, nor didthe authorities fund, the collection of the new and bet-ter economic source data essential for improving realsector statistics. No overall agreed country strategy orplans emerged during this period.

2000–03—Training, Coordination andPlanning Attempts, Limited Collections, aStatistical Law, and Further Donor Meetings


3.40 Donor and country plans for improvement inthis period focused largely on extensive training forstatisticians in many fields and study visits to othercountries; attempts at better coordination among thestatistical office; the development of a work program;efforts to produce more and better statistics at thecountry level by initiating some new data collectionsand drafting a State statistics law; and holding a cou-ple of further donor meetings later in the period. In ad-dition, the authorities committed to improving macro-economic statistics in the context of a reform programsupported by the IMF.

3.41 Programs that had been discussed in generalterms at the November 1998 donors’ meeting were de-veloped in more detail in the early years of this period,and implementation commenced. The EC’s PHAREprogram was the largest and from 2000 to 2002 wasdesigned to train statistical staff, both in-country andabroad. It also provided numerous consultancies invarious fields of statistics, especially national accounts,prices, and foreign trade. The program was also in-tended to support frequent meetings of entity and Statestatistical staff. Three semiresident coordinators wereprovided for, one in each of the statistical offices, tomonitor and coordinate the PHARE program. Also, inthis period Eurostat planned to develop a draft Statestatistics law and seek its acceptance.

3.42 Early in this period, various donors planned amajor set of household surveys, reflecting in part sta-tistical offices’ requests for support for householdbudget and labor force surveys. Also, an LSMS wasadded as a donor requirement. These three surveyswere planned to be undertaken consecutively becauseof limited country capacities for such large surveys.Donors planned a master sample to select specificsamples for each of the surveys. Italy committed fund-ing of an HBS, designed to provide household con-


31

sumption data for use in compiling expenditure-basedGDP and new weights for CPIs.

3.43 The World Bank/UNDP/DFID proposed fund-ing for the LSMS, and the UNDP/World Bank pro-posed funding an LFS. All three donors, together withthe EC and the Swedish International DevelopmentAgency (SIDA), funded the development and imple-mentation of the master sample. Donors and statisticaloffices differed in their views about the desired orderof priority of the three surveys. Those donors and sta-tistical offices especially concerned to improve macro-economic statistics preferred that the HBS be first,while others preferred that the LSMS be first becausethe microdata were to be used to assist in developingpoverty statistics.

3.44 Sweden planned consultancies and training. TheWorld Bank provided a grant for developing a masterplan for statistics, for some of the work on sample im-plementation, for a long-term expert for national ac-counts, and for some English training. Turkey plannedstudy visits and consultancies on price and industrystatistics. The DFID, one of the donors for the LSMS,also planned to fund extensive annual follow-on lon-gitudinal surveys (“panel surveys”).

3.45 The IMF’s multisector statistical advisor was tocontinue supporting the development of statisticallegislation, capacity building in all fields of macro-economic statistics, on-the-job training of statisticians,dissemination of macroeconomic statistics in IMFpublications, and coordination with other donors inmacroeconomic statistics. The IMF’s plans for im-proving macroeconomic statistics were continuouslyreviewed and updated throughout this period.

3.46 The BHAS hosted two donor meetings, in 2001and in June 2002. The main objectives were to urgedonors to speed up action on implementing the HBSand LFS and to provide funding for the HBS sample(not part of the ISTAT project), which appeared to beavailable and not needed for the LFS. The EC dele-gation expressed concern at the lack of progress indeveloping new and better statistics, especially at thecountry level, and indicated that its future fundingwould support that objective. BiH statistical officessought further funding for high-priority projects, butnone was offered.

3.47 An EC Community Assistance for Recon-struction, Development and Stabilization (CARDS)

regional program for BiH was planned to begin in2002. It later changed to 2003 and covered pilot proj-ects for foreign trade statistics, small- and medium-size business statistics, and the BiH’s participation inthe European Comparison Project.

3.48 On May 31, 2002, in a Memorandum of Eco-nomic and Financial Policies for BiH prepared in thecontext of a Stand-By Arrangement supported by theIMF, State and entity authorities committed to im-proving a wide range of statistics. These included thefollowing: by end-2002, GFS, an HBS, foreign tradecommodity data, and quarterly BOP statistics; and bySeptember 2004, production of national accounts onan expenditure basis, real GDP, nonobserved economy(NOE) estimates, a revised CPI and an industrial pro-duction index based on updated weights, improvedforeign investment data, and the initiation of a popula-tion census.4 On December 4, 2002, in a Supplemen-tary Memorandum, the authorities committed to anHBS and to strengthening business registries, imple-menting the State statistics law, and continuing workon new national accounts, the CPI, and foreign tradestatistics.5

3.49 The first successful efforts of the statistical of-fices to agree to a program of high-priority, new, andimproved statistics, mainly macroeconomic statistics,began in mid-2002, with the support of the IMF andEC. The heads of all the statistical offices developedand endorsed a proposed three-year work program oftheir highest priority projects in June 2002 and laterdeveloped a one-year detailed program based on thethree-year plan. The EC presented the program to ahigh-level meeting of heads of relevant internationalorganizations and bilateral donors, which endorsed theprogram.

3.50 The one-year plan focused mainly on establish-ing working groups to harmonize existing statisticsbetween the entities and the HBS; to begin work on apopulation census; to develop strategies for future de-velopment of new systems of economic censuses, sur-veys, and price statistics; to conduct feasible “light”economic censuses in selected industries (mining andenergy); and to plan the censuses of other industries.

4Bosnia and Herzegovina, Memorandum of Economic and Finan-cial Policies (attached to Letter of Intent, May 31, 2002, paragraph22). Available at www.imf.org.

5Bosnia and Herzegovina, Supplementary Memorandum on Eco-nomic and Financial Policies, December 4, 2002, paragraphs 9 and23. Available at www.imf.org.


In the absence of any credible economic source datafor real sector statistics, these proposed economic cen-suses were seen as an essential step in obtaining bench-mark data for national accounts and other economic in-dicators and as a basis for establishing regular businesssurveys.


Statistics law

3.51 Early in the period, Eurostat proposed a draftState statistics law, which was reviewed by otherdonors and the statistical offices. It would have pro-vided a legal basis for the BHAS but with very lim-ited powers, for example, no power to directly collectdata and no power to coordinate entity statistical of-fices. Also, it proposed a statistical council largelyconfined to statistical officials. The COM eventuallyendorsed the draft law in July 2001. Later, one houseof the Parliament endorsed it. However, for about ayear, it awaited approval of the second house of Par-liament, after being rejected several times.

3.52 With OHR backing, IMF and CBBH support,and the agreement of other international organiza-tions, a new and much more modern statistics law wasdrafted based on sound international practices and UNprinciples. Noting the failure of Parliament to enact alaw and the importance of progress in statistics to sup-port economic reform, the OHR, with the support ofthe governor of the CBBH, imposed the statistics lawin October 2002. It became effective in November2002. As part of its decision imposing the law, theOHR instructed that the entities and the State mustenter into negotiations within a year, with a view to-ward combining entity and State statistical offices intoa single statistical service under State control.

3.53 The new statistics law provided for new posi-tions of director and deputy director of the BHAS,with the director having full decision-making powers(i.e., not shared among three as with the former man-agement board). It created a statistical council includ-ing outside users (the governor of the CBBH, the Stateminister of finance and treasury, and three others fromthe fields of economics, business, or statistics). Thecouncil advises the COM on work programs and re-sources, coordination, and other issues. A rolling four-year work program is specified, and an annual reportis published. The BHAS is given powers to collectdata directly if necessary. The small Brcko Bureau of

32

Statistics became a branch of the BHAS. Substantialpenalties are provided for breaches in confidentiality,nonresponse, and false response. The OHR’s decisionrequired the entities to harmonize their laws with thenew State law.

3.54 Implementation of the law was slow. Not untilJune 2003 were the director and deputy director po-sitions advertised. Nevertheless, with OHR directionand support of the CBBH and IMF, work was under-taken with the entity statistical offices to harmonizetheir laws with the new State law. Both entities by theend of 2003 had enacted new laws: the RS law wasmore closely harmonized with the State law than wasthe Federation law. Although there was still no statis-tical council or a new four-year work program by theend of 2003 (these awaited the appointment of a newdirector of the BHAS), the provision of a modern lawon statistics was an essential achievement for BiHstatistics.

Coordination

3.55 The donor objective of building the capacity ofthe BHAS to lead effectively the development of asound statistical system had limited success duringthis period. BHAS resources remained at only 9 stafffor most of this period, increased to 12 only in 2002,but fell to 10 in 2003. The complex managementarrangements established for the BHAS created con-tinuing difficulties for decision making and achievingprogress. The entity statistical offices retained theirlargely independent operations, and coordination re-mained limited. Very limited country-level statisticswere produced regularly by the BHAS. One soundachievement for the BHAS in macroeconomic statis-tics was, with IMF assistance, the release of the firstcountry-level national accounts bulletin in 2000.

3.56 From mid-2002, with IMF and EC assistance,the BHAS coordinated a series of lengthy meetingswith entity statistical offices with the objective of de-veloping a three-year program of high-priority macro-economic statistics and a more detailed program forthe first of the three years. All heads of statistical of-fices endorsed this program, the first such initiative inBiH. A primary objective of the one-year programwas the EC’s request that it be prepared as the basisfor support of an EC PHARE program that the ECeventually decided not to fund. Nevertheless, the de-velopment of these agreed work programs reflected animportant step in the creation of country priorities forstatistics. It was also the country’s first attempt at a


33

medium-term strategy for real sector statistics; unfor-tunately, the attempt lapsed because of inadequatedonor support.

