in emerging economies - imf elibrary...budget system reform in emerging economies : the challenges...

120

Upload: others

Post on 05-Oct-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary
Page 2: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

O C C A S I O N A L P A P E R 245

Budget System Reform in Emerging Economies

The Challenges and the Reform Agenda

Jack Diamond

INTERNATIONAL MONETARY FUND

Washington DC

2006

©International Monetary Fund. Not for Redistribution

Page 3: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

© 2006 International Monetary Fund

Production: IMF Multimedia Services Division

Cataloging-in-Publication Data

Diamond, Jack, 1946–Budget system reform in emerging economies : the challenges and the reform agenda /

by Jack Diamond — Washington, D.C. : International Monetary Fund, 2006.

p. cm. — (Occasional paper ; 245)

ISBN 1-58906-474-7

1. Program budgeting. 2. Accrual basis accounting. 3. Expenditures, Public. 4. InternationalMonetary Fund. I. Title. II. Series : Occasional paper (International Monetary Fund) ; no. 245

HJ2031.D43 2006

Price: US$28.00(US$25.00 to full-time faculty members and students at universities and colleges)

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

©International Monetary Fund. Not for Redistribution

Page 4: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Abbreviations and Acronyms vi

Preface vii

I Overview 1

II The Strategy for Budget System Reform in Emerging Market Economies 3

Stages in Budget System Development 3A Three-Track Path to Reform 4The Steps toward Reform 7Three Preconditions for Reform 9Institutional Framework for Devolved Management 12Managing the Change Process 13The Wider Framework for Budget Reform 13Is There Need for a New Legal Framework? 14

III Moving from Program to Performance Budgeting 17

The Road from Old to New Performance Budgeting 17Issues of Design in a Program Structure 21The Wider Strategic Framework for the Program Structure 23Introducing Medium-Term Budget Frameworks 24

IV Introducing a Performance Management Framework 31

Different Measures of Performance 31Defining a Framework for Performance Measurement 33Experience in Using Performance Measures 36Guidelines for Using Performance Measures 38Establishing a Performance Information System 39Developing a Performance Management System 40Concluding Remarks 43

V Establishing Basic Public Expenditure Management Thresholds 44

Internal Control Systems 44Internal Audit 45Financial Management Information Systems 49Costing Systems 52Assessing the Performance of the Budget System 54Concluding Remarks 55

Contents

iii

©International Monetary Fund. Not for Redistribution

Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Page 5: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI Moving from Cash to Accrual Accounting 57

Cash Versus Accrual Accounting 57Accounting Models for Different Management Needs 59Accrual Accounting as a Component of Wider Budget

System Reform 61Preconditions for the Move to Accrual Accounting 64A Five-Stage Transition to Accrual Accounting 65Preliminary Guidance for Accrual Accounting 68The GFSM 2001 Standard for Fiscal Reporting 70Concluding Remarks 73

VII The Institutional Framework for Budget System Reform 75

The Impetus for Performance Budgeting 75Different Approaches to Less-Centralized Budget Management 75The Five Ds in Practice 77Concluding Remarks 88

VIII Managing the Reform Process 90

What Can Be Learned from the OECD Experience? 90The Overall Approach to Reform 91Elements of a Reform Strategy 92Providing an Enabling Environment 98The Need for a Compatible Legal and Regulatory Framework 100The Key Lessons for Budget Reform 101Concluding Remarks 103

References 104

Boxes

1. Pros and Cons of a Traditional Line-Item Budget 42. OECD Practices: Flexibility in Budget Management 53. Russian Budget System Reforms 74. Chinese Budget Reforms: Strengthening Basic

PEM Systems 105. Major Types of Fiscal Policy Rules 146. Summary of Good Practices in Institutional Transparency 157. Elements of the Brazilian Fiscal Responsibility Law, 1999 158. The Steps toward the “New” Performance Budgeting 189. OECD Practices: Use of a Program Structure 2110. The U.S. Government’s Program Assessment

Rating Tool (PART) 2211. Limitations of a “Bottom-Up” Planning Process 2312. Connecting Planning with Budgeting 2513. OECD Practices: Use of MTBFs 2614. Process for Agreeing on Program Budget Format 2815. General Guidelines on the Design of Programs 2916. Desirable Properties of Outputs 3417. Desirable Properties of Outcomes 3418. Typical Service Quality Characteristics 3519. OECD Practices: Use of Performance Information 3620. The U.K. Government’s Public Service Agreement (PSA)

Framework 3921. Key Principles for Reporting Performance Information 4022. Elements of the Strategic Plan 41

CONTENTS

iv

©International Monetary Fund. Not for Redistribution

Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Page 6: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

23. The Operating Plan 4124. OECD Practices: Performance Management in Government 4325. The Attributes of a Well-Designed FMIS 5026. Changes in Information Needs to Accommodate

Performance Budgeting 5027. Attributes of a Costing System 5228. Basic Steps of a Program-Based Costing System 5329. Critical Objectives of PEM Systems Measured through a

Standardized Assessment 5530. Pros and Cons of Cash-Based Accounting 5831. OECD Practices: Reorienting toward Accruals-Based

Accounting Systems 6032. Usefulness of Accrual Accounting for Government

Operating Units 6533. Design of the Capital Use Charge 6634. Steps Involved in Moving to Modified Accruals 6735. Key Features of GFSM 2001 7036. Stages in Accommodating GFSM 2001 Reporting 7237. The Pros and Cons of Devolved Budget Management 7638. Different Types of Autonomous Public Bodies 7839. Seven Broad Types of Performance Contract 8040. Pros and Cons of Contracting 8141. OECD Practices: Degree of Commercialization in

Budget Management 8242. OECD Practices: Performance Management in

Government Agencies 8343. Tanzania’s Experience in Creating Executive Agencies 8544. Commercialization of Government Services 8645. Guidelines for Contracting In and Outsourcing

Government Services 8746. The Malaysian Modified Budget System 9347. The Steps for Creating Major Institutional Change 9348. Thailand: Line Agencies Reform Contract 9549. Formal Performance Agreements between Central Authorities

and Agencies in Latin America and the Caribbean 9650. The Experiences of Australia and Sweden with Using

an Efficiency Dividend 9751. The U.S. Government Performance and Results Act, 1993 10052. Turkey’s Public Financial Management and Control Law 10253. Checklist for Assessing the Risks in Delivering TA for

Budget Reform 103

Figures

1. The Concept of Performance in Different Budget Systems 322. Technical and Economic Efficiency 33

Contents

v

©International Monetary Fund. Not for Redistribution

Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Underline
Page 7: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

ABC Activity-based costingBO Budget officeBOP Balance of paymentsB-O-T Build-Operate-TransferCBO Congressional Budget Office (U.S.)CE Chief executiveCoA Chart of accountsCOFOG Classification of the functions of governmentCRM Constituent/customer relationship managementECOFIN Council of Economics and Finance Ministers of the European UnionEU European UnionFAD Fiscal Affairs Department (of the International Monetary Fund)FRL Fiscal responsibility lawGAO Government Accountability Office (U.S.), formerly General

Accounting OfficeGFMIS Government financial management information systemGFS Government finance statisticsGFSM Government Finance Statistics ManualGPRA Government Performance and Results Act (U.S.)IMF International Monetary FundIRA Independent revenue authorityIT Information technologyLM Line ministryMoF Ministry of financeMBS Modified Budgeting SystemMTBF Medium-term budget frameworkMTEF Medium-term expenditure frameworkOECD Organization for Economic Cooperation and DevelopmentOMB Office of Management and Budget (U.S.)PART Program assessment rating toolPBC People’s Bank of ChinaPEFA Public Expenditure and Financial Accountability Program (World

Bank, European Commission, IMF, and others)PEM Public expenditure managementPFM Public financial managementPPBS Planning, programming, and budgeting systemPPP Private-public partnershipPSA Public service agreementSDA Service delivery agreementSGP Stability and Growth Pact (EU)SMART Specific, measurable, achievable, relevant, and timedSNA System of national accountsSPO State planning officeSUNAT State revenue authority of PeruTA Technical assistanceTSA Treasury single accountUN United Nations

Abbreviations and Acronyms

vi

©International Monetary Fund. Not for Redistribution

Page 8: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Over the past two decades, many countries in the Organization for Economic Coop-eration and Development (OECD) have introduced fundamental changes in budgetmanagement involving increased emphasis on performance and results achieved fromthe use of public resources. With increasing frequency over the past decade, the FiscalAffairs Department (FAD) of the IMF has been called upon to assist middle-incomecountries, especially emerging economies, in adopting these types of budget reforms.Repeatedly, technical assistance (TA) missions have been required to offer advice onhow to introduce or sustain such reforms. In doing so, missions typically have addresseda number of recurring questions. What has been the experience of OECD countries?Are there any general lessons to be learned? Can, or should, the same general reformstrategy be applied to non-OECD countries? Just how universal is this reform para-digm? How should countries first begin these reforms, and how should they be subse-quently sequenced? Do all countries have the management capabilities within theirgovernments to implement such reforms?

Given that the same questions are raised in many countries, it was considered usefulto attempt to provide the answers, perhaps more comprehensively, based on a reviewof our experience providing TA to middle-income countries. Not surprisingly, this com-prehensive view revealed the basic underlying strategy that was being recommendedfor budget system reform, and this is reported in this study. Not only is this an impor-tant input to FAD’s ongoing efforts to review and improve its TA advice, but it is hopedthat it also will be useful to policymakers and administrators in emerging economieswho are contemplating such reforms.

Given the nature of the study, based on a review of TA advice, this is very much aproduct of the numerous FAD TA teams that have worked on a wide range of relatedissues. The fundamental contribution from all these colleagues, especially in the twopublic finance management divisions of FAD, is gratefully acknowledged here. The au-thor is particularly appreciative of the opportunity to work with and to learn from PiyushDesai, Geoff Dixon, Ole Hovland, Tony Olliffe, and Vijay Ramachandran, whose con-tributions to this study are significant. Special thanks are due to Jim Brumby, MarcRobinson, and Holger van Eden who commented on—and considerably improved—previous drafts. As always, the author bears responsibility for any remaining errors.Special mention should be made to Miriam Villarroel, Raquel Malamud, and VictoriaMacchi, who labored with such good humor on a number of previous drafts of thisstudy. Linda Griffin Kean edited the manuscript and coordinated production of thepublication.

Preface

vii

©International Monetary Fund. Not for Redistribution

Page 9: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

This page intentionally left blank

©International Monetary Fund. Not for Redistribution

Page 10: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

I Overview

1

Among the major challenges faced by transitionaland emerging market economies is the need to ad-

just institutions to function in an increasingly market-oriented and global environment. Among the reformsthat middle-income countries have looked to emulateare the budget reforms that have been introduced inOECD member countries since the 1980s. These re-forms have reoriented budgeting from the traditionalfocus on inputs to a new focus on the results derivedfrom these inputs. This latter focus has often beentermed performance or results-based budgeting. How-ever, the resulting budget systems embody more thana change in the process of budgeting. They reflect afundamental change in budget management, awayfrom traditional, centralized control systems to moredecentralized management models.

It is argued in Section II that this reform processcan be characterized as following three fundamentaltracks: first, to allow managers greater flexibility inmanaging resources; second, to give them greater cer-tainty in resourcing; and third, to introduce a system ofrewards and penalties to pressure them to perform, inthe sense of achieving the stated objectives of govern-ment policy. The pursuit of this three-track reformprocess in turn has fundamentally altered the account-ability relationships within government by replacingdetailed central controls with greater flexibility forbudget managers operating at “arm’s length.” Thesenew accountability relationships are designed to im-pose discipline on this new performance managementframework, are oriented toward results rather than to-ward inputs, and can be viewed as the cornerstone ofthe new performance management model. Not sur-prisingly, their introduction has generally required con-siderable effort to restructure the budget system.

Usually, the first step is to improve the definition ofgovernment programs and clarify their objectives, toensure that programs are prioritized according to astrategic policy framework. While the need for a mean-ingful link between policy and budgeting has long beenrecognized, it has also consistently proved elusive, asindicated by a brief review of the history of the perfor-mance budgeting approach. Officials in most emergingeconomies accept that the central role of the programstructure is to translate broad policies into activities andprojects that can be costed, with the identified resourcerequirements approved as budget appropriations. At thesame time, many also appreciate that this cannot be

viewed as a one-time annual exercise. Increasing num-bers of countries are adopting medium-term budgetframeworks (MTBFs) to assist in capturing the fullcosts of activities and projects over time, and hence tobetter plan programs, to improve prioritization amongthem, and to provide some overall discipline over theirresource use. However, as discussed in Section III, de-veloping and maintaining an MTBF is not always easy.

The next step in creating the new accountabilityframework is to link inputs with program outcomes, andthen to make performance information relevant to man-agers by tying this information to resource allocationdecisions. As discussed in Section IV, the definition andmeasurement of program outputs (and, even more so, ofprogram outcomes) is often problematic. Also, there areadded requirements, namely to produce such perfor-mance information on a consistent basis, to provide in-centives for managers to use this information, and tomonitor the managers’ performance against these stan-dards. Not surprisingly, it is likely to prove difficult andcostly for many emerging economies to implement afull-blown performance management system such asthat found, for example, in New Zealand or Australia.

The remainder of this study discusses the type ofchanges required to facilitate adoption of the new bud-get management model. Based on OECD country ex-periences, it argues that, first, certain basic safeguardsshould be put in place to ensure that public expendituremanagement (PEM) systems are able to accommodatethe new demands. Second, the new accountabilityframework generally requires complementary changesin the way government operations are institutionallyorganized. This in turn often requires parallel changesin the legal framework. Third, a considerable invest-ment must be made, both politically and financially, tomanage the reform process.

Three aspects of the PEM system that often are weakin emerging countries and require upgrading are dis-cussed in Section V. First, there is the need to strengtheninternal controls. Before attempting to give agencieswider responsibilities in resource allocation, it is essen-tial to ensure that they are operating within an effectivefinancial management framework. Good internal con-trol is an important feature of this framework. Withoutsatisfactory controls, management may not detect se-rious errors and irregularities, and the work of thecentral oversight agencies, as well as external audit,becomes more difficult. Second, internal audit is a

©International Monetary Fund. Not for Redistribution

Page 11: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

I OVERVIEW

central component of internal financial controls aimedat protecting the government’s financial interests.Third, there is typically a greater need to apply infor-mation technology. Against the background of theever-growing volume and complexity of governmentfinancial operations, timely management informationreports in a usable form are clearly of critical impor-tance to fiscal managers. Tailoring such reports tomanagement needs through a computerized financialinformation system has been a general PEM reformundertaken in various parts of the world. This has notbeen easy, and the new performance information re-quirements are likely to make this more difficult.Fourth, preparing this new performance informationrequires that institutions have the capacity to capturethe full cost of programs and activities so that these canbe related to performance measures to judge programperformance. Unfortunately, such cost accounting ex-pertise is often scarce in government.

Section VI reviews in greater depth one of themost discussed tools for strengthening governmentperformance—upgrading the government accountingsystem from a cash to an accrual basis. The typicalgovernment accounting system, even if conceptuallywell defined and internally consistent, is a cash-basedsystem or, at most, a modified cash system. The em-phasis is on matching approved items of spending withactual cash outlays, the last stage of spending. This maybe adequate for the compliance and stabilization objec-tives of the ministry of finance (MoF) but it is less rel-evant for budget managers in government departments.It has been argued, therefore, that the move to the nextstage of budget system development requires a switchin government accounting toward an accrual-based sys-tem. Those countries that have attempted to shift to ac-crual accounting have found it difficult and costly. Anargument can be made that many of the benefits of im-proved costing derived from accrual accounting can beobtained by other means, and accordingly, that perhapsan intermediary move to accruals is all that should beattempted by emerging economies.

The performance budget management approachrequires introducing enhanced accountability withimproved transparency, further emphasizing theevaluation of outputs in relation to inputs, and stream-lining control mechanisms to balance control needsand new efficiency requirements arising from moredecentralized budget management. Moving to thismanagement model has created parallel pressures toreorganize the way that governments do business. Thisaspect of budget reforms is taken up in Section VII. Toprovide a framework for discussing the wide range ofinstitutional innovations, these are ordered into fivemain groups, each characterized as a progressive shift

2

away from the traditional, centralized budget manage-ment model. These groups are characterized as thefive Ds—deconcentration, decentralization, delega-tion, devolution, and divestment. The latter three aremost closely associated with the new performancemanagement approach.

Two aspects of these new arrangements that havereceived the most attention internationally are high-lighted—the role of contracting in government and therole of more autonomous devolved management units,the so-called agency model. The study warns that theseapproaches can be problematic, especially initially ifadministrative and accountability systems are not suffi-ciently robust. Full output and performance contractingis very resource-demanding and has been fully realizedin only a few fairly advanced countries. Not surpris-ingly, even more limited movement in this direction islikely to strain the administrative systems in emergingeconomies. Similarly, the demands on PEM systemsare quite heavy, especially with regard to reporting re-quirements and the skills required of managers. Similarconcerns are expressed with regard to various strategiesfor divesting or sharing government responsibility forproviding public services, through such mechanismsas contracting and private-public partnerships (PPPs).

The fact that it is difficult to find emerging econo-mies that have made rapid progress on all three tracksof the reform process indicates that there are importantimplementation problems. Section VIII suggests thatmany of these difficulties have arisen from the lack ofbasic management infrastructure within the budgetsystem. The capacity to successfully link policies withprograms through strategic planning mechanisms re-quires the skills to construct MTBFs with well-definedprograms. It requires analytical skills and adequate re-porting systems to make clear the relationship betweenthe resources used by a program and its outputs and/orpolicy results (outcomes). As indicated, the basic workof program design and program costing is unlikely tobe straightforward and is likely to demand skills, suchas cost accounting, which are often in short supply. Atthe same time, this new approach to budget manage-ment also requires parallel administrative proceduresto activate the strategic plans and the program struc-ture for decision making. This move inevitably mustreflect the reform capacity of the country and, to besuccessful, should not be implemented too rapidlyand should be carefully sequenced. As argued in Sec-tion VIII, the transformation from traditional budget-ing procedures is likely to require a substantial effortto manage the change process, a dimension which hasoften been overlooked. Thus, for many countries,human resource constraints may impede, or at leastslow, the move to performance budgeting.

©International Monetary Fund. Not for Redistribution

Page 12: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II The Strategy for Budget System Reformin Emerging Market Economies

3

This section examines the process of budget sys-tem reform required to move from traditional,

centralized input-oriented systems to more modern,devolved, performance-based systems. It identifies astrategy based on recent attempts at this reform inemerging market economies and stresses the requiredsequencing of steps in this process. The strategy out-lined involves a three-pronged approach, requiringmeasures first to increase the flexibility in the agencyoperating environment, then to provide more certainresourcing for spending agencies, and finally to exertpressure on agencies to improve their performance.This section stresses the need to progress on all threefronts simultaneously and to base these reforms on afirm platform of agency management skills. It high-lights the need to attain some basic PEM thresholds inthe area of restructuring the budget and its classifica-tion system, to improve the accounting system, to de-velop a financial management information system(FMIS), and to strengthen internal financial manage-ment skills. At the same time, it explains that it mayalso be necessary to introduce changes to the institu-tional and regulatory framework in which budget man-agers operate.

Stages in Budget System Development

Among the major challenges faced by emergingmarket economies is the need to adjust institutions tofunction in an increasingly market-oriented and globalenvironment. One of the most important of these insti-tutional reforms has been restructuring the budget sys-tem.1 The term “budget system” should be interpretedquite widely to encompass the institutional frameworkas well as the administrative procedures that determinethe means whereby resources are transferred to gov-ernment; how their use is prioritized and directed toagreed policy objectives; and how their use is subse-quently managed, controlled, monitored, and reported.It is generally agreed that a modern budget system

should be able to meet three main requirements: first,to ensure control over expenditures so that they areconsistent with the budget law; second, to stabilizethe economy through timely and efficient adjustmentin fiscal aggregates; and third, to promote efficiencyin service delivery through procedures that provideincentives for greater productivity.2 The EuropeanUnion recently adopted such a framework for review-ing the quality of budgeting in its member countries.3

Typically, budget systems evolve by progressivelyassuming and placing different emphasis on thesethree requirements. The most basic systems tradition-ally focus on the first objective—ensuring compliancewith the annual budget law—and are generally associ-ated with detailed line-item budgeting. While meetingbasic compliance requirements, traditional budget sys-tems usually are later modified to accommodate thegovernment-wide stabilization objectives and the needto control fiscal aggregates. This usually involves atop-down approach to ensure that fiscal policy can beharmonized with monetary policy, through the intro-duction of procedures to enable the government both toplan, control, and monitor spending effectively and toadjust fiscal aggregates to meet fiscal targets which areincreasingly set in a multiyear framework. After bud-get systems are capable of handling compliance andstabilization objectives, increasing emphasis is typi-cally placed on the third objective, ensuring the efficientand effective use of government resources. Putting inplace mechanisms to meet this objective has progres-sively become the focus of interest for OECD budgetmanagers over the last two decades. The challengehas been to ensure enough flexibility so that budgetmanagers are given ample scope to manage, withoutsacrificing compliance and overall macro stabiliza-tion objectives.

The move to this third phase has been spurred bytwo separate developments. First, the limitations oftraditional compliance-oriented budget systems be-came more apparent when faced with the need to con-strain an ever-expanding public sector. Second, theprogressive use of more flexible budget management

1A parallel and related area, the reform of tax systems in transi-tional economies, has received much more attention. See Tanzi andZee (2000); Martinez-Vazquez and McNab (2000).

2See discussion by Schick (1998, 2001).3See European Commission (2004, pp. 166ff).

©International Monetary Fund. Not for Redistribution

Page 13: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

procedures brings about an increased realization thatstabilization and efficiency objectives do not necessar-ily compete and can be complementary. As a conse-quence, traditional budget systems, based on short-termand detailed control of inputs, have generally been dis-credited as a tool for promoting public sector perfor-mance, which by definition should focus on the outputsor results from these inputs (Box 1).

The increased suspicion that fiscal stabilization ob-jectives were being achieved at the expense of perfor-mance led OECD countries to modify their budgets toserve as performance management instruments and notjust as instruments for macro control. When they intro-duced performance management into their budget sys-tems, the associated increase in managerial freedomhelped make it evident that greater managerial flexibil-ity could be viewed as a tool not only for improving ef-ficiency but also for achieving expenditure targets. Thisderived partly from a realization that managers of indi-vidual programs are in the best position to decide themost appropriate mix of inputs to use for executing

4

their programs and that providing them greater man-agerial freedom could assist them in achieving tighterbudgetary limits—that is, improved efficiency in re-source use could support stabilization targets.4

Most emerging economies are only now entering thisthird stage of budget system development. They findthemselves facing a common set of problems whichstem from their past attempts to ensure that budgetaryprocesses delivered a satisfactory aggregate fiscal out-come in support of macroeconomic stabilization. Evenwhen care was taken to incorporate some degree ofbudget flexibility, the budget systems typically wereleft with a complex set of restrictions. These were in-creasingly recognized to diminish the allocative andoperational efficiency of budget execution, usually withbudget managers operating with limited responsibilityfor results. Accordingly, there has been a growing ac-ceptance that the next stage of reform is to providegreater inducement for managers to focus on possibleimprovements in allocative and operational efficiencyat the program delivery level.

A Three-Track Path to Reform

The shift within OECD countries over the past twodecades toward performance or results-based budget-ing has involved more than merely changing the waygovernments prepare budgets. While it is true thatthese reforms place much greater emphasis on perfor-mance information in the budget process, the innova-tion—almost a revolution—has been the abandonmentof highly centralized forms of budget management andthe adoption of more decentralized forms of manage-ment. The degree of decentralization has varied amongcountries, but all have attempted to create an environ-ment that will ensure better performance in resourceuse by providing budget managers incentives that letthem manage and also make them manage. In thisway, the reforms seek to optimize the perceived trade-off between tight fiscal control and effective programexecution. In a large number of OECD countries, theresult has been a fundamental shift away from tradi-tional, centralized decision making toward decen-tralized decision making based on a reformulatedaccountability framework that focuses on objectivesand specifies expected performance in terms of outputsor, in some cases, outcomes.

An examination of the trends in OECD budget re-forms reveals that countries have pursued a three-trackpath in creating this new budget management model.

Box 1. Pros and Cons of a Traditional Line-Item Budget

Pros

• Usually provides compliance control, allowing pay-ments to be limited to voted appropriations

• Provides database for across-the-board cuts to controlfiscal aggregates

• Is an uncomplicated system, easy to understand, focused on “the bottom line”

• Easy to apply in a tight timetable, i.e., annual; mini-mizes discussion on programs/objectives

• Adaptable to different economic situations (being incremental as well as decremental)

• Reporting is not demanding; accounts are preparedafter year-end for statutory review and audit

• Cash-basis accounting is relatively easier to imple-ment and maintain

Cons

• Budget is divided by spending unit/activity centersthat may have several programs

• Prepared annually, with year-end rush to spend• The typical incremental approach favors existing pro-

grams, regardless of priority• Based on line items of expenditure; control is on in-

puts rather than outputs or outcomes• Reports are for compliance purposes, by institution

and approved cost, and are usually detailed• Reports for stabilization purposes and control over ag-

gregates need additional economic classification• Supporting accounting system is cash based, focused

only on the payment stage of the spending process

4In one OECD study, greater managerial flexibility was associ-ated with a positive aggregate fiscal outturn. Allowing managersgreater freedom resulted in less emphasis placed on aggregatespending caps and increased the success of spending cutting exer-cises. See Strauch (2000).

©International Monetary Fund. Not for Redistribution

Page 14: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

A Three-Track Path to Reform

5

tures have been freed, there is still central control overstaffing.

Reduction of line-item specificity in the budget hasbeen a general feature of such reforms, although in-creased flexibility has been provided in different ways.A first step is usually consolidation of detailed appro-priation items into wider bands of expenditure cate-gories. Portfolio budgeting, a more pronounced formof this approach, gives ministries greater flexibilityand incentive to reallocate resources within portfolioallocations and to reflect changing priorities. A fewcountries have also introduced greater end-of-yearflexibility by allowing a carryover of unused appro-priations and the use of net appropriations to encouragethe generation of nontax revenues to finance specifictypes of spending. For example, the Australian andNew Zealand systems allow the sale or acquisition offixed assets. However, these approaches are still re-garded as radical by most countries. All such ap-proaches are founded on two main preconditions: atight budget constraint which agencies observe; andadequate capacity in the central budget office to mon-itor developments and intervene if necessary.

Track Two: Greater Certainty in Resourcing

The goal is to provide greater certainty about the op-erating environment of spending agencies, particularlyin regard to the availability of funds. There is usually

Box 2. OECD Practices: Flexibility in Budget Management

Percentage of OECD Countries

1. Government organizations receive one appropriation for all their operating expenditures 262. Government organizations receive separate appropriations for salaries and operating expenditures 263. No restrictions on transfers between appropriations 94. Transfers are allowed but only with MoF approval 365. Transfers between programs allowed with MoF approval 306. Transfers between capital and other expenditures allowed with MoF approval 317. Ability to carry over unused appropriations for operating costs at year’s end:

a. Without limit 19b. Up to a maximum percentage 19c. Allowed case by case by MoF 23

8. Ability to carry over unused appropriations for investments at year’s end:a. Without limit 17b. Up to a maximum percentage 8c. Allowed case by case by MoF 29

9. Ministries and government organizations maintain cash accounts separate from the MoF:a. Not allowed 48b. Fully separate 22c. Balances are swept to the treasury account periodically 22

10. Each ministry is in charge of its own procurement (operating within certain central limits) 8811. Internal audit is organized at ministry/department level 65

Source: OECD (2003, Tables 3.2.a, 3.2.b, 3.3.b, 3.4, 4.1).

5While agencies in the United States do not have the same flexi-bility as departments in Australia, it should be recognized that flex-ibility does exist through various mechanisms such as the authorityto reprogram funds up to specified limits without congressional ap-proval, the use of revolving funds, and multiyear appropriations.

Track One: Flexibility in Managing Resources

This involves providing increased flexibility tospending agencies in their access to budget funds andthe uses to which funds can be put. The focus of in-creased flexibility has mainly been on the ability ofspending agencies to reallocate funds within controlson budget line items. Changes are directed at givingorganizations and managers greater freedom in opera-tional decisions and removing unnecessary constraintson resource management. In return, organizations andmanagers are more directly accountable for results.The degree to which this approach has been adoptedamong OECD countries is indicated in Box 2. Aus-tralia, New Zealand, and the Nordic countries have ledreforms in this direction, and Canada and the UnitedKingdom reflect the same approach. There are still anumber of OECD countries that resist giving increasedpowers to public servants on the suspicion that over-spending will be the result. Thus, in the United States,Congress continues to exert firm control over spend-ing by the executive branch,5 and in some countriessuch as France, while controls on operating expendi-

©International Monetary Fund. Not for Redistribution

Page 15: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

the need to take parallel steps to increase the certaintyof budget funding. Forward medium-term expenditureframeworks (MTEFs) represent an important step inthis direction. Medium-term budget planning has beena major development in OECD countries, with fiscaltargets set on a three- to five-year basis rather than ona traditional one-year basis. Different approaches areevident. Australia, as a pioneer in this area, has perhapscarried the approach further than most in developing aforward estimate system in which refined forward es-timates become estimates of future spending ratherthan simply forecasts of existing policies. In this way,there has been a transition to medium-term budgeting.A comprehensive medium-term budgeting systemcomplements the increased flexibility given to agen-cies by allowing carryover of funds between years.6

However, in some countries budget planning occursonly at the aggregate—rather than at the agency orministerial level— where plans are just that and are notused to give indicative allocations or limits to individ-ual agencies and programs.7

Track Three: A System of Rewards and Penalties

The goal is to put increased pressure on agencies topursue improvement in program results. Within theOECD, there has been an increasing reorientation ofcentral budget offices away from compliance issues,based on detailed control over inputs, toward moreperformance-oriented managerial issues, focusing onoutputs. In part this reflects greater concern for improv-ing resource allocation than for meeting fiscal stabi-lization objectives, but it also reflects a recognition thatimproved program performance is as important formacro control as detailed controls over inputs. Hence,in recent years, there have been increased attempts tointegrate budgeting with other management processes;require agencies to measure performance and evaluatethe results of their operations; develop new guidelinesand methods for holding managers accountable; anddevelop the information bases and reporting systemsthat can enforce this accountability. As a whole, thisapproach is often termed performance budgeting, orbudgeting for results.