Household surveys

3.57 Provision of a master sample for the three plannedhousehold surveys was undertaken with advice fromSwedish and EC experts and funding from the WorldBank, UNDP, and DFID. The cost was significant (anestimated 860,000 euros of donor funding alone) andwas completed in 2001. As indicated earlier, the issueof which of the three surveys should be first conductedwas debated among the IC and statistical offices.Those (including the IMF and the statistical offices)who urged much-needed improvements in macro-economic statistics favored the HBS. But others pre-ferred the LSMS with its emphasis on microdata. TheLSMS was chosen as the first survey, although it wasnot in the priority work programs of statistical offices.Official results were not published until August 2003.

3.58 Several aspects of the LSMS’s operation war-rant comment. First, the survey was largely under-taken by donors working directly with entity statisti-cal offices, with very little BHAS involvement. Thisled, in effect, to two separate and inconsistent surveys,rather than to a coordinated single survey. Second,after the survey there was significant criticism fromsome donors and statistical offices that the sample wasnot representative and, in particular, that rural house-holds were underrepresented. It is difficult to know theextent of this problem without some basis of compar-ison (e.g., with a population census). While the LSMSprovided a large database of microdata for BiH house-holds, the quality of the results remains in question.

3.59 The follow-up “panel” surveys, which are longi-tudinal surveys using a subsample of LSMS samplehouseholds, were conducted in three waves from 2001to 2003, funded wholly by the DFID. Although fundedto undertake specific operational tasks of the survey,the statistical offices had no decision-making author-ity for the surveys. They did not consider these costlysurveys a high priority, as they generally were not in-cluded in their work programs.

3.60 Since the HBS was planned to follow on quicklyfrom the LSMS, it could have commenced early in2002. This high-priority project was the first projectthat donors had funded for macroeconomic sourcedata since 1996. Statistical offices supported early im-plementation and made use of ISTAT’s expertise in

conducting household surveys. However, commence-ment was delayed until mid-2003. After the HBS wasdelayed, the LFS—the last of the trilogy of householdsurveys, which was expected to be at least in the plan-ning stages by this point—was itself deferred.

IMF TA

3.61 IMF TA led to a number of demonstrable out-comes. First, the new State statistics law was estab-lished. The IMF worked with the OHR on implement-ing aspects of the law, including the terms of referencefor the new positions of director and deputy directorof the BHAS and the harmonizing of the entity statis-tics laws with the new State law.

3.62 Second, regular short-term IMF missions inmonetary statistics, combined with day-to-day assis-tance from the statistical advisor and training of staffat the JVI, dealt with numerous changes in the bank-ing environment and improved the quality and scopeof monetary statistics. The quality, including timeli-ness, of BiH monetary statistics improved greatly. Atime series of monetary data was disseminated on theCBBH website in June 2003, together with the firstBiH advance release calendar for official statistics.6

After May 2003, monthly monetary data were dis-seminated immediately on the website rather thanbeing held up for publication in the next quarterly bul-letin. With IMF assistance, the statistical organizationof the CBBH was rationalized and made more effi-cient for producing monetary statistics.

3.63 Third, quarterly BOP statistics were introducedstarting 2001, with IMF mission and statistical advisorassistance, and were generally disseminated withinone quarter. Extensive improvements were made to thequarterly and annual BOP estimates, and more meta-data were provided. In particular, detailed and largeadjustments for coverage, valuation, and timing weremade in 2002 for merchandise trade, back to 1998.These adjustments related to undervaluation of a largeproportion of imports and considerable smuggling ofimports (and to a smaller extent exports). Adjustmentswere also made to exclude from customs data importsby international organizations, embassies, and otherforeign government offices in BiH. With improvingcustoms and tax administration and introduction of aState border service, appropriate trends in those ad-justments were built into the estimation methodology.

6Available on the CBBH website, www.cbbh.ba.


However, the underlying customs data on exports andimports, despite the strengthened customs administra-tion and the Automated System for Customs DataAdministration (ASYCUDA) computer systems, re-mained of doubtful quality and needed substantial im-provement. Financing arrangements for the significantunrecorded trade in goods also needed to be estimatedin parallel with adjustments to trade in goods. A newestimation methodology for workers’ remittances re-sulted in substantially increased amounts, more in linewith other countries in the region. A time series ofBOP estimates was disseminated on the website, dataand metadata were provided for the IMF Balance ofPayments Statistics Yearbook (BOPSY), more detailedmetadata were published locally, and an advance re-lease calendar was introduced.

3.64 Fourth, the previous IMF statistical advisorworked extensively with statistical offices to introducethe first country-level national accounts publication,with metadata, in October 2000.7 National accountsbecame the third field of macroeconomic statisticsbased on internationally accepted methodologies to bedeveloped with IMF assistance. However, the mainsource data for the national accounts remained poor.Also, no estimates were included in the official statis-tics for the assumed large NOE. The general sourcedata problem could have been substantially overcomeonly if donors had supported economic censuses ofthe main industries. The failure to include NOE esti-mates was due in part to poor and limited source data,but mainly to a reluctance of statistical offices to makesuch adjustments. With no significant improvementsin source data for the national accounts, the GDP esti-mates throughout the period lacked credibility both inlevels and in changes. The number of staff compilingthe national accounts and their capability (aided bytraining from IMF and other donors) improved dur-ing the period.

3.65 Fifth, with improved macroeconomic statisticsthe IMF was able to introduce a country page for BiHin the monthly IFS starting July 2001. Though lim-ited in range of data, the page provided users world-wide with better access to BiH statistics that werebased as far as possible on consistent internationalmethodology.

34

3.66 Sixth, as noted above, the IMF assisted in re-organizing the CBBH’s statistical capabilities with thecreation of a new statistics division, with a section eachfor monetary statistics, BOP, GFS, and real sector sta-tistics. This resulted in new leadership, increased staffnumbers, and an organization more focused on im-proving the macroeconomic statistics for which theCBBH is responsible. New three-year work plans withclear objectives and planned outcomes have been in-troduced, and staff have received deeper and broadertraining. These factors, together with senior manage-ment support for statistics, provide a good basis forsound continuing progress in statistics.

3.67 Seventh, the foundation has been created for ini-tiating the development of GFS, including govern-ment debt statistics for BiH based on internationalstandards. The foundation includes establishing a sec-tion in the CBBH for this work, preparing detailedwork plans, and training staff. The task of developingthese statistics has begun, with support of short-termIMF missions. TA from the United States in estab-lishing modern treasuries and computer systems forthe State and entities led to the introduction of newcharts of accounts and classifications based on theIMF’s 1986 Government Finance Statistics Manual(GFSM) in 2001. This will now serve as a good basisfor the actual production of GFS statistics, with theCBBH having assumed responsibility for that role.

3.68 Eighth, the three statistical offices introducedwebsites during 2002, ensuring better access to theirstatistics. The CBBH had already introduced a web-site in 1998.

Other TA

3.69 The EC’s extensive PHARE program com-menced in late 2000. Originally planned to cover 18 months, it was extended to over two years. Its aimwas to “support the development of a sustainable sta-tistical system” and to “support the entity statisticalinstitutes to be in a position to provide . . . BHAS . . .data for the production of reliable, up-to-date, re-levant . . . statistics.” The program provided trainingcourses for statistical staff and many study visitsabroad. The PHARE program also provided threesemiresident foreign coordinators—one in each sta-tistical office—to coordinate and monitor the trainingactivities and outcomes and to assist the BHAS to de-velop its capacity. It also financed travel costs to fa-cilitate meetings of entity and State statisticians, andfinanced computers, software, and other office equip-

7 Agency for Statistics of Bosnia and Herzegovina (Sarajevo),Statistical Bulletin 3, October 2000.


35

ment for the entity and State statistical offices. Theextensive training provided by the PHARE programat times contributed to the shortage of staff availableto work on producing new and improved statistics,posing a difficult trade-off for managers.

3.70 In mid-2001, the World Bank consultant pre-pared a master plan for the future development of BiHstatistics; however, this was not pursued. SIDA andTurkey provided some training courses, consultancies,and study visits.

3.71 After mid-2001, the World Bank national ac-counts expert worked intensively with the statisticaloffices for a year and recommended changes, such asincluding NOE estimates in GDP and estimating con-stant price GDP by using available volume measures;however, these were not implemented. The expert em-phasized that only limited progress was possible untilbetter source data were collected.

3.72 The CARDS regional program had, by late 2003,mainly involved meetings abroad for BiH statisti-cians planning the extensive work needed for partic-ipating in the European Comparison Project over thecoming years. The project will require BiH to collectextensive price data over a period. But because BiHlacks expenditure-based GDP, it is difficult to envi-sion results from this project except in the longer term.Expenditure-based GDP would be essential if volumecomparisons of GDP (expenditure approach) were tobe made based on purchasing power parities.

3.73 In summary, donors intended the period 2000–2003 to show major improvement in economic andother statistics, including a population census, togetherwith the building of a sound State statistical office. In-deed some major successes did occur, such as the pro-vision of a State statistics law, substantial training,well-equipped offices, and improved monetary andBOP statistics. There was little improvement in realsector statistics throughout the period.

2004 Onward—Implementation of Law,Priority Surveys, Declining TechnicalAssistance, and Possible OrganizationalChange

Strategies, plans, and implementation

3.74 A top priority for the future is the full imple-mentation of the State statistics law. A statistical

council needs to be created and a medium-term workprogram developed. Entity statistics laws, broadlyharmonized with the State law, should provide im-portant support for consistency of statistics.