Some Anglo-Saxon countries have attempted to em-ploy formal contractual arrangements to ensure per-formance (e.g., New Zealand’s purchase agreements

6

for outputs between ministers and agencies, or theUnited Kingdom’s annual performance agreementsbetween a minister and an executive agency).8 This isan attempt to connect budget provision directly to per-formance. Recognizing the many methodological andpractical problems of performance measurement, othercountries (Canada, Denmark, Finland, Sweden, andthe United States) have introduced initiatives to pro-mote improved performance without tying this directlyto budgeting. These have taken the form of developingperformance indicators, formalizing requirements forprogram evaluation, and enhancing the role of externalreview agencies such as supreme audit institutions. Ithas been recognized that there are limitations to theability of performance indicators to capture all relevantfactors that dictate performance. For example, it is eas-ier to devise and identify outputs rather than outcomes,which are more relevant to performance. Nor is it al-ways clear what should be the budget implications ofpoor performance—how to achieve the linkage be-tween performance and budgeting remains largelyunresolved.

An Integrated Strategy

What does appear clear is that each of the threetracks is essential and should be followed to ensure theoverall success of the reform process. In particular, in-creasing flexibility for spending agencies without pres-sure to improve performance could increase ratherthan reduce inefficient use of budget funds. Similarly,provision of a more certain operating environmentwithout increased pressure on agencies to deliver re-sults could reduce aggregate fiscal control withoutgenerating improved program outcomes. It is there-fore considered essential to proceed down these threetracks in a highly coordinated manner.

The aim of this three-track strategy is to allowagencies to use scarce budget funds more flexibly andeffectively, while also putting pressure on them to doso. This encourages a reallocation of funds from lesseffective to more effective uses within each agency.It also has advantages for the central agencies, enablingstronger aggregate expenditure control through theability to identify and more easily withdraw fundsfrom unproductive activities currently hidden in agen-cies’ baseline budgets. But perhaps most important,these steps are also directed at setting up a clear ac-countability framework for budget managers that isthe basis of the new performance budgeting model.Agency heads are given the degree of managerialautonomy they need to achieve the tasks assigned tothem, and in return they are required to meet certain

6This circumvents the perverse incentive of “use it or lose it” thatis inherent in annual budgeting.

7On a somewhat parallel line, a few countries (Italy, Norway, andthe United States) have published intergenerational accounts usingthe methodology developed by Auerbach, Kotlikoff, and Leibfritz(1999). This approach relies on a number of rather difficult as-sumptions, and there appears to be considerable reservation aboutthe clarity of the message that is conveyed. 8See OECD (1994, 1995); Davis, Sullivan, and Yeatman (1997).

©International Monetary Fund. Not for Redistribution

Page 16: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Steps toward Reform

7

conditions. This accountability framework is based onsome key elements: a clear ex ante specification of theperformance expected of each agency head; agreedarrangements for collecting all the information re-quired to assess performance; incentives and sanctionsto encourage agency heads to act in the government’sinterests; and a clear performance assessment processinvolving ex post reporting of actual performanceagainst the initial specification.

Operationalizing this accountability framework, inturn, requires putting in place a reformulated PEMsystem. First, this requires some restructuring of thebudget to clearly define programs with identifiable ob-jectives that are set in a strategic framework. Second,it requires resolving the many technical problems re-lated to measuring these objectives and hence agree-ing on indicators of performance. Third, mechanismsneed to be put in place for creating service contracts;procedures need to be established for monitoring per-formance against contracted obligations; and a systemof penalties and rewards must be agreed. Many of

these elements are to be found in recent Russian re-form initiatives (Box 3).

The Steps toward Reform

How then to move traditional budget systems to-ward these new objectives, along the three reformtracks? What are the options for introducing thesechanges as part of a coherent program of budget sys-tem reform? To answer these questions, what followsis an outline of the approach followed by some emerg-ing economies, which will be explored in greaterdepth in this study.

Flexibility in the Agency OperatingEnvironment (Track One)

Moving away from detailed line-item budgets is noteasy to engineer. Line-item specificity exists for a num-ber of good reasons. It reduces the scope for spending

Box 3. Russian Budget System Reforms

Reforms in budget procedures were introduced with theBudget Process Reform Concept Paper for 2004–2006,whose main principles were agreed in April 2004. The de-clared objective of the reforms is to move from a system ofbudget management focusing on costs to a system of man-agement by results.

There are five main elements of this reform:

• Reform of the budget classification and accounting.

This involves a move away from excessive detail in bud-get nomenclature, and toward the approval of main itemscontained in the Budget Code. This is consistent with a sin-gle chart of accounts (CoA) for government institutions,closer to internationally approved standards, that will allowa phased transition to accrual accounting. The main changesproposed to the budget classification affect the functionalclassification and are intended to reduce the current 27 cat-egories to 11.

• A clear differentiation between existing and new policycommitments when preparing the budget.

Recognizing that 90–95 percent of commitments are al-ready “locked in” and that new programs only occupy5–10 percent of the budget, a differentiated treatment is pro-posed for approval of a baseline budget and approval of newpolicies, which will further lock in the budget in the future.

• Introduction of medium-term budget planning as part ofthe 2006 budget covering the period 2006–08.

A move away from the annual approach to budgeting tothe approval of an indicative rolling budget framework forthree years, broken down by government department.

• A streamlining of the budget process.

Most notably, the number of readings for the budget isreduced from four to three. The first reading will only dis-cuss the broad parameters and the committed “old policy”budget; the second reading will discuss new policy; and thethird reading will be a general review of the entire budgetwith detailed annexes for each main budget institution. Theexisting practice of approving the detailed budget classi-fication under a separate federal law is replaced by a lawfixing only the main codes of the classifications that aremandatory for all budget levels.

• Introduction of program and performance budgetingmethods.

Introduced as a two-year “experiment,” budget institu-tions were invited to prepare and execute their budgets ona results-oriented basis for the 2005 and 2006 budgets.Specifically, there is a requirement for budget institutionsto report on results and to develop mechanisms for moni-toring efficiency in resource use. This implies that federal“earmarked programs” for large investment, research, andstructural projects will be restructured to reflect efficiencyobjectives (and their number will be reduced). Departmentswill also introduce “targeted programs,” smaller in scaleand focused on extension of services or projects, along withagreed procedures that emphasize objectives, indicators forthese objectives, and measures of results. Budget institu-tions will be required to submit annual reports on the resultsof their main activities, with an assessment of the efficiencyof spending by previously determined indicators and withagreed powers and responsibilities for different manage-ment units for each activity.

©International Monetary Fund. Not for Redistribution

Page 17: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

units to make unproductive reallocations, either inpursuit of bureaucratic interests or under the influ-ence of personal inducements.9 Detailed controls alsoreduce the chance that appropriations are inadver-tently overspent in the aggregate. Finally, a high levelof specificity enables central agencies to identify un-derspending at the end of the budget year at the de-tailed operational level and to reduce funding or takeremedial action.

Given the case for detailed budgets, specificityshould be reduced only where spending agencies havethe internal financial and management skills to takeresponsibility for control in these areas. The devolu-tion of financial flexibility should be viewed as an im-portant carrot in persuading agencies to developimproved internal resource allocation procedures andcontrols.

Further freedom in transferring funds within the ex-isting line-item structure, or moving that structure to-ward broader expenditure category bands, should notbe granted until a review has been made of the qualityof internal financial control, coordination, and plan-ning mechanisms in spending agencies. This shouldinvolve a case-by-case examination of the capabilitieswithin the accounting, budgeting, and planning unitsof spending agencies and the interaction among theseunits.

The granting of further flexibility in budget alloca-tion should be made conditional on clearly definedcapacity-strengthening activities in spending depart-ments which will benefit from such flexibility (thecarrot and stick principle). Further flexibility shouldbe granted on an agency-by-agency basis, as specifiedlevels of internal control standards are met by eachagency.

Providing a More Certain Environment forSpending Agencies (Track Two)

There is often an initial reluctance in middle-income countries to commit publicly to medium-termfiscal plans, because of the danger that they may beregarded as a commitment of future resource avail-ability and use. One cannot dispute that this is a risk,but it is one that other countries have faced and largelyovercome. The important point is that if managers areto be empowered to manage effectively, they cannotbe constrained to a one-year time horizon. It is usu-ally agreed that the typical rush to spend before theend of the fiscal year and the resultant waste of re-sources, which are features of traditional budgets, areto be avoided. Moreover, one advantage of investingtime to prepare more detailed forward estimates is

8

that the government may be better able to plan or ar-ticulate in advance the details of how it will achieveits fiscal targets—improving both credibility and fis-cal management.

Even from the central management viewpoint, thereappears to be a need for a medium-term planningframework. Typically, emerging economies have hadto undertake extensive investment programs aimed atcompeting globally by upgrading their institutions,human capital, and basic infrastructure. Such invest-ment produces substantial debt service burdens and fu-ture recurrent spending, at the same time that thesecountries typically must also resolve the legacy of manymedium-term budget commitments arising from enti-tlements and other laws. Preparing and operating anMTBF appears beneficial simply because it spurs aprocess of assessing the medium-term macroeconomicconsequences of such commitments, checking theconsistency of the resource envelopes, and meaning-fully discussing future priorities.

While the introduction of an MTBF is recommended,this falls far short of preparing forward estimates, be-cause there is no implied legal commitment of re-sources and because an MTBF is produced primarilyfor information purposes. The MTBF should be a co-herent and sustainable forward plan for spending twoto three years ahead, based on present budget policiesand expected new policies. It should be revised on arolling basis and should form the basis of public dis-cussions on future budget policies (although it couldbe only a spending plan, conceivably with no revenueestimates). This framework could then be used in anumber of ways:

• In developing the forward estimates of expendi-ture, the first stage should be on the forward esti-mates of existing policy commitments, and only,say, for the third year should an attempt be madeto program in new policy commitments.

• The scope for proactively using the MTBF for cut-ting into baseline allocations should be activelypursued. This could involve the application of anannual “efficiency dividend” to personnel servicesand maintenance and operating expenditures,10 orthe assignment of selective savings targets to indi-vidual departments through a reduction in specificcomponents of their forward estimates. The result-ing pool of savings could be assigned to fund newpolicy measures or applied to reducing the deficit,as the need arises.

• A high priority should be given to ensuring thatthe cost analysis underpinning future outlays pro-grammed into the MTBF be sufficiently rigorousto be of “budget quality” (assisting the transfer

9With broader allocations, there is more scope for underspendingin one area to be dissipated through spending sprees in other areas. 10See Box 50.

©International Monetary Fund. Not for Redistribution

Page 18: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Three Preconditions for Reform

9

between the MTBF and annual budget preparationprocesses).

As a long-term goal, there should perhaps be theaim of automatically loading the forward estimatesof expenditure into the starting point of successiveannual budget preparations. This will require action torestrict the existing forward estimates to previouslyapproved expenditure measures. It will also requirethat budget deliberations be confined to changes inthe forward estimates (rather than the reconsiderationof expenditures already contained in the estimates).Estimates of the out-year costs of approved mea-sures will need to be firmed up to budget quality stan-dards by excluding “gambit claims” or overgenerousresourcing.

Increasing the Pressure on Agencies toPerform Better (Track Three)

Along with providing budget managers with a moreflexible operating environment and a more stable re-source base, there is a concurrent need to pressuremanagers to take advantage of this new environmentin a positive way—that is, to improve their perfor-mance. The concern is to ensure that agencies actuallyuse their increasing management discretion to improveprogram outcomes rather than either to maintain thestatus quo or to pursue goals of a more bureaucratic orpersonal nature.

At the most fundamental level, putting pressure onagencies to perform better under the new arrangementsrequires a high degree of transparency in their opera-tions. This raises questions about the adequacy of gov-ernment financial information systems, an issue mademore pressing by any move toward decentralizingspending authority. A beginning step could be to re-quire agencies to report on their achievements in termsof performance indicators against budget targets. How-ever, such reports are unlikely to have much impact ifthere is no enforcement mechanism and if resourcesare neither increased nor withdrawn on the basis ofperformance. As indicated, connecting performance tobudgeting is not easy, and a dual approach is perhapsmore effective: applying pressure from above as wellas increasing capacity from below. This implies that theMoF would encourage a reevaluation of agency activ-ities and put pressure on managers to economize theirresource use and, at the same time, boost the capacityof agencies to manage their resources more flexiblyand efficiently.

Three Preconditions for Reform

Adopting this three-track approach requires threepreconditions: first, an adequate level of PEM capac-ity; second, reorganizing government institutions to

work in this new environment; and third, providing afacilitating external environment.

Some Basic PEM Thresholds

Budget system reform requires that the existingPEM system have some basic threshold capacities. Itis possible to interpret recent Chinese budget reformsas geared to reaching these basic thresholds (Box 4).Some of these requirements can be identified.

Restructuring the Budget and Its Classifications

Traditional budgets are usually characterized asline-item budgets—budgets that focus on cost itemsor inputs such as wages and salaries, other goods andservices, and spending on capital items (see Box 1).Such budgets are typically subdivided on the basis ofdepartments or activity centers, each of which mayoperate several programs. Budgets are constructed(usually on an annual basis) by comparing the pre-vious year’s expenditures and adding increments (orsubtracting decrements). There tends to be a bias infavor of existing programs, and new programs mustcompete with each other for limited resources. Budgetcontrol is primarily exercised on inputs, rather thanoutputs. Recurrent and capital expenditures often arebudgeted separately. The focus is on compliance withthe details set out in appropriations. To enforce suchcompliance, the budget classifications also reflect de-tailed inputs and the corresponding departments andcost centers that utilize them.

Restructuring the budget involves a shift in empha-sis away from inputs and toward the outputs associ-ated with these inputs. This approach, broadly termedprogram budgeting, attempts to identify amounts al-located to individual services and programs. It is usedto counter the irrationality of using the previous year’sbudget as the basis for the current budget and for onlyarguing for or against increments (decrements). Itcounters splitting the budget into recurrent and capi-tal spending by recognizing, first, that both types ofspending can contribute to the same program, and sec-ond, that capital expenditures generate future recur-rent costs. Budget control is thus reoriented, awayfrom the narrow perspective of limiting expendituresto those amounts and items included in the annual ap-propriations and toward ensuring a more efficient andeffective use of resources over the medium term. Tosupport the latter objective, it is usually necessary toadd a meaningful program classification system to thebudget. Thus one important precondition for movingto the next stage of reform is a suitable budget struc-ture with a supporting program classification. It is alsorecommended that other classifications be compatiblewith international government finance statistics (GFS)standards.

©International Monetary Fund. Not for Redistribution

Page 19: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

10

and reports; and effectiveness and efficiency of oper-ations. In the first two areas, problems are typicallyclustered around control over the payroll and effectiveprocurement procedures, which usually account for alarge part of an agency’s expenditures and presentspecial problems of internal control. In strengtheningand ensuring quality control over internal financialmanagement, the emphasis is usually on collecting in-formation needed to support internal control and onensuring its relevance, timeliness, and objectivity. In-ternal audit has also been recognized as making a sub-stantial contribution to quality control in internal finan-cial management.

Improving the efficiency and effectiveness of agen-cies’ operations usually involves a change in mind-setand takes more time. Under a traditional type of bud-get system, there is little emphasis on performance.

Box 4. Chinese Budget Reforms: Strengthening Basic PEM Systems

• Review the legal and regulatory framework for thebudget.

The 1994 organic budget law, recognized as having weak-nesses, is under review. While imposing identical budgetprocedures at all five levels of government, the proceduresare hierarchical across levels, causing considerable delaysin approving lower-level budgets. Current and develop-ment expenditures are split, with oversight under differentagencies and with different rules for their preparation. Thereis no medium-term budget framework and no performanceinformation.

• Improve budget preparation and classification systems.

Budget remains a small summary document. Recently,this increased in size and was supplemented with detailed de-partmental budgets, including the expenditure of “self-raisedfunds.” The MoF has enlarged and improved its projectsdatabase, with a focus on new projects. Classification doesnot specify costs to a sub-item level and is insufficient for de-tailed cost analysis and accounting control. Recent reformsinvolve deepening economic objects classification, with fi-nancing items moved below the line. Although still someway from international standards, there is improvement.

• Strengthen the technical basis of budget preparation, in-troducing a performance orientation.

There is a push for greater transparency. with depart-ments splitting their budgets between “basic” expendituresand “projects.” Budgets are compiled using manpower andother input-based norms, with project evaluations includingexpected outputs. Emphasis is on planned activities ratherthan last year’s budget, with input norms developed for clus-ters of similar operations. The approach is termed “zero-based.” A spur to improved budgeting is increasing the

11This is discussed more fully in Diamond (2002b).

Strengthening Internal Financial Management

Before attempting to give agencies wider respon-sibilities in resource allocation, it is essential to en-sure they are operating within an effective financialmanagement framework. Good internal control is animportant feature of this framework.11 Without satis-factory controls, management may not detect seriouserrors and irregularities, and both the work of the cen-tral oversight agencies and any external audits be-come more difficult. An internal control frameworkshould be designed to provide reasonable assurancesthat three basic objectives are achieved: compliancewith laws and regulations; reliability of financial data

importance of “preliminary scrutiny” of the Finance and Eco-nomics Committee and its subcommittee. The MoF now pre-pares annual three-year fiscal rolling outlooks, but these arenot yet integrated into the budget preparation.

• Establish a treasury system, based on a Treasury SingleAccount (TSA).

A new Treasury Department was created in 2000, replac-ing the People’s Bank of China’s (PBC’s) treasury systemwhich was based on a multitude of bank accounts and wasviewed as a source of corruption. A parallel PBC system con-tinues, with the budget executed using credit limits at spend-ing units’ bank accounts that are reimbursed through the PBCbranch network (“quota system”). The Treasury Departmentalso developed its own direct payments system through aTSA at the PBC. The latter has caused elimination of thou-sands of bank accounts and a reduction of government float.Capital construction expenditures are executed centrallythrough the Accounting Division of the MoF Budget De-partment using an account at the China Construction Bankthat transfers funds to the accounts of spending units that paysuppliers. Along with improved central cash management,the emphasis has been on improved agency financial man-agement, focusing on internal controls and procurement.

• Introduce a government financial management informationsystem (GFMIS).

There is no integrated information system across the cen-tral government. In parallel with the increasing use of theTSA, the MoF has developed specifications for a countrywideGFMIS and by mid-2003 had introduced a pilot system basedon a mixture of five systems developed in-house and three Or-acle Financial™ packages (general ledger; payment account-ing; and cash management, which is used for reconciliation ofpayment orders). The authorities are working on an appropri-ate information technology strategy as the basis for a more ad-vanced GFMIS in the future.Source: Brumby (2005).

©International Monetary Fund. Not for Redistribution

Page 20: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Three Preconditions for Reform

11

With the budget focused on inputs and with detailedline-item controls on expenditures, agencies focus onobtaining spending approvals and utilizing them withinthe year. There is little emphasis on nonfinancial per-formance and few sanctions for poor performance (aswell as few rewards for good performance). Managersin line agencies, therefore, are not used to subjectingnew policy proposals to critical analysis, nor to man-aging expenditures to achieve a budgeted outcome be-cause these functions are tightly controlled by thecenter. There is a need, therefore, to develop suchevaluation skills among agency managers and to es-tablish a supporting information system based on in-dicators of output and performance.

Developing an FMIS

With the ever-growing volume and complexity ofgovernment financial operations, it is clear that fiscalmanagers have a critical need for timely managementinformation reports, in a usable form. Producing suchreports requires improved classification systems toidentify programs so that costs can be allocated to in-dividual activities and compared with the outputs ofthese activities. Similarly, moving to an accrual-basedaccounting system has obvious benefits to resourcemanagers by providing a more complete picture of anagency’s financial position. Tailoring such reports tomanagement needs through a computerized account-ing system has been a general PEM reform undertakenby a number of countries in various parts of the world.

The need for an integrated and computerized FMISbecomes more urgent as the transition is made fromcentralized controls over spending to a devolved fiscalmanagement regime. If agencies have relative free-dom in determining the inputs required to achieve theoutputs expected from them, then managers must beable to track inputs used and to develop output indi-cators. Furthermore, as budget horizons extend underan MTBF, the information system will have to supportagencies’ needs to align their resource allocation strat-egies with the performance targets over a multiyearexpenditure plan. These are ambitious reforms formany emerging economies but reflect current practicein most OECD countries. For these reforms to suc-ceed, a crucial prerequisite is a well-conceived FMISthat provides access to a shared information databaseby all participating financial managers.

Establishing a Cost Accounting Capacity

Apart from the financial aspects of a budget institu-tion’s operations, additional information will be re-quired from an FMIS in a performance-based budgetsystem. Greater emphasis needs to be paid to programsand activities approved by the legislature, and resultsfrom the use of resources allocated to these activities

must be described in physical and not just financialterms. This has resulted in an increased emphasis onthe costing methodology so that the full costs of pro-grams and activities can be captured and meaningfullyrelated to their outputs and outcomes in order to derivemeasures of program performance. Unfortunately, suchskills are often scarce in government and must be de-veloped. At the same time, there are limits on the ex-tent to which costing systems, however sophisticated,can provide meaningful information unless backedwith an adequate accounting system that captures andrecords the total costs of operations. Government ac-counting systems, typically cash based, have often beenconsidered deficient from this viewpoint.

Improving the Accounting System

The typical government accounting system, even ifconceptually well defined and internally consistent,has the drawback of being cash based. The traditionalemphasis on voted costs, the last stage of spending,may be adequate for the compliance and stabilizationobjectives of the MoF at the center, but they are lessrelevant for budget managers in individual govern-ment departments. A precondition for moving to thenext stage of budget system development is to upgradethe accounting system.

Traditionally, a government accounting system isrequired to serve two distinct objectives. First, the sys-tem must compile a formal/legal annual financial re-port of the government’s position in the form andcontent prescribed by statute and amenable to externalaudit. Second, the accounting records of primary finan-cial transactions must form a part of a comprehensiveand fully classified database capable of providing con-temporary, accurate, and appropriate management re-ports required by financial managers in the governmentdepartments, including central MoF departments, tomonitor and control government fiscal operations.Recognizing these two basic objectives, upgrading agovernment accounting system has typically followedtwo paths.

The first upgrade is typically to reformulate the ac-counting system to be able to capture all stages of theexpenditure process in the accounts, namely: (i) ap-propriation; (ii) obligation incurred; (iii) verificationof supply of goods/services; (iv) obligation due forliquidation; and (v) obligation liquidated (i.e., cashpayment). A pure cash-based system only recognizesthe final stage (v) with controls placed on ensuringthat cash payments do not exceed the levels indicatedin the first stage (appropriation). The problem of ar-rears, where the government has unmet obligationsdue for payment (stage iv), which has arisen whencontrols are placed solely on cash spending, has causedgovernments to focus on prior stages of spending, andin particular to realize there can be no true expendi-

©International Monetary Fund. Not for Redistribution

Page 21: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

ture control without controlling commitments made(stage ii). Thus typical amendments introduced topure cash reporting include the provision of infor-mation on payment arrears and various noncashtransactions to provide a better indication of the totalresource use by the government and its future cashrequirements.

The second major reform of the accounting systemrecognizes that tracking the different stages of spend-ing is not enough. For a comprehensive approach tobudget management, which stresses performance, thereis also a need to account fully for government assetsand liabilities, to include all costs of providing services,and so to move government accounting closer to pri-vate sector accounting, with similar types of financialstatements. Australia, Iceland, and New Zealand havealready adopted accrual-based accounting for theirwhole-of-government financial statements and bud-gets, and others, such as the United Kingdom, havemoved substantially in this direction. Some, such asCanada, Finland, Sweden, and the United States, haveadopted accrual accounting principles for whole-of-government financial reporting, while others are adopt-ing accrual accounting standards on a pilot basis on alimited scale—for example, Germany, Ireland, andthe Netherlands. Still other countries, including someemerging economies, already prepare limited balancesheets, but these typically leave out a large componentof the government’s assets in common infrastructureand “heritage assets.”

In the next stage of reform, but over time, these coun-tries should move to a full accrual system, where all as-sets are included, progressively, in the balance sheet,and their valuation is shifted from an indexed historicalcost to a replacement basis. The migration from cashbasis to accrual basis through modified accruals will bediscussed in some detail in Section VI, particularly,how this makes it possible to meet the reporting stan-dards of the 2001 Government Finance StatisticsManual, or GFSM 2001 (IMF, 2001a), which formallyadopts an accrual-based recording to supplement, ratherthan substitute for, the cash-based system used in mostcountries. Despite the importance of the accrual systemin terms of properly stating the magnitude and timing offinancial operations, it should be acknowledged that it isdifficult to value certain government assets and thus de-termine net worth, as well as that it is impracticable formany countries to implement this change.

Institutional Framework for Devolved Management

There are two defining characteristics of performancebudgeting: the greater emphasis on performance datain making resource decisions, and the more decentral-ized form of budget management. Adopting this new

12

management orientation creates parallel pressures forreorganizing government operations, resulting in anumber of innovative institutional arrangements. Forexpository convenience, in this study institutionalarrangements are ordered by degree of decentralization,here characterized as the five “Ds”—deconcentration,decentralization, delegation, devolution, and divest-ment. Each of the Ds involves a progressive shift of de-cision making away from the centralized traditionalmodel, and each in turn has involved different institu-tional arrangements for managing public resources.The latter three Ds are perhaps the most characteristicof the new performance budgeting approach.

Two aspects of these new arrangements are high-lighted in this study: the role of contracting withingovernment to establish accountability for man-agers at arm’s length and the choice and design ofdecentralized institutions, with special focus on therole of the “agency.” Both these institutional inno-vations have their problems.

Problems with making the contract approach ef-fective often stem from difficulties in defining per-formance and devising operational measures of thatperformance. Even if these problems can be resolved,additional constraints often include the increased levelof management expertise required to make the contractapproach effective and the limited availability of ade-quate management tools. Similarly, the effectivenessof devolved agencies cannot be divorced from the ex-isting level of PEM expertise. The case for the agencymodel is enticing: it separates policymaking tasks fromimplementation, better defining responsibilities andleading to better service delivery. As always, the prac-tice has been less simple than the theory, with concernsover establishing adequate accountability. The needfor “institutional clarity” on the exact status of thesedevolved institutions has often been neglected, even inOECD countries. That experience suggests the need todefine very carefully the degree of autonomy of theseinstitutions, as well as to clarify their accountabilityarrangements. It is important not only to buttress thiswith strengthened PEM procedures, especially withregard to their reporting requirements and the skills re-quired of managers, but also to ensure that the level ofgovernance in the country is sufficient to support thislevel of devolved decision making.

Countries have also experimented with various op-tions for divestment, using models of devolved man-agement that go closer to, but stop short of, outrightprivatization. These arrangements all aim for a higherlevel of commercialization in the provision of publicservices by giving customers a greater purchaser role.Two of the most popular approaches are competitivetendering and the use of concessions or private-publicpartnerships (PPPs), both of which also present chal-lenges from a PEM viewpoint. While efficiency gainsare generally associated with contracting in or con-

©International Monetary Fund. Not for Redistribution

Page 22: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Wider Framework for Budget Reform

13

tracting out, these often must be counterbalanced byhigh transaction costs, problems of quality control, andthe need to ensure that the benefits to society are ade-quate and sustainable. Concessions and PPPs gener-ally raise the level of PEM requirements even higherthan contracting. This arises from the problems of de-fining outputs, issues related to concession design, andthe need to set the form, level, and structure of pay-ments, as well as to create the mechanisms to regu-late the concession. Needless to say, the successfuloperation of these arrangements is heavily dependenton the government having the necessary specializedskills for their negotiation and supervision.

While the pace of such reforms has varied by coun-try, OECD countries generally have adopted a phasedapproach in introducing these more devolved forms ofresource management. This is partly in recognition thatinstitutions should be progressively prepared for theshift to performance budgeting. This preparation gen-erally entails significant upgrades to management andcontrol systems and changes in the internal culture ofgovernment institutions, which inevitably take time.Often reforms cannot proceed until there is a changein an organization’s legal status, which again takestime. Of course, the longer the time horizon, thehigher the chance that reforms are derailed and thegreater the importance of effective steering and man-aging of the reform process.

Managing the Change Process

The three tracks of the reform strategy should stayin phase to ensure that the devolution of financial free-doms does not run ahead of the ability of agencies toput these new freedoms to good effect. Typically, theemphasis is on the top-down management of the per-formance improvement process. However, experienceshows that there is a parallel need for refocusing ef-forts on “bottom-up” capacity building in line agen-cies. Unfortunately, the task of building this capacityis often well beyond the existing technical resources oftraditional budget offices. External assistance fromother agencies is likely to be necessary if the coherenceof the reform strategy is to be preserved.

The wide-ranging nature of budget system reform,which involves not just central budget units but allgovernment spending agencies, implies a significantresource cost not only in enhancing capacity but alsoin monitoring the responsiveness of departments andministries to the enhanced flexibility to ensure that thereforms are sustained. The cost in terms of scarce man-agement skills is perhaps the most significant foremerging economies. Even within OECD countries,reform programs have had to be engineered—a reformplan formulated, an implementation strategy agreed,and the implementation process successfully man-

aged—to achieve the objectives and sustain the reforminitiative. This raises the questions of whether there issufficient capacity to engage on all three tracks simul-taneously and whether there exist adequate change-management skills within government.12 In manyemerging economies, a major reservation about the ef-ficacy of the budget reform process often relates lessto the overall approach to reform, or to the nature ofthe specific changes proposed, than to doubts con-cerning a viable “execution strategy.” To counterthese doubts, it has often been considered important tonest budget system reforms within wider governancereforms stressing transparency and reinforcing thelegal framework of budget management—as indeedhas been the approach of pioneering OECD countries.

The Wider Framework for Budget Reform

Increased Emphasis on Fiscal Transparency

Transparency in government operations is increas-ingly regarded as an important precondition for goodgovernance and sustainable economic growth; how-ever, for the emerging economies, it is also an essen-tial aspect of sustaining confidence in governmentand, through this, sustaining support for the demo-cratic system and economic advancement. The latterhas become particularly acute with the exposure of fis-cal policies to international financial markets arisingfrom the increasing use of market financing of deficits.Such exposure has required many governments tomodify past policies and so has assisted in achievingfiscal discipline and improved resource allocation and,in turn, has supported reform of the budget system.13

One source of fiscal discipline is more transparentfiscal targeting. Box 5 indicates a wide variety of fis-cal targets that have been adopted.14 This is already afact of life for most OECD countries. Their currentbudgetary goals are typically expressed as specificquantitative targets rather than general statements ofintent. The benefits derived from providing an unam-biguous message about the government’s financial

12This aspect of the reform process, the exercise in change man-agement, is dealt with in Diamond (2001).

13“Global commercial liberalization and the free flow of capitalare exerting new pressures on systems of public governance. Re-cent experience shows starkly that the quality of public institutionsand the trust in which they are held by economic players can havevery demonstrable effects on the behavior of these markets. Publicsector governance systems that induce loss of market trust imposecosts not only directly on their domestic economies, but moregenerally as they reduce global growth rates below potential.”(Brumby, 1999)

14The approach of different countries, with a discussion of theusefulness of fiscal rules, is contained in Kopits (2001).

©International Monetary Fund. Not for Redistribution

Page 23: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

objectives to the markets is felt to outweigh the curtail-ment of the government’s ability to subsequently ad-just its strategy for expediency, especially when thetargets are specified for several years ahead.15 Mosttargeting has focused on the budget balance, withsecondary emphasis on stabilization or reducing thelevel of debt, and less often on reducing outlays or thelevel of taxation. The convergence requirements ofthe 1993 Maastricht Treaty on European Economicand Monetary Union (deficit not to exceed 3 percentof GDP and gross debt not to exceed 60 percent ofGDP) are fiscal targets that drive much of the fiscalstrategy of EU members. In a few countries, fiscaltargeting is focused on balanced budget requirements(for example, the United States and Switzerland),while many governments have followed the “golden-rule” approach—no borrowing except for capital pur-poses, supplemented by a limit on this borrowing ordeficit as a percentage of GDP. New Zealand also tar-

14

gets net worth (as well as the budget balance and ex-penditure, taxation, and debt as a percentage of GDP).

Fiscal transparency is difficult to enforce if a largeproportion of government transactions takes place“off budget.” Improved resource allocation has beenassisted by addressing a number of issues to increasethe coverage of the budget. Many countries havelarge extrabudgetary funds set up with earmarked taxrevenues, which escape budget discipline. For exam-ple, a large number of countries have public pensionschemes, partly financed by employer contributions (apayroll tax) and employee contributions, which oper-ate outside the budget. Many of these schemes give adistorted position of the government’s true position iftheir balances are measured on a “pay as you go” cashpayments basis and not a full accrual (actuarial) basis.There is a similar need to identify and include quasi-fiscal transactions. A wide range of government ac-tivities often are not captured in government spendingbut rather are channeled through government financialinstitutions (e.g., concessional lending), central bankprofits (or losses) on foreign exchange currency trans-actions, government bailouts of insolvent financial in-stitutions, and “social expenditures” of the state-ownedenterprises (e.g., employing more staff than commer-cially justified). The fiscal impact of public enterpriseson the budget is another area that often requires greatertransparency.

In many countries, public enterprises in such“strategic” areas as defense, transport, and govern-ment finance impose significant costs on the govern-ment, with neither the total costs nor the individualfinancial position of such undertakings clearly re-ported. Taxation concessions (or tax expenditures)should be regarded as a form of expenditure, but theseoften escape scrutiny, as do the resource costs of reg-ulating the private sector, which are often a non-transparent alternative to public expenditure programs.Usually, there is no estimate of regulatory costs im-posed on the private sector nor are estimates madewhen deciding whether to use such an instrument ratherthan a more transparent, direct government program.The pursuit of greater fiscal transparency, therefore, inaddressing these problems has provided an importantimpetus to reform budget systems and is an importantchallenge for emerging economies seeking to institutesimilar reforms. Box 6 summarizes good practices inthis area.

Is There Need for a New Legal Framework?

A number of countries’ budget system reforms arelegally based, often in the form of fiscal responsibilitylaws, to force governments to commit credibly and toassuage governance concerns. The basic idea is to cre-

Box 5. Major Types of Fiscal Policy Rules

Balanced budget or deficit rules

• Balance between government revenue and expenditure(i.e., prohibition on government borrowing), or limiton government deficit as a proportion of GDP

• Balance between structural (or cyclically adjusted)revenue and expenditure, or limit on structural (orcyclically adjusted) deficit as a proportion of GDP

• Balance between current revenue and current expendi-ture (i.e., borrowing permitted only to finance capitalexpenditure)

Borrowing rules

• Prohibition on government borrowing from domesticsources

• Prohibition on government borrowing from centralbank, or limit on such borrowing as a proportion ofpast government revenue or expenditure

Debt or reserve rules

• Limit on stock of gross (or net) government liabilitiesas a proportion of GDP

• Target stock of reserves of extrabudgetary contin-gency funds (e.g., social security funds) as a propor-tion of annual benefit payments

Source: Kopits and Symansky (1998).

15Transparency is increasingly regarded as important in improv-ing the quality of forecasts. Economic forecasting is an area wherethere is now general acceptance of such transparency. The eco-nomic forecasting models used in Australia, Sweden, and theUnited Kingdom are public. Independent forecasting panels whichmeet in public are now a common institutional arrangement in Ger-many, Sweden, the United Kingdom, and the United States, en-couraging contestability in forecasts.

©International Monetary Fund. Not for Redistribution

Page 24: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Is There Need for a New Legal Framework ?

15

and fiscal reporting. The need for greater transparencyin fiscal management was an overarching objectiveof this legal framework. Not only are governmentalbodies required to specify and adopt agreed fiscal tar-gets, but they are required to explain the policiesunderlying their pursuit and, when deviating from thispolicy commitment, are required to explain not onlythe rationale but also planned future corrective action.Much emphasis has been placed on best practice pro-cedures and much less on numerical objectives. TheAustralian Charter of Budget Honesty and the UnitedKingdom’s Code for Fiscal Stability have also adoptedthis approach, albeit in different forms. Recently, someemerging economies have adopted this approach, in-cluding Brazil (Box 7). The adoption of such legalframeworks is useful in anchoring budget reforms butshould not be considered a precondition for successfulreform.

A key element in these frameworks is a transpar-ent medium-term fiscal strategy. This adds credibil-ity to fiscal policy by demonstrating its sustainability.Thus, Australia issues a Fiscal Strategy Statement,New Zealand publishes an annual Fiscal Strategy Re-port, and the United States publishes an Economicand Budget Outlook. For EU countries, a number ofmedium-term reporting requirements are embodiedin the Maastricht Treaty and the Stability and GrowthPact. Since 1993, EU member states have been obligedto publish medium-term convergence programs, whichare scrutinized by the Council of Economics and

Box 6. Summary of Good Practices in Institutional Transparency

Overall structure and functions

• Clear demarcation of functions between public andprivate sectors

• Delineation of the boundaries of the operations ofstate-owned nonfinancial enterprises and financial in-stitutions from those of the general government andprovision of information on the costs of quasi-fiscalactivities performed by such enterprises and institu-tions, as well as any financial resource operationsfunded by the government

• Clear assignment of responsibilities and resourcesamong national and subnational levels of government(limiting the scope for case-by-case negotiation)

• Clear statement of the rationale and extent of extra-budgetary fund operations

• Independent review agency with wide investigativeauthority over government operations

Budget process

• Detailed public explanation of fiscal targets and prior-ities in draft budget

• Open legislative debate and authorization• Transparent execution and control (including procure-

ment, contracting, employment)• Public disclosure of performance and financial audits

Tax treatment

• Explicit statutory basis (instead of discretionary taxconcessions or negotiated tax liabilities)

• Clear administrative procedures, information require-ments, taxpayers’ rights and obligations, tax officials’code of conduct

• Estimates of tax expenditure budget

Financing operations

• Disclosure of the terms (interest rates, maturity) andsources of government deficit financing

• Specification of the policy criteria as well as terms andconditions of government lending decisions

Regulation

• Open legislative and administrative process (hearing,approval)

• Clear and simple statutes and implementation• Estimates of regulatory costs

Source: Kopits and Craig (1998).

Box 7. Elements of the Brazilian FiscalResponsibility Law, 1999

• Specifies the principles of responsible fiscal manage-ment.

• Benchmarks focus on:– Restrictions on public indebtedness– Disciplining the growth in long-term spending, espe-

cially on social security– Limits on increases in personnel expenditure– Prudential limitations on credit operations and

financial and asset management• Defines necessary fiscal procedures to meet these

principles (i.e., how restrictions, discipline, and lim-its will be imposed), consolidating existing laws andregulations.

• Requires fiscal reporting and enables transparency toensure principles are upheld.

• Requires explanation of deviations from principles, fu-ture corrective action to remedy deviations, and anypenalties to ensure accountability of governing au-thorities and managers.

• Covers all levels of government (federal, state, andmunicipal).

ate rules and procedures that impose costs on govern-ments for deviating from fiscal responsibility. The pi-oneer in this field was New Zealand, which passed theFiscal Responsibility Act in 1994. This laid down prin-ciples of responsible fiscal management as a frame-work and ensured that these principles were upheldby defining necessary fiscal procedures, fiscal targets,

©International Monetary Fund. Not for Redistribution

Page 25: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

II THE STRATEGY FOR BUDGET SYSTEM REFORM IN EMERGING MARKET ECONOMIES

Finance Ministers of the European Union (ECOFIN)and which contain medium-term projections of coun-tries’ fiscal and debt positions, their main assumptions,and the measures that are contemplated to achieve thestated objectives in the Stability and Growth Pact.

To recognize the need for this wider framework forsuccessful budget reform is to recognize that the lattershould not be regarded merely as a technical exercise.Budget system reform represents a fundamental insti-

16

tutional shift not only in techniques of public expendi-ture management—in the procedures used to plan,control, and use public resources effectively—but inthe wider context of a society’s view of the role of thestate and the citizens’ rights and expectations vis-à-visthe state. Insofar as the budget system reflects soci-ety’s view of such fundamentals, its reform cannot bepushed further or faster than society’s perceived needfor change.

©International Monetary Fund. Not for Redistribution

Page 26: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III Moving from Program to Performance Budgeting

17

Many emerging economies are trying to improvetheir budget processes and move to performance-

based budgeting modeled on OECD-type reforms. This section first presents a brief historical perspectiveof these reforms by reviewing the evolution of the“new” performance budgeting model being applied inindustrial countries. It identifies the main componentsof this budget management model as a means to deter-mine the challenges facing emerging economies whenconverting their existing budget systems to this model.This review highlights the critical importance of a goodprogram basis for budgeting, which is not easy and typ-ically requires four important reform elements to be inplace. First, any existing program structure must be setin the wider context of strategic budget planning. Sec-ond, such strategic planning should be anchored on amedium-term budget framework. Third, this usuallyimplies redesigning and refining existing programstructures. Fourth, and perhaps most difficult, a newsystem of accountability and budget incentives needsto be introduced. For emerging economies, these fourelements should be viewed as the prerequisites for asuccessful introduction of performance budgeting.

The Road from Old to NewPerformance Budgeting

Budget reforms that were introduced in industrialcountries following World War II, under the rubric ofperformance budgeting, have been introduced in manyguises and generally have endured in some form inmost countries. Unfortunately, owing to these manyvariants, the term itself has been interpreted differ-ently at different times and in different countries. Atthe broadest definitional level, the term is associatedfirst with a budget presentation that emphasizes theoutputs rather than the inputs associated with govern-ment operations16 and, second, with a restructuring ofgovernment operations on the basis of programs andactivities producing these outputs. As a consequence,

the term is often used synonymously with programbudgeting.

There is a vast body of literature on program budget-ing,17 which unfortunately also tends to be somewhatconfusing. There is no agreed form of program budget-ing, and the term program budgeting is often used in-terchangeably with related terms, such as performancebudgeting, planning-programming-budgeting systems(PPBSs), and output-based budgeting. Accordingly,to impose some order in this terminological confusion,this section attempts a stylized characterization of eachof these budgeting concepts and their specific features.As indicated in Box 8, the progression toward the newperformance budgeting follows distinct steps.

Origins of Performance Budgeting

Elements of program budgeting were in evidence inthe United States prior to World War II,18 but the termperformance budgeting is more clearly associated withthe 1950s reforms undertaken in the United States.These focused on developing performance informa-tion for budgeting and reorienting the U.S. federalbudget process from its focus on inputs to one that alsoincluded the outputs derived from using those inputs.The approach really took off when the Hoover Com-mission on the organization of the executive branch ofthe government promoted this approach and encour-aged its widespread implementation in 1949.19 TheBudget and Accounting Procedures Act of 1950 re-quired agency heads, in coordination with the Bureau

16Or as Burkhead (1956) put it, “that emphasizes the things whichgovernment does, rather than the things which government buys.Performance budgeting shifts the emphasis from the means of ac-complishment to the accomplishment itself.” (p. 133)

17For example, Wildavsky (1969) and Henke (1992).18An early definition by the United States Bureau of the Budget

quoted in Axelrod (1995): “A performance budget is one which pre-sents the purposes and objectives for which funds are required, thecosts of the programs proposed for achieving these objectives, andquantitative data measuring the accomplishments and work perfor-mance under each program” (p. 266). The Tennessee Valley Au-thority’s program budget system of the mid-1930s, while a “forprofit” agency, is a classic example of the approach, and other fed-eral agencies followed by developing activity schedules withinprogram budgets during the war years.

19The Hoover Commission also publicized the concept of a per-formance budget in recommending “that the whole budgetary con-cept of the federal government should be fashioned by the adoptionof a budget based on functions, activities, and projects; this we des-ignate a performance budget.”

©International Monetary Fund. Not for Redistribution

Page 27: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

of the Budget, to support their budget requests with in-formation on performance and program costs by orga-nizational unit. The U.S. budget for the 1950–51 fiscalyear was the first to show the effects of this changedperspective. Even from these beginnings, the HooverCommission used the terms program budgeting andperformance budgeting almost synonymously.20

In these early days, performance budgeting wasnot viewed as a fundamental change in budget deci-sion making, but rather as a useful adjunct that wasperhaps relevant for only some service-based opera-tions. These reforms were seen as being severelyhandicapped by inadequate accounting systems thatcould not identify the full costs of government oper-ations, the equally constraining problem of develop-ing measures of performance, and, most particularly,the lack of interest of decision makers in using per-formance information (Burkhead, 1956). In practice,program and performance budgeting proved time-

18

consuming and required much administrative sup-port, namely, improved accounting, costing, activity,and performance-measurement systems.21 Nor was itvery popular with appropriation committees, who weremore comfortable with detailed line-item controlsthan with providing lump-sum appropriations in re-turn for specified levels of performance. Budgetersalso often preferred that information on operating tar-gets, workload, and performance standards not be ex-plicit but remain largely hidden in detailed objects ofexpense.

The Planning-Programming-Budgeting System Period

Nevertheless, the application of program and per-formance budgeting persisted in the United States atthe federal government level. The PPBS, as a com-

Box 8. The Steps toward the “New” Performance Budgeting

Step 1. Planning, programming, and budgeting systems(PPBSs)

Program, output, and “new” performance budgetingoccur only after the first two elements of PPBS have beenachieved.

Elements:• Identify and examine goals and objectives in each major

area of government activity.• Analyze the output of a given program in terms of ob-

jectives.• Measure total program costs, not just for current year but

also for several years ahead.• Formulate multiyear expenditure programs.• Analyze alternatives to find the most efficient and effec-

tive means of attaining program objectives.• Establish these procedures as a systematic part of the bud-

get review process.

Step 2. Program budgeting

Program budgeting leaves the PPBS’s higher-level strate-gic planning functions out of the budget process. It entailsan interactive process for refining cost assignments and out-put definitions.

Elements:• Group organizational units within common functions and

subfunctions.

20“A program or performance budget should be substituted for thepresent budget, thus presenting in a document of much briefercompass the government’s expenditure requirements in terms ofservices, activities, and work projects, rather than in terms of thethings bought” (A.E. Buck in the Hoover Commission’s TaskForce Report).

21As a consequence, Schick (1971) complained of hybridization,whereby various aspects of the performance system were assimilatedinto traditional budget procedures and thereby watered down. Hewrites: “once the initial allure of performance budgeting had wornoff, it was viewed with indifference. Performance budgeting wasnot salient to the interest of budget participants, nor was it regardedas an important reform.” (p. 62)

• Identify costs of a function and subfunction.• Given these costs, decide what the unit’s output should be.

Step 3. Output budgeting

Elements:• Group together all costs of achieving a particular output,

regardless of the number of agencies involved.• Emphasize full costing, including overhead assignment.• Define outputs in terms of measurable indicators and as-

sess the quality of goods and services delivered, analo-gous to what is done in the private sector.

• Compare with actual output to gauge efficiency.

Step 4. New performance budgeting

New performance budgeting contains all elements of out-put budgeting. It also includes higher-level accountabilitywith associated rewards and sanctions, i.e., moves to a per-formance budget management system.

Elements:• Define outcomes/impacts in terms of measurable indica-

tors, and compare these with actual outcomes to gaugeeffectiveness.

• Incorporate explicit performance measures and systems ofperformance assessment.

©International Monetary Fund. Not for Redistribution

Page 28: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Road from Old to New Performace Budgeting

19

plete system of budget making, was first adopted bythe Department of Defense in 1961, and later appliedto all agencies. It subsequently spread to states andlocal governments and also to some other countries.

The PPBS process—as its name implies—had threebasic phases which linked planning with budgetingthrough programs. The planning phase sought to iden-tify present and future objectives and to evaluate dif-ferent possible ways of achieving these objectives.The programming phase then took the proposals ofthe planning phase and integrated them into programsarranged by a hierarchy of priorities that would besubject to decision making at different levels of thepolitical hierarchy. For example, the setting of broadpriorities (program categories) would be a govern-ment or cabinet-level responsibility, while the opti-mization within individual programs would be thetask of sector agencies. The third phase of budgetingwas the translation of each multiyear program into aset of specific annual actions, by determining whodoes what and then assigning required resources. Thiswas a difficult stage of the process. Since the admin-istrative structure of the budget could be differentfrom that of the program, it was difficult to apportionresources among programs. Moreover, the U.S. fed-eral budget process was strictly regulated by a com-plex set of laws that had evolved over time and whichgoverned not only the budget structure but its presen-tation, approval, and execution.

By 1971, the system was largely abandoned. Inpart, the system was a victim of its own ambition. AsAxelrod (1995) argues, “many of the problems of thePPBS in the federal government stemmed from its si-multaneous introduction across the board in agenciesthat were ill-prepared to implement it” (p. 288). Thediagnosis of what went wrong holds lessons for thefuture adoption of a full-scale budget programmingapproach.

Despite its name, the PPBS was not a single sys-tem; each agency had to develop its own version andagency directors often did not appreciate its value andwere reluctant to undertake such work. Central leader-ship of the reform process was lacking. The reluctanceto fully convert to the new system was fueled by thefact that the Bureau of the Budget continued to reviewcost and performance on the traditional basis in mak-ing its recommendations. Its adoption was furtherundermined by a lack of commitment by legislators:Congress insisted on the traditional budgetary presen-tation and on preserving the existing appropriationstructure. Additionally, there were important practicalproblems that were never fully resolved related to pro-gram definition, how to develop a program or subpro-gram around each objective, and how to allocate costs.In many ways, as will be evident from subsequent sec-tions in this study, many of these questions still beg an-swers. Further, there was a question of coverage. The

system was probably not useful for dealing with largesegments of government expenditures such as entitle-ments, transfer payments to state and local govern-ments, debt servicing, tax expenditures, and off-budgetexpenditures.

In the end, the critics were damning. Wildavsky(1969) characterized the system as “a vast amount ofinchoate information characterized by prematurequantification of irrelevant items.” The system im-posed what was considered an unbearable burden ofcalculation that choked the budgetary process and im-peded budgetary decision making instead of expedit-ing it. Schick (1971) was equally damning, focusingon what he felt was a chasm between planning andbudgeting—the ideal of analytical policy analysis, asseen in the PPBS, was stifled by the routines of bud-getary process reality.

Consolidation and Refinement of Program Budgeting

Despite the demise of the full-blown PPBS, pro-gram budgeting continued to be applied in the UnitedStates in a less ambitious form. The passage of theCongressional and Impoundment Act of 1974 ad-vanced multiyear budgeting; budget classifications bymission, function, and program; the use of sophisti-cated analytical techniques by the Congressional Bud-get Office (CBO) and the General Accounting Office(GAO);22 the development of performance indicators;and the improvement of accounting and informationsystems. This approach persists to the present day, re-sulting in a situation where it is possible for federalbudget decision makers to have data on program andperformance as well as on objects of expense.23 More-over, in the meantime, the program approach to bud-geting spread to other countries.

With the 1965 publication by the United Nations ofa Manual for Program and Performance Budgeting, theuse of program and performance budgeting was rein-forced as a tool of development planning. In the 1960s,nearly 50 countries introduced variants of program andperformance budgeting, although never on the scale ofthe United States. This then spread to developingcountries, so that by the end of the 1960s, nearly allLatin American countries, several Asian countries,and some African countries had introduced versions ofprogram budgeting.24 On the whole, the impact on

22Now the Government Accountability Office23However, to some critics, the data on programs and perfor-

mance are just window dressing, with decisions being taken basedon traditional line-item objects of expense (Schick, 1966).

24Waterson (1965) identifies 16 countries, 10 in Latin America,that had experimented with performance budgeting; Premchand(1977) provides brief case studies of similar budget innovations inAsia.

©International Monetary Fund. Not for Redistribution

Page 29: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

budget decision making has been disappointing25—perhaps indicating that for many developing countries,its adoption was premature. The UN manual, heavilybased on the U.S. experience, did lay down precondi-tions for successful performance budgeting in develop-ing countries: sound budgetary operating procedures,financial discipline, efficient methods of recording andreporting financial and physical data, and close coordi-nation between the central budget agency and othergovernment agencies. All, or most, were typically ab-sent in developing countries.

Output and the New Performance Budgeting

In the United States, after the PPBS system, therewere brief initiatives, such as Management by Ob-jectives (1973) and Zero-Based Budgeting (1977),but these were limited in impact, being pushed by theexecutive branch with little or no involvement ofCongress.26 However, between 1990 and 1996, therewas a spate of new laws that emerged from Congressthat created a government-wide platform for the newperformance-based management. The term new per-formance budgeting was used by Mikesell (2003) tocategorize the following reforms in the United Statesof the 1990s: the Chief Financial Officers Act of 1990,the Government Performance and Results Act of 1993,the Federal Acquisition Streamlining Act (FASA) of1994, and the Clinger-Cohen Act of 1996.27 All repre-sented an attempt to convert the federal budget processto a more results-based one with specific references tothe use of performance information for budgeting.28

The 2003 Program Assessment Rating Tool (PART) isthe most recent initiative (see Box 10).

However, in many ways this new legislation wasonly the means for the United States to catch up withinternational developments in PEM reform. Followingthe early lead of New Zealand, whose public sector re-forms culminated in the passing of the Fiscal Respon-sibility Act in 1994, other countries, such as Australiaand the United Kingdom, were quick to follow, and inturn their attempts were mimicked in many northernEuropean countries as well as in the United States. Thepenetration of the new ideas to lower levels of govern-

20

ment has been almost universal in these countries.29

The themes of these new reforms—often termed theNew Public Management—were the primary empha-sis on outputs or outcomes, performance in budgeting,and particularly the use of performance information inthe budget process. This was coupled with other majorPEM reforms, including the introduction of greaterflexibility for government bodies (agencies), the intro-duction of contractual arrangements, and the moveaway from cash accounting to the use of accrual ac-counting and budgeting along the lines of the privatesector.30 The first aspect, the new performance ap-proach, has been the most universally accepted inter-nationally. A 2001 OECD survey found that 70 percentof OECD members included performance information intheir budgets, whereas the remainder included suchinformation for a limited set of programs. Moreover,40 percent of the countries surveyed claimed theirperformance measures distinguished between out-puts and outcomes, with 20 countries using system-atic annual reporting on performance against outputsand 15 countries on performance against outcomestargets.31

However, as noted, performance budgeting is morethan simply introducing performance information intothe budget process. A key feature of the new perfor-mance budgeting approach is the recognition that, ifperformance is to matter, the objectives of the budgetmanagement system must be integrated with overallaccountability, so that good budgetary performanceis rewarded, and poor performance is penalized. Al-though countries have taken different approaches tothis, there are some common elements that can beidentified from OECD countries’ experience.

While using a program structure to introduce a per-formance orientation, it is clear that, to be effective, aprogram budget format must also be integrated into awider model of budget management. Experience sug-gests that the benefits of a budget program format willsoon be lost unless departments (and most centralagencies) continue the momentum and purpose ofthe reform by moving on to develop standards of ser-vice delivery and to search for ways by which thesestandards can be continuously improved and servicesdelivered more effectively. That is, government ser-vice delivery needs to be redefined in a more results-oriented manner. How to do this? The most effectiveway, at least from the experience of a few pioneeringindustrial countries, involves a fundamental revisionof accountability relationships within the public sec-tor. The implications for budget management are pro-found: first, performance needs to be specified and

25These problems led Dean (1989) to conclude that “the historyof performance budgeting is one of high hopes and disappointingachievement” (p. 123).

26For a fuller description of these approaches, see Anderson(2004, pp. 5ff), which also discusses the impact of the 1981 Presi-dent’s Private Sector Survey on Cost Control (The Grace Commis-sion).

27FASA included provisions for agencies to establish cost, per-formance, and schedule goals for major acquisitions and related payto performance; Clinger-Cohen included requirements for agenciesto develop performance measures for information technology usedor acquired by the agency.

28See Mikesell (2003).

29See Bellamy and Kluvers (1995, p. 39).30For a review of the experience and difficulties in implementing

this approach, see Schick (1996).31For full details, see Kristensen (2002).

©International Monetary Fund. Not for Redistribution

Page 30: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Issues of Design in a Program Structure

21

reported in a way that is operational for budget man-agers; second, government agencies need greatermanagerial autonomy and freedom from tight inputcontrols, so that they can determine their most effi-cient delivery of results; and, third, changes are re-quired to the range of incentives and sanctions facingdepartmental managers.

Accordingly, while a good program structure is thebasis for the new performance budgeting approach, ex-perience indicates that it will not be sufficient to ensurethe success of budget reform. If the program structureis to yield benefits, it must be linked to other, muchwider reforms, which are taken up in subsequent sec-tions of this study.

Issues of Design in a Program Structure

From the previous, necessarily brief, review of theevolution of performance budgeting, some key featuresemerge. First, an essential element in all systems is theconcept of a program and the restructuring of the bud-get on this basis. At first glance, the principles deter-mining programs are fairly clear: a program can beviewed as any suitable and meaningfully integratedgroup of activities and projects, under a single manager,which consumes resources to contribute to a specifiedpolicy objective. The operational objectives of each pro-gram and activity can then be identified. Budgeting andaccounting can be carried out on this basis, so that theseparate costs and revenues of each program/activityare made clear to decision makers. It is apparent fromthis description that later stages in the evolution ofthis program approach maintain the fundamentals butadd some elements. Thus, output budgeting involvesmeasuring the outputs and performance of activitiesso that they can be related to their costs and opera-tional objectives. In turn, the new performance bud-geting involves adding the use of the resultant data toestablish standards and norms, so that costs and per-formance can be evaluated and government resourcesused more efficiently/effectively through an incentivesystem that enforces accountability. In all approaches,however, the basis of performance budgeting is ameaningful program structure.

The second essential element is that the program ap-proach to budgeting is an enduring one which, despiteits critics, is viewed by budget practitioners as havingvalue added. Specifically, program budgeting is seento counter the alleged deficiencies of traditional line-item budgeting, including its short-term focus, its in-cremental nature, its emphasis on inputs rather thanprogram outputs, and its excessive detail which ob-scures alternative means of reaching the same objec-tives (Babunakis, 1976). At the same time, there islittle agreement on the guiding principles of program

classification, and as a consequence, the term is usedin a variety of ways—in some cases a program corre-sponds more closely to a broader concept of a function,and in others to a line item of expenditure. Because itis policy oriented, a program structure is inevitablycountry specific.

The third essential element is that its successful im-plementation has proved to be universally difficult re-gardless of the country setting and the form attempted.Some of these problems have already been indicated:difficulty in providing meaningful information, par-ticularly relating costs to performance data; suspicionof information overload; legislators’ reluctance toyield detailed line-item controls associated with theconcern that somehow budget compliance will becompromised; and failure to implant this in widerbudget management reforms, particularly increaseddelegation—indeed, perhaps some view it as a way ofcircumventing centralized controls. Despite the diffi-culties, however, program budgeting is undoubtedly acentral element of OECD budget practice (Box 9).

If one views the progression from old to new per-formance budgeting as stages in the development ofbudget management, it is possible to characterize manyemerging market economies as having reached thestage of program budgeting, with some already ex-perimenting with the next stage of output budgeting.To successfully move on to the next stages, emergingmarket economies must introduce a program structureinto their budgets, and where this already exists, mustfundamentally redesign their program structures toaccommodate the new performance budgeting ap-proach. Specifically, it requires reviewing programs

Box 9. OECD Practices: Use of a Program Structure

Percentage of OECD Countries

• Legislature approves budget appropriations:– Disaggregated appropriations

at program level 25– At program level 17– Split between personnel and program

spending at program level 17• Expenditure classification includes

program classification, with governmentpolicy objectives and individual programbudgets 44

• Expenditures linked to strategic goals:– For all goals 19– For some goals 31– For a few goals 23

Source: OECD (2003, Tables 2.9, 5.3, 5.4.c).

©International Monetary Fund. Not for Redistribution

Page 31: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

to ensure that there is a clear policy statement or listof objectives that adequately defines the purpose of theprograms, so that the results expected from the pro-grams can be assessed, measured in some way, and re-ported. In turn, this allows introduction of the morerigorous accountability mechanisms that characterizethe new performance budgeting approach. Followingis a further exploration of these considerations, whichstresses four main features:

1. The program structure should be set in a widerstrategic framework. The program structure is basicallya way of describing the expenditure plan of the budgetin terms of objectives. To be relevant, the programstructure must be anchored in a wider strategic viewthat allows it to describe the contribution of govern-ment operations to achieving nationwide objectives. Touse the language of strategic planning, as pioneered inNew Zealand, the national objectives or “strategic re-sult areas” are translated into detailed and fully costed“key result areas” which are the responsibility of spe-cific government units. The latter should then for-mulate the required detailed implementation plansrequired for preparing the agency’s budget bid. Thewider strategic view has typically been addressed inmany countries by adopting a medium-term budgetframework, as discussed below, but also through a con-

22

stant review of program design to ensure its alignmentwith government objectives. This is very evident in theU.S. government’s Program Assessment and RatingTool (PART) (Box 10).