3.75 The main HBS, which began in January 2004,should lead to improved real sector statistics, provid-ing data on household consumption expenditure foruse in preparing expenditure-based GDP, serving as acheck on official production-based GDP estimates,and providing a basis for up-to-date weights for entityand country-level CPIs.

3.76 A third in the series of donor-financed house-hold surveys, the LFS, would provide urgently neededdata on employment/unemployment and other labormarket indicators.

3.77 The CBBH, with the support of IMF TA, willprogressively develop government external debt andother government finance statistics, which, apart fromproviding essential fiscal data, will also yield signifi-cant input to the measurement of GDP on an expendi-ture basis. Also, the CBBH will complete the initialexploratory survey of foreign investment, followed byregular quarterly surveys, for BOP purposes.

3.78 IMF missions will work with entity and Statestatistical offices to develop for the first time a StateCPI using the latest international methodology andexisting weights. They will also develop new entityCPIs based on the same updated methodology and ex-isting weights and price collections.

Setting 2004, Conclusions, and Lessons Learned

3.79 In this final section, some lessons learned andconclusions from the postwar period of TA in statis-tics for BiH are discussed, but first a summary is givenof the setting in 2004.

The Setting 2004

Economic and political

3.80 Eight years after the war’s conclusion, BiH’soverall political economic setting is much improved.Governments have been elected for the first time bydemocratic processes, peace has been maintained(although some ethnic tensions remain), SFOR strengthhas been reduced progressively to about 12,000, and


some 940,000 people have returned to their prewarlocations. Supporting this process, international donorshave spent US$5–6 billion in various forms of civilassistance to BiH. The rate of economic growth hasbeen respectable, and inflation has fallen to typicalWestern European levels. Fiscal stabilization hasoccurred, with the help of the central bank’s strictimplementation of the currency board arrangement.The establishment of a stable banking system hasraised confidence. Foreign investment has begun togrow substantially, albeit from a small base. Never-theless, GDP per capita stands at the lowest level forany European country. Unemployment remains veryhigh. Many challenges remain for BiH to recover andemerge fully from the postconflict arena.

Statistical

3.81 In the statistical arena, significant changes haveoccurred during the last eight years in statistical infra-structure and organization. The BHAS was created in1998 to coordinate development of country-level sta-tistics and negotiate with foreign donors. The State sta-tistics law was enacted in October 2002 and is slowlybeing implemented, with an initial focus on harmo-nizing entity laws with the State law. Concerning re-sources and capabilities, much-improved office facili-ties, including computer equipment and software, arenow in place; statisticians have received much train-ing in international statistical methodologies; staffresources have increased over 90 percent since 1996;budgets are close to typical EU countries in terms ofstatistical expenditure as a proportion of GDP (althoughofficial GDP estimates for BiH are thought to be sig-nificantly understated); but staff numbers are wellbelow typical EU levels on a population basis. Coop-eration and coordination between State and entity sta-tistical offices have improved, with frequent meetingsof heads of statistical offices and working groups es-tablished to improve statistics. In particular, for thefirst time, the statistical offices agreed on a three-yearwork program of high-priority statistics as the basis forseeking donor funding.

3.82 In terms of specific fields of statistics, notableimprovements are seen in the range of macroeco-nomic data over the period and in dissemination prac-tices. Monetary statistics were the first to be devel-oped and are now at a high standard. BOP statisticswere introduced from 1998 and are being progres-sively improved in quality and detail. GDP estimateson a 1993 SNA basis became available in 2000. Prox-

36

ies for CPIs are available at the entity level (withCBBH estimates at the country level). A start has beenmade on developing GFS including government debtdata, based on international standards.

3.83 Nevertheless significant shortcomings remainto be addressed. No credible GDP estimates yet exist,largely owing to inadequate source data. No country-level price indices are produced. CPIs at the entity levelare not based on consistent international methodol-ogy, and the weights are seriously outdated. PPIsexist for only one entity. Foreign trade statistics donot include estimates for smuggling and undervalua-tion and suffer from a lack of transparency withrespect to their differences with customs data. GFS—consolidated and based on international standards—are not yet available. BOP statistics, though very muchimproved, suffer from significant inadequacies insource data, especially inadequate customs data andthe lack of regular surveys of foreign direct invest-ment, services, foreign aid, and other items. An In-ternational Investment Position (IIP) is not yet avail-able. Credible labor market data are unavailable. Therehas not been a census of population and householdssince 1991, before the massive effects of the war.There are no comprehensive external debt data—thegaps relate to sub-State and nongovernment nonbankdebt.

Lessons Learned

3.84 The postwar history of statistical progress re-flects in many ways the wider political, entity/State,and ethnic climate in BiH as well as the highly de-centralized structure of government. The key lessonsthat can be identified for transition and postconflictcountries like BiH are the following:• Effective leadership is important for progress in

statistics. Such leadership in the CBBH was criti-cal in the progress of macroeconomic statistics. Onthe other hand, for real sector statistics, the poorlyresourced State statistical office, with a complexmanagement structure, is a major reason for the lackof much progress since 1996.

• Adequate legislative basis for compiling statisticsis needed. Again, where the CBBH has had suchlegislative backing, until very recently the State sta-tistical office has not, severely limiting its ability tocoordinate country-level statistical development.

• Appropriate organization of statistics for thecountry is essential. Reflecting the general decen-tralization of government powers in BiH, real sec-


37

tor statistics are mainly compiled by the entities andthe Brcko District. The State is largely left to therole of seeking to coordinate the development ofcountry-level data, rather than primary data collec-tion. It is therefore highly dependent on cooperationfrom the entity statistical offices. A combined sta-tistical service that meets the needs of both entityand State governments would achieve substantialeconomies of scale. Also, by eliminating the time-consuming and frustrating tasks of seeking coordi-nation and consistency among disparate adminis-trations, it would yield an organization much moreconducive to progress in statistics.

• Governments must set overall strategies for sta-tistical capacity building, with clear and priori-tized work programs. Once this is accomplished,preferably on the advice of a statistical council,countries can seek donor support for the priorities.This is one of the most important lessons from post-war TA in BiH. Reflecting in part the structure ofits governments, the BiH did not develop a coher-ent strategy for statistics and generally accepted alloffers of TA in statistics.

• Donors should coordinate their TA strategies. Ifa country finds it difficult to create a set of statisticalpriorities, donors should coordinate their programsamong themselves, avoiding duplication and bal-ancing aid appropriately among types of statistics(e.g., economic and sociodemographic statistics).Apart from the household surveys, coordinationefforts in BiH were largely ineffective.

• An appropriate balance should be struck amongthe various focuses of TA (i.e., TA focused on di-rect production of statistics and collection of reliablesource data, or on staff training or on improving sta-tistical infrastructure such as office premises andequipment). This balance would be determined bythe appropriate phasing of TA inputs. For example,regarding BiH real sector statistics, the phasing wasfrustrating for the staff because they were amplytrained in the technical aspects of producing statis-tics, but they had little source data with which toapply this training actually to produce the statistics.

• A pragmatic and flexible approach to institu-tional responsibility for different sectors of sta-tistics may lead to faster progress. For example,reflecting its relatively better resource situation andstrong interest in economic statistics, the centralbank (at least in the early years of statistical de-velopment) may improve macroeconomic statisticsmore rapidly if it takes formal responsibility forcompiling them.

• The selection of high-quality experts is critical.The selection of consultants who have had little orno experience in compiling official statistics willlikely lead to limited results.

• Governments should provide adequate statisti-cal staff to absorb TA, since donors generally donot fund salaries for statistical staff. This has beena constraint for the BHAS and RSIS in particular.

• IMF lessons are critical. The main lessons fromthe IMF’s contribution to TA in BiH have been thefollowing: (i) the flexibility of IMF TA was highlyimportant in being able to move quickly (as with thedevelopment of monetary statistics in 1996) to as-sist in creating new fields of statistics and buildinglong-term capability; (ii) a supportive host institu-tion (e.g., the CBBH) is essential if timely action isto occur and the capability is to be sustained; (iii) the IMF’s approach of focusing on concrete ac-tivities that directly lead to development and dis-semination of new and better statistics led to posi-tive results for users; (iv) the IMF’s interest inbuilding local in-depth capability in all fieldsof macroeconomic statistics contributed to cross-sectoral consistency; and (v) a long-term residentmultisector statistical advisor can be especiallyuseful where there are many serious gaps in statis-tics, a lack of suitable infrastructure, and many insti-tutional issues that need day-to-day attention.

• Estimates should be made for omitted or poorlycovered activities. When country statisticians makeestimates, it is important to accompany the estimateswith a clearly stated revisions policy for official sta-tistics, so that the estimates can be revised crediblywhen more and higher-quality statistics becomeavailable.8 In many countries, especially in transi-tion countries, these reforms require changes in cul-ture to be able to correct or revise statistics withoutcreating suspicion of wrongdoing. In the case ofBiH, this has been a significant factor for nationalaccounts statistics, which suffer from the failure toinclude estimates of what is generally regarded as alarge NOE. This failure to include the NOE alsoencourages international and other organizations toproduce their own estimates (in the case of GDP,these range up to several times the official figures),thus creating confusion for policymakers, potentialforeign investors, and others.

8See Carol S. Carson, Sarmad Khawaja, and Thomas K. Morri-son, 2004, “Revisions Policy for Official Statistics: A Matter ofGovernance,” IMF Working Paper 04/87 (Washington: Interna-tional Monetary Fund).