2. The program structure should be defined in a wayto support political decision making and prioritization.Following from the need to relate programs to widerpolicy and planning priorities, it is necessary to makeclear the relationship between the resources used by theprogram and the proposed outputs and/or policy results(outcomes). This implies that programs are providedwith sufficient resources to meet the objectives as-signed to them, and in turn, that there be adequate cost-ing of the program’s outputs. The latter has oftenproved technically difficult.

3. The program structure should ensure account-ability. Programs and subprograms should be disag-gregated to activities and projects in such a way as tosupport clear managerial responsibility in attaining theproposed outputs or outcomes. This implies each pro-gram has an appropriate size for efficient managementand clear managerial responsibility, usually within asingle organizational unit. It also implies that programsshould be designed and “owned” by the organizationalunit and not by the MoF or budget office.

Box 10. The U.S. Government’s Program Assessment Rating Tool (PART)

In 2002, the Office of Management and Budget (OMB)developed PART, a questionnaire to collect information todevelop a consistent program rating. On this basis, theOMB and agencies rated 20 percent of programs for the2004 budget, added 20 percent more for 2005, and plan toadd another 20 percent for 2006.

Objectives of PART

• Systematize performance information into the budgetpreparation process to better align objectives with budgetstructure.

• Make agencies link information on program purpose anddesign with strategic planning and program managementand results.

• Work with agencies to improve the operating and finan-cial performance of their programs.

• Reach overall assessment of program effectiveness bybetter outcome measurements and more systematic eval-uations.

Features of PART

• Comprises 30 questions on program design, strategicplanning, program management, and program results.

• Represents a clear attempt to link the program’s purposewith its design, reveal the strategic planning behind its im-plementation, and generate performance-based programmanagement and results information.

• Questions differentiate types of program: grants, regulatory-based programs, capital assets, service acqui-sition, credit, and research and development programs.

• Rates and systematically assesses programs as part offormulating the president’s budget.

• Measures and diagnoses program performance and indi-cates areas requiring attention.

• Assists agency and budget decision makers on manage-ment and regulatory improvements, legislative proposals,and budget requests.

Unanswered questions

• How can PART be integrated with the GPRA (see Box 51)?PART is clearly designed to serve the president’s inter-ests, but the GPRA process is designed to be consulta-tive and to seek agreement on goals and measures.

• How can support from Congress be sought? Congres-sional involvement is critical if PART is to be used be-yond the executive branch. While some congressionalmanagement committees have shown interest, the appro-priation committees have few incentives to change thestatus quo and use the performance information.Source: GAO (2004); Fantone (2004); Anderson (2004).

©International Monetary Fund. Not for Redistribution

Page 32: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Wider Strategic Framework for the Program Structure

23

4. The program structure should be integrated intoa wider performance-enforcing budget managementmodel. Gaining maximum benefit from the programstructure in terms of improved performance and al-lowing accountability to be enforced requires widerbudget management reforms. It entails that responsi-bility for implementing each program/subprogramshould be aligned to budget appropriations as much aspossible, providing a clear link between budget andpolicy outputs/outcomes. This implies that programsare directly linked to one budget appropriation man-ager, as described further below.

The Wider Strategic Framework forthe Program Structure

The core building block of performance budgetingis a meaningful program structure for government op-erations. However, underlying this reorientation aretwo wider objectives: to improve the quality of bud-geting, by providing a link with policy, and to helpensure the relevance to wider sectoral policies ofgovernment operations supported by budget appro-priations. However, designing an operational programstructure for the budget should not be regarded simplya way of reclassifying the government’s expenditureplan. Rather it should involve some more fundamen-tal rethinking of government activities and the waythey are managed. That is, budget managers should begiven the opportunity to—

• reconsider their respective roles and missions.• review the way to best organize themselves to

carry out their missions.• redesign cost allocation and financial reporting

systems to improve the efficiency of management.• review the efficiency and effectiveness of opera-

tions and delivery of goods and services.The starting point in such a review of government

activities is to connect them to underlying governmentobjectives. This can be achieved by linking the opera-tions of each spending agency with the implementa-tion of its sector’s policies. Broad sector functions canbe divided into subfunctions and further into programsthat comprise a number of activities, of both a recur-rent and capital (project) nature. These activities andprojects are then considered as cost centers for whichan economic, or object of spending, classification canbe prepared. In this way, the program structure is de-pendent on sectoral policies, and these in turn are de-pendent on the system of prioritization among sectors.The set of programs will then reflect the major priori-ties of the government. The program structure is thusthe link between the budget and overall strategic plan-ning, which reflects the policy framework for govern-ment operations. Only in this way is the programstructure a tool for policy analysis and not simply an-

other way of classifying expenditures. It is a way toconsider what should be funded, the level of thefunds that should be applied, and the policy impactof those funds. If effective, therefore, program budget-ing should represent a substantial change in the way thatbudgets are prepared, approved, managed, reported, andevaluated.

This would imply a more top-down approach tobudget planning, in contrast to the bottom-up approachof more traditional budgeting. The drawbacks of thebottom-up approach to budget planning are well ap-preciated (Box 11). Despite the attraction of develop-ing a more proactive, top-down approach, this is noteasy because it implies having a strategic view of thestate’s role in society, with important implications forbudget policy. First, it implies a sharper differentiationbetween private and public sector activities and a clearview of what government should and should not do.Second, it requires a determination of the maximal sizeof the resource envelop available to the governmentsector. Third, it demands a clearer view of where andfor what purposes the government resources are opti-mally deployed within this resource envelop. The im-petus for performance budgeting comes from this latterperspective.

For many countries, the implications of this reorienta-tion have been profound. The more precise specificationof activities to be carried out by government has, for manyindustrial countries, resulted in a considerable downsizingof government and other fundamental public sector re-

Box 11. Limitations of a “Bottom-Up”Planning Process

• The spending plans may not be sustainable. The bottom-up aggregate demands may result in an unsustainablelevel of resource use because no one has an overallaggregate view. The absence of an effective resourceconstraint results in poor budget discipline.

• There may be counterproductive resource alloca-tions. People working in individual sectors do notfully understand and/or allow for complex develop-ment linkages. Projects and programs may not be fullycompatible and mutually supportive.

• There is a tendency for bottom-up demands to reflectpowerful interest groups and not necessarily generalwelfare considerations. One might suspect the interestsof the weak or disadvantaged may not be adequatelyheard.

• There may be significant gaps in the coverage of suchplans. Important areas may not be covered because noone acts as a “product champion” by accepting respon-sibility for formulating a needed project, or becauseambiguities in the public’s understanding of the role ofthe state result in neither the government nor the privatesector undertaking some types of activity (for example,some types of research or information gathering).

©International Monetary Fund. Not for Redistribution

Page 33: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

forms. It also has resulted in a restructuring of the rela-tionship between government and the private sector,which typically involves a greatly reduced role for thegovernment in influencing private sector activities. Forthe government, it has usually involved borrowing thetechniques of the private sector for planning and deliver-ing services, including the adoption of “strategic plan-ning” in marked contrast to more traditional developmentplanning methodology. This can be viewed, followingthe New Zealand model, as a top-down process in threestages: first, a determination of the overall sustainablelevel of government resource use in the economy; sec-ond, within this envelope, a definition by the governmentof its core activities (“strategic results areas,” in the ter-minology of strategic planning); and third, translation ofeach of these basic core activities into action plans foreach department or agency involved in that core activity(its “key results areas”). This is schematically repre-sented in Box 12.

Are there any guidelines for formulating thesestrategic plans? First, it is easier if there are relativelyfew strategic results areas and if they represent thoseactivities or achievements considered to have criticalbearing on a country’s development.32 For this reason,they should perhaps center on a government’s core“public good” function: law and order, infrastruc-ture, education, and health facilities as investmentsin human resources. They should, to the greatest ex-tent possible, be presented in output-oriented terms.33

Within this framework, the next step is to translateeach strategic results area into a set of key resultsareas for each department which are represented in thedepartments’ budget vote, incorporated in their busi-ness plans, and reflected in their level of budget fund-ing. In this somewhat idealized system, the strategicplans are translated into budget allocations—set in anMTBF to ensure sustainability. The role of the budgetthen is to provide the link between higher-level plan-ning and formulation of more detailed implementa-tion plans.

Introducing Medium-Term Budget Frameworks

A realistic MTBF is often viewed as the corner-stone of a performance-oriented budget approach,linking resources to the policy objectives that defineperformance. As indicated, this approach involves areorientation of the budget process to employ system-atic use of medium-term forecasts of economic and fis-

24

cal aggregates to assist in developing a multiyearframework for spending policies. However, the MTBFdoes not imply multiyear budgeting in the sense of ap-propriations extending beyond a single budget year.34

Rather, the first-year outturn estimate becomes thestarting point for preparing the budget for the follow-ing year. While most OECD countries have adoptedsome type of MTBF, the exact forms vary from coun-try to country. Box 13 summarizes OECD experiencein this area. However, most of existing MTBFs includeall or several of the following key characteristics:

• A clear fiscal policy statement which enables theMoF to set a medium-term path for a specific pub-lic expenditure aggregate;

• A medium-term macroeconomic forecast withinwhich the annual budget and multiyear budget es-timates can be presented to the legislature eachyear;

• The capacity to produce realistic revenue esti-mates for the plan period (a crucial element);

• A formal requirement that ministries maintainestimates of expenditures covering several yearsbeyond the proposed budget year;

• A budget preparation process that assesses andgives formal status to the “forward” or “out-year”estimates as well as the estimates of outlays forthe budget year;

• A process by which the budget figures for the in-dividual spending ministries and agencies becomehard budget constraints expressed in cash terms.

In introducing this approach to budget preparation,middle-income countries have perhaps not fully ex-ploited the MTBF’s role as part of a wider shift inbudget management to simultaneously limit and takestrategic choices at the macro level while introducinga performance perspective at the micro level. Also, ithas not always been fully appreciated that a numberof preconditions are required for success. For exam-ple, given that an important objective of an MTBF isto lengthen the time horizon of fiscal policy, a degreeof macroeconomic stability is important, especially onthe revenue side. Simply put, if planned resources arenot forthcoming, even for priority activities, no greatattention will be given to either the defining prioritiesor the resulting framework.

Furthermore, the required degree of transparent pri-oritization at the political level can create problemsfrom different perspectives. First, introducing a trans-parent process like an MTBF that makes explicit pol-icy choices presumes an adequate level of governanceand transparency. Second, to introduce an MTBF isalso to introduce a process of budget rationalization.

34Although multiyear appropriations are used by some countriesfor some expenditure items, such as capital spending, others havethe flexibility of carrying forward unspent budget appropriations tothe next year.

32These will be country specific because what constitutes a “pub-lic good” is debatable, and governments need not necessarily de-liver all such public goods.

33That is, the expected results should be explicitly stated andshould be quantified where appropriate and relevant.

©International Monetary Fund. Not for Redistribution

Page 34: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Introducing Medium-Term Budget Frameworks

25

Portfolio rationalization, especially where it involvesreductions in ministries’ activities, is a difficult andoften painful process with direct political implica-tions. Third, there are other political considerations:the timetable for budget preparation must be changed,and this may need some legal backing. Most OECDcountries use up to one year to prepare the next year’sbudget and MTBF in order to allow for extensive in-teraction between the MoF and individual ministries,to facilitate a common agreement of the economic sit-uation and the scope for government spending, andthus to enhance the legitimacy and creditability of thebudget preparation process. Last, the MTBF must besupported by adequate capacity in the rest of the PEMsystem. In isolation, an MTBF cannot compensate for

inadequate financial planning, project/program man-agement, access to information, and forecasting skills.This theme is pursued further in subsequent sections.

A more detailed description of the key issues in-volved in the design of an MTBF can indicate the scaleof the challenges of operationalizing this approach.

• Clarity in the policy statement: At a minimum,there should be an explicit statement of fiscal pol-icy with a target fiscal deficit for the budget yearand some planned track for the deficit in the twoto three following years. On the basis of revenueforecasts, that target can be translated into amedium-term path for an appropriate public ex-penditure aggregate. Such policy statementsshould be stable (not changed every year or so),

Box 12. Connecting Planning with Budgeting

Strategic Planning Process Time Period Budget Process

Whole GovernmentStrategy and Analysis

Strategic Goals

Strategic Action Plan

Performance Plansand Indicators

Long-TermView

(5+ years)

Long-TermBudget Strategy

“StrategicResults Area”

Medium-TermView

(3-5 years)

“Key ResultsAreas”

Short-TermView

(1 year)

MTBF/MTEFFiscal Objectives

StrategicResource Plans

Sector Plans

Annual BudgetPolicy Statement

andMinistry/AgencyAnnual Budgets

Sustainable Levelof Government

Activity

©International Monetary Fund. Not for Redistribution

Page 35: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

realistic, and transparent. Accordingly, the suc-cessful implementation of an MTBF requires thatat least two types of procedure be institutional-ized: first, a system of regular reporting of actualfiscal developments compared to planned perfor-mance, with reasons for any deviations; and sec-ond, a requirement that decisions be made andacted upon for required policy adjustments incurrent and future years in light of the medium-term targets. Country experience shows that thisrequires a commitment by all units within thebudget process to make the system work. Of par-ticular importance is that the aggregate target beendorsed by the cabinet and the legislature. Insubsequent policy proposals, the governmentand/or legislature should be obliged to show howthese proposals can be undertaken within the normestablished for the current year and the mediumterm.

• The distinction between new and old policy: AnMTBF generally requires that a clear distinction bedrawn between expenditures associated with newand existing policies. As emphasized previously,in this way the MTBF supports strategic planningat the highest level and allows for transparent pri-oritization between sectors and within sectors.Accordingly, expenditures can be divided into ex-isting policy and “agreed new policy,” with a re-

26

serve left for the admission of new policy propos-als in later years. This requires all types of spend-ing to be clearly identified and costed in astandardized way in advance, typically necessitat-ing more detailed knowledge of the composition ofgovernment expenditures. This, in turn, requiresimproved budget classification, usually by devel-opment of a more functional economic classifica-tion that includes subfunctions and clearly relatesprograms in a hierarchical and unique way. Sec-ond, there is a need for improved reporting: min-istries will have to maintain and report not just onpast expenditure levels but also on planned futureexpenditure levels. The MTBF should also leadto more detailed budget proposals for strategicplanning purposes, including why governmentintervention is required, specific outcomessought, expected outputs required for those out-comes, and multiyear estimates of the costs.Third, there must be clear differentiation be-tween those expenditures that are obligationsarising from other laws (statutory spending) andthose expenditures that are truly discretionary.

• Reliability in the macroeconomic framework: Thegovernment must have a macroeconomic fore-casting capacity and set realistic envelopes for theaffordable total government expenditure withinthe medium term. Clearly, the more stable the

Box 13. OECD Practices: Use of MTBFs

Percentage of OECD Countries

1. A medium-term framework exists with targets/ceilings for expenditures, deficits, and revenues: 63a. Ceilings for expenditures, but targets for revenues and deficits 20b. Ceiling for deficits, targets for expenditures and revenues 36

2. An MTBF is a legal requirement 503. An MTBF covers:

a. 3 budget years 40b. 4 budget years 24c. 5 budget years 24

4. Based on a growth assumption with a margin for “prudence”:a. Margin less than half a percent lower than forecast 17b. Margin more than half a percent lower than forecast 4

5. Multiyear cost estimates are made for:a. All new spending 56b. Only for selected expenditures 18

6. Made available to the public 647. Legally required to be made available to the public 368. Budget documents contain a statement of medium-term fiscal policies 819. Budget documents contain a forecast of fiscal aggregates for two future years 58

10. Budget documents contain formal rolling medium-term (3- to 5-year) estimates of revenues and expenditures 50

Source: OECD (2003; Tables 2.2.b, 2.3, 5.2).

©International Monetary Fund. Not for Redistribution

Page 36: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Introducing Medium-Term Budget Frameworks

27

main parameters, such as inflation and the ex-change rate, the easier the task. However, even foreconomies experiencing some instability, the MTBFis a useful approach for investigating the areas of un-certainty and setting a realistic path, while allow-ing for uncertainty through a planning reserve(discussed below). To be effective, implementa-tion of an MTBF requires enhanced planning andforecasting capability not only in the MoF butalso in the ministries, as well as a stronger reviewcapacity in the MoF. This may require capacitybuilding in the MoF in macroeconomic forecastingtechniques, particularly with regard to revenues,and capacity building in most ministries for sup-plying information regularly to the MoF on theprognosis for current and future expenditure lev-els. When projected expenditures seem likely toexceed budget estimates, ministries will have theresponsibility of finding savings within their ownareas and proposing new policies to the MoF to re-spect the estimates within their appropriations.

• Allowance for cyclical factors: A number of coun-tries also make some allowance for cyclical factorsin their MTBFs by accommodating the variationin certain well-defined expenditure programs. Suchan approach allows for the operation of automaticfiscal stabilizers on the expenditure side. It alsoimposes discipline during upswings when, say,unemployment expenditure programs are tem-porarily low and policymakers are tempted to usethose resources for other expenditure programs.When first developing macro forecasts it is recom-mended that, at least initially, no specific allowancebe made for cyclical factors.

• Accommodating the effects of inflation: The ques-tion arises whether expenditures should be forecastin terms of cash rather than volume (or constantprice) terms. Several countries that started witha real-expenditure-based norm have switched toa cash planning system. This is due to its trans-parency as well as to a reluctance to set the normin real terms, which provides ministries with com-plete protection against inflation at a time when itmay be desirable to cut spending. Using a cash-based system does not mean the impact of inflationis ignored, but if inflation is higher than expectedthere is no presumption of an accommodating in-crease in line ministry appropriations. By setting aclear cash limit, giving greater flexibility to man-agers, and establishing a degree of performanceaccountability, governments using this approach inthe MTBF have provided stronger incentives toline managers to use their budget resources moreefficiently.

• The coverage of the MTBF: There is a tension be-tween planning for those expenditures that arewholly under the government’s control (which

may only cover parts of central government ex-penditure) and a wider expenditure aggregate, suchas general government expenditure (which may beless subject to central government control but moremeaningful for macroeconomic analysis and thusmore relevant to establishing the acceptable size ofthe deficit and tax burden). While practice varies,most OECD countries now employ an aggregatethat includes the activities of the secondary and ter-tiary levels of government—although generallyexcluding public sector enterprises, except fortransfers and net lending to them.

• Choice of time horizon: While practice has varied,there is a growing consensus for MTBFs of threeto four years, with the majority of OECD coun-tries now using three years. Of course, for somepurposes, the government should look much fur-ther forward—for example, to answer questionsabout the longer-term consequences of existingpension or health policies as the population ages.For such questions, the MTBF needs to be sup-plemented by periodic longer-term analyses.35

• The inclusion of planning and contingency reserves:A planning reserve can be defined as a reserve in theforward years available to be allocated for provi-sion to priority expenditures in forthcoming annualbudget negotiations. A contingency reserve is usu-ally thought of as a reserve for dealing with unex-pected, but unavoidable, expenditures during thebudget year. Reserves accommodate uncertain-ties of several types. They accommodate changesin macroeconomic assumptions, and so countrieswith higher or more unpredictable inflation pathsmust have higher reserves. They accommodateexogenous shocks such as natural disasters andchanges in the costs of programs, for example,when the projection of demand for governmentservices underestimates volume requirements.They also accommodate new and changing policypriorities. The aim should be to set a reserve largeenough to accommodate all of these uncertainties,without it becoming so large that it negates effi-cient budgeting. For example, a figure much above5 percent of general government expenditurewould be hard to justify except for a high debt/highinflation economy. In setting reserves, a simpli-fied 1-2-3 rule of thumb may be a useful startingpoint: if the unallocated contingency would be setaround X percent in the current fiscal year, it wouldrise to 2X percent in the second, and 3X percent in

35One technique that has been employed for this purpose is inter-generational accounting, developed in the early 1990s by Auerbach,Gokhale, and Kotlikoff (1999). The United States was the first topresent such accounts, in 1993, with other OECD members fol-lowing, including Germany, Italy, New Zealand, Norway, and Swe-den. For a fuller description see OECD (1997b).

©International Monetary Fund. Not for Redistribution

Page 37: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

the third, reflecting the fact that uncertainty in-creases over time. Of course, use of the contingencyreserve in the current year would require approvalby the MoF, which should grant it only reluctantly.A record should be kept of the agreed claimsagainst this reserve, which should be published atthe end of the fiscal year. Above all, it should bemade clear that the contingency need not be fullyutilized.

Formulation of Program Structure in Budget Format

Emerging market economies typically must funda-mentally redesign their program structure to allow anew performance budgeting approach. As argued inthe previous section, this needs to occur in the contextof a wider strategic plan for government operations tostrengthen the links between policy and planning andbudgeting. In this context, the review of programstructures should involve a clear policy statement, orlist of objectives, that adequately defines the purposeof the programs and the results expected from the pro-grams so that they can be assessed and measured insome way. The process for this review is outlined inBox 14. In this way, the central characteristic of a pro-gram is that a given collection of government activi-ties shares a common set of objectives—that is, theprogram structure is based on policy.

While the program structure should be seen as themeans whereby the budget is linked to strategic ob-jectives, the issue still arises how programs should bebest designed to meet these objectives. There appearto have been two basic approaches to this task, largelyreflecting the trade-off between viewing the programeither as a tool for planning or as a tool for budgeting.The route pioneered by the United States is to makethe program structure agency specific—take a spend-ing agency (say, a ministry) and design the programonly within that ministry’s activities, hence anchor-ing the program to a specific budget. Other countrieshave designated broad policy areas and have identi-fied programs on this basis, which means that indi-vidual institutions can end up contributing to onlypart of a program. The agency-specific approach con-strains the logic of the policy basis of programs by theprevailing organizational structure of government,and it is possible to end up with cost centers that aremeaningless from a policy analysis viewpoint. Thisapproach does, however, make accountability easierto specify, monitor, and enforce. The broad policy ap-proach, while purer from a policy viewpoint, dependson having adequate classification and accountingcapability to capture all inputs associated with pro-grams, regardless of where they arise. This can be aproblem in countries without this capacity or wherebudgets are still firmly rooted in approved inputs, and

28

this makes accountability more difficult to specify,monitor, and enforce.

In general, the agency-specific program design ispreferable for most countries, but in exceptional cases,multi-agency programs may be unavoidable. A usefulstarting point in determining the policy framework forprogram design is to use the functional classification ofgovernment expenditures. In presenting categories ofgovernment expenditure useful for international com-parisons, as well as a means of presenting budget datathat is meaningful to the general public, functional clas-sifications have been standardized internationally (forexample, the UN’s COFOG system). Because the func-tional classification aggregates data from a range ofgovernment agencies that are engaged in similar activ-ities, functional classifications provide a comprehen-sive view of government expenditures in such broadareas as defense, health, social welfare, and educationand thereby provide a useful framework for construct-

Box 14. Process for Agreeing on Program Budget Format

Identify strategic results areas.

• Define policy objectives in the agency’s area of oper-ations (sector).

• In light of this, identify the key changes required forthe agency to deliver its strategic goals.

Realign strategic action plan.

• Review each spending unit’s role.• Redefine the spending agency’s current mission (may

change from year to year).

Define key results areas.

• Given these key results areas, decide whether theagency is organizing itself in the best way to achievethese results.

• Decide on the final distribution of accountabilities,i.e., who is responsible for doing what.

Create responsibility centers.

• Identify accountable officers.• With their supporting staffs, accountable officers

should prepare implementation plans that elaborate indetail how, when, and by whom key results will beachieved/delivered.

Prepare budget format and costing.

• Finalize budget formats and cost-allocation proceduresso that agencies are able to prepare mock-ups of finan-cial and accounting reports to be produced by theagency or the MoF systems.

• Agree with MoF on any new codes required for newactivities and on nonfinancial performance measures.

©International Monetary Fund. Not for Redistribution

Page 38: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Introducing Medium-Term Budget Frameworks

29

ing a program structure. However, within this frame-work, the way that functional categories can be brokendown into programs and directly related to policy willvary from country to country. General guidelines onhow this might be done are presented in Box 15.

When reviewing country experience in adoptingground rules for the design of program budgets, it ispossible to identify three key issues: the agreed rulesfor aggregation to a program structure; the institutionalcoverage of programs; and the existence of adequatemanagement information.

Agreed Rules for Aggregation to a Program Structure

This is an important precondition for making pro-gram budgeting a success. These rules should reflectthe level of aggregation of basic activities and/or proj-

ects that is most useful for defining a program. If theprograms are too disaggregated, then some programsmay have only one subprogram, calling into questionthe usefulness of this distinction. However, if the pro-grams are too aggregated so that their implementationcuts across institutions, it will be difficult to define spe-cific program outputs and assign specific accountabil-ity. Care should also be taken to agree on the specificityin program definitions: Should programs be expressedin terms of a specific major policy area, rather than as acatch-all for a loosely related set of activities? Is it pos-sible for each subprogram to capture activities thatproduce homogeneous outputs? Can there be a clearspecification of the outputs sought in terms of timeli-ness and accuracy, and in a way that can be measured?

The Institutional Coverage of Programs

In redesigning programs, the following range ofquestions typically needs to be addressed: How closelyshould programs be aligned with the organizationalstructure? Can programs extend beyond single min-istries? Can a work unit contribute to more than oneprogram? A related accountability issue is the align-ment of programs with appropriations: How shouldappropriations be allocated to programs?

When originally introduced, program budgeting de-scribed programs irrespective of organizational linesof responsibility. It was not unusual for responsibilityfor the design of the program structure to be given tothe central budget agencies. Sometimes, the line min-istries were not heavily involved, especially at the de-sign stage, even though they had the most detailed andcomprehensive view of the functions and objectives oftheir entities and were responsible for implementingthe programs. Many countries have tried this approachsince the 1960s, and few have truly succeeded in mak-ing it work. Increasingly, the emphasis is on a modi-fied approach to program budgeting, which focuses onoutputs and performance with improved accountabil-ity—the output budgeting and new performance bud-geting approaches. However, this should not detractfrom the conclusion that introducing a good programstructure should involve substantial changes in theorganizational structure of government. The reality is,of course, that this is usually not a practical option.

A dominant reason for arranging for a given outputto be managed within one organizational unit is to as-sign clear accountability for performance. In practice,the bulk of outputs are produced within a single orga-nization, for example, medical treatment at hospitals,teaching services in schools. Problems arise, however,where a program seeks complex outcomes, such ascrime prevention or tourism development, which nec-essarily require outputs from a range of agencies.When programs are conducted jointly by two or moreadministrative units, responsibility is also divided, so

Box 15. General Guidelines on the Design of Programs

• Programs are “unifunctional,” so that each program islinked to only one function.

• Programs are hierarchically constructed, so that eachprogram has a number of subprograms, and each sub-program can be decomposed into a number of activitiesand projects. Each subprogram is related to only oneprogram; likewise each activity and project is related toonly one subprogram.

• Each program has an appropriate size for efficientmanagement. This varies from country to country, butoften implies that for broadly defined and resource-intensive programs, the main unit for accountabilityand performance specification is at a lower level (say,at the subprogram rather than the program level).

• Programs and subprograms should be defined in a waythat supports political decision making and prioritiza-tion by making clear the relationship between theresources used and the expected outputs and policyresults (outcomes).

• Programs must consider all related activities (includ-ing regulatory ones) and projects that together assist inachieving their objectives. This means that capital andrecurrent spending should both be considered in judg-ing program performance relative to its objectives.

• Greater levels of disaggregation to activities and proj-ects will be used to support, and be designed to facil-itate, management’s pursuit of the subprogramobjectives and results.

• Accountability for subprograms should be clearly as-signed to particular managers, usually and preferablywithin a single organizational unit.

• Responsibility for implementing each particular pro-gram should almost always align by administrativeunit to one chapter (or “vote”) in the budget. Wherethis is not possible, it is important to assign lead rolesto a particular chapter.

©International Monetary Fund. Not for Redistribution

Page 39: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

III MOVING FROM PROGRAM TO PERFORMANCE BUDGETING

that specific activities and end products are difficult toassign to individual organizations. In such cases, cost-ing becomes difficult and requires elaborate costingsystems, discussed further in Section V. At the sametime, it is unwise to reorganize government agenciesin the interests of introducing a simplified programstructure. Instead, in designing a program structure, itis necessary to recognize a variety of operating units:some are merely supervisory, others implement theirown programs, and some do both; some fund projectsimplemented by other levels of government; and somehave responsibility for public enterprises and quasi-autonomous bodies. Each type may require a differentapproach.

Existence of Adequate Management Information

Improved accountability is dependent on adequateinformation to judge performance. The design of pro-grams therefore should take into account the practical-ity of constructing a comprehensive and regular

30

information and reporting system to provide relevantdata in a timely manner. A major user of a performancebudget should be the program management. However,a government consists of vast hierarchical systems ofmany levels and organizational relationships. Design-ing information systems for such structures is com-plex: At what levels should costs be collected? Canperformance indicators be provided for those cost col-lection centers? From what sources should originatingdata be collected? How can the reliability of this databe ascertained and improved? How should data be re-ported to higher levels? What information do differentmanagers need, and what regular reporting systemwould supply that information?36 What are the impli-cations for the redesign of government financial man-agement information systems? The establishment ofa performance budget management system requiresfinding answers to these questions, and how to do so isdiscussed in subsequent sections.

36The particular problems of performance measurement are ad-dressed in Section IV.

©International Monetary Fund. Not for Redistribution

Page 40: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV Introducing a Performance Management Framework

31

Many emerging economies have completed thefirst step of the budget reform process by moving

away from the traditional emphasis on managers’stewardship of public resources and on compliancewithin strict detailed appropriations. As described inSection III, this usually involves implementing someform of program management and budgeting, alongthe lines of the reforms introduced in OECD countries,where there is greater emphasis on achieving efficientand effective outputs and outcomes and where mea-sures of performance have tended to play a key role.

This section argues that, while it might be temptingfor emerging economies to press forward to adopt afull-blown outputs and outcomes performance man-agement framework, there are some risks involved.Such a change in orientation is only possible oncemanagers have had adequate experience in refiningthe definition of programs and objectives and, on thisbasis, have experience in developing a comprehensivesystem of performance measurement. It is argued inthis section that developing a comprehensive perfor-mance management system requires three criticalsteps: first, to clearly define how to measure “perfor-mance”; second, to overcome a number of technicalissues in the design and use of measures of that per-formance; and third, to make performance informa-tion relevant for resource allocation decisions. Basedon international experience, this section reviews eachof these three hurdles in moving toward a perfor-mance management framework.

Different Measures of Performance

A general consensus is emerging in OECD countriesthat performance measurement improves the quality ofmanagement by helping officials make better decisionsand use resources more effectively.37 Quantifying per-formance allows managers to monitor changes, iden-tify potential problems, and take timely correctiveaction. Having data that describes performance interms of where the agency has been, where it is today,

and how it compares with similar agencies allows man-agers to direct resources to, and officials to support, ac-tivities that are successful. Performance measures alsoinfluence employee behavior. When employees knowthe basis on which they will be assessed, they are morelikely to meet their performance goals and more likelyto find their work rewarding. For government as awhole, it can be argued, performance recognition buildstrust and enhances credibility with taxpayers.

In accepting the need for performance measure-ment, it is important not to underestimate the problemsin its implementation. There is general agreement thatthe critical first step is to define performance. Tryingto measure performance demands that organizationsclearly articulate their objectives and then translatethem into measurable desired results, against whichperformance will be measured. This process itself canbe invaluable—the very act of defining the desiredresults has enormous power to focus the energies ofan organization. Unfortunately, this first step is oftenquite difficult for government agencies. Performanceis system specific and depends ultimately on the goalsof the wider budget management system. If the budgetmanagement system is a traditional one, performancewill be defined by measures of compliance and stew-ardship. On the other hand, if the budget managementsystem is outcomes focused, with success gauged interms of impact on society, performance will be de-fined by measures of the effectiveness of outputs pro-duced. The need to refer to the budget system’s basicorientation when defining performance is illustrated inFigure 1, which also attempts to define some specificperformance terminology.

Figure 1 describes in somewhat abstract terms a gov-ernment “production process.”38 The process com-

37“Organizations that measure the results of their work . . . findthat the information transforms them” (Osborne and Gaebler, 1992,p. 63). See also Kristensen, Groszyk, and Bühler (2000).

38The use of a simple production model to characterize the oper-ation of a government unit is usually regarded as the foundation forassessing performance and has a long tradition; see Brace and others(1980); Jackson and Palmer (1989); Boyne and Law (1991); andHyndman and Anderson (1995). However, others have questionedits adequacy, especially for complex public programs; see Osborneand others (1995). Others have been more critical of the funda-mental production model, for example, stressing the differential ac-cess to information of ruling groups, hence using “political models”of performance assessment (Pollitt, 1993) or “organizational” mod-els (Kotter and Heskett, 1992).

©International Monetary Fund. Not for Redistribution

Page 41: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV INTRODUCING A PERFORMANCE MANAGEMENT FRAMEWORK

mences from the point of acquiring inputs, progressesto the use of these inputs in a production process, andends at the production of outputs that have an outcomein the sense of meeting some defined objective of gov-ernment policy. Arising from this process are a rangeof possible indicators of performance. Traditional bud-get systems focus on inputs, the amount of resourcesactually used, usually expressed as the amount offunds or the number of employee years or both. Thekey concept is economy, or the aggregate control ofinput costs at the margin. In output-focused budget sys-tems, inputs are related to an agency’s output to produceindicators of efficiency or productivity.39 In outcomes-focused budget systems, an agency’s outputs are re-lated to the achievement of its ultimate goals,producing indicators of effectiveness. In such systems,costs are often compared with the final outcomes togive measures of cost-effectiveness, sometimes termedvalue-for-money indicators.

In performance budgeting, therefore, there are anumber of ways in which spending can fail to meetexpected performance, and in order to specify the curefor such performance failure, it is important to differ-entiate its source. Four types of performance failurecan be identified:

32

• Technical inefficiency: Resources are not beingemployed in the technically best way to produce agiven output or service level. The attainment oftechnical efficiency implies it is impossible to re-duce the physical level of any input without re-ducing the level of output. The failure to attain themaximum possible output level with given inputshas been referred to by Liebenstein as X-inefficiency(1966, p. 394), and must be resolved by improvingand changing the internal functioning of organiza-tions and the units that comprise them.

• Economic inefficiency: Resources are not beingemployed in the most economically efficient way,so that a higher return in the form of a higher pro-vision of service can be obtained without increas-ing costs by switching spending between resources.The attainment of economic or allocative efficiencyimplies that it is impossible to substitute one inputfor another without increasing total costs of a givenoutput or service level.

• Technical ineffectiveness: Expenditures are noteffective in the sense that although resources areallocated efficiently (both in a technical and aneconomic sense) to provide a certain service, theservice itself does not satisfy the objectives it wasdesigned to meet.

• Economic ineffectiveness: Expenditures can be ef-ficient (in the sense that resources are allocated toproduce the maximum output of a certain service

Figure 1. The Concept of Performance in Different Budget Systems

Performance Concepts

Economy

Costs

Inputs

Process

Outputs

Tech

nica

lly E

ffici

ent

Econ

omic

ally

Effi

cien

t

Valu

e fo

r M

oney

Effectiveness

Outcome

Outcomes-Focused BudgetsBased on what is to be achieved.

Emphasis on outcomes and impact.Outputs not constrained, input types constrained

(i.e., the effectiveness of outputs produced).

Output-Focused BudgetsBased on services delivered and products produced.

Emphasis on relating inputs to outputs.Outputs constrained, inputs not constrained

(i.e., efficiency).

Traditional BudgetsBased on standards and rules onhow inputs should be allocated.

Emphasis on economy.Input types constrained, outputs not constrained

(i.e., least-cost inputs).

Style of Budget Management

Perf

orm

ance

Bud

getin

g

39Where the ratio of input to output defines efficiency, and thereciprocal ratio of output to input defines productivity.

©International Monetary Fund. Not for Redistribution

Page 42: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Defining a Framework for Performance Measurement

33

at least cost) and effective (in the sense that theoutput has the desired outcome), but overall ef-fectiveness in the use of public resources could beincreased by cutting some expenditures and re-allocating the resources to other services, i.e., be-coming more allocatively effective.

Some of these performance concepts are illustratedin Figure 2, where curve HH′ shows all the differentcombinations of labor and other inputs to produce agiven output of health service, say, rural disease pre-vention. The line AB shows the available budget forthis service, which can buy OA units of labor, or OBunits of other inputs, or any combination along the lineAB. The slope of AB reflects the relative prices oflabor and other inputs. If the service delivery is oper-ating at point F, this is technically inefficient. All thepoints above HH′ use more of at least one input than isrequired. Technical efficiency can be increased by amore efficient allocation of resources to reach point D.However, even though at point D the service deliveryis technically efficient, this may not be economicallyefficient since costs can be reduced by reallocating themix of labor and other inputs to reach point E. Withthis combination of labor and other inputs, it is possi-ble to deliver the same service at least cost.

Suppose the objective is to prevent disease in ruralareas by providing the services of rural health clinics.While the service delivered may be technically andeconomically efficient at point E, it may not be techni-cally effective in the sense it may not actually preventdisease in rural areas. For example, the rural health clin-ics may not be located within easy access of disease-ridden areas. This would be a technical flaw, to beresolved by health specialists and not economists.However, even if service delivery were technically ef-

fective in eliminating disease in the rural population,this does not mean it is economically effective. Itmay be possible, for example, to attain the same level of disease prevention in a more cost-effective way,through less costly improvements to the rural watersupply. In this case, the government would attain bet-ter value for its money.40

There are variations in the extent to which differentdimensions of performance can be improved throughbetter systems of budget management. As indicated,technical inefficiency is symptomatic of more generalshortcomings in the internal organization and proce-dures of the producing unit. This type of inefficiency,Liebenstein argues, is likely to be the greatest sourceof inefficiency, particularly in the public sector. It isalso the one that is more amenable to improved budgetmanagement. To some extent, improved budget pro-cedures can promote economic efficiency. At the sametime, the ability of improved budget management tocompensate for a lack of effectiveness of expenditureprograms is necessarily circumscribed. The task oflinking resource allocations to objectives more clearlyfalls in the realm of policy, and the role of improvedbudget management tends to be supportive—to facili-tate and improve the process of making such policychoices through greater transparency and the provisionof timely and relevant performance information.

Defining a Framework forPerformance Measurement

Moving from the above, rather schematic descriptionof the dimensions of performance to the practicalproblem of developing a framework for measuringperformance in government requires addressing somekey issues, which are discussed in turn: output versusoutcome; process indicators; the quality dimension;and relating performance to inputs.

Output Versus Outcome

There is an ongoing debate about the value of focusingon outputs or outcomes (i.e., the impact of the outputs),and the term performance budgeting is used to refer toboth output- and outcome-focused budget systems.The desired properties of these concepts are describedin Boxes 16 and 17, respectively. The debate over out-

40From this brief review of the main concepts and issues in theperformance literature, it is perhaps possible to find sympathy forthe conclusion that “the notion of performance—often bereft ofnormative standards, invariably full of ambiguity—is, in theory andpractice, both contestable and complex” (Carter, Klein, and Day,1992, p. 50).

Figure 2. Technical and Economic Efficiency

LaborInputs

Line AB = available budgetD = technical efficiencyE = economic efficiencyF = service delivery

OtherInputs

HF

D

E

A

O B

H�(100 units ofhealth services)

©International Monetary Fund. Not for Redistribution

Page 43: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV INTRODUCING A PERFORMANCE MANAGEMENT FRAMEWORK

put versus outcome tends to be at two levels. First,there are the evident practical measurement problems.Many outcomes are difficult to measure directly (e.g.,greater national security) or are complex, for example,involving interlinkages among a number of differentprograms and subprograms (e.g., lower morbidity rates).Second, at a higher level, there is controversy over whatmanagers should be held accountable for. On practicalgrounds, an output is generally under an agency’s con-trol, while an ultimate outcome is often determined byexternal factors that are unpredictable. Also, observedoutcomes can be interpreted in different ways and canbe considered to include side effects, whether intendedor not. Thus the Office of Management and Budget(OMB) in the United States focuses on intended out-comes,41 while other countries such as Australia reportex post on the total outcome, intended and unintended.This has led some to differentiate outcome, which is de-fined by intended effects, from impact, which includesthe total effect of the agency’s actions. Others differen-tiate among different levels of outcome or, like Hatry(1999), differentiate between intermediate and end out-comes (pp. 15ff). For example, some countries includ-ing the United States and the United Kingdom make adistinction between different levels of outcomes de-pending on the time horizon.

34

Although performance in government is clearlysubject to different interpretations, it does appear thatin the OECD countries there is increasing recognitionthat the output approach has limitations and may de-flect the attention of agencies away from the impact oftheir programs. From the government viewpoint, im-pact is the more relevant concept.42 As a consequence,more and more OECD member countries are adoptingan outcome-focused approach, though many precededthis move with extensive use of output measures.43

In preparing the budget, it is necessary to defineand agree on performance targets for each outputmeasure. These should be chosen selectively, with pri-

Box 16. Desirable Properties of Outputs

Outputs are generally under an agency’s control tosome extent. To be effective for performance budgeting,outputs to be measured should—• be a good or service provided to individuals/organiza-

tions external to the agency.• be clearly identified and described.• be for final use and not for an internal process or an in-

termediate output.• contribute to the achievement of planned outcomes.• be under the control (directly or indirectly) of the

agency.• generate information on attributes of performance—

price, quantity, and quality.• generate information that is a basis for performance

comparisons over time or with other actual or poten-tial providers.

Example:Output: policy advicePerformance measure: briefs or submissions preparedQuantity: numberQuality: satisfaction of minister and staff; other assur-

ance tests

Box 17. Desirable Properties of Outcomes

Outcomes are generally complex, are determined bya number of factors, are difficult to measure, and involveelements outside an agency’s control. To be effectivefor performance budgeting, outcomes to be measuredshould—• adequately reflect the government’s objectives and pri-

orities.• be indicated by the impact on the community.• be differentiated from the agency’s strategies to which

they contribute.• clearly identify the target groups, if so focused.• be achievable in the specified time frame.• be suitable for monitoring and progress assessment.• be the result of an identifiable causal link with the

agency’s output.• be clearly defined and described so as to be easily re-

ported externally.

Example:Outcome: Ensuring young homeless people have access

to appropriate accommodation.Performance target: 90 percent of young homeless peo-

ple have access to appropriate accommodation within24 hours.

Operationalized by clearly identified target group,definition of “appropriate housing,” and causal linkbetween agency action, such as assistance through a subsidy.

41For a full description of the Government Performance and Re-sults Act (GPRA) performance management framework and itsspecific terminology, see Groszyk (2001) and Anderson (2004).

42See Kristensen, Groszyk, Buhler (2002, p. 1). At the same time,several writers, notably Pollitt (1986), suggest that the emphasis onoutcomes over outputs (and the broader emphasis on economy andefficiency rather than effectiveness) may reflect the political inter-est of a government that is primarily concerned with cost-cuttingrather than performance evaluation.

43“Australia, Netherlands, and New Zealand began by concentrat-ing on outputs and are now moving to an outcomes approach. Francerecently passed a law that requires the production of outputs and out-comes in budget documentation for the majority of programmes.”(OECD, 2004, p. 7) It should be noted that while Australian statesbegan with a focus on outputs, the federal government specifically re-jected output budgeting, preferring to move directly to outcomes.

©International Monetary Fund. Not for Redistribution

Page 44: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Defining a Framework for Performance Measurement

35

ority on clarity in their interpretation, so that they pro-vide an undisputed indicator of the degree of output(and even outcome) achieved in any time period. Theapproach is to specify the output as a rate and to spec-ify the frequency with which the target will be met.Additionally, it is useful to set out the information onhow each target will be estimated, perhaps based onactual performance data for previous periods, as wellas to identify the implications for the cost of the outputif there is a significant divergence between expectedand actual target achievement. The latter will providea platform from which the operating unit can analyzeand explain any variances in performance.44 TheUnited Kingdom, in designing its Public Sector Agree-ments (see Box 20), has emphasized the importance ofsetting meaningful targets alongside agencies’ bud-gets. This, it is argued, assists managers in improvingtheir agency’s performance, allows the center to iden-tify policies and processes that work, and enhancespublic accountability (Hill, 2004). To use the prevalentacronym, these targets should be “SMART”: specific,measurable, achievable, relevant, and timed. However,with the shift in focus from outputs to outcomes, thetechnical problems of measurement have increased,and setting targets for some outcomes has become in-herently more difficult.45

Process Indicators

It is also important not to neglect the process that cre-ates outputs and outcomes. Indicators of performance,such as workload, throughput, and work rate, give im-portant measures of the technical efficiency of anagency’s operations. Workload indicators include theamount of work that comes into a program and workin progress but not yet completed.46 While amounts ofwork by themselves are not outputs or outcomes,workload data can be used as inputs to produce outputdata: for example, the amount of work not completedcan be a proxy for delays in service to customers, aquality dimension of output (discussed below). Thereare some activities for which it is difficult to measureoutputs or outcomes but for which process-related

measures may be useful in assessing performance.The major gains to be made in budget system reformare often made in improving processes and eliminat-ing the so-called invisible X-inefficiency in govern-ment operations.

The Quality Dimension

Quality is an important component of the efficiencyof service provision, where efficiency is measured bythe ratio of outputs to inputs and therefore can be improved by reducing the quality of the output. Ifoutcomes are tracked, a more accurate indicator of ef-ficiency becomes possible by including a quality di-mension, which can be broken down into variouscharacteristics, as shown in Box 18. Some regardquality characteristics to be an output, others (Hatry,1999) designate them as a special type of intermedi-ate outcome, since quality characteristics indicatehow well the service was delivered but do not indi-cate the results of the service after delivery. All agree,however, that quality of service is an important di-mension to measure and to track, although the modeof tracking is often disputed. Some governmentshave questioned the value of internal measures ofquality (e.g., Denmark) and have emphasized insteadclient surveys of the level of service satisfaction.Both approaches are important: client surveys cer-tainly provide an important “reality check” on formalindicators of quality that are generated internally.Moreover, an important question remains: since thereare many dimensions to quality, who decides whatqualitative characteristics are important for a partic-ular service?

Relating Performance to Inputs

Ultimately, resource allocation decisions require thatoutput and outcome measures be related to costs. Thisis often difficult for four reasons. First, accountingsystems are not set up to bring cost data together on aprogram basis and hence data on inputs is not readily

44For a much fuller discussion, with examples, see OECD (2000).45Henderson (2004) contrasts a straightforward target such as “re-

duce substantially the mortality rates by 2010 from heart disease byat least 40% in people under 75; from cancer by at least 20% in peo-ple under 75 . . .” (Department of Health); with, for example, “Im-prove effectiveness of the UK contribution to conflict preventionand management as demonstrated by a reduction in the number ofpeople whose lives are affected by violent conflict and a reductionin potential sources of future conflict . . .” (FCO, Ministry of De-fence, DfID).

46Performance measurement is even applicable to internal sup-port services, but the outcomes from internal support occur withinthe organization, and it is impossible to estimate the impact theseservices have on outcomes of external services.

Box 18. Typical Service QualityCharacteristics

• Timeliness in service provision.• Accessibility and convenience.• Accuracy of the assistance.• Courtesy in service delivery.• Adequacy of information dissemination to potential

users.• Condition and safety of facilities.• Customer satisfaction.

Source: Hatry (1999, p. 17).

©International Monetary Fund. Not for Redistribution

Page 45: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV INTRODUCING A PERFORMANCE MANAGEMENT FRAMEWORK

available on a timely basis.47 Often in emerging econ-omies, especially in transitional economies, this isaggravated by the fact that cost accounting systems arenot developed and there is little expertise to carry out thiswork.48 Second, the work is often not straightforward.Allocating costs to programs can be a major problem inagencies with large central overhead budgets and for ac-tivities involving delivery of more than one service.49

36

Third, there may be limitations in the accounting sys-tem, especially if it is cash based, that prevent the fullcosts of capital from being recorded (i.e., if deprecia-tion is not included).50 Fourth, there is a need to focuson marginal rather than average cost, and this is typicallymore difficult to measure. Confusing average efficiencywith marginal efficiency may cause extra resources to beallocated to programs with proven average efficiency butfor which additional resources will have little impact onoutput. These practical costing problems are importantbecause feasible performance measures are those thatcan be established at reasonable cost, which may well de-pend on the ready availability of reliable cost data. How-ever, despite all such concerns, it is clear that the OECDcountries have placed considerable emphasis on perfor-mance information in their budget processes (Box 19).

Experience in Using Performance Measures

Performance measurement is a key tool in the processof improving the delivery of public services, but itshould be viewed not as an end in itself but as part of a

Box 19. OECD Practices: Use of Performance Information

Percentage of OECD Countries

1. Nonfinancial performance data routinely included in budget documentation:a. All programs 32b. For less than 25% of programs 20

2. Performance information includes performance targets:a. All programs 21b. For less than 25% of programs 25

3. Type of performance targets:a. Mostly outcomes 27b. Combination of outputs and outcomes in most programs 27

4. Process by which performance targets are set for ministries:a. Agreed as part of the budget process 52b. Decided by ministry/department 48

5. Formal responsibility for achieving the targets:a. The relevant minister 60b. The relevant civil servant 16

6. Regular monitoring of performance against targets:a. Internally in relevant ministry 56b. By the MoF 24c. By the legislature 20

7. Reporting of actual performance against targets for most programs 508. Performance results made available to the public:

a. Government-wide report on performance 24b. Published by individual ministries 44c. As part of other government-wide documents 24

Source: OECD (2003, Tables 5.4.a, 5.4.b).

47The problem of the measurement and allocation of costs is nec-essarily limited when government units use cash accounting princi-ples. In the United Kingdom, one of the problems encountered inenforcing accountability in the Next Steps Agencies, created in the late1980s and early 1990s, was the lack of information on unit costs. Thisarose from their lack of commercial-style accounts as well as from theundeveloped nature of their costing systems. A related problem wasthat the measure used as the cost object was rarely the ultimate result,but rather was an intermediate output measure related to activity. Afull discussion is contained in Pendlebury, Jones, and Karbhari (1994).

48Discussed more fully in Section V.49It has also been argued that the increasing complexity of gov-

ernment operations has significantly contributed to the problem.Traditional cost accounting is adequate when processes are simple.However, as technology advances in scale and complexity, the costprofile of government organizations becomes significantly morecomplicated. Costs that traditionally have been considered over-heads now represent activities critical to the delivery of governmentservices. It is increasingly difficult to associate these costs directlywith individual programs or customers. See Gearhart (1999). 50Discussed more fully in Section VI.

©International Monetary Fund. Not for Redistribution

Page 46: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Experience in Using Performance Measures

37

wider performance management reform process. Per-formance measurement has become so popular it hastended to lead the reform, and while many benefits flowfrom the mere act of trying to measure performance,to be fully effective, performance measurement mustbe integrated into a performance management system.Failure to move from performance measurement toperformance management is cause for some concernbecause of the dangers of overreliance on performancemeasures; inappropriate measures; misuse and mis-interpretation; and information overload and lack ofselectivity.

The Danger of Overreliance on Performance Measures

Performance measurement has at least three limita-tions.51 First, performance data do not by themselvestell why the outcomes occurred—that is, the extentto which the program caused the measured results—because there is unlikely to be any practical counter-factual. A performance measure assumes some causalitybetween a program’s activities and its outcomes, butone which can only be confirmed ex post through aproperly designed evaluation study.52 Second, someoutcomes cannot be measured directly. The most ob-vious example is success in preventing undesirableevents. Third, the information provided by perfor-mance measures is just part of the information that isneeded by managers and elected officials to make re-source allocation decisions. Performance informationmay provide a clearer understanding of what is ex-pected and what has been accomplished, but othersupporting information is required. For example, thecosts of providing goods and services by alternativeapproaches can be derived from agency financial andresource information or from benchmarked cost data.

In sum, it might be best to view performance mea-surement as necessary but not sufficient to determinewhat should be done. This implies that managers can-not rely solely on performance measurement for deci-sion making.53 It is interesting to note that the privatesector is moving away from a narrow focus on out-comes measurement toward a broader approach, as ev-

idenced in the “balanced scorecard” (Kaplan and Nor-ton, 1996) which has also attracted some interest withinOECD governments.54

The Danger of Inappropriate Measures

Measuring performance is not an easy job, and it getseven more difficult as measurement moves from pro-gram outputs to program outcomes. In this paper, therehas been a deliberate use of the term performance mea-surement rather than performance indicator. Whendealing with processes and outputs, there is greater as-surance over consistency in the interpretation of themeasures of performance. When one deals with out-comes, however, direct measures are often difficult,hence the measures can only indicate the outcomerather than directly measure it. Often, it takes morethan one indicator to adequately capture an outcome,further complicating interpretation.55 Not surprisingly,performance indicators tend to be subject to a greaterdegree of interpretation and hence are more open toabuse as well as to the danger of being inappropriatefor the purpose at hand. A concern about the use of per-formance measures is that the measures displace theactual outcomes to become an agency’s objectives.This “goal displacement” can often lead to emphasiz-ing the wrong activities (e.g., those that are easier tomeasure) and to an agency’s energies being spent inimproving “the numbers” without improving actualoutcomes.56 An obvious example is the usual bias offocusing performance measurement on the short term,for ease of measurement, despite the distinct dangerthat short-run beneficial outcomes may result in in-creased costs over the longer run, with an undesirableshifting of costs into the future.

The Danger of Misuse and Misinterpretation

Performance measures require careful interpretationby those with adequate knowledge of the different

51See Hatry (1999, pp. 4ff); Perrin (1998, p. 374).52“Performance indicators are no substitute for the independent,

in-depth qualitative examination of the impact of policies whichevaluations can provide” (OECD, 2004, p. 15). Due to the costs in-volved, these necessarily have to be used sparingly and guided bysome cost-benefit principle.

53As a consequence, it is often recommended that any ongoingmonitoring of performance be supplemented with periodic programevaluation or reviews. That is, “an effective performance informa-tion system will include an appropriate mix of on-going perfor-mance information and periodic evaluations” (Australian NationalAudit Office, 1996, p. 3). The latter allows a wider range of infor-mation and stakeholder perceptions. From the U.S. perspective, seeJoyce (1993, p. 10).

54The balanced scorecard looks at a wide range of measures, in-cluding difficult-to-measure factors such as a company’s focus oninnovation and learning. For example, Osborne and others (1995),would include as an important dimension of performance in socialprograms indicators of the process of “learning lessons” and “em-powerment” of communities (pp. 31ff).

55Perhaps not surprisingly, it is often found that the developmentof performance indicators has been fastest in the least problematicareas where government units have clearly defined and rather narrowfunctions, and that the problems of measuring performance increasewith the complexity of government activities. See Carter, Klein, andDay (1992).

56The PEM literature is replete with examples of poor or dys-functional performance measures: a tax office measuring the costper application of revenue collected, which might encourage the of-fice to leave difficult cases aside; a hospital using cost per occupiedpatient bed, which could encourage managers to retain patients toensure no beds are unoccupied; rating a school’s performance basedon examination success rates, which might lead to schools discour-aging low performers from joining the school.

©International Monetary Fund. Not for Redistribution

Page 47: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV INTRODUCING A PERFORMANCE MANAGEMENT FRAMEWORK

factors that affect the measures. This need for inter-pretation brings its own dangers. No matter howclearly defined, performance indicators are invariablyrecorded and interpreted in very different ways bydifferent people. Moreover, when people feel the fu-ture of a program is dependent on an indicator, a pos-itive bias comes into play—they inevitably interpretthe indicator in the way that is most favorable to theagency. Perhaps of more concern is the danger thatperformance measurement is introduced without themanagers’ commitment to the new performance man-agement system, which often occurs when such sys-tems are imposed from above or from the outside(e.g., by the legislature). Experience has shown thatwhen managers and staff are required to collect andcompile performance measures that they deem in-appropriate or of little use to the agency, they becomecynical about performance measurement and morelikely to seek ways to use it to their own ends. Often,the greater the pressure from above to use performancemeasures, the higher the stakes become and the greaterthe incentives for people to identify self-serving per-formance indicators as well as to report misleadingperformance data. Hence the need to provide managerswith incentives to buy in to the new performance mea-surement system, for example, by offering increasedmanagerial flexibility.

The Danger of Information Overload and Lack of Selectivity

Performance indicators can be irrelevant for decisionmaking, even if they are accurate. The essence of per-formance measurement is to reduce the multidimen-sional aspects of the program’s outcome to a fewmeasurable indicators.57 This often leads to oversim-plification and to counterpressure to employ multiplemeasures of performance. However, from a decision-making viewpoint, this may not be a solution—moredata makes decisions more informed but not necessar-ily easier.58 This may also result in confusion amongdifferent types of indicators and their use. For exam-ple, the use of a weighted average of different mea-sures may lead to misleading aggregate indicators ascritical subgroup differences are lost in the aggregationprocess.

38

Guidelines for Using Performance Measures

How to guard against the above dangers? The experi-ence of practitioners in performance measurementpoints to some guidelines.

• Aim for clarity in purpose. It is necessary to besure about who will use the performance informa-tion, how they will use it, and why they will use it.The potential users range from the program man-ager, to the central evaluator in the MoF, to themember of a legislative watchdog committee, tothe final consumer of the government service.Each of these potential users has different objec-tives, for example, making better resource alloca-tion decisions, improving services, increasingaccountability. Measures should only be chosenafter users are identified and their objectives prop-erly defined. The ultimate aim should be to helpthe users reach more informed conclusions and tomake better decisions about service performance.

• Focus on core information. Once the purpose hasbeen clarified, there should be a conscious effort toprioritize among different performance measuresto avoid information overload. It is usually recom-mended that in the first instance the focus shouldbe on the priorities of the entity, i.e., its core ob-jectives and service areas. Often it takes substan-tial time and effort to establish success criteria forthese priorities and objectives without wideningthe scope of performance measurement.

• Align with the practical needs of the agency. Makeperformance information as relevant as possible tothe agency. This requires that performance mea-surement be aligned with the objective-setting pro-cedures and the performance review process ofthe agency. In this way, those who collect andprocess the data can appreciate and relate its use tothe agency’s procedures. Of course, this does notmean ignoring the multiple uses of performance in-formation; there should be links among the perfor-mance measures used at an operational level, thoseused by agency’s management, and those used byhigher-level government officials, say in the MoF.

• Balance the perspective on performance. Recog-nize that different activities create different needs forperformance measurement. Narrow program activ-ities perhaps can be covered by a very few mea-sures, but complex program activities may need awide range of performance measures balancedover the different aspects of the services. It may bedesirable to have a mix of internally generated andexternal measures. Obviously, meeting this needmay involve trade-offs with other criteria, such asfocus and alignment.

• Review the performance measures. Selected per-formance measures should be regularly reviewed

57“Performance measurement should aim at developing a limitednumber of well-chosen, stable measures, so that a track record of anorganization’s performance over time can be built up. This does notmean that performance measures are defined once and for all; theymay need to be modified to take into account changes in the man-agerial context and the environment in which the organization ex-ists.” (OECD, 1994, p. 14)

58“There are limits on how much information decision-makerscan make good use of: people have ‘bounded rationality’ and so doorganizations” (OECD, 2004, p. 4).

©International Monetary Fund. Not for Redistribution

Page 48: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Establishing a Performance Information System

39

and updated to reflect changes in priorities andshifts in the focus of public policy over time. Agencyobjectives change, priorities among objectives change,different users emerge, and hence the required bal-ance of performance information necessarily changes.A performance measurement system that is static isnot likely to fulfill its function as part of a perfor-mance management system.

• Ensure the robustness of basic information. Per-haps most important, performance measuresshould be based on reliable and timely data. Thebasic raw data should be robust, in the sense ofbeing derived in a way that is verifiable, free frombias, and preferably comparable over time andamong different organizations. Data should alsobe capable of being derived in a time frame com-patible with the needs of decision makers.