3.85 In conclusion, BiH statistical offices and donorsare at a crossroads. One path leads to a continuing lackof real progress in real sector statistics and the dangerthat the statistical offices will become irrelevant formeeting the minimum statistical needs. The other pathleads toward integration with Europe and the neededmajor progress in real sector statistics. This path wouldbe based increasingly on local initiative, adequate bud-gets, enhanced government interest in using statistics for

38

decision making, and continuing TA in priority fieldstargeted on new and better source data. It would improvenational accounts, prices, and other statistics, preferablywith a single statistical service serving State and entitygovernments and with a well-focused and prioritizedwork program. The CBBH, based on experience to date,will continue to play its part in ensuring sound progressin the fields of macroeconomic statistics for which it isresponsible.

Setting

Institutional Setting (Capacity, StatisticalOutputs, Shortcomings, and Challenges)

4.1 Four months after the signing of the Dayton PeaceAccord in December 1995, the first of a series ofmonetary and financial statistics missions from theIMF’s Statistics Department began work in Bosnia andHerzegovina (BiH). Countrywide monetary and finan-cial statistics were urgently sought by the BiH author-ities and the IMF to support efforts to establish a newcentral bank and a countrywide currency, as called forin BiH’s constitution.1

4.2 BiH presented a particularly challenging envi-ronment for delivering TA in macroeconomic statis-tics, not least because the conflict had severely af-fected data collection and processing systems and theorganization structures established under the formerSocialist Federal Republic of Yugoslavia (SFRY).For instance, computer equipment at official institu-tions and commercial banks was generally inopera-tive or severely outdated; thus, many commercialbanks were relying on manual methods for recordingtransactions. In addition, banks’ balance sheet data,typically used to compile monetary and financial sta-tistics, were drawn up according to plans of accountsand accounting rules that had evolved differently inthe three regions of BiH (see Map 4.1). Several cur-rencies were being used for transactions and savingspurposes—one private currency issued by commer-cial banks, and three other currencies issued by for-eign central banks—thereby adding an element of so-called currency co-circulation to the local economy.

4. Bosnia and Herzegovina:Monetary Statistics

Graham L. Slack*

39

*The author was senior economist in the IMF Statistics Depart-ment during the period of this study. He is grateful to Keith Dublin,Gérard Raymond, Xiuzhen Zhao, and Silvano Spencer—fellowmembers of the IMF monetary and financial statistics missions toBiH—for comments on this case study.

1 In addition to the IMF’s Statistics Department, the European,Monetary and Financial Systems, and Fiscal Affairs Departmentsof the Fund were closely engaged in providing policy advice andtechnical assistance to BiH.

4.3 Added to these difficulties and reflecting the post-conflict regional/political boundaries was the fact thatthree institutions undertook monetary authority func-tions in BiH (see Table 4.1): the Federation Ministryof Finance (FMoF) in the city of Mostar; the NationalBank of Bosnia and Herzegovina (NBBH) in Sarajevo;and the National Bank of Republika Srpska (NBRS) inthe city of Banja Luka, Republika Srpska (RS). Pay-ment bureaus in each region of BiH also undertooksome monetary authority functions, such as releasingcurrency into the local economy.2 As with the com-mercial banks, the payment bureaus’ charts of accountshad evolved differently in each region of BiH. Finally,responsibility for compiling statistics on the bankingsystem was shared by the NBBH, the FMoF, and theNBRS; each of them was collecting balance sheet datafrom the commercial banks under their respective con-trol. The key elements of BiH’s banking system as of1996 are summarized in Table 4.1

The Plan of Action and Outcomes

Plans for Improving Statistical Capacity,Technical Aspects, and Institutional and Systemic Issues

4.4 The first step in defining a plan of action was toestablish what data were being collected by whichagencies. Extensive consultations were undertakenwith national/regional authorities and agencies, as wellas with IMF departments and bilateral and other inter-national agencies engaged in BiH. Meeting national/regional officials was not straightforward, however—not least because they were generally reluctant to traveloutside their jurisdictions, and key agencies were dis-persed around the country. Also, initial technicaldiscussions were complicated by the need for transla-tion between languages and differences in financialterminology. Nevertheless, with the help of patient

2 Payment bureaus controlled the electronic payment system inBiH.


40

Map 4.1. Bosnia and Herzegovina

BosanskiBosanskiBrodBrod(Srp.Brod)(Srp.Brod)Prijedor

Bihac

Tuzla

Srebrenica

Zenica

Mostar

BosanskiBrod(Srp.Brod)

BrckoBrcko ˇ Brcko Banja Luka

Trebinje

Gorazde

SARAJEVO

REPUBLIKASRPSKA

REPUBLIKASRPSKA

FEDERATION OFBOSNIA AND HERZEGOVINA

SERBIA ANDMONTENEGRO

CROATIA

ALBANIA

CROATIA

Adriat ic Sea

16°E

45°N

44°N

43°N

45°N

44°N

43°N

17°E 18°E 19°E

16°E 17°E 18°E 19°E

BOSNIA andHERZEGOVINA

This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other informationshown on this map do not imply, on the part of The World BankGroup, any judgment on the legal status of any territory, or anyendorsement or acceptance of such boundaries.

IBRD 34043

MAY 2005

0 10 20 30 40

0 10 20 30 Miles

50 Kilometers

BOSNIA AND HERZEGOVINA

CITIES VISITED BY IMF MISSION

CITIES VISITED BY IMF MISSION

OTHER SELECTED TOWNS

NATIONAL CAPITAL

DAYTON AGREEMENT LINES

INTERNATIONAL BOUNDARIES

3 A top-down approach was initially used by the authorities butdid not provide sufficiently reliable statistics for analysis of thecomponents and counterparts of broad money growth.

translators, key elements of the situation were broadlyclear by the end of the first mission.

4.5 The mission reviewed available source data andassessed whether they could be used to compilereliable countrywide monetary and financial statis-tics. A key decision was whether a “top-down” or“bottom-up” approach should be followed for com-piling countrywide monetary statistics. Under a top-down approach, data compilers use aggregate bal-ance sheet data as the primary source together withsupplementary information to allow certain groupsof accounts to be allocated to sector and instrument

classifications. This approach can yield monetarystatistics quite quickly but often at the price of trans-parency and long-term reliability, not least becauseit is approximate and over time generates unallocatedresidual balances.3 Under the bottom-up approach,accounting balances at the most disaggregated level(e.g., at the five-digit level in a five-digit plan of ac-counts) are classified, account-by-account, accord-ing to their attributes (such as holding sector and type

4 • Bosnia and Herzegovina: Monetary Statistics

41

of instrument). This approach requires intensivework, especially if these attributes cannot be dis-cerned from the account coding structure or from thetext descriptors.4

4.6 Although application of the bottom-up approachwas further complicated by the various plans of ac-counts that existed in BiH, this approach was adoptedbecause it provided a transparent and reliable frame-work and facilitated transfer to the authorities and fu-ture development work. In view of policymakers’ de-mand for countrywide monetary statistics in the nearterm, however, it was clear that an intensive periodof work would be needed to put in place the compi-lation system, and that IMF staff would meanwhileneed to be the principal compilers of BiH’s monetarystatistics. Because of the lack of functioning database

systems, it was also clear that a spreadsheet systemfor compiling monetary statistics would have to bedeveloped. Given the multiple data sources and thedifferent charts of accounts under which the datawere reported, careful design of the spreadsheet architecture and thorough documentation were es-sential in transferring the compilation system to theauthorities.

Implementation

4.7 Implementing the plan of action required a seriesof TA missions. As shown in Table 4.2, an initialphase of intense TA occurred during 1996 to estab-lish the system for compiling monetary statistics. Thiswas followed by a consolidation phase during 1997and 1998 aimed at improving the system and enablingthe national authorities to compile countrywide mon-etary statistics. The third phase was aimed at devel-oping the authorities’ capacity to further improve thesystem and to develop related statistics, such as rep-

Table 4.1. Overview of Key Aspects of the Banking System in BiH in 1996

Banking System

Monetary Authority Commercial Plan of Region Entities Currency Composition Monetary Authority Functions Banks Accounts

Sarajevo

Mostar Federation

Banja Luka

1 Issued by the NBBH.2 Including one bank that had its license revoked in March 1996.3 Issued by the National Bank of Croatia.4 Through banks’ required reserves.5 Through overdraft facilities at the Payment Bureau.6 Issued by the NBFRY.7 Private currency issued by commercial banks between May 1995 and February 1996.8 Through selective interest rate controls, credit ceilings, and banks’ required reserves (prior to April 1996).

Federation

RepublikaSrpska

Bosnian dinar1

Deutsche mark

Croatian kuna3

Deutsche mark

New Yugoslav dinar6

Cashier checks7

(May 95–Feb 96)Deutsche mark

1. NBBH, Sarajevo2. Payment Bureau

1. FMoF, Mostar2. Payment Bureau

1. NBRS, Banja Luka2. Payment Bureau

1. Currency issue2. International reserves management3. Financial relations with the Fund4. Credit and monetary policy5. Bank regulation6. Fiscal agent of government7. Banker to banks

1. International reserves management2. Credit and monetary policy4

3. Bank regulation4. Banker to banks5

1. Currency issue7

2. International reserves management3. Credit and monetary policy8

4. Bank regulation5. Fiscal agent of government6. Banker to banks5

302

7

12

1995 Ministry of

Finance

1990 formerSFRY

1994 Serbian

BOSN

IA A

ND

HER

ZEG

OV

INA

4 Even where the text descriptors appear to provide sufficient in-formation to accurately classify the account, it is necessary to care-fully validate with the account manager(s) that the descriptor accu-rately portrays the account’s attributes.


resentative interest rate statistics.5 Hands-on and for-mal training was provided throughout the three phasesof TA.

To provide context for describing in more detail thethree phases of TA in BiH, selected methodologicalissues are first described below, followed by an over-view of the tools used to deliver the TA.

Selected Methodological Issues

Reliability of Source Data

4.8 A key check on the reliability of the source dataused to compile monetary statistics is to determinewhether reported bilateral positions among banks(interbank positions) are broadly consistent. Once thefirst set of countrywide monetary statistics was com-piled, the focus was put on identifying the causes ofinconsistencies in the bilateral positions. Some of thekey factors identified included the following:

42

• After the creation of BiH as an independent countryin 1992, the NBBH assumed a liability to local com-mercial banks for frozen deposits held at the Na-tional Bank of the Federal Republic of Yugoslavia(NBFRY) and entered a matching claim on theNBFRY in its balance sheet. A review of relevantaccounts showed that the liability recorded in theNBBH balance sheet was not equal to the sum of thecommercial banks’ claims on the NBBH. In addi-tion, in the commercial banks’ balance sheets, liabil-ities to citizens for frozen deposits were not alwaysequal to the counterpart claims on the NBBH. Re-solving these inconsistencies formed a key part ofthe work plan drawn up for the authorities.

• Locating the source of interbank discrepancies inthe monetary statistics for RS was complicated be-cause no distinction was made in the plan of ac-counts between commercial banks’ transactions withthe NBRS and with the other commercial banks.

• Parts of one bank’s branch network declared them-selves independent banks and ceased to recognize li-abilities to the former parent bank, which continuedto recognize in its balance sheet claims and liabilitiesrelating to the branches. Also, some banks’ branchoperations located outside of BiH were included inthe source data used to compile monetary statistics.

Table 4.2. Time Line of IMF Monetary Statistics Missions and Key Outcomes

Mission Dates Key Outcomes

1996April 4–23June 4–23November 6–26

1997August 13–28

1998January 14–28September 23–October 8

1999June 23–July 8

2000October 31–November 15

2001December 3–18

Note: Part of a multitopic mission covering balance of payments, government finance, and national accounts statistics, as well as monetary statistics.

Facts established and Plan of Action drawn up.First set of countrywide monetary statistics compiled, and new commercial bank report forms introduced.Authorities take on responsibility for producing regional monetary statistics on a regular and timely basis.

Compilation procedures refined. Plans developed for authorities to compile countrywide monetarystatistics.

Authorities ready to begin compiling countrywide monetary statistics.Development of countrywide interest rate statistics and data on exchange rates and international

liquidity. Ongoing improvements in the monetary statistics.

Draft country page prepared for BiH for inclusion in IFS. Timeliness of monetary data improved.

Ongoing improvements to the monetary statistics and agreement on IFS country page for BiH.

Improvements in dissemination of monetary and financial statistics, further methodologicalimprovements, and development of capacity to analyze monetary statistics.

5 This description is somewhat stylized for illustrative purposes.In practice there was some overlap of tasks and outcomes across thethree phases of technical assistance.


43

Data on Domestic Bank Financing

4.9 An important contribution of monetary statisticsto formulating and monitoring economic policy is thetimely and frequent information it provides on the gov-ernment’s recourse to net financing from the domesticbanking system. This facet of monetary statistics wasespecially valuable in BiH in the absence of completeand reliable fiscal statistics. As fiscal statistics steadilyimproved, the 1998 multisector statistics mission toBiH assisted in reconciling the two data sets on do-mestic bank financing. The mission’s work led to mod-ification of the monetary and fiscal statistics to pro-mote closer consistency among the two data sets.

Payment Bureau Accounts

4.10 Although operating under different plans of ac-counts, the payment bureaus in each of the three mainregions of BiH undertook similar functions: (i) effectingcashless payments among legal entities, (ii) releasinglocal currency into the economy on the instruction of thecentral bank, and (iii) acting as accountant for the pay-ment system (thereby generating data used in monitor-ing bank system liquidity and financial statements for alllegal entities). A key task of the initial missions was de-lineating these functions and identifying which accountsin each of the three payment bureaus’ plans of accountswere mirror images of accounts recorded by the banksthemselves. The exercise was immensely time consum-ing because a detailed understanding of the unique ac-counting codes and structure was needed.

Cashier’s Checks

4.11 Following the Federal Republic of Yugoslavia’simposition of sanctions on RS in 1994, the supply ofYugoslav dinars (YUDs) from the NBFRY to RS de-clined sharply. In the absence of sufficient YUD cashbalances to service deposit withdrawals, RS commer-cial banks issued transferable cashier’s checks through1996, representing a form of private currency issue.These checks were transferable and used widely ineveryday transactions and were therefore included inthe countrywide measure of currency in circulation.

Co-Circulation of Currencies

4.12 As noted earlier, BiH used several foreign curren-cies for transactions and savings purposes. While thebalance sheets of the banking system and payment bu-reaus captured foreign currencies deposited with these

institutions that were therefore included in source dataused to compile monetary statistics, the source data didnot capture corporate and household foreign currencycash held outside of the banking system. In the absenceof reliable information on the amount of foreign cur-rency in circulation in BiH, estimates had to be made.6

Currency Board

4.13 The currency board that was introduced in 1997under the auspices of the Central Bank of Bosnia andHerzegovina (CBBH) required some modifications tothe analytical presentation of BiH’s monetary statistics.In particular, to facilitate analysis of the currency boardwithin the context of broader monetary developmentsin BiH, the analytical accounts of the monetary author-ities needed to separately identify issuances of localcurrency and the Deutsche Mark (subsequently, euro)reserves that backed the currency.

Tools of Technical Assistance in Bosnia and Herzegovina

Technical Assistance Reports

4.14 At its conclusion, each mission to BiH prepareda summary technical assistance report (translated intothe local languages) and discussed it in detail with therelevant authorities. The mission discussed issues atvarious levels of seniority, discussing detailed tech-nical issues at the working level and broader policyand strategic issues at meetings with senior officials.Shortly after returning to IMF headquarters, the mis-sion finalized comprehensive mission reports, com-plete with detailed appendixes and technical notes,and sent translated versions to counterparts in BiH.The reports presented a step-by-step discussion ofrecommendations and time-lined actions to be takenby the authorities, set within an overall medium-termwork plan agreed with the authorities.

Training

4.15 Hands-on training was the initial method used toenable a core group of officials to produce monetarystatistics in BiH. This training consisted of their infor-mally discussing issues as they arose and sitting down

6For information on the measurement issues involved, see Rus-sell Krueger and Jiming Ha, 1995, “Measurement of Co-Circulationof Currencies,” IMF Working Paper 95/34 (Washington: Interna-tional Monetary Fund).


with staff in front of computer spreadsheets for guid-ance on the mechanics of the data compilation pro-cesses. Officials involved in compiling monetary sta-tistics also took advantage of various formal coursesoffered by the IMF, both at its regional training centersand in Washington, DC.

IMF Statistical Advisor and Resident Representative

4.16 The work of the missions was greatly facilitatedby the appointment of an IMF resident statistical ad-visor and an IMF resident representative to BiH. Notonly did they facilitate meetings with senior countryofficials, but their active involvement in the mission’swork reinforced to the authorities the importance of asound statistical base for policy work. Also, the sta-tistical advisor followed up between missions onprogress made by the authorities in implementingrecommendations, provided general day-to-day TA tothe authorities, and relayed to IMF headquarters tech-nical difficulties that required expert advice.

Multisector Mission

4.17 The IMF multisector mission in 1998 provideda valuable opportunity to check consistency amongthe monetary statistics and other macroeconomic sec-tors (i.e., government finance, balance of payments,and national accounts statistics). In particular, recon-ciliation between monetary and fiscal statistics wasundertaken during this mission.

Phases of Technical Assistance

Implementation Phase (1996–97)

4.18 The bottom-up approach to compiling monetarystatistics required close collaboration with the account-ing and operational departments of the central andcommercial banks and other institutions. Discussionswere initially held with the staff of the largest com-mercial banks and a sample of other banks to providean overview of the scope of their banking activities,accounting systems, and database capabilities. Impor-tantly, this approach also provided an opportunity toexplain the context of the statistical work that wasbeing undertaken and its relevance to the broader eco-nomic and financial issues facing BiH as well as anopportunity to listen to the concerns and suggestionsof these key data providers.

44

4.19 Several rounds of review and discussion werenecessary to identify economic attributes (e.g., type offinancial instrument, sector of holder, maturity, cur-rency of denomination) of each account specified inthe various charts of accounts used by the commercialbanks and in the trial balances of the monetary author-ities. A useful tool in this process was to assign alpha-numeric codes to each account identifying its eco-nomic attributes and thereby its classification in theanalytic presentation.7 This systematic approach fa-cilitated not only explaining the compilation systemto the BiH officials charged with preparing and ana-lyzing the monetary statistics but also automating thecompilation process.

4.20 The process of consultation was especially im-portant in designing and implementing a standard set ofbalance sheet report forms (so-called call reports) forthe commercial banks. These call reports were neces-sary to enable the compilation of meaningful country-wide monetary statistics.8 IMF staff conducted a semi-nar to explain the motivation for introducing the reportforms and the broad macroeconomic concepts they em-bodied and to provide a line-by-line review of how theforms should be completed.9 The seminar was followedup with bilateral discussions to assist the banks maptheir accounting balances onto the report forms.