Establishing a PerformanceInformation System

If performance measurement is to be more than simplyan add-on to the traditional budget process, it shouldbecome an essential part of the process of performancebudgeting. In turn, performance budgeting must beviewed as an integrated method of allocating re-sources. The central idea is to link resource allocationexplicitly to outputs or outcome—performance—andthis should be recognized as involving more than just

budgeting. Rather, it encompasses the entire budgetmanagement process in supporting and demandingbetter performance: budget planning, budget execu-tion, reporting, and auditing. In this way, performancemeasurement must be fully integrated into a perfor-mance management system, which in turn entails es-tablishing a performance information system. Box 20outlines the U.K. experience with its Public ServiceAgreements, which highlights the importance of sucha system.

An effective performance information system re-quires that managers develop their performance in-formation to ensure the following management needsare met:59

• Performance measures should be clearly linked tointended objectives and should enable ready as-sessment of performance in terms of effectiveness,efficiency, and service quality. Needless to say,they must be feasible, accurate, and derived in acost-effective way.

• As far as possible, a core set of performance datashould be identified for routine collection. Thegoals should be to minimize the cost of perfor-mance management and ensure that performance

Box 20. The U.K. Government’s Public Service Agreement (PSA) Framework

Beginning in 1998, along with other reforms, the PSA setmeasurable targets for a full range of government objectives.Departmental PSAs include targets on inputs, outputs, andoutcomes. Since its inception, the number has been substan-tially reduced, from 387 in 1998 to 130 in 2002, and a pro-gressively higher proportion have been outcome focused.

Objectives of the PSA

• Explain what departments plan to deliver.• Set out national targets.• Reflect the government’s key priorities in terms of out-

comes.• Represent agreement between the government and the

public.

Five features

• Begin with an aim that sets at a high level the role of thedepartment.

• Objectives set out in broad terms what the department in-tends to achieve.

• Performance targets set clear outcome-focused goals forpriority objectives.

59The need to amend management information systems to meetthe needs of managers working in a performance environment isdiscussed in Section V.

• Establish a value-for-money target for each department,focused on improving cost-effectiveness.

• State who is responsible for the delivery of these targets(usually the relevant secretary of state).

Supplemented by

• Technical Notes: detailed documents that set outhow the PSA targets are defined, measured, and interpreted.

• Service Delivery Agreements (SDAs): These were in-troduced in 2000, with a high-level summary of deliveryplans to meet targets in 2004; replaced by Service Deliv-ery Plans in 2003. More detailed department plans focuson actions required to meet targets, role of key deliveryagents, resources required, risks, how interim progress isto be measured and monitored, and strategies with for-ward projections of progress.

• Department Reports: Accountability is enforced by de-partments reporting annually in the spring on their spend-ing and performance against PSA targets. Recently,Autumn Performance Reports were also introduced. AWeb site has been created as a single portal to all depart-ment performance documents.Source: H.M. Treasury (2004).

©International Monetary Fund. Not for Redistribution

Page 49: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV INTRODUCING A PERFORMANCE MANAGEMENT FRAMEWORK

information is relevant and understandable to theorganization and its stakeholders.

• The continued appropriateness of performance in-formation should be regularly assessed, consider-ing such factors as relevance, cost, value, andusefulness to decision makers.

• Responsibilities for performance measurementand reporting need to be clearly defined and un-derstood, including whether services are deliveredby the agency, outsourced, or produced by otherthird-party arrangements.

• Monitoring and periodic evaluation of perfor-mance should be balanced with other operatingdemands and used appropriately. It should be ex-pected that performance will be monitored on acontinuing basis and complemented by periodicevaluation, generally on at least a five-year cycleand preferably on a three-year cycle.

• Performance management activities should be in-tegrated into business planning processes that canbe fully supported by the performance manage-ment information system.

Two requirements of such an information systemare worth emphasizing: the need for a reporting struc-ture and the need to ensure data quality.

Reporting Structure

Performance management requires not just measur-ing performance but reporting on it. From the outset,there is a need to consider the structure of the data cap-ture and the ultimate presentation of the data as man-agement information to different levels of users whocan be expected to have different reporting needs. Un-fortunately, performance measures are often designedand used by agency managers to meet their own needs,without regard for the information needs of otherusers, such as citizens or their elected representatives.To encourage accountability and better governance,citizen participation is to be encouraged. Lack of citi-zen involvement can undermine the value of perfor-mance measurement by minimizing its importance inthe eyes of elected officials.

Some general principles about reporting on perfor-mance are indicated in Box 21. The performance mea-sures should be reported regularly and consistentlyand should allow performance to be measured againstthe standards chosen. This means data could cover:actual performance compared to the plan, actual per-formance compared to a predetermined standard, ac-tual performance compared to actual performance ofpeers, or actual performance compared to perfor-mance in previous periods. There are great benefits inbenchmarking in this way, especially in allowing ex-ternal comparisons to some predetermined standardof good practice or to a relative benchmark such aspeer group activity.

40

A System of Data Verification

Performance measurement must be credible. It is notenough to introduce performance measurement with-out the added assurance of internal controls and dataverification.60 When elected officials and adminis-trators trust the reliability of performance informa-tion, they are more likely to use that information fordecision-making purposes. The value of performanceinformation is undermined if its accuracy cannot beverified. Given the need, who will carry out this dataverification? The most obvious candidates are inter-nal audit staffs or external auditors, and typically acooperative effort is recommended. This verificationshould be performed at least annually on a rotationalbasis so that all programs are covered in, say, a three-year cycle.

Developing a PerformanceManagement System

It takes time to introduce a comprehensive system ofperformance management. The ultimate objective is toput in place a system to match costs with activities, tomeasure performance of these activities, to developstandards of performance, and to compare costs andperformance levels with those standards. The challengeof this approach is to link performance information tothe budget process and the allocation of resources.International experience shows that until this connec-tion occurs, performance can be merely a regular re-porting requirement, not directly relevant to day-to-daymanagement and budgeting.

Box 21. Key Principles for Reporting Performance Information

• Be Open: Feed back the results and explain the reasonsfor collection of data and the use(s) to which it will beput.

• Be Selective: Do not report all measures to everyone,so as to avoid overloading users with information thatis not relevant to them.

• Be Focused: When a specific issue is under review, itis necessary to report only the measures relevant to thatissue.

• Be Proactive: Take action to indicate when a responseor an action is required as a result of the informationbeing provided.

• Be Pragmatic: Concentrate on what can be influenced.• Be Reasonable: Take or suggest action that is reason-

able in the circumstances.

60This issue of quality control in performance measurement isdealt with more fully in Wholey (1999).

©International Monetary Fund. Not for Redistribution

Page 50: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Developing a Performance Management System

41

How to develop such a performance managementsystem? Six key steps are described here, all of whichrequire developing and strengthening performancemeasurement capacity.

1. Clarify the objectives of programs and hence bet-ter define how their performance will be judged.61

2. Provide a stronger link between inputs into theproduction process and monitoring the achieve-ment of outcomes.

3. Present information on planned performance(usually in budget documents) and actual perfor-mance (in documents such as annual reports) ona consistent basis.

4. Make performance information relevant in re-source allocation decisions.

5. Provide incentives for managers to use perfor-mance measures in making day-to-day manage-ment decisions.

6. Monitor the performance of management inreaching their program objectives to hold themaccountable and to allow for corrective action.

Step One: Improve the Definition of Programs and Their Objectives

As argued in Section III, well-defined programsmust be anchored in a strategic plan, which should in-corporate the aspects indicated in Box 22 and shouldbe formulated for implementing units.

The term of the strategic plan should be at least fiveyears, consistent with an MTBF, and should be period-ically updated. As indicated in Section III, annual im-plementation plans, or operating plans, should be basedon the strategic plan to provide a direct link between thelong-term goals of the strategic plan and the goals iden-tified in budgets and to serve as a point of reference forannual progress evaluations. The main features of anoperating plan are described in Box 23. In the initial pe-riod of a performance management framework, theremay be a role for a central advice unit to assist agenciesin making the required changes outlined here.

Step Two: Provide a Stronger Link betweenBudgetary Inputs and Program Outcomes

Since the budget remains the chief motivating forcein determining the allocation of resources to meet theobjectives of program management, the performancetargets should be matched with full annual budgetarycosts. This provides not just the information but the in-centives for budget managers to make efficient trade-offs in allocating resources to meet the program targets.

Each agency should be required to identify the link-ages between financial and nonfinancial inputs to theirproduction of goods and services and the outcomesthey have identified, or are in the process of identify-ing in redefining their programs. Examining out-comes associated with programs in isolation from thedirect inputs required for their daily operation re-duces the relevance of performance measures. Agen-cies do not necessarily have to identify all the outputsof the production process up front; this will comelater as managers gain more experience in measuringperformance.

The linking of inputs to program outcomes shouldbe a joint exercise of the MoF and the staff of the rel-evant ministry because only they have the detailedknowledge to provide this essential connection. Theoutcomes associated with the programs should then beagreed with the head of that ministry.

61See the interesting discussion on the importance of defining ob-jectives as one of the main obstacles found when measuring per-formance of U.S. federal agencies (U.S. Congressional BudgetOffice, 1993).

Box 22. Elements of the Strategic Plan

• Comprehensive description of the agency’s mission,including the organization’s main functions and oper-ations.

• General or strategic goals and objectives for the orga-nization’s main functions and operations.

• Description of the guidelines to be followed to attainthe goals and objectives.

• Identification of external factors crucial to the organi-zation which are beyond its control and which couldhave a significant impact on the attainment of generalgoals and objectives.

• Description of program evaluation.• Procedures used to define or revise general goals and

objectives.

Box 23. The Operating Plan

The operating plan should be formulated annuallyand should include the following:• Operating objectives defining the targeted level of pro-

gram implementation, i.e., outputs.• Output goals expressed in an objective, quantifiable,

and measurable manner.• Description of how annual goals or operating objec-

tives will relate to the general goals of the strategicplan.

• Indicators for use in assessing the value of relevantproducts, levels of service, and the results of each pro-gram activity.

• Bases for comparing program results with establishedimplementation targets.

• Methods to be used in checking and validating themeasurements obtained.

©International Monetary Fund. Not for Redistribution

Page 51: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

IV INTRODUCING A PERFORMANCE MANAGEMENT FRAMEWORK

Step Three: Make Performance Information Relevant

It is common to find program structures with definedmeasures of performance that have little impact on theallocation of resources. Goals or objectives may be set,but this often occurs within the service ministry, afterthe budget has been passed. As a result, there is littleimpact on the allocation of resources among majorfunctional areas of the budget, such as education andhealth. Furthermore, goals or objectives within a min-istry often only impact the allocation of resources atthe margin, with the bulk of resources allocated with-out regard to performance.

To force a break with this approach, a phased intro-duction of the new system is recommended. In the firststage, major ministries and service agencies shouldprovide an agreed set of outcomes of their programssix months prior to the processing of the budget, withclear linkages to the associated financial and non-financial inputs. This allows the MoF to examine thetotal level of resourcing in light of the proposed out-comes of the programs and to feed that informationinto the formulation of the budget. Although this datamay not have a significant impact on the determinationof the budget in the absence of associated performancemeasures, this is an important transitory phase. In thenext stage, all agencies should provide similar informa-tion six months prior to the budget, together with a fullset of performance measures that clearly shows howeach agency has performed during the past 12 months.This would be the first point at which the budget wouldreflect the relative allocation among ministries and ser-vice agencies and would be directly impacted by theirperformance.62

Step Four: Present Performance Informationon a Consistent Basis

The process of publicly releasing performance in-formation six months prior to the budget must be com-prehensive, so that all agencies release information ona full set of programs and associated outcomes, withclear linkages between inputs and performance indi-cators. The central budget documentation should in-clude a summary of performance information publishedby agencies and a clear indication of where resourcinghas been modified based on good or poor performance.It is recommended that the publication of such infor-

42

mation be an annual process to raise the accountabil-ity of agencies and their managers for the efficient andeffective use of public funds. This accountabilityshould not, however, be restricted to an annual report-ing process. All documents released publicly by agen-cies about their operations should include performancemeasures from the date they begin to be used. Along-side this regular reporting, the results of periodicevaluation should also be published. Indeed, it is rec-ommended that all agencies conduct periodic reviewsof their programs, as previously indicated, preferablyso that all programs are reviewed every three years.

Step Five: Provide Incentives for Managers to Use Performance Information

To clearly establish performance targets as a per-manent feature of public sector management, a pro-cess should be put in place for concluding performanceagreements with agencies. Agencies’ performanceagreements should be included in their business plans,commencing with the major ministries and serviceagencies and then progressively including all agen-cies. However, incentives do not work if they arepurely negative, for example, budget cuts. There mustalso be an appropriate set of rewards for good man-agers. Those managers that are innovative in the pro-vision of goods and services and at the same timeprovide savings to the budget should be identified andappropriately compensated. This may take the form ofallowing them more flexibility in decision making.Initiatives in the introduction of performance-basedpay should also be encouraged. Employees should begiven the opportunity to enter into workplace agree-ments with management, whereby performance bonusesare paid for an agreed level of performance or achieve-ment of some operational or strategic objectives. Thiscould extend to individual employees, reducing theirbase rate of pay and replacing it with performance-based pay.

Step Six: Develop a System to Monitor Program Management

In this new performance management system, theMoF should define the general features of the moni-toring system to be adopted by the units and serviceagencies that manage budget programs and should alsooutline the information to be reported periodically.Such data include progress in attaining objectives; costsincurred; and the most significant physical and finan-cial discrepancies vis-à-vis indicator-based estimates.There should also be an explanation and examinationof the causes of discrepancies, classified as internal ifattributable to management, or external in all othercases. Ideally, this should be accompanied by a de-scription of corrective measures to be adopted or pro-

62During this stage, ministries could also use these performancemeasures to determine the allocation of up to 10 percent of re-sourcing among service areas. It may also be useful to recommendthat the percentage of resource allocation subject to satisfactory per-formance of managers be increased progressively by at least 5 per-cent each year, in order to provide a clear message to managers thatperformance information is important and relevant in deciding onthe allocation of budget resources.

©International Monetary Fund. Not for Redistribution

Page 52: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Concluding Remarks

43

posed, as appropriate. Performance measures shouldhave clearly defined characteristics to ensure the pro-vision of sufficient and relevant information for takingmanagement decisions. Some current OECD practicesin the use of performance measurement in budget de-cision making are summarized in Box 24.

Concluding Remarks

These steps toward a performance managementsystem are directed to setting up a clear accountabil-ity framework for budget managers that will lie atthe heart of the new performance budgeting model.This accountability framework is based on these keyelements:

• a clear ex ante specification of the performanceexpected of each agency head;

• agreed ex ante arrangements for the collection ofall the information required to assess performance;

• incentives and sanctions to encourage agencyheads to act in the government’s interests;

• a clear performance assessment process involvingex post reporting of actual performance againstthe initial specification; and

• devolution of decision-making authority to giveagency heads the degree of managerial autonomythey need to achieve the tasks assigned to them.

This reformulated accountability arrangement rep-resents a fundamental change toward a more devolvedsystem of budget management whose effects are quitefar-reaching. However, for this reorientation to befully effective requires that certain preconditions aremet, and its implementation may need to be facilitatedby a reformulated institutional structure. Discussion ofthese complementary changes are the subject of the re-mainder of this study and will emphasize the follow-ing points. First and foremost, before the new budgetmanagement model can be introduced, certain basicsafeguards should be put in place to ensure that PEM(public expenditure management) systems are ade-quate to deal with the new demands that will placedupon them. This typically requires a major upgradingof management systems, which will be discussedmore fully in the following two sections. Second, thenew devolved system of decision making may requireparallel changes both in budget institutional arrange-ments and the budget system’s regulatory framework.This aspect of budget reforms will be taken up in Sec-tions VII and VIII.

Box 24. OECD Practices: Performance Management in Government

Percentage of OECD Countries

1. Expenditures specifically linked to each output or outcome target:a. For all targets 15b. For some targets 35

2. Performance results used in determining budget allocations:a. Within government organizations and programs 31b. Within ministries 35c. By MoF to decide allocations between programs 42

3. Performance against targets linked to pay:a. No 65b. Sometimes 23c. Yes 12

4. Politicians use performance measures in decision making:a. Minister with responsibility to deliver performance target 58b. Cabinet 12

5. Rewards/sanctions applied if performance targets met/not met:a. Reflected in heads of ministries responsible for delivering target 17b. Reflected in the size of the budget of department/government organization 21

6. Future plans to introduce performance targets into the budget system:a. Introduce output targets in the budget in next five years 12b. Already output targets in budget and no plans to include outcomes 20c. Already output and outcome targets with plans to improve measure and use 56

Source: OECD (2003, Table 5.4.c).

©International Monetary Fund. Not for Redistribution

Page 53: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

44

V Establishing Basic Public ExpenditureManagement Thresholds

In the initial stages of the transition to performance-oriented budgeting, reforms in public expenditure

management have largely focused on what may betermed the “macro” aspects of PEM reform—namely,putting in place new systems for budget preparationsuch as introduction of MTBFs and better-designedand more policy-relevant program structures. Thebroad thrust of this macro approach, which was de-scribed in the previous section, is to emphasize re-forms within the central agencies of the government.Often, however, the viability of this new approach de-pends on supporting changes at the “micro” level of thebudget system—that is, within the line ministries andother agencies of government that are required to mod-ify their internal management procedures to success-fully operate this new model of budget management.

There is increasing recognition within emergingeconomies that OECD-type budget systems must bebased on a solid platform of financial managementwithin such government institutions, which has led toa greater emphasis on the micro basis of broader bud-get system reforms. At the same time, however, manyof these countries have been forced to acknowledgethat this aspect of their PEM systems is problematic.Three areas related to internal management controlsystems have generally been found weak: internalaudit, management information systems, and costingsystems. This section explores some of the problemsfaced by emerging economies in each of these areasand reviews the generally recommended solutions tothese problems. First, by reviewing the nature of inter-nal control systems, this section stresses their impor-tance not only for an effective PEM system but alsomore widely for ensuring good governance in themore devolved performance-based budget manage-ment model. Second, by reviewing international stan-dards for government auditing and accounting, thissection emphasizes that internal audit is a central com-ponent of internal financial controls aimed at protect-ing the government’s financial interests. Third, byemphasizing the need to create suitable managementinformation systems to facilitate and anchor reforms atthe micro level, this section emphasizes the need forproviding the performance information required to en-sure accountability at that level. Fourth, by examiningin more depth the required costing arrangements, this

section stresses the challenges faced in generating ad-equately precise and meaningful cost information ongovernment programs that is required for comparisonwith output information in order to judge performance.

Internal Control Systems

Internal management control systems encompass arange of management tools aimed at various broad ob-jectives: first and foremost, to ensure compliance withlaws and regulations; second, to ensure the reliabilityof financial data and reports; and, third, to facilitate theefficiency and effectiveness of government operations.In this way, a sound internal control framework is de-signed to assure the public that these operations attainsome basic fiduciary standards: in guarding against themisuse, and inefficient use, of financial and human re-sources; safeguarding assets; achieving budgetary ob-jectives as set out in government policies and spendingplans; countering fraud and error; and maintaining sat-isfactory accounting records to enable the organizationto produce timely and reliable financial and manage-ment reports.63 As such, internal controls can be re-garded as one of the foundations of good governanceand the first line of defense against improprieties. Theyalso provide the public with “reasonable assurance”64

that if improprieties do occur, they will be made trans-parent and appropriately addressed.65

In many countries, the heritage of traditional budgetsystems is a weak internal management system. Undertraditional systems, detailed spending priorities tend tobe set at the central level, rather than by the operational

63See Allan and Tommasi (2001), which defines internal controlas “the organization, policies and procedures used to help ensurethat government programs achieve their intended results; that theresources used to deliver these programs are consistent with thestated aims and objectives of the organizations concerned; that pro-grams are protected from waste, fraud and mismanagement; andthat reliable and timely information is obtained, maintained, re-ported and used for decision-making.” (p. 260)

64This concept is important, since there can be no absolute guar-antee against wrongdoing or honest error. Rather, a control systemshould be designed to reduce that risk to a “reasonable” level com-patible with the cost of implementing the control system.

65See IMF (2001b, especially pp. 56ff).

©International Monetary Fund. Not for Redistribution

Page 54: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Internal Audit

45

units, and therefore the latter have little incentive to de-velop their management systems in order to judge theefficiency with which inputs generated outputs or de-livered needed services. Instead, allocations are drivenby rules—or budget norms—at a detailed budget linelevel. Good management at the operational level is di-rected at preserving or increasing historical levels offunding in each budget line, rather than at makingthese funds more productive in terms of final outputsor outcomes. Managers have no incentive to spendtheir funds better. Indeed, being more cost-efficientwould most likely result in future reductions in fund-ing. In this environment, there is little incentive todevelop internal management accounting, reporting,and controls. Adopting a more performance-orientedapproach requires changing this set of incentives, andhence increases the pressure to strengthen internal lineagency management systems.

The range of management competencies and pro-cedures required for good PEM at the agency level isquite wide, including the following, among others:

• planning budgets, setting their priorities, and re-formulating them in the required detail for man-agement control purposes;

• putting in place a financial management systemthat enables appropriate categories of costs to beclearly identified, accounted for, and reported on;

• establishing internal procurement procedures toensure that purchasing is directed to the most eco-nomical source, meeting acceptable standards ofquality and timeliness;

• putting in place an accounting system to allowcorrect recording of transactions, and generatingrequired management reports ensuring control of,and transparency in, the agency’s operations;

• developing specific management procedures forimportant items of spending, such as capital assetsand human resources;

• strengthening the agency’s internal control mech-anisms to prevent any new financial freedom,brought about by the dismantling of central con-trols, from resulting in less effective use of publicfunds. These internal controls—based on a stronginternal audit function—should also be viewed asperforming a watchdog role within agencies to en-sure the effectiveness of the other internal man-agement systems.

Obviously, the way that internal PEM is organizedand effected varies considerably among countries andultimately reflects the country’s predominating bud-get management philosophy. For emerging economiestransforming their budget management and basic insti-tutions to accommodate a more performance-orientedapproach, the strengthening of internal control systemsshould be emphasized as one of the basic precondi-tions for the success of this reform process. Threeareas are highlighted here, partly reflecting their im-

portance but also reflecting their previous neglect:internal audit, management information systems, andcosting systems.

Internal Audit

International Standards

In moving a budget system toward OECD stan-dards, the role of internal audit (IA) will generally needto be better defined. Recently, there has been progressin reaching a consensus on the audit standards thatgovernments should meet. Specifically, both the Inter-national Organization of Supreme Audit Institutions(INTOSAI)66 and the Institute of Internal Auditors(IIA) have issued standards to guide the auditing andaccounting professions. While these are not compul-sory, they are generally regarded as “best practices,”and so it is expected that as countries develop their ownpublic sector auditing standards, they will try to keepthem consistent with these international standards.

The standards adopt a broad view of the role of IA,placing it as a more central element of PEM ratherthan as a narrow function of compliance or financialregularity. The emphasis is on IA as a managementtool and as an integral part of both management con-trols and information and communications processes.From this perspective, the purpose of IA is to review,appraise, and report to budget managers on the sound-ness and adequacy of internal controls (e.g., safe-guarding assets, ensuring reliable records, promotingoperational efficiency, monitoring adherence to poli-cies and directives).67 The IA is thus seen as perform-ing a watchdog role to ensure the effectiveness ofinternal management controls.

The standards stress four aspects of IA, which un-fortunately appear hard for many countries to attain,including many emerging economies:

• Independence to make objective judgments: Thisimplies that the auditor will have no direct man-agement responsibility for what is being audited,will be free to choose any transaction/topic foraudit, and will be allowed access to all necessaryinformation to come to an informed judgment.Unfortunately, in many countries, systemic gov-ernance problems often imply real difficulty inensuring the auditor’s independence.

66INTOSAI (1995); Internal Auditing Standards Board (2001).See also OECD (1999).

67The IIA defines internal audit as “an independent, objective, as-surance and consulting activity designed to add value and improvean organization’s operations. It helps an organization accomplishits objectives by bringing a systematic, disciplined approach toevaluate and improve effectiveness of risk management, control,and governance processes.” (IIA, 1999)

©International Monetary Fund. Not for Redistribution

Page 55: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

V ESTABLISHING BASIC PUBLIC EXPENDITURE MANAGEMENT THRESHOLDS

• Professional proficiency: This assumes an appro-priate audit methodology, technical competence,and sufficient levels of resourcing for the IAfunction. In many countries, skilled auditors are inshort supply, professional proficiency is very low,and/or the government’s pay scales are insufficientto attract or retain suitable staff. These factorsoften represent an important constraint on attemptsto strengthen the IA function.

• Wide scope for IA: The scope of the IA describedin these international standards is based on thebroader view that IA is a tool of management,helping to close the loop in the agency’s PEMmanagement cycle by ensuring the efficient andeffective use of resources.68 This, in turn, assumesa mechanism for follow-up and action on audit re-ports. In many parts of the world, including inmany emerging economies, IA continues to bedefined rather narrowly—focusing on financialcompliance and regularity, rather than on broadermanagement issues. Moreover, governance prob-lems and a lack of professional competence alsoconstrain the internal auditor to this narrower roleand hinder his or her ability to generate timely andrelevant reports.

• Effective management of the IA function: In manycountries, management of IA is poor—poor workpractices, lack of planning and personnel manage-ment, and little support from the external auditagency. Additionally, management is constrainedby the institutional arrangements for IA, whichoften compromise the role of the auditor as an aidto internal management.

IA in the OECD Countries

Even among OECD countries, organization of theIA function varies greatly, for example, from central-ized to decentralized approaches. In the centralizedmodel, the MoF not only plays a key role in budgetingand allocating funds to line ministries, but also directlyintervenes in ex ante controls, placing its own staff inthe line ministries. In this environment, the IA is con-centrated in a specific organization performing certaincontrol functions, traditionally a centralized ex ante fi-nancial control organization, an inspectorate general,or a treasury external audit service. In the decentral-ized approach, each line ministry takes full responsi-bility for spending its own budget and for ensuringappropriate checks and safeguards. In this environ-ment, the IA is focused on the ministry’s overall sys-tem of organization, controls, rules, procedures, andregulations set up to ensure the most economic, effi-cient, and effective use of resources. To do this, the IA

46

control system includes a range of ex ante controls,systems, performance, and IT audits. Other models ap-pear to mix internal and external audit functions. Forexample, in Germany, IA is not part of a governmentagency’s control system, but can be viewed as a com-ponent of external audit.69

Despite the variety of IA approaches among OECDcountries, there are some common general principles.First, IA is viewed as a central component of internalfinancial controls aimed at protecting the government’sfinancial interests. The important concept is the inter-nality of this executive function, distinguishing itfrom external audit. Second, although IA activitiesinclude traditional compliance and regularity opera-tions, they can be defined quite widely to includesubstantive tests and systems, performance, and ITaudits.70 Third, to function effectively, the internal au-ditor must be functionally separated from the day-to-day management of an organization (otherwise theaccountability of designated managers will be di-luted), but at the same time, the auditor must haveinput to top management to ensure that his or her find-ings and recommendations result in corrective action.Fourth, internationally recognized auditing standardsshould be upheld.

Developing the IA Function in Emerging Economies

In developing the IA function, the most importantstep for emerging economies is to determine the roleof the IA in the country’s budget management system.Establishing such an overall framework involves twobasic design issues:71

• whether to adopt a control or a management ori-entation—a question of the objectives to be pur-sued by the IA function; and

• the degree to which IA is centralized—a questionof the organization of the IA function.

Once these overarching issues are resolved, thereare further issues related to how this will be imple-mented at the agency level:

• the relationship with external audit—a question ofresponsibilities and coordination; and

69While the IA cadre operates within agencies, they are subject totechnical and professional guidance, as well as supervision, by theGerman supreme audit institution, the Federal Court of Audit. Theyreport only to the Federal Court of Audit and perform a “pre-audit”role rather than a traditional IA role. See discussion in Diamond(2004).

70“IA is the total sphere of activities of ex post verification by anorganization (located within the organization to be audited but in-dependent of the management functions of that organization) ofwhether management and control systems comply with budgetspecifications, objectives, rules and standards and, more generally,to the principles of sound financial management” (Larsson andMadsen, 1999, p. 5).

71These are discussed more fully in Diamond (2002b).

68See Gray, Jenkins, and Segsworth (1993). The external audit in-stitution does this for the government as a whole.

©International Monetary Fund. Not for Redistribution

Page 56: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Internal Audit

47

• how to restructure work practices in agencies—aquestion of operational effectiveness.

These are addressed in turn.

Deciding the Objectives of IA

There are a variety of interpretations of the role ofIA. At one extreme, there is the centralized view of IAas a support function, assisting the MoF in monitoringministry and department compliance with MoF finan-cial regulations, instructions, and accounting proce-dures. The emphasis is on compliance and control.Alternatively, under a decentralized approach, IA isviewed as assisting budget managers in the effectivedischarge of their responsibilities by providing feed-back on their use of public resources through the sub-mission of reports and, when justified, the provisionof recommendations for corrective action. The em-phasis is on efficiency and effectiveness in the use ofresources and the delivery of services. Between thesetwo approaches exist many variants, including the useof a central inspectorate to set standards and to assistdecentralized IA units in the line ministries with del-egated responsibilities.

The overall design of the IA function should begeared to the specific priorities of the country. Forcountries with governance problems, the foremost ob-jective should be to ensure compliance with financiallaws and regulations. For those emerging economiesfacing a high degree of fiscal stress, the need to attainmacroeconomic objectives should be paramount. Forthose countries that can ensure compliance with thelaw and have reached a fair degree of macroeconomicstability, more attention can be paid to ensuring the ef-ficiency and effectiveness of resource use, as currentlyemphasized in the OECD countries.

Deciding the Degree of Centralization

A fundamental design issue is the degree of central-ization in the organization of the IA function. The cen-tralized approach has often been viewed as better froma capacity-building viewpoint, because it is argued thatthis approach—

• allows easier maintenance and better developmentof the proficiency of internal auditors. With ascarcity of skilled manpower, a decentralized ap-proach often implies the diversion of IA staff toother duties that will reduce the proficiency of thestaff. However, if the MoF develops a specialcadre, it will be able to concentrate scarce auditingresources and so maintain proficiency, ensure theauditors’ specialization, and develop centralizedstandards and training programs for them.

• fosters greater independence. The audit should beconducted with adequate independence. The cen-tralized option is better in this regard, it is argued,

since the IA is managed by the MoF outside the di-rect control of line ministry managers. However,the necessity of independence is in direct conflictwith the necessity for the MoF’s close cooperationwith other departments for budget management.