4.21 The commercial banks began using the newreport forms in July 1996. Their willingness to adoptthese new report forms was facilitated when thenational/regional authorities agreed to review theoverall reporting burden on the banks. In particular,they had concerns that many of the existing data re-quests were no longer relevant and were wasteful ofresources. Reducing the banks’ reporting burden wasalso important in the context of improving the accu-racy and reliability of data reported on the core statis-

7 While analytical presentations can be derived directly from thesource data, a two-step approach was used here that involved trans-ferring the accounting data into a (sectoral) balance sheet presenta-tion from which the analytical accounts could be derived. This ap-proach proved useful because it facilitated transparency in themethodology and thereby explanation to the authorities. Also, suchan approach facilitates wider application of the data, for instancecompilation of financial accounts statistics for the national accounts(see the IMF’s Monetary and Financial Statistics Manual (MFSM2000)).

8 These forms were later adopted as the centerpiece for a newsuite of prudential supervisory report forms introduced under thesupervision of the United States Agency for International Devel-opment (USAID).

19 Reflecting the goodwill that had been developed with the com-mercial banks, more than 100 commercial bank representativesattended the seminar.


45

tical and supervisory returns.10 The successful intro-duction of the new report forms was also facilitated byadopting a flexible approach to historic data series—while historic data compiled in accordance with thenew report forms was desirable, it was given a lowerpriority to reporting current balance sheet data.

4.22 A particularly sensitive issue to be addressed wasthe designation of an official agency to be responsiblefor compiling countrywide monetary statistics. Notwanting to risk upsetting the reporting and compilationprocedures that were by now in place, IMF staff initiallytook on the responsibility of preparing countrywidemonetary statistics by consolidating the regional mon-etary data compiled by the NBBH, FMoF, and NBRS.11

4.23 Aside from the organizational and technical as-pects, training of BiH officials on the concepts andprinciples used in macroeconomic statistics was alsoneeded. The initial objective was to enable officials toadvise commercial banks on the appropriate classifi-cation of accounting balances on the new report forms,with more general training in macroeconomic statisticscoming later.

Consolidation Phase (1997–98)

4.24 Regional monetary statistics for the Federationand RS continued to be compiled by the NBBH12 andNBRS, respectively, and countrywide statistics con-tinued to be assembled by the IMF, until the CBBHwas established in August 1997. The Research andDevelopment Department (RDD) of the CBBH wascharged with compiling and analyzing countrywidemonetary and balance of payments statistics.13

4.25 Establishing the CBBH involved equitably dis-tributing key functions across the Federation and RS.Under the regional decentralization plan for CBBH de-partments, the RDD was located outside of Sarajevo,in the town of Pale in RS. The RDD enabled theauthorities in BiH to assume responsibility for com-piling countrywide monetary statistics, but because thestaff of the NBBH14 was reluctant to travel to RS, andthe staff of the NBRS was reluctant to relocate fromBanja Luka to Pale, a further round of training was re-quired. The staffs of the NBRS and NBBH were ex-tremely helpful in this regard and met with RDD staffto discuss methodology and compilation issues. Dataflow and compilation arrangements nevertheless con-tinued to be rather complicated (see Figures 4.1 and4.2), reflecting the decentralized nature of official in-stitutions in BiH. These arrangements were simplifiedsomewhat when the RDD was eventually transferredto the CBBH’s main unit in Sarajevo.

4.26 Under the supervision of an expert group headedby USAID staff, work was also underway to developand implement a unified chart of accounts for the com-mercial banks. Important in this regard was the close in-volvement of CBBH staff to ensure that the new chartcontained sufficient detail to compile monetary statis-tics. In view of the privatization program that was un-derway in BiH, separate identification of banks’ finan-cial relations with public and private entities was animportant consideration for the new plan. The involve-ment of the monetary statistics compilers in reviewingthe CBBH’s new chart of accounts proved to be simi-larly useful in ensuring that it met the needs of mone-tary statistics.

Facilitating Phase (1998 to Date)

4.27 From late 1998 onward, TA focused on devel-oping the capacity of the authorities to further im-prove the compilation system and develop relatedstatistics. This phase was aided by the day-to-day as-sistance of an IMF resident statistical advisor ap-pointed to BiH in 1999. With the development ofcountrywide interest rate statistics and data on BiH’sinternational liquidity position, a country page forBiH was published in the July 2001 issue of theIMF’s International Financial Statistics (IFS). As-sistance also extended to developing official statisti-cal publications of the CBBH.

10USAID provided considerable on-site technical assistance aimedat improving the reliability of data reporting by the commercial banks.

11 The technical details of the consolidation process involved incompiling monetary statistics are described in the MFSM 2000,paragraphs 242–44.

12The NBBH emerged as the lead agency for compiling monetarystatistics in the Federation and extensive assistance was provided tospecify the functions, responsibilities, and staffing for a new statis-tics department in the NBBH.

13 The IMF’s Monetary and Financial Systems Department (for-merly Monetary and Exchange Affairs Department) extended con-siderable technical assistance to the authorities in establishing theCBBH. The Central Bank Law stipulated that the CBBH was thesingle unified monetary authority in BiH and would act as a cur-rency board. Its primary functions are (i) the issuance of domesticcurrency (the Convertible Marka (KM)) in exchange for DeutscheMarks (subsequently, euro) on a one-to-one basis; (ii) the manage-ment of the official foreign exchange reserves; (iii) levying requiredreserves on commercial banks’ KM deposits; and (iv) acting as fis-cal agent for the State government (limited to maintaining depositaccounts of the government since the CBBH is prohibited by lawfrom extending loans). The CBBH main unit (headquarters) occu-pied the Sarajevo premises of the NBBH. 14 Most of the NBBH staff were employed by the CBBH.



• Prerequisites: Statistical legislation and adequateresources are generally thought of as necessary pre-requisites to deliver reliable statistics.15 In the im-mediate postconflict setting of BiH, however, a cul-ture of data reporting among the commercial banksand goodwill among the banks and statisticians sub-stituted for a legalistic approach to data reporting.

46

Similarly, careful staff selection, planning of work-loads, and dedicated officials compensated for short-ages in staff numbers and technology that mightotherwise have impeded the development of themonetary statistics. Thus, while statistical legisla-tion and adequate resources16 in particular are neededto support sustainable improvements in statisticaloutput, shortcomings in these areas in the early yearsof development of a statistical system need not pre-

15 See, for instance, the IMF’s Generic Data Quality AssessmentFramework (DQAF) available from the IMF’s data quality refer-ence site: http://dsbb.imf.org/Applications/web/dqrs/dqrshome.

16 Determining what constitutes adequate resources usually in-volves a heavy dose of judgment based on extensive experienceacross countries and with local circumstances.

Main UnitSarajevo

CommercialBanks

Regional andCountry-wide

Monetary Statistics

CENTRAL BANK OF BiH (CBBH)

Research &DevelopmentDepartment

(Sarajevo)

AccountingDepartment

(Sarajevo)

Main UnitMostar

Main Bankof RS

Banja Luka

CBBH BranchBrcko District

Main UnitSarajevo

FEDERATION

REPUBLIKASRPSKA

BRCKODISTRICT

Main UnitMostar

Main Bankof RS

CBBHBranch

CommercialBanks

CommercialBanks

CommercialBanks

SOURCES OF MONETARY DATA

Figure 4.1. Institutional Arrangements for Data Collection and Compilationof Monetary Statistics for BiH(As of July 2001)


47

MONETARY STATISTICS COMPILATION FOR BiH(beginning July 2001)

MA ANALYTICAL

ANALYTICALPRESENTATIONS

DATASTORAGE

RESEARCH & DEVELOPMENTDEPARTMENT

SOURCEDATA

Central BankSurvey

DMB ANALYTICAL

SURVEYBiH

BiH MA

CBBH

Monthly periods(one per sheet)


CBBH CBBHCBBH

(Central Bank of BiH)

Sarajevo AreaDMBs

CommercialBanks

Mostar AreaDMBs

CommercialBanks

RS DMBsCommercial

Banks

Brcko DistrictDMBs

CommercialBanks

REPORT FORMS BALANCE SHEETS

SARAJEVO DMB

MOSTAR DMB

FED DMB

RS DMB

BD DMB

BiH DMB

SARAJEVO DMB


MOSTAR DMB


RS DMB


BD DMB

OtherDepository

CorporationsSurvey

DepositoryCorporations

Survey

SOURCE DATA

FEDERATION

REPUBLIKASRPSKA

BRCKODISTRICT

Current IFS TerminologyMonetary Authorities (MA)Deposit Money Banks (DMB)Monetary Survey

New TerminologyCentral Bank Survey (CBS)Other Depository Corporations Survey (ODCS)Depository Corporations Survey (DCS)

Figure 4.2. Spreadsheet Architecture for Data Compilation


vent production of at least limited reliable statisticsin the near term if there is a focus on overcomingmore immediate shortcomings.

• Pace of technical assistance: The authorities needsufficient time to absorb TA. While intensive TAwas initially necessary in BiH and ran ahead of theauthorities’ capacities to absorb it, the consolidationphase and training that followed allowed the author-ities to catch up before additional TA was provided.Also, while having a medium-term framework fordelivering TA is important, the framework shouldevolve over time and be flexible enough to allow theintensity and frequency of the TA to respond tochanging local circumstances. The extended periodand changing pace of TA to BiH, with each missionbuilding on the previous mission, greatly facilitatedthe gradual enhancement of the authorities’ capacityfor compiling reliable and timely monetary statis-tics. An environment that encourages candid post-mission assessment by the mission team of theauthorities’ capacity to absorb TA and the success orfailure of a particular mission is clearly important toadjusting the pace of TA to suit local circumstances.

• Balance in technical assistance: Describing to theauthorities what needs to be done, how to do it, andactually implementing the advice are important ele-ments of TA. Finding the right balance is not easy,but it is important for developing country owner-ship of the project. In BiH, the initial balance wasweighted heavily toward getting the compilationsystem up and running, with the IMF taking the leadin setting up and running the system (a hands-onapproach). Thereafter, the balance shifted towardidentifying tasks that needed to be done, makingrecommendations about how to do them, and leav-ing the authorities to implement them (a hands-off approach), with support provided via e-mail and

48

the IMF resident statistical advisor and resident rep-resentative, as needed. An important link betweenthese two approaches was having mission represen-tatives sit down with the data compilers in front ofcomputer spreadsheets and walk through the nitty-gritty of how the statistics were put together.

• Comprehensive documentation: Essential to thehands-off approach is comprehensive and timelydocumentation that not only sets out clearly thetechnical aspects of the recommendations but alsoprovides a comprehensive and time-lined work plangeared toward senior- to mid-level country officials.In short, the plan should summarize in nontechnicalterms what needs to be done, by which depart-ment(s), by when, and why it is important. Transla-tion of the mission reports into the local language(s)is essential in this regard.

• One size does not fit all: While experience fromother country settings is valuable in developingrecommendations, a formula approach is unlikelyto yield successful results—a flexible approach isnecessary. For example, while striving for the beststatistical practice, it will often be necessary to set-tle for second- or third-best alternatives during aninterim period to achieve results in the near term.Recommendations should be shaped to fit localconditions, and a prioritized action plan should beagreed to with the authorities, taking into accountlocal constraints.

• Coordination with other donors: Having anoverview of the activities of other donors can gen-erate synergies with the statistical project at hand.For example, the timely USAID work on upgrad-ing the commercial bank accounting and data re-porting systems rapidly improved the accuracy andreliability of the sources used to compile monetarystatistics.

Setting

5.1 Following the dissolution of the Soviet Union,Ukraine gained its independence in December 1991.On September 3, 1992, it became a member of theIMF. In May 1991 the National Bank of Ukraine (NBU)was established, replacing the State Bank (Gosbank)of the Soviet Union.

5.2 In 1991, the banking sector of Ukraine consistedof the NBU (functioning both as a central bank and acommercial bank), the Savings Bank (the only ins-titution that accepted household deposits), and theExport-Import Bank for financing foreign trade oper-ations. In 1992, commercial activities were separatedfrom the central bank functions, and commercial banksemerged.

5.3 As in other countries of the former Soviet Unionand centrally planned economies in eastern Europe,monetary data were limited to those on credits anddeposits in national currencies. Allocation of credit tothe economy had been centrally planned, and mone-tary data were used for determining how the plan wasto be fulfilled. Thus, data on credits and deposits wereclassified by purpose, maturity, and debtor/creditorbranch (ministry) of the economy. However, data onbanking operations in foreign currencies and detailedbalance sheets of banking institutions were consid-ered top secret and therefore not used for compilingmoney statistics.

5.4 Other types of financial instruments had not beenused, a sectoral breakdown of domestic institutionswas unavailable, and the concept of residency wasunknown. The monetary data had been compiled onmainframe computers, and personal computers were

5. Ukraine: Monetary Statistics

Jaroslav Kucera*

49

*The author is a senior economist in the Financial InstitutionsDivision I of the IMF Statistics Department.

unavailable to the NBU staff. The NBU’s staff wereunfamiliar with compiling monetary statistics in ac-cordance with international standards.

Statistical Capacity, TechnicalAssistance, and Institutional Issues

Stage One: Fact Finding and Outlining the Plan

5.5 The IMF’s TA to Ukraine in 1992 identified prob-lems and initiated the compilation of monetary statis-tics using the analytical framework that the IMF de-veloped for the countries of the former Soviet Union.The first step in the action plan was to establish a sys-tem for collecting accounting data from all bankinginstitutions for compiling monetary statistics on amonthly basis. A significant challenge was collectingmonthly data from the Savings Bank, which previ-ously had supplied only quarterly data. An NBU unitresponsible for compiling monetary statistics was iden-tified; its staff was trained during the missions, as wellas in subsequent IMF courses on money and bankingstatistics.

5.6 To introduce and apply international methodologyto Ukrainian monetary statistics, it was necessary tobegin by examining the existing methodology and planof accounts to see how these data sources could beadapted. The first set of monetary statistics that broadlyaccorded with international methodology (developedby the NBU staff working with IMF staff) used the ex-isting accounting data. The NBU and all other bankinginstitutions in Ukraine used a unified plan of accounts,based on the former Gosbank accounting plan. At theoutset, it was necessary to study this accounting plan(consisting of about 1,000 three-digit accounts) andmap its components to the monetary statistics cate-gories. Since this accounting plan did not provide allthe necessary data for compiling monetary aggregates,some data had to be estimated. In particular, most ac-counts denominated in foreign currencies, even whenthey belonged to residents of Ukraine, were classified


as positions with nonresidents. These shortcomingscomplicated the accurate determination of the externalposition of the banking system.

5.7 The TA missions provided guidance on the con-solidation of banking system accounts into a monetarysurvey. The missions also made recommendations tomodify the plan of accounts so that banks could com-pile data in accordance with the residency of the ac-count holder, rather than on the basis of the currencyin which the account was denominated. On this basis,it was expected that the authorities would be able tobegin compiling monetary statistics in accordancewith the framework established by the TA missionsand that they could start reporting to the IMF on a reg-ular monthly basis.

Stage Two

Implementation of initial recommendations andfurther steps to improve monetary statistics

5.8 The IMF mission in late 1993 found that the ini-tial recommendations had not been implemented. Inparticular, the separation of resident and nonresidentaccounts had not been completed. Improving the datahad been hampered by lack of cooperation among thethree departments involved in collecting and compil-ing monetary statistics (the Accounting, MonetaryPolicy, and Economic Analysis and Statistics Depart-ments) and by insufficient monitoring of the account-ing changes that were crucial to the analytical frame-work for the compilation of monetary statistics.

5.9 Owing to a lack of coordination within the NBU,confusion arose over how to establish the officialmonetary statistics for Ukraine. Different monetarystatistics were compiled by the Economic Analysisand Statistics Department and the Monetary PolicyDepartment. The Economic Analysis and StatisticsDepartment compiled monetary statistics according tothe analytical framework developed by IMF TA mis-sions. However, it had not been able to adjust theframework, because new accounts were continuouslybeing created by the Accounting Department. In ad-dition, there was a problem with the proper distinctionbetween on–and off–balance sheet items, with a por-tion of international reserves being recorded as off–balance sheet accounts.

5.10 Responding to an urgent need for monetary sta-tistics for financial programming purposes, the Mon-etary Policy Department developed procedures for

50

compiling monetary statistics that were sufficient forinitial monitoring of the economy in Ukraine. How-ever, the Monetary Policy Department’s methodologyfor compiling the data differed in many respects fromthe internationally accepted methodology. In particu-lar, (i) foreign currency deposits of residents were in-cluded in foreign liabilities, (ii) insurance premiumswere classified as broad money, and (iii) local gov-ernment and pension fund deposits were classified asmoney, even though local governments and pensionfunds were not independent and obtained financingresources mainly through redistributions and transfersfrom the central government.

5.11 TA missions during 1994 and 1995 aided the au-thorities in revising the analytical framework so thatmore detailed monetary statistics could be based onthree-digit accounts. As a result, staff of the NBU be-came more familiar with the classifications (by eco-nomic sector and liquidity) underlying the analyticalframework. To strengthen communication among de-partments involved, a working group for compilingand monitoring monetary statistics was formed in re-sponse to mission recommendations.

5.12 Because of frequent changes in the plan ofaccounts, missions recommended that the NBU pro-vide extensive training and assistance to commercialbanks to ensure the proper introduction and classi-fication of new accounts. They suggested that thisassistance be provided through training seminars forcommercial bank staff, NBU staff visits to majorUkrainian banks, and written instructions for intro-ducing the new accounting codes prepared by theNBU’s Accounting Department.

5.13 It was also recommended that the NBU compileonly one monetary survey, which would be based ona proposed analytical framework that would be re-fined over time. The single set of official monetarystatistics to be produced by the NBU’s EconomicAnalysis and Statistics Department would aim at pro-moting meaningful interpretation of economic devel-opments in Ukraine, as well as fostering the futurepublication of monetary statistics.

Substantial progress

5.14 The monetary and banking statistics mission inFebruary 1997 concluded that the NBU had made sub-stantial progress over the previous four years in devel-oping the capability to compile and publish monetary

5 • Ukraine: Monetary Statistics

51

statistics in accordance with international standards.This progress reflected the strong commitment to thiswork by the senior management of the NBU and a highdegree of cooperation between the NBU staff and theIMF’s TA missions. It also reflected the dedication andexpanding skills of the staff of the Economic Depart-ment (successor of the Economic Analysis and Statis-tics Department), which maintained effective qualitycontrol over the monetary statistics.