There are also some disadvantages to centraliza-tion, namely that it—

• weakens accountability of the line ministry man-agement. It could be argued that the prime respon-sibility for internal control should rest with, and be“owned” by, the line ministry management. How-ever, the centralized option divides responsibilitybetween the line ministry management and theMoF, obscuring the ownership (or accountability)of this control mechanism. The line ministry’smanagement may be only too happy to considerthe responsibility for internal control as belongingto the MoF.

• is of limited effectiveness because of weak trans-parency. In many countries, the flow of informationto external officials (including internal auditorsfrom the MoF), is typically limited and untimely,constraining the effectiveness of the IA.

• fails to foster close cooperation with other depart-ments. Close cooperation among departments isessential for efficient IAs.72 However, the central-ized approach does not promote such coopera-tion—the internal auditor is viewed as the “spy” ofthe MoF, rather than as a member of the line min-istry management team.

In weighing these two options—a centralized or adecentralized design for the IA—there are considera-tions that suggest the answer will be country specific.First, for many countries the danger is very real that inan entirely centralized approach, the MoF will assumeresponsibility for the rectitude of financial manage-ment in budget institutions, undermining the basic ac-countability of budget managers.73 Second, in somecountries, the risk of political interference with routinebudget management is high, so that the budget man-ager’s accountability is undermined from above, andtherefore a centralized system is justified. Third, wherethe administrative capacity to perform IA functions islow, with regard to the recruitment and retention ofcompetent staff, a centralized system controlled by theMoF is also recommended. Given the time it takes toestablish a professional corps of internal auditors,

72Although at the same time, operating at arm’s length securessome independence from day-to-day operations.

73The European Commission recently decentralized responsibil-ity for spending to avoid weakening accountabilities of Commis-sion managers. That is, “the explicit prior approval of a separatefinancial control service has been a major factor in relieving Com-mission managers of a sense of personal responsibility for the op-erations they authorize . . . while doing little or nothing to preventserious irregularities. . . .” OECD (2001, p. 272)

©International Monetary Fund. Not for Redistribution

Page 57: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

V ESTABLISHING BASIC PUBLIC EXPENDITURE MANAGEMENT THRESHOLDS

including in many emerging economies, this is a mostrelevant consideration.

This presents a dilemma. Taking due account ofthe above considerations, a centralized approach foremerging economies is often recommended, at leastinitially, as the most prudent approach, although thisruns counter to the basic decentralized institutionalmodel that underlies many OECD IA systems that sup-port performance budgeting. As argued previously, al-though the decentralized budget management model,focused on performance, is the system that emergingeconomies will eventually move toward, until budgetmanagement capacity has reached an adequate thresh-old, in the interim, a more centralized IA system maybe required.

Strengthening IA at the Agency Level

Ensure the independence of IA An important ob-jective in restructuring the IA function is to give someassurance of its independence from day-to-day man-agement and hence of greater objectivity in its evalua-tions. Obviously, the degree of IA independence is notthe same as for external audit, which reports to parlia-ment. Rather, the IIA (1999) defines IA independenceas follows: “Internal auditors are independent whenthey can carry out their work freely and objectively. In-dependence permits internal auditors to render the im-partial and unbiased judgments essential to the properconduct of audits. It is achieved through organizationalstatus and objectivity.” (p. 11)

Ideally, the internal auditor should be responsible tothe minister or the chief executive of the ministry oragency. In a decentralized model, the internal auditorwill report directly to this top official. In the central-ized approach, having the centralized audit office re-porting directly to the minister of finance helps ensurethe independence of the IA.74 The internal auditor is re-sponsible to the head of the ministry/agency and is partof that agency’s staff and part of the chief executive’smanagement team. However, care must be taken not toinfringe the cardinal rule of audit: an auditor shouldnot audit him- or herself. Typically, this is avoided byseveral institutional mechanisms:

• A clear and agreed definition of the internal audi-tors’ tasks: This is a way of clarifying the place ofthe IA function in the work of the budget institu-tion, dispelling ambiguities and avoiding disputes.

• Establishment of line ministry audit committees:These should be formed from the top managementof the institution and technical experts in the ac-counting and budget fields. The aim is to act as asteering committee for the work of the IA, identi-

48

fying problems as well as the corrective or pre-ventive action. Not only does this act to strengthenthe role of IA within the agency in enforcing fi-nancial discipline, but it also creates some distancebetween the agency’s regular operations and IAevaluations.

• Create a central audit committee in MoF: Simi-larly, to enforce this distance from day-to-daymanagement and to offer some external support,it is also recommended that a central audit com-mittee in the MoF review the findings of the IAunits and pursue remedial actions. Another possi-ble mechanism is to have outside professionalsconduct an independent external review of IApractices every two or three years, thereby coun-tering any tendency for agency managers to inter-fere with the IA.

• Clear demarcation of responsibilities in relation toexternal audit: In some ways, this can be addressedby a clear and well-documented definition of theduties of internal auditors. At the same time, the re-lationship between the two functions should berecognized as symbiotic—it is important for the IAthat there be a strong external audit, and vice versa.The external audit should use the work of the IA,and the IA should be guided by the findings of theexternal audit.

Restructure work practices A strategic decision tobe taken in many emerging economies is where best todeploy scarce audit skills. There are ways to economizeon the use of these scarce IA resources:

• Prioritize to extend the scope of IA: Improvedwork practices can often offer significant savings—say by moving away from extensive pre-audit ofvouchers to a sampling approach—as can im-proved management of the audit function throughfocusing on priority areas and key weaknesses.One area that can typically benefit from IA reviewis the evaluation of internal controls. An importantfunction of the IA should be to examine and eval-uate the adequacy and effectiveness of internalcontrols in existing systems, as well as in any newsystems before these are introduced. This clearlyimplies that the entire system of internal controlhas to be reviewed for each ministry, department,or agency as well as by function. This needs to beemphasized because, with strong internal controls,the system will automatically have its own checksand balances to minimize the possibility of errors,irregularities, and fraudulent manipulations.

• Create special teams: In examining the IA of tran-sitional economies, it is not unusual to discovermany functions that are either not being performedor for which coverage is superficial because of in-adequate staff, lack of specialized skills, etc. Oftenthe most productive use of limited IA staff is incentral teams earmarked for conducting special

74Of course, it is necessary for the MoF to maintain its own IAand be subject to the close scrutiny of the external audit institution.

©International Monetary Fund. Not for Redistribution

Page 58: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Financial Management Information Systems

49

audits in government agencies with the assistanceof IA staff already stationed there.

• Better formulate work plans: Existing operationalstandards for IA require that the internal auditoradequately plan, control, and record his or herwork. Such planning should be done not only forindividual audit assignments, but also for varyingtime periods such as a quarter, a year, or evenlonger periods.

While the above approaches can make IA more ef-fective, undoubtedly this requires proper resourcing,which may be difficult. However, the returns could besubstantial insofar as a sound IA function is considereda precondition for introducing a more devolved per-formance budgeting model.

Financial Management Information Systems

The term financial management information system(FMIS) usually refers to comprehensive computeriza-tion of PEM processes. At its most general, an FMISincludes budget formulation, budget execution, andaccounting with the help of a fully integrated systemfor financial management of the line ministries andother spending agencies. The full system also shouldbe securely integrated and in communication withother relevant information systems.75

Although it is no panacea, an FMIS provides sub-stantial benefits. First, improved recording and pro-cessing of government financial transactions allowsprompt and efficient access to reliable financial data.This supports enhanced transparency and accountabil-ity of the executive branch to the legislature, the gen-eral public, and other external stakeholders. Second,an FMIS strengthens financial controls, facilitating afull and updated picture of commitments and expendi-tures on a continuous basis. Once a commitment ismade, the system should be able to trace all the stagesof the transaction, including budget release, commitment,purchase, payment request, reconciliation of bank state-ments, and accounting of the expenditure. Third, morecomprehensive financial information on current and pastperformance assists budgetary control and improveseconomic forecasting, planning, and budgeting. As aconsequence, establishment of an FMIS has becomean important benchmark for a country’s budget reformagenda, often regarded as a precondition for achievingeffective management of budgetary resources and as arequirement for introducing performance budgetingprocedures.

As the name implies, there are, and should be, threeguiding characteristics for an FMIS:

• The FMIS is a management tool. When developingan FMIS, it is important that it cater to the needs ofmanagers—not just in the central agencies but alsoin the line agencies. Moreover, it should support themanagement of change. It must be viewed as an in-tegral part of budget system reform—and henceshould be designed to meet not only present budgetsystem requirements, but also those that are likelyto arise as parallel budget reforms are implementedelsewhere.

• The FMIS should provide a wide range of financialinformation. As a tool of management, it shouldprovide the information required for decision mak-ing. It should be anchored in the government ac-counting system and be designed to perform allnecessary accounting functions as well as to gen-erate custom reports for internal and external use.However, this does not mean that it should con-centrate exclusively on financial information.Managers will require other, nonfinancial infor-mation, including for example, personnel infor-mation (i.e., numbers of employees, their gradeswithin the organizational structure, and their ratesof remuneration). Performance information will beimportant to managers, such as the identificationof program objectives or outcomes, the types of goods and services produced, as well as indica-tors by which to judge programs’ efficiency andeffectiveness.

• The FMIS is a system. Its role is to connect, accu-mulate, process, and then provide information toall parties in the budget process on a continuousbasis. All participants therefore need access to thesystem and need to be able to derive the specificinformation they require to carry out their differ-ent functions. The converse is also true: if theFMIS does not provide the required informa-tion—that is, does not have the correct function-ality—it will not be used and will cease to fulfillits central function as a system. Box 25 broadlydescribes the attributes usually required of anFMIS.

Existing information systems will need to be modi-fied to accommodate a performance information sys-tem, the need for which was outlined in the previoussection. Generally, more modern government FMISsrequire a progressive shift from a focus on annual,cash-based budgets toward a more medium-term, fi-nancial planning approach, based on a clearer policyframework in terms of performance and focused on thesustainability of the associated resource implications.Also, as discussed more fully in Section VII, institu-tional PEM arrangements are an integral part of thisnew performance orientation and will most likely needto be modified. For example, they must accommodate

75Because of the integration requirement, the FMIS is commonlycharacterized as an Integrated Financial Management InformationSystem (IFMIS).

©International Monetary Fund. Not for Redistribution

Page 59: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

V ESTABLISHING BASIC PUBLIC EXPENDITURE MANAGEMENT THRESHOLDS

the trend toward decentralization, devolution, andthe adoption of competitive contracting arrangementsfor the delivery of public sector services. Therefore,any FMIS in this new environment needs to be suffi-ciently flexible to address the financial managementneeds of a variety of organizational service deliveryarrangements.

The FMIS must also accommodate and facilitate anumber of other reforms, which have already beendescribed and will be discussed more fully in subse-quent sections. Among these are: the transition tomodified cash/accrual accounting and eventual pro-gression toward full accrual accounting for at leastsome parts of government; the phased introduction ofperformance budgeting with clear definition of in-puts, outputs, and expected outcomes; introduction ofcompetitive markets (purchaser/provider splits); de-centralization (delegation of authority and responsi-bility for resource management); more responsivenessto beneficiary/client requirements and more inclu-sive and participatory approaches; and improved ap-proaches to strategic management and human resourcemanagement. Needless to say, the new performancebudgeting orientation adds new tasks and requiresmore information, focusing on: improved analyticaland policy formulation capacity; technical capacity tomove toward more performance- or results-orientedplanning, budgeting, and accounting systems; andimproved control, monitoring, and performance re-view of the government’s budget (Box 26).

50

Box 25. The Attributes of a Well-Designed FMIS

A well-designed financial management information sys-tem will—

• be modular and capable of progressive upgrading to caterto future needs.

• offer a common platform and user interface to the stake-holders in different agencies responsible for financialmanagement to allow them to add to and access the infor-mation database (in its absence, each agency will have theincentive to develop “its own” FMIS to meet its needs).

• maintain a historical database of budget and expenditureplans; transaction data at the highest level of detail; cashflows and bank account operations including checks is-sued, cancelled, and paid; cash balances; and floats.

• have dedicated modules to handle forward estimates ofrevenues, expenditures (prepared by the agencies), andresulting cash flows on a monthly, rolling, short-term(one- to three-month), and longer-term (three-month toyear-end) basis.

• have built-in analytical tools to offer trend analysis ofvarious elements of fiscal operations to permit a for-

Box 26. Changes in Information Needs toAccommodate Performance Budgeting

Traditional financial management needs:

• institutional or expenditure heads, as detailed in thebudget approved by parliament

• release of funds against expenditure heads by detailedline item

• recording of expenditures and reconciliation with bankstatements

• periodic financial statements on cash-based accounting• recording of debt and other liabilities

New performance-related tasks added to theseinformation requirements:

• programs and activities as detailed in the budget ap-proved by parliament

• allocation of funds based upon particular, desired re-sults described in physical and financial terms

• budget documents that describe the expected outcomesand performance measures to monitor achievement

• responsibility for achievement identified and physicaltargets tracked

• broader-based performance reports that documentphysical as well as financial measures of performance

• data required to administer a set of performance-basedrewards and sanctions

ward look at the emerging events bearing on the fiscalstance.

• compile formal government accounts from the database ofthe agencies’ authorizations and cash allocations, primaryrevenue and expenditure transactions, and of treasury op-erations, avoiding the need to duplicate data entry for ac-counting purposes.

• enable real-time reconciliation of parallel but relatedstreams of transaction data—at the agency level, to rec-oncile checks issued with those paid by the banks; at thetreasury level, to reconcile receipts from banks with thechecks paid by taxpayers; and at the agency level, to rec-oncile cash balances reflected in the agency ledgers withthe cash balances in the banks.

• mechanize all possible routine tasks at the central and spend-ing agencies—generating various forms/authorizations,checks, outputting hard copies of key registers and state-ments, and the like.

• be flexible enough to provide user-defined managementinformation from the database, aggregated at the desiredlevel of detail.

©International Monetary Fund. Not for Redistribution

Page 60: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Financial Management Information Systems

51

All such tasks require the use and development ofimproved IT and communication facilities to meet theneeds for improved analysis and efficient informationprocessing. The backbone of this effort will continueto be accounting and financial information systemsthat enable the production of complete and timely pe-riodic financial information on revenues, expendi-tures, and financing, disaggregated by organizationalresponsibility, nature of account (expenditure, rev-enue, asset, liability), activity, project, and sector oreconomic category, as appropriate. The system shouldincorporate recognized accounting standards and sup-port enhanced levels of public accountability.

Development of an FMIS

Introducing an FMIS has proved to be a challenge inevery environment.76 Based on the international expe-rience of FMIS projects, a number of implementationissues should be stressed to ensure success:

• Identify the owner(s) of the system: The impor-tance of this phase of the project may not be inher-ently obvious. An owner, or set of owners, needsto be identified because there will be many criticalissues that arise during the course of the projectthat require quick and decisive resolution. It is rec-ommended that the ownership role not be spreadtoo thinly, for instance across a panel of users,since this may compromise the required decision-making process and its legitimacy.

• Select a dedicated project team and steering com-mittee: Many FMIS developments have failed, orproduced unsatisfactory results, because insuffi-cient investment was placed in the team that man-aged the project. Using a part-time team is notrecommended, because such an approach invari-ably produces an FMIS that is late or fails to ade-quately satisfy user requirements. The team needsto be separated from day-to-day work responsi-bilities, but should not be too isolated from theultimate users of the system, who can providevaluable feedback on its development. It is advis-able that the project be driven by the users on theteam rather than the IT members of the team.

• Establish a high-level steering committee: A steer-ing committee is needed to provide strategic guid-ance throughout the life of the project and to assistthe owners in making crucial decisions that sig-nificantly affect the direction and final characterof the system. Major users of the system shouldbe represented on this steering committee, whichshould address such critical issues as: specifyinguser requirements; decisions on in-house or con-

76For some specific issues arising in developing countries, seeDiamond and Khemani (2005).

tractor development and the issue of locking incontractors; the use of standard software packages;and the extent of customization. These issues willbe discussed more fully below.

• Clearly specify the development milestones: Toensure that the system is developed on time andthat any contractor payments are only made asparts of the system are successfully delivered, it isessential to clearly define milestones. It may bedifficult to do this during the formulation stage ofthe project, but the key stages should be listed withdetails on the delivery components to be agreed ata later date. No contractor payments should bemade until the contractor can justify the comple-tion of a milestone, with late delivery subject topenalty clauses, with partial delivery subject to dis-counts, and with both specified in the contract.Successful completion of a milestone should besupported by appropriate testing and audit reports.

• Produce system and user documentation: Throughoutthe development of the FMIS, system and user doc-umentation must be produced on a regular basis.There is a risk in large system development proj-ects that required documentation is left to thefinal stages of the project, but regular user docu-mentation assists in the development of trainingmodules and thereby allows the essential task oftraining the ultimate users to commence while thesystem is being developed. This not only providesa pool of knowledgeable users to test the final sys-tem before it comes on-line, but also providesvaluable feedback on the system modules as theyare produced. Regular system documentation isessential to avoid reliance on a single contractor,or group of contractors, for maintenance of thesystem.

• Provide appropriate training: It is recommendedthat the training function be handled in-house. Therisk of outsourcing this function is that the trainingprovided may be superficial and not at the appro-priate level for those users who will be using thesystem extensively, the so-called power users. Dif-ferent types of users should be identified, and notall users should receive the same level of training.Power users should be trained first and should haveaccess to follow-up training throughout the devel-opment process. Other users should be able to in-dicate the level of training they require and shouldreceive such training as close to the roll-out of thesystem as possible to reduce the need for follow-up training.

• Negotiate an agreement for long-term mainte-nance and integration of enhancements: The issueof contractor lock-in also extends to the mainte-nance of the system after its implementation. It isessential that the long-term maintenance agree-ment be subject to a full tender process, as was the

©International Monetary Fund. Not for Redistribution

Page 61: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

V ESTABLISHING BASIC PUBLIC EXPENDITURE MANAGEMENT THRESHOLDS

original contract. Obviously, the developer or de-velopers would have a comparative advantage inbidding on the maintenance agreement, but a fulltender serves the interests of securing a competi-tive price and high level of service.

• Conduct an independent review of system devel-opment: It is recommended that a full, indepen-dent review of the FMIS be conducted within sixmonths of the system’s introduction. This shouldbe in addition to the audit reports on the completionof each module of the system. The review shouldinitially be provided to the nominated owners of thesystem, and then to a panel of users for review andcomment.

Costing Systems

The costing methodology for performance-basedbudgeting has often been neglected. That is, the newemphasis on outputs or outcomes does not remove theneed for input identification, which remains the mostwidely used means for annual budgeting and account-ing. For the program structure to be implementable,there must be adequate mechanisms to fully cost pro-grams, so that the outputs of programs can be relatedto their budgetary costs, and then ultimately to theirbenefits in order to judge program performance. Costscan be allocated on many different criteria.77 From aprogram budgeting viewpoint, the more pertinent cri-teria are the direct or indirect association with theprogram—i.e., based on the principle of assignment.Costs also can be assigned to different cost objects: op-erating units, cost centers, outputs, and projects. What-ever criteria are used, they should be clearly specifiedin a government-wide costing system that not only canproduce accurate cost information on programs, butcan also assign those costs correctly and is neither toocomplex nor too resource-intensive to apply (Box 27).Because these characteristics can be hard to satisfysimultaneously, important trade-offs exist.

To be accurate, the cost allocation must avoid twoimportant types of distortion. The first is a flow distor-tion, when the cost bases used to describe program ac-tivities do not accurately capture their consumption ofresources. The solution is to devise more accurateassignment mechanisms or to decompose programactivities into more homogenous categories. Bothsolutions, however, involve increased complexity andhigher costs. The second type of distortion is a stockdistortion, when the costs assigned to program activityoutputs do not reflect the actual quantity of resourcesconsumed. Simple costing systems often assume that

52

all resources aggregated in an operating unit are con-sumed proportionately by all outputs that are producedby the unit. If this is not the case, then quantity distor-tions occur. The solution is to assign costs by more ap-propriate operations and to allow each operating unitto use multiple assignment mechanisms.

Cost-allocation systems are designed to provide costinformation along various dimensions, such as organi-zational units, cost centers, outputs, or projects, whichare described as the cost objects of the system. Thechoice of cost objects affects the administrative com-plexity and expense of operating the cost-allocationsystem. There is an obvious trade-off between thedegree of detail by cost object and the resources re-quired to operate the system. There are dangers atboth extremes: providing a lot of cost information todecision makers is expensive, but an overly simplecost-allocation system may be of poor quality, mayincrease the risk of cross-subsidization, and may givemisleading information to decision makers. Unfortu-nately, the increasing complexity of government op-erations has added considerably to the difficulties ofcost allocation.78

For the purposes of program budgeting, the mannerin which resources are consumed by programs, activi-ties, outputs, or cost centers determines whether theyare allocated as direct or indirect costs. To increase theaccuracy and relevance of total program or outputcosts and, at the same time, to reduce arbitrary cost al-locations, it is necessary to categorize as many of theresources as possible as direct costs.79 Of course, itmay not be possible to directly allocate some costs tocost objects when they are common to more than oneobject. However, such indirect costs may still be ap-portioned to cost objects on the basis of consumption,

77For example, by their nature (variable or fixed); productionfunction (output or nonoutput); decision making (investment, sunk,opportunity cost); accountability (controllable or noncontrollable).

Box 27. Attributes of a Costing System

• Relevant: Provide reasonably accurate information ina form relevant for decision making.

• Clearly assigned: Offer a clear assignment of cost toeach government activity, avoiding the risk of cross-subsidization of activities.

• Cost-effective: The expense of managing the costingsystem must not be onerous in terms of resources (in-cluding staff time).

78As pointed out by Gearhart (1999), “. . . technology invest-ments and the expenses associated with increased overheads, in-creased services, complex processes, and changing roles ofgovernment have significantly complicated the cost profile of gov-ernment organizations” (p. 13).

79There are many ways to do this, e.g., time recording systems,coding actual transactions, and judgment by management.

©International Monetary Fund. Not for Redistribution

Page 62: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Costing Systems

53

i.e., on the extent to which they contribute to, or are in-curred because of, the cost object.80 Much depends onhow closely the costing system follows an existing or-ganizational structure. If each operating unit or costcenter is contributing to only one output (program oractivity), the need for allocation in the second stagewill be minimized. However, if costs accumulated in acost center must be allocated to two or more outputs,or if some units within an organization provide ser-vices to other units, then a more complicated two-stageallocation (or step-down) process is required. Underthis method, the costs accumulated in cost pools arethen allocated down to other cost pools or cost objects,using a variety of allocation bases, until all costs have

been allocated to final cost objects.81 The process isshown schematically in Box 28.

A variant, more in line with the output budgetingapproach, is activity-based costing (ABC). In this ap-proach, programs are broken down into their basicactivities—the steps or sequences of processes thatconvert inputs into outputs. As a first step, this ap-proach also starts with apportioning an organization’sexpenses to a set of cost pools, but then ABC compilescost information on all activities performed in an orga-nization, including activities that do not vary with out-put volume. ABC is oriented to allocating overheads

80Where costs cannot be directly assigned to cost objects, agen-cies must identify appropriate bases for assigning them, reflectingin some way the consumption relationship between the costs andcost objects.

81For example, the allocation process could begin with deprecia-tion of buildings and fixtures. These costs would be distributedacross all remaining cost pools. The amount to be distributed fromthe next pool (building maintenance) would now include not onlyits own costs but also the amount allocated to it from the previousstep. This total would then be allocated to all the remaining pools,and so forth. (Thompson, 1998, p. 387)

Box 28. Basic Steps of a Program-Based Costing System

1. Break down the program into the activitiesthat produce the program’s objective, ser-vice, or product, and identify the work unitsundertaking each activity.

2. Identify all resources used and their associ-ated costs in terms of the basic work units de-livering program outputs/outcomes.

3. Categorize and measure direct costs—those

Work Unit Costs

Costs on Behalfof Work Unit

Government-Wide andDepartment-Wide Costs

Direct Costs Indirect Costs

Activity1

ActivityN

Resourcesconsumed

Translatedinto costs

Cost allocationmethodology

Allocated toactivities

Program

CostingSystem

Activity2

Aggregated toprograms

arising from the operations of a work unit orthe costs incurred on behalf of the unitthrough the operations of other work units—and those government-wide or department-wide indirect costs, i.e., overheads.

4. Assign or allocate all direct and indirectcosts to specific activities, using an agreedcosting methodology.

©International Monetary Fund. Not for Redistribution

Page 63: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

V ESTABLISHING BASIC PUBLIC EXPENDITURE MANAGEMENT THRESHOLDS

with a primary focus on identifying the full costs ofproviding government services.82 Organizations thatare most likely to benefit from ABC are those withhigh overhead costs and widely diverse operating ac-tivities or service lines—i.e., most government agen-cies. This approach is typically most applicable at theoperational level on a service-by-service basis, en-abling service managers to better understand the truecosts of an activity. In this way, ABC enables costs tobe better related to performance measures and allowsmanagers to budget more confidently for target activ-ity levels and to break down the budgetary impact ofeach level of activity.83

There is common agreement that the uniform mea-surement of the full costs of end products or activitiesacross all government programs is logically requiredin order to judge performance from a policy perspec-tive. There are many practical problems to overcome,however. First, there is the difficult problem of indirectcosts, or overheads. The activities of any agency willinclude certain common staff functions such as per-sonnel management, accounting, and general manage-ment, which are difficult to break down and apportionto specific activities. Indeed, some would argue thatthe usefulness of the program structure could be sub-stantially diminished in certain government sectors bythe presence of such indirect costs. Second, the mea-surement of program costs is further complicated bythe existence of joint products and activities by thesame basic unit. In the private sector, such cost mea-surement is the special domain of managerial or costaccountants geared to servicing internal managementdecision making. In the government sector, these skillsare typically underdeveloped and often nonexistent.Last, but not least, it is often argued that capturing fullcosts ideally requires an accrual system of accounting,as used in the private sector, where the whole rationaleis to ascertain the true cost of “outputs”—i.e., throughan output-oriented accounting system. The case formoving to accrual accounting in government, at leastin principle, is a strong one because it enables the fullcosts of output production to be ascertained each yearand compared among years. However, in practice, thecase may be less strong, as will be discussed more fullyin Section VI.

In the first stages of implementing performance bud-geting, it may be prudent to adopt systems that do notrequire complex product costing. As indicated, one ofthe important principles in designing costing systems

54

is their cost-effectiveness (see Box 27). It should berecognized that the allocation of indirect/overheadcosts could involve potentially high costs, and for thisreason alone costing structures should not be overlyambitious. For example, “full” cost attribution for pro-grams or even outputs, that is, distribution of all indi-rect and capital costs, may not be required in all cases.If the manager is determining his potential activitylevel within a given budget constraint, then variableand marginal costs are more relevant, since full cost-ing likely contains elements that would not change ifthe level of output changes.84 Of course, if the decisionmaking affects numerous budgets and programs, thenfull costing becomes more relevant. This brings intooperation the first principle in designing costing sys-tems, that it be relevant (see Box 27). Based on thisprinciple, it can be argued that more sophisticated cost-ing methodologies need to be deployed selectivelyrather than indiscriminately throughout the budget sec-tor, and that for countries with more limited financialand human resources, costing systems, at least ini-tially, should serve modest objectives.

Assessing the Performance of theBudget System

A central message of this section is that before a per-formance management framework is introduced, thereshould be some assurances that it will rest on a solidbasis of public expenditure management. The questionthen arises how to judge whether a PEM system is ro-bust enough to accommodate the previously discussedchanges required for the introduction of performance-oriented budget management. Answering this questionin turn requires an assessment of the overall perfor-mance of the PEM system.

From this perspective, it is important to note somerecent work directed at measuring the performance ofthe PEM system as a whole.85 This effort has arisenfrom the fact that donors and multilateral developmentagencies increasingly direct aid to countries throughgeneral budget support rather than through projects.Given that expenditure is fungible, these donors anddevelopment agencies need to look at governmentspending in a more comprehensive way and, for fidu-ciary purposes, to examine the performance of PEMsystems as a whole. Work is ongoing to replace themultiple diagnostic tools used by bilateral and multi-lateral agencies with common tools for measuring theperformance of budget systems. This effort has al-ready resulted in the development of a standardized

82In this way, it supports what is often termed “activity-basedmanagement” which focuses on the management of an activity tocontinuously improve its efficiency and effectiveness. For a casestudy of its application, see Abrahams and Reavely (1998).

83Although the ABC approach is not used by higher levels to setoverall budgets, it can lead to greater knowledge about unit costsand can provide a basis for setting benchmarks for government activities.

84This also applies to joint costs, i.e., costs that contribute to mul-tiple products but do not vary with the changes in the productionlevel of only one of those products.

85See Allen, Schiavo-Campo, and Garrity (2004).

©International Monetary Fund. Not for Redistribution

Page 64: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Concluding Remarks

55

assessment framework, the Public Financial Manage-ment (PFM) Performance Measurement Framework,which is designed to measure the performance ofcountries at all levels of PEM development from botha strict fiduciary risk perspective and a developmentrisk perspective.

The PFM Performance Measurement Frameworkwas developed by the World Bank with the assistanceof its development partners, including the IMF, and isaimed at streamlining and condensing the previous di-agnostic tools.86 It takes six critical objectives of awell-functioning PEM system (Box 29) and derivesa small number of high-level indicators to assess per-formance against these objectives.87 Each indicator ismeasured against a four-point ordinal scale (A–D),and guidance notes have been developed on what levelof performance would meet each score for each indi-cator. This has then been applied in pilot countries fortesting.

The result is a standardized assessment tool bywhich countries’ budget systems are assessed againstbenchmarks based on international good practices in avariety of different dimensions. Even though the tool

is a synthesis of assessments generated through exist-ing instruments, the objective in the medium term is toreplace the existing instruments and to harmonize rec-ommendations for improving and strengthening PEMsystems.

The ultimate objective is to make aid more effective.However, this approach also has potential as a guide tobudget system reform. The power of the approach is itsrecognition of the PEM system as a system in whichdifferent phases of the process are interrelated. Thisintegrated approach to performance assessment isbeneficial in identifying key weaknesses that causeproblems elsewhere in the system, and this assists inprioritizing remedial actions, which can be especiallyuseful for capacity-constrained public administrations.The emphasis has been placed on proactively usingthis tool for capacity development, first by diagnosingweaknesses in the PEM system and then by develop-ing a strategy and action plan jointly with the donorcommunity and the government to correct these weak-nesses. Using this approach should make it possible tostrengthen PEM systems to a minimum level that ac-commodates performance management reforms.