5.15 Even though staff still compiled two sets ofmonetary data, TA in monetary statistics narrowed thegap between the two data sets, bringing each set closerto the IMF’s methodology. It remained for both de-partments to coordinate and harmonize their work onmonetary statistics with a view to producing a singleset of monetary statistics that would fully conformwith the IMF’s methodology.

5.16 Other achievements included the introduction ofnew accounts to separately identify banks’ nonresi-dent positions and thereby improve the sectorizationof residents’ accounts; computerization of the inter-bank settlements system; streamlining of the NBU’sprocedures for recording foreign exchange transac-tions; and changes in banks’ accounting procedures toreduce asymmetries in the interbank accounts.

5.17 The improvements in Ukraine’s monetary sta-tistics were essential for the creation of a Ukrainepage in the IMF’s International Financial Statistics(IFS), which was introduced in July 1996. Thereafter,the reporting of monetary data to the IMF by the Eco-nomic Department became reliable and timely.

Stage Three

New plans of accounts

5.18 The blueprint for developing the new plans of ac-counts for the central bank and other banking institu-tions was initiated in 1993, in cooperation with theWorld Bank. This approach was fully supported by theIMF’s TA missions, which emphasized that the abilityto obtain more meaningful analytical data would belimited unless new plans of accounts for the centralbank and other depository corporations were imple-mented. The new plans of accounts were to containthe classification of financial instruments by eco-nomic sector and liquidity, which was needed forcompiling reliable analytical accounts to support thepolicy needs of the NBU. In particular, the TA mis-sions emphasized that new plans of accounts should

disaggregate accounts by residency, instrument, eco-nomic sector, and liquidity category.

5.19 The authorities had planned to introduce the newplans of accounts without a formal transition period.The IMF’s TA missions suggested, however, that theaccounting reform should contain a transition periodduring which banks would use both the old and newcharts of accounts. If the old charts of accounts wereabandoned without a transition period, the introductionof the new plans of accounts would expose the Ukrain-ian banking system to unnecessary risks. Only with atransition period would it be possible to ensure conti-nuity in the monetary data and to verify the consistencyof the monetary statistics derived from the old and newplans of accounts. Also, potential disruption in the flowof monetary data could result from a switch to the newaccounting system that was only partially successful.

5.20 The TA mission in 1997 reviewed the ongoingwork on the design of the new plans of accounts andnew reporting forms. The main conclusion was that thelatest drafts of the new charts of accounts would makeit possible to derive the monetary statistics directlyfrom the balance sheets of banks (as was the case whenthe old accounting plan had been used). Thus, the mon-etary statistics would continue to be derived from afully integrated set of accounts for which the reliabil-ity of the accounting data could be monitored.

5.21 The new plans of accounts for the NBU andcommercial banks were introduced successfully inJanuary 1998, followed by a short transition period.Even though monetary data were not reported to theIMF for a few months, the staff of the NBU were ableto prepare a conversion to the new compilation frame-work on their own.

5.22 The new plans of accounts provided internatio-nal accounting guidelines to Ukrainian banks, to en-sure that the monetary data would be compiled in ac-cordance with international standards. This applied inparticular to the accounting and the monetary statis-tics methodology to include accrued interest in theunderlying monetary aggregates.

Improvements in the framework for thecompilation of monetary data based on new plans of accounts

5.23 TA provided by the IMF from 1999 to 2001 ini-tially concentrated on reviewing in detail the new


framework for the compilation of monetary statisticsand recommending improving data quality by compil-ing monetary statistics directly from accounting re-cords. Improvements in monetary statistics were alsoachieved by resolving the most important outstandingmethodological issues relating to monetary aggre-gates. Later, TA was provided to implement the IMF’sMonetary and Financial Statistics Manual (MFSM).

5.24 In general, it was confirmed that procedures forcompiling monetary statistics, based on the new plansof accounts, were broadly in accordance with inter-national accounting guidelines. These procedures werebased largely on balance sheet data of the NBU andcommercial banks, but supplementary data were stillneeded for proper sectorization of bank loans, securi-ties holdings, and deposit liabilities.

5.25 Detailed analysis of the data and compilationprocedures revealed that even though staff were com-piling the monetary data in accordance with inter-national accounting guidelines, shortcomings in thecomposition of some monetary aggregates persisted.Accordingly, TA recommendations included introduc-ing additional accounts within the plans of accounts ofthe NBU and commercial banks, thereby further facil-itating the compilation of monetary statistics directlyfrom accounting records. These recommendations pri-marily dealt with the compilation of foreign assets,international reserves, and claims on the central andlocal governments. To improve the quality and relia-bility of monetary data, it was also recommended thatthe NBU’s and commercial banks’ accounting datashould be regularly analyzed from the perspective ofmeeting the needs of the monetary statistics.

5.26 These efforts contributed to the formulation andmonitoring of economic policy by providing timelyand frequent data on the banking sector’s activity withindividual sectors of the domestic economy (in par-ticular, with the government sector) and nonresidents.To improve the analytical quality of the monetary datacompiled by the NBU, the IMF recommended severalchanges in methodology, the most important of whichwere (i) expanded institutional coverage to include allfinancial institutions that accept deposits or issue otherliabilities viewed as deposit substitutes and (ii) theproper treatment of (a) Ukrainian bank branches oper-ating abroad, (b) data on banks with revoked licenses,(c) classification of certificates of deposits issued bythe NBU and commercial banks, (d) classification oflocal government deposits, (e) classification of ac-

52

counts of the banks’ employees, and (f) treatment ofshares and equity.

5.27 Most of the recommendations on methodologywere implemented, and in the process the hands-onTA and on-the-job training contributed significantlyto the NBU staff’s understanding of monetary statis-tics methodology. Overall, the implemented recom-mendations on methodological issues have greatly im-proved the quality of Ukrainian monetary statistics.

5.28 TA in 2002 concentrated on the implementationof the MFSM. TA missions conducted seminars for theNBU staff and held discussions with individual staffmembers of the Economic Department concerning theMFSM chapters dealing with sectorization of the econ-omy, classification of financial instruments, and com-pilation of monetary aggregates. Issues concerning theadaptation of the IFS to the MFSM methodology havealso been discussed. It was concluded that close co-operation of the Economic Department and other NBUunits is necessary to introduce the system for collect-ing accounting data from commercial banks and, pos-sibly, all other financial institutions if the reporteddata are to be consistent with the methodology in theMFSM. The IMF mission in April 2002 also suggestedthat the authorities establish a plan for implementingall of the recommendations of the MFSM.


5.29 During the past 10 years, the IMF’s TA toUkraine achieved most of the objectives agreed withthe authorities, notwithstanding the hope that the re-sults might be achieved in a shorter period. However,these hopes had to be tempered by the realization thatfundamental reforms were needed in both the bankingsystem and the rest of the economy. In addition, theculture of the NBU had to evolve during the 1990s ina way that supported a more proactive environment ofinitiating and implementing reforms.

5.30 Rapid progress in the development of monetarystatistics was hampered by cumbersome administra-tive procedures, as well as by frequent changes in thestaff involved in compiling monetary data. For exam-ple, the IMF mission in February–March 2000 foundthat all previous staff of the Monetary and BankingStatistics Division of the NBU—the unit responsiblefor compiling the analytical accounts of the monetaryauthorities and the commercial banks—had retired asa result of the Law on Public Service and the presi-

5 • Ukraine: Monetary Statistics

53

dential decree on the reduction of the staff of govern-ment institutions.

5.31 Cooperation and coordination were essential forimproving monetary statistics. The strong leadershipof NBU management, a high degree of interaction withIMF TA missions, and the dedication of compilers ofmonetary statistics gradually improved the workingenvironment and the cooperation of all parties in-volved in collecting and compiling monetary statistics.

5.32 Through interactions with IMF missions, the NBUhas made the two sets of monetary statistics almostidentical. Actions were undertaken to introduce an in-tegrated monetary database for Ukraine, wherein staffwould compile and report a single set of monetary sta-tistics to the IMF. The single set of monetary statisticswill be used also for official publication within Ukraineand for the NBU’s conduct of monetary policy.

5.33 A consistent and sustained strategy of IMF TAover a medium- to long-term period contributed todeveloping monetary statistics in Ukraine in accor-dance with international standards. Regardless of ini-tial problems and difficulties in cooperation amongNBU units, it was important that the IMF continuouslyprovide TA and train staff in monetary statistics meth-odology. The fact that the IMF was the major TA

provider in monetary statistics and hence did not gen-erally need to coordinate with other donors simplifiedissues and avoided the confusion that sometimes ariseswhen many donors are involved in the same sector.One area of coordination that was necessary and thatwent smoothly was ensuring that the World Bank ad-visors incorporated in the new plan of accounts theclassifications and sectorizations needed for compilingmonetary statistics directly from accounting records.

5.34 Proper legislation and adequate budgetary andhuman resources have also been positive factors fa-cilitating progress in the development and improve-ment of monetary data. In the early stages of devel-opment, the obligatory reporting culture applicable inthe former Soviet Union overcame possible problemsin the collection of data. The current law for the NBUentitles it to obtain all necessary information requiredfor the compilation.

5.35 Over the past decade, monetary statistics inUkraine have improved significantly. On January 10,2003, Ukraine became the fifty-second subscriber to theSpecial Data Dissemination Standard (SDDS) and thefirst among the Commonwealth of Independent States.Subscription to the SDDS represented an importantmilestone in the decade-long refinement of Ukraine’smonetary and other macroeconomic statistics.

Statistical Capacity Building - IMF -- International … capacity building : case studies and...

Documents

Transcript of Statistical Capacity Building - IMF -- International … capacity building : case studies and...