Concluding Remarks

A central message of this section is that the widescope of budget system reforms required at the microlevel to accommodate a performance-oriented ap-proach stands as a warning not to underestimate eitherthe magnitude of the task or the time that this reformmay require. These new management procedures needto be effected throughout all levels of the governmentsector. Unlike macro reforms, which have been fo-cused on a few central agencies, the micro reforms po-tentially affect a large number of institutions and hencea much larger number of people. Accordingly, thescale of the task raises important issues related to ad-ministrative capacity—for example, the need to de-velop internal auditing capacity when there is a generalshortage of accountants in the economy and especiallyin the government sector, or the lack of specializedcomputerization skills, which is a significant constrainton development of modern accounting and informa-tion systems.

Another important dimension of administrative ca-pacity is the preparedness of agency managers to adoptchange. Certainly, the budgeting and financial man-agement methods used in the past are likely to requiremodification to serve the new performance-orientedsystem. Hence those who advanced to managementpositions in the old system may not be the best suitedfor the new. Given the evident skills required to im-plement the procedures described above, at a mini-mum one would have to recognize the need for anextensive and necessarily lengthy training process.

Box 29. Critical Objectives of PEM SystemsMeasured through a Standardized Assessment

• Budget realism: Budget is realistic and implementedas intended in a predictable manner.

• Comprehensive, policy-based budget: Budgetcaptures relevant fiscal transactions and is preparedwith due regard to government policy.

• Fiscal management: Aggregate fiscal position andrisk are monitored and managed.

• Information: Adequate fiscal, revenue, and expendi-ture records and information are produced, maintained,and disseminated to meet decision making, controlmanagement, and reporting purposes.

• Control: Arrangements are in place for the exerciseof control and stewardship in the use of public funds.

• Accountability and transparency: Arrangementsfor external transparency and scrutiny of public financesare operating.

Source: Public Expenditure and Financial Accountability(PEFA) Program (2005).

86The Public Expenditure and Financial Accountability (PEFA)Program is a joint program of the World Bank; European Commis-sion; IMF; development agencies from France, Norway, Switzer-land, and the United Kingdom; and the Strategic Partnership withAfrica (SPA). The four-year program was established in December2001, and its secretariat is based in the World Bank.

87While the approach concentrates on a standard set of indicatorsconsidered relevant to all country circumstances, it is recognized acountry may add further indicators to meet specific needs.

©International Monetary Fund. Not for Redistribution

Page 65: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

V ESTABLISHING BASIC PUBLIC EXPENDITURE MANAGEMENT THRESHOLDS

One also might have need for a more basic restructur-ing of incentives in support of budget system reforms.Specifically, larger-scale administrative reforms maybe required that allow government pay scales to offeradequate incentives to attract and retain specializedskills. Discussion of this topic is beyond the scope ofthis study.

This section has also emphasized the connection be-tween the macro and micro levels of reform, whichraises the issue of how to ensure a balanced approach.When reforming budget systems, there may be no al-ternative to adopting a top-down approach and intro-ducing the macro framework first; at the same time,these macro reforms should not get out of gear with thecapacity of the system to implement them. There is aconcern that this has been the case in the past and thatthere have been instances in which the centralized in-stitutions, by moving too far too fast, have seen re-forms flounder and stall due to the lack of managementcapacity at the agency level.

56

These considerations argue for coupling budget sys-tem reforms with management improvement at theagency level and underline that the lack of basic man-agement skills in government, especially at the microlevel, can be a binding constraint on the speed at whichbudget systems can be reformed. Traditional budget-ing procedures required administrators, not man-agers. In the old systems, controls were centralizedand imposed from the top down. Reforms envisagedat the micro level, as discussed above, require a morebottom-up focus and depend on adequate manage-ment capacity at lower levels to be successful. Not onlydoes this take time to create, but the well-documentedbureaucratic resistance to change will only be over-come with time (and special skills). This implies asecond order of management skills—the need forcentral agencies when introducing budget system re-forms to invest in developing a change-managementstrategy and in change-management skills. Thisissue will be explored further in Section VIII.

©International Monetary Fund. Not for Redistribution

Page 66: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI Moving from Cash to Accrual Accounting

57

This section reviews the role of accounting in bud-get system reform from the perspective of emerging

economies who wish to adopt the OECD performancebudgeting reforms.

Emerging economies and several economies in tran-sition look to the OECD member countries as modelsfor reforming their PEM systems. Accordingly, theyseek to replicate the accounting reforms in the moreadvanced OECD countries—usually by moving gov-ernment accounting from a cash to an accruals basis.88

In fact, this conversion has been fully achieved in onlya handful of OECD countries, although others havehad to modify their pure cash-based systems to mod-ified accrual systems to accommodate accrual-basedreports.89 Accrual-based reporting has received con-siderably more attention with the recent change ininternational government finance statistics (GFS) re-porting standards which recommend adding accrualinformation to existing cash data and reporting cashdata consistent with an accruals basis.90

Many emerging economies have too eagerly ac-cepted this reorientation and have overlooked a num-ber of important issues. This section examines one ofthe more important of these issues: the need to recog-nize the relationship between reforming the govern-ment’s accounting system and wider budget systemreforms. First, moving to an accrual accounting sys-tem, given the costs involved, is perhaps only worth-while in the context of wider public sector managementreforms. Second, since the reform process inevitablytakes time, it should be implemented in phases, andpossible stages for moving to accrual accounting aredescribed here. Third, many countries are constrainedfrom moving in this direction in the near future andmay require intermediate solutions for realizing someof the benefits afforded by accrual accounting in cash-based systems. Fourth, although accrual accountingsystems are more comprehensive and provide a wealth

of financial information, it is important not to overstatetheir benefits nor to overlook the typical requirementsof budget management for supplementary informa-tion. Finally, this section reviews the approach takenby countries that have adopted the new accrual-basedGFS reporting requirements, noting the difficulties ofmeeting the new requirements without adopting fullaccruals and describing possible stages for adoptingthe new international reporting requirements whichparallel the phased move to accrual-based accounting.

Cash Versus Accrual Accounting

To understand the attraction of accrual over cash ac-counting, it is best to first clarify the differences be-tween the two systems. The fundamental distinction isthe so-called basis of accounting, which describes thebasic measurement rules for when accountants regis-ter, or “recognize,” the effects of an entity’s transac-tions and is based on the timing of registering thetransaction’s effects. In the accrual system, recognitionoccurs with the exchange of liabilities and/or assets,which is usually before the cash flow and closer to theactual economic impact of the transaction (Chan,1998, pp. 362ff). In contrast, the cash basis, which isused in most traditional budget systems, records out-lays and receipts only when they involve cash transac-tions. The simplicity of a cash-based system providesboth advantages and disadvantages (Box 30).

Despite its simplicity, the use of cash accounts canpresent many problems in practice. Many of the prob-lems arise not from the accounting system as such, butrather from its abuse. Cash accounting is often consid-ered more objective and less subject to manipulationthan accrual accounting, since it requires less judg-ment to record cash transactions than to recognize andmeasure assets and liabilities. However, cash account-ing also requires judgment, for example, to deter-mine whether transactions are on or off budget or thetiming in which they are recorded, or recognized, inthe government accounts. Also, one strength of cashaccounting—that it can serve as a constitutionalsafeguard because the legislature must approve allmonies spent and collected—can also be its princi-pal weakness in countries with governance prob-lems, particularly where there is political pressure tohide rather than reveal.

88About half of OECD member countries have adopted accrualsto some degree. There are great variations, however, in the extentof this movement: some reforms are limited to agency- and ministry-level financial reporting; some encompass the whole of governmentfinancial reporting; and some involve only budgeting. See OECD(2002a).

89Such as those required by the Maastricht criteria.90The IMF’s standards for the reporting of GFS emphasize ac-

crual concepts as a statistics standard and not as an accounting stan-dard; the IMF standards are designed for macroeconomic analysisof fiscal data. See IMF (2001a).

©International Monetary Fund. Not for Redistribution

Page 67: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

Apart from such practical problems, there are legit-imate limitations and drawbacks to cash accountingwhich can only be resolved by ubiquitous “memo-randa items.” In particular, to the extent that goodsand services are not paid for and revenues are not col-lected during a particular year, carryovers in the formof payables (a liability) and receivables (an asset)should be reported in the balance sheet at year’s end.However, there is no full balance sheet in a cash sys-tem. In addition to the problem of identifying such ar-rears, there are a number of financial transactions,including on-lending operations and government loansto enterprises, among others, that are difficult to cap-ture in cash accounts because the only ledger balancecarried forward is the cash balance.91 There is nomechanism in the formal accounts to carry forwardother parts of the ledger balances, in particular finan-cial assets and liabilities, as opening balances in a newfiscal year. The accrual basis does account for theseinter-period effects and, in doing so, provides a more

58

comprehensive measurement of the government’sfinancial position.

Not surprisingly, adopting an accrual basis comes atthe price of introducing considerably more complexityinto the accounting process. For example, there is lat-itude in deciding the degree of accrual to be used (thatis, in determining which assets and liabilities shouldbe included and which should be excluded), and thereis usually considerable subjectivity in applying therecognition criteria (for example, whether a contingentliability is to be recognized in the accounts or whetherit is recorded off the balance sheet).92 Also, the recog-nition of revenues can be quite elastic: Should cash andcurrent financial resources be included? noncurrent fi-nancial resources? Can a government recognize rev-enues as soon as it acquires a legal claim, that is, themoment a taxable transaction takes place, or should thegovernment be treated like a business and recognizerevenues only when they are “earned” or received? Toavoid too much subjectivity, such questions must beanswered according to well-defined government ac-counting standards.

Spending recognition is, of course, a critical as-pect of budget management. The budget spendingprocess goes through distinct stages: after the budgetis approved and appropriations are determined, gov-ernment agencies make commitments against theseappropriations; goods are received and verified, atwhich point the obligation to pay is usually recog-nized; the payment order is issued and then eventuallycashed; and the government’s account is debited.Some argue that a liability should be recognized at thecommitment stage, when the government first takescommand of resources, rather than at the verificationstage, when goods and services are supplied. However,not all commitments are the same. Some commitments,such as wages, are converted into cash with minimaltime lags; others, such as investment spending associ-ated with large-scale, multiyear projects, involve con-siderable lags. Furthermore, it is necessary to recognizethat not all commitments will be fulfilled, because notall contracts will be honored. As a result, for recog-nition purposes, the verification stage is generallypreferred as the point when the liability is reliablycreated.93 The budget manager, however, must track allthe stages of the spending process, and regardless of thebasis of accounting (cash or accrual), sound budgetmanagement requires that attention be paid to accrual

Box 30. Pros and Cons of Cash-Based Accounting

Pros

• Easily comprehended by users.• Allows judgment on compliance with budget appro-

priations, traditionally cash based.• Simple to implement, with ease of compilation

facilitating reports that are timely, reliable, and com-parable.

• Costs are low due to the lower level of accounting skills required.

Cons

• Limited scope, not good at showing the impact oftransactions resulting in cash flows outside the currentreporting period.

• Cannot meet the demands for information on assetsand liabilities, impact of current consumption of thestock of assets held by government, transactions inkind, and arrears.

• Focuses solely on cash flows and ignores other re-source flows which may also affect government’sability to provide goods and services now and in the future (e.g., accumulated borrowings and otherliabilities).

• Limits accountability to use of cash, and ignores ac-countability for management of assets and liabilities.

Source: International Federation of Accountants (1998).

91Even government debt has to be recorded through subsidiaryregisters and balances carried forward.

92Also, for example, generally accepted accounting principlesleave off the balance sheet investment in human capital as well asprojected receipts, the tax base, or the power to tax (all assets).

93Hence, on an accrual basis, flows should be recorded when eco-nomic value is either exchanged or transferred (“economic trans-actions”) or created, extinguished, or transformed (“other economicflows”).

©International Monetary Fund. Not for Redistribution

Page 68: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Accounting Models for Different Management Needs

59

concepts, such as payables and contingent liabilities,and not just cash flows. Therefore, the speed at whichcommitments are made and goods received and verifiedis important for planning future cash needs, but the dif-ference between outstanding orders due for paymentand those overdue for payment is also very important.

Accounting Models for DifferentManagement Needs

As argued earlier, accounting systems should servemanagement needs. Cash accounting has some advan-tages from a management viewpoint—especially forthe central agencies of the MoF—when the principalobjective is to enforce compliance with the annualbudget law and the primary requirement is to prepareend-of-year accounts that show the amounts approvedin the budget for detailed items of spending and theactual spending on those items. Traditional budgetsystems are thus more focused on inputs and the costsof those inputs, and cash accounting serves this orien-tation quite adequately.

As budget systems have evolved to meet stabiliza-tion requirements, they have had to accommodate ad-justments to the approved budget that may be requiredduring the financial year due to unforeseen changes ineconomic circumstances.94 Recognizing that govern-ment spending decisions have an impact on the econ-omy before cash payments are made in turn highlightsboth the economic impact of budget expenditures andthe need to monitor the different stages of spending.This has led to demands for noncash information—forexample, commitments made, commitments comingdue, and outstanding obligations to be met in cash.Thus cash accounting and reporting have been supple-mented by what may be termed “modified cash ac-counting,” a term for which there is no authoritativedefinition and which is used here to signify cash ac-counting with some supplementary accrual informa-tion. These procedures are typically not regarded aslong-term solutions, but rather as ad hoc and selec-tive, arising from particular problems and avoidingthe rigor of a well-constructed overall accountingframework.

Having ensured compliance with their budget sys-tems and put in place improved mechanisms to ensuremacroeconomic stabilization objectives, in the lastthirty years OECD governments have turned to focuson efficiency and effectiveness objectives.95 In this en-

vironment, accrual accounting has certain attractions.The degree to which these countries have reorientedtheir accounting systems is summarized in Box 31.This shift did not arise because there was no need forcash accounting and reporting, but rather because theadditional accrual information offered many advan-tages. Specifically, an accruals system can and mustgenerate cash information, so that it is equally able tomonitor short-term effects of fiscal policy. At the sametime, an accrual system recognizes financial flows thatdo not result in cash flows—such as depreciation,transfers, write-offs of physical assets, and accruedinterest—which are necessary to consider from an ef-ficiency viewpoint because they indicate the full costsof providing government services and also allow awider assessment of government operations.

There are four major advantages to accruals:• Improved resource allocation: Because of its abil-

ity to provide more comprehensive information onthe costs of operations, accrual accounting allowsdecision makers to make better resource allocationchoices. A common problem of cash accounting isa lack of good information on the full costs of pro-viding services. For example, the cost of utilizingcapital assets paid for in prior years (depreciationcosts) is frequently ignored. Perhaps more impor-tant, where fixed asset values are based on currentcosts, the depreciation amount can also provide anindication of the future expenditure required to re-place existing assets at the end of their useful livesand maintain current activity levels. It also can bedifficult for the government to evaluate capital pur-chase proposals under cash accounting systems ifneither the government nor the operating unit hasgood information about what assets are owned byagencies and therefore cannot judge whether thereis any scope to finance a new asset by liquidatingsome existing assets. Although this informationcan be recorded in subsidiary ledgers under a cashsystem, it is an important focus in accrual ac-counting systems. Similarly, with accrual accountsthe pensions of public servants are included as anexpense and the additional liability is shown on thebalance sheet. This can be important if annual cashoutlays differ significantly from the accruing lia-bilities. In particular, if accruing liabilities signifi-cantly exceed cash payments, there could bequestions about the sustainability of the non-pay-as-you-earn (non-PAYE) pension arrangements,in addition to broader issues of intergenerationalequity.

• Strengthened accountability: A side product of theavailability of accrual information is greater precisionin determining management performance. Man-agers are more accountable when they become re-sponsible not only for cash inflows and outflows,but also assets and liabilities. With the introduction

94It is also argued that accruals improve the reporting of the un-derlying fiscal position by removing year-to-year variations arisingfrom the timing of large cash receipts and payments. See IMF(2001b).

95See OECD (1994, 1995).

©International Monetary Fund. Not for Redistribution

Page 69: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

of accrual accounting, agencies need to more pre-cisely specify performance criteria because the fullresource costs of delivering services is an elementof the financial and nonfinancial information usedto manage the government and its spending agen-cies. One of the shortcomings of existing cashaccounting regimes is that attempts to measureagency performance can be divorced from con-sideration of whether the agency involved can beheld accountable for those results. Under accrualaccounting, agencies account only for those pub-lic resources for which the government wants tohold them accountable. It is possible, for example,that resources administered by agencies on behalfof the central government need not be reported ondepartmental financial statements.

• Enhanced transparency about the total resourcecosts of government activities: Accrual accountingprovides a greater range of financial informationon the operations of government and therefore hasthe potential to enhance transparency and improvefiscal responsibility. Governments generally havesignificant assets and liabilities, and the disclosureof this information is an essential element of fiscal

60

transparency and accountability. This is facilitatedby an accrual accounting framework, which re-quires a full statement of assets and liabilitiesand revenues and expenses and, most importantly,integrally links these with one another—the differ-ence between revenues and expenses directly af-fects the net assets position, and movements incash balances are reconciled to results of opera-tions and movements in other assets and liabilities.This complete picture provides key information formacro fiscal management and also serves as a pow-erful tool for strategic planning and financial man-agement. Noncash transactions, which are left outof cash accounting but which have an obvious eco-nomic impact, are routinely and systematically re-ported under an accrual framework. This wouldinclude such transactions as off-set arrangements,creation and settlement of liabilities (including pay-ments due), assumption or settlement of debts ofstate-owned enterprises, and off-budget funding bydonors. In this way, the true cost of government op-erations becomes transparent.

• A more comprehensive view of government’simpact on the economy: Similarly, because of the

Box 31. OECD Practices: Reorienting toward Accruals-Based Accounting Systems

Percentage of OECD Countries

1. If budget is on nonaccrual basis, what transactions are treated on an accrual basis:a. Interest on government debt 42b. Civil servant pensions 25c. Wages and salaries 17

2. Plans to change the accounting basis of the budget:a. Present additional accrual basis information 33b. Introduce a full accrual basis 7c. Full accrual basis considered and rejected 4

3. Accounts integrated into revenue and expenditure accounts to prepare financial statements:a. Assets 72b. Liabilities 72c. Government equity 68

4. Basis of accounting used for whole of government financial statements:a. Full cash basis only 37b. Full accrual basis only 7c. Both full cash and full accrual 4d. Cash basis with exceptions 7e. Accrual basis with exceptions 15

5. Full cash basis used with main exceptions:a. Interest on government debt 33b. Expenditure on fixed assets subject to depreciation 22

6. Full accrual basis used with main exceptions:a. Tax receipts 7b. Capital expenditure treated as ordinary expenditure 4c. Land and natural resources 4

Source: OECD (2003, Table 4.2).

©International Monetary Fund. Not for Redistribution

Page 70: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Accrual Accounting as a Component of Wider Budget System Reform

61

integrated nature of accrual accounts, differencesbetween above- and below-the-line information infiscal reports needed for fiscal management can beeliminated or significantly reduced. The cash flowstatement clearly shows the reconciliation betweenthe net cash flow and the change in the cash andbank balances as reported on the balance sheet.Moreover, the integrated framework can improvethe quality of cash flow data, allowing for betterquantification and tracking of the cash impact ofgovernment activities by classifying the flows asbeing related to operating, investing, and financialactivities. For example, just as in the private sec-tor, it is easier to see whether operating activitiesare generating a surplus or at least breaking even,whether operating surpluses are being plowedback into investments, and whether sales proceedsfrom investing activities are being utilized to retiredebt or pay for operating activities.

Accrual Accounting as a Componentof Wider Budget System Reform

The benefits to be derived from accrual accountingclearly can be wide-ranging and can affect resourcedecisions at different levels. For example, politicianscan increase their control over resources through ac-crual reporting on a whole-of-government basis. Sim-ilarly, the central policymaking agencies can use morecomprehensive and meaningful data to develop poli-cies, to judge their sustainability in the longer term,and to undertake more comprehensive financial analy-sis of government operations. Accrual accounting isalso useful for the management of assets, particularlynonfinancial assets and liabilities. Arguably, however,the main benefits of accrual accounting are to be seenat the operational level, where managers can use morecomplete information to make decisions, to enhancetheir accountability, and to assess their performancemore comprehensively and consistently.96

The case for moving to accrual accounting seemscompelling. However, a change in the accounting sys-tem should follow rather than lead more general bud-get system reforms. Indeed, gaining the full benefitsof accrual accounting requires parallel changes in thePEM system.97 As explained, the new performance

management reforms rest on this reformulated ac-countability framework.

While different countries have adopted their own in-terpretations of accountability and performance, twoimportant features have emerged. The first is that per-formance is redefined in terms of outputs and out-comes. Accrual accounts give managers the means tofocus on the resources used and the prices paid forthose resources. In this way, accrual accounting is notan end in itself, but rather, a means of shifting the em-phasis of the budgetary process away from cash inputs,toward outputs and outcomes. The expectation is thatthis not only will result in greater management effi-ciencies and hence better outcomes for governmentsand the communities they serve (Carlin and Guthrie,2000, p. 3), but also will facilitate comparisons be-tween government agencies and potential private sec-tor (and voluntary) service providers. The latter hasbeen the basis for the move to delegate, devolve, anddivest from direct government provision of public ser-vices that has been the cornerstone of recent OECDbudget reforms.

The second feature rests on a distinction between thepurchasing and ownership interests. In the public sec-tor, the government is typically both the owner of in-stitutions and the purchaser of services from theseinstitutions on behalf of its citizens. The informationderived from accrual accounting makes this tensionmore explicit and forces the government to considerthe trade-offs between its ownership and purchasinginterests and to settle on a price for the use of its assets.

The implications of this general reorientation in bud-get management are far-reaching, and no consensus hasyet been reached on the net benefits of many of these:specifically, the need for changes in financial reportingto ensure accountability, the need for parallel changesin the budget process, the need for greater emphasis onefficient asset management, and the need for parallelchanges in the PEM system. These are examined turn.

Required Changes in Financial Reporting toEnsure Accountability

To strengthen the new performance accountabilityrelationships, performance measures and reportingshould mirror those of the private sector, and hence itis argued, government should adopt the private sec-tor’s form of accounting. Under accrual accounting,reporting requirements for government are basicallythe same as for the private sector. Three main reports—statement of financial position, operating statement,and statement of cash flows—are supplemented by anumber of other reports, including statements of ob-jectives, service performance, commitments, contin-gent liabilities, and accounting practices. Often, theargument that the public sector should adopt generallyaccepted private sector accounting practices is taken

96Hood (1991, 1995), described the New Public Management ascentrally involving, first, a lessening or removing of differences be-tween the public and private sector, and, second, a shift in empha-sis from process accountability to accountability for results.

97“As reporting is but one aspect of the whole management cycle(planning, short-term budgeting, implementation, performancemonitoring and reporting), it is a truism that management planningand budgeting in the public sector should be undertaken using thesame basis of accounting to ensure the benefits of reforms areachieved.” (Mellor, 1996)

©International Monetary Fund. Not for Redistribution

Page 71: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

too far, lending support to those that have taken a con-trary view. They would argue that in the private sector,annual financial accounts identify the financial impactof a company on its individual shareholders. Accrualaccounting is designed to put the focus on the bottomline—the concept of profit—and to satisfy the pre-dominantly financial accountability of the manage-ment of a private sector company to its shareholders.98

In public sector organizations, the same concept, profit(or surplus), cannot be expected to govern decisionmaking. For example, a surplus need not equate withgood management, but only that the service was over-funded.99 Perhaps more important, for governmentthere is a wider political accountability to the elec-torate and its elected representatives that should alsoguide financial reporting.100 In the government sector,financial information is primarily used for makingpolitical rather than economic decisions.

Accountability requirements for government do notpreclude those applicable to the private sector, but arelikely to be much broader. For the public sector, ac-countability requires that government accounts iden-tify the financial consequences of government policiesfor individual citizens. Adopting a narrow financialviewpoint refocuses and dilutes the stewardship func-tion of financial accounting in government and oftencan result in financial statements that do not include acomparison of actual income and expenditure with theamounts authorized in the annual budget—a crucialconcept of the accountability framework for govern-ment. In the government sector, the budget-settingprocess is still seen as a fundamental aspect of finan-cial control. A government’s financial accountabilityarises from the budget-setting process by which agree-ment is obtained on the level of taxation and the fund-ing allocated to provide various services. Thus thebudget outturn report is the prime document by whichgovernments are held to account for the regularity andprobity of their financial management. Unfortunately,both the Accounting Standards Board (2003) and the

62

International Federation of Accountants (2003a, 2003b)have deemed the budget outturn reports one of the op-tional additional statements that public sector orga-nizations produce.

Parallel Changes in the Budget Process

If accrual accounting is adopted, is there also a needto move government to accrual budgeting? Logically,the budgetary process should be consistent with the expost reporting format.101 If this position is carried to itslogical conclusion, there are two main budgetary im-plications of accrual accounting. First, ex ante and expost performance information should be harmonized.Agencies should move both their ex post reporting andex ante budgetary documentation to an accrual basis.This would require agencies to prepare prospectivefinancial statements as well as prospective serviceperformance (output) information for inclusion inthe estimates. Second, the major budgetary constrainton agencies—the appropriation process—should bein a format that is consistent with the performancesystem. The appropriation process under an accrualbudgetary regime would appropriate for:

• the purchase of goods and services (outputs) fromagencies, other government-owned entities, andprivate sector entities;102

• capital injections to increase the government’s netasset holding in a department or agency; and

• other payments of benefits and grants.A handful of countries appropriate on an accrual basis,

but having an accrual-based, performance-orientedPEM system does not necessarily imply accrual appro-priations (Athukorala and Reid, 2003, pp. 51ff). It isconceivable that performance projections can be es-tablished on an accrual basis, the cash requirements as-sessed, and appropriations made accordingly for thecash required for the period.103 This approach, how-ever, could result in dual, and potentially conflicting,performance objectives: on the one hand, accrual per-formance against accrual expectations, and, on theother, compliance with cash appropriations. In short,the appropriation process would determine, rather thanreflect, the criteria for agency performance. Moreover,focusing on cash would again fail to give the legisla-

98The user of financial statements is assumed to be “the investorthat has supplied risk capital in expectation of a financial return oncapital.” Thus “the defining class of user for financials statementsof public benefit entities is the funders and financial supports.”(Accounting Standards Board, 2003, p. 29)

99As Chan (2003) puts it: “Governments uniquely provide publicgoods and finance them through taxation . . . These characteristicssever the link between service delivery and revenue recognition,making it impossible to match revenues and expenses.” (p. 5)

100As articulated by U.S. Federal Accounting Standards AdvisoryBoard (1993): “The Federal government derives its just powersfrom the consent of the governed. It therefore has a special respon-sibility to report on its actions and the results of those actions. Thesereports must accurately reflect the distinctive nature of the federalgovernment and must provide information useful to the citizens,their elected representatives, federal executives and program man-agers.” (para. 8)

101The use of accrual appropriations is now a feature of the Aus-tralian “accrual output budgeting” approach. For a description ofthe general outlines of the system, see, for example, Australian De-partment of Finance and Administration (1999).

102Ideally, the appropriation for output purchases should be madeon the basis of a contestable market price, with agencies invited toquote prices for the provision of the goods and services the gov-ernment is contemplating buying and those prices compared to al-ternative quotations from private sector suppliers where available.In practice, this is likely to occur only in exceptional cases.

103This is the approach of the recent French reforms; see France,Ministère de l’Economie, des Finances et de l’Industrie (2001).

©International Monetary Fund. Not for Redistribution

Page 72: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Accrual Accounting as a Component of Wider Budget System Reform

63

ture effective control over the true consumption of re-sources and negate one of the great advantages of mov-ing to an accrual system.

Although accrual accounting records past eventsand provides more informed public reports on the trueuse of resources, it perhaps does not serve well for thediscrete ex ante resource decisions that are taken in thecourse of budget approval. The budget is a system ofauthorization and control that benefits from the sim-plicity provided by recording resource costs up front,and arguably is more readily measured by cash. In thisway, the essential budget process requirement is to in-clude all costs for budget decision making. Apart fromits direct costs, a program could be charged for its in-direct costs, including support goods and services,joint costs such as building maintenance and security,and the costs of accruing employee pensions and re-tiree health care. Even for capital costs, some capitalcharge could be assigned, as well as some rent for theuse of premises.

Greater Emphasis on Efficient Asset Management

The new performance-oriented accountability placesmore emphasis on efficient use of assets. However, theconcept of government assets has not proved an easyone to accommodate in accounting principles de-signed for the private sector. The Accounting Stan-dards Board adopted the private sector definition of anasset as “rights or other access to future economicbenefits controlled by an entity as a result of pasttransactions or events” and then extended this to in-clude future service potential. This concept is furtherstretched to include heritage assets. The fact is thatdefinition of assets in government is very complex.Some public facilities may be provided by natural en-dowment or gift and so there is no equivalent of his-torical cost incurred by the purchase of these assets.Some facilities must be maintained and preserved forfuture public use, and there may be no private marketavailable to realize the economic benefits from theiruse. Additionally, there may be restrictions imposed sothat disposal of the facilities is prohibited or requiresgovernment/political approval.

Even so, too much can be made of these definitionand valuation problems. Perhaps more significant isthe focus on the maintenance of assets. In the privatesector, the recognition and valuation of the company’sassets have a significant impact on the bottom-lineprofit that is reported. In the public sector, the valua-tion of assets is much less important; reporting theextent to which public sector capital assets are main-tained is far more important than reporting their totalvalue. Recognizing this, the U.K. Treasury adopted acost-benefit trade-off in determining which assets areto be valued in its resource accounts. Nonoperating as-

sets such as the crown jewels, Stonehenge, and paint-ings in the National Portrait Gallery are all given anominal value, but new additions to the latter collec-tion will be included in the balance sheet at their pur-chase price or value when acquired.

Parallel Changes Required in the PEM System

The most important implications of the move to anaccrual-based accounting system are for the PEM sys-tem. Accounting for agency activities is probably bestdecentralized to the agencies themselves, and the cen-tral accounting functions of the MoF would then focuson provision of bookkeeping for payments and re-ceipts of the core central government functions; con-solidation of departmental information; promulgationof accounting policies for the government; and pro-duction of financial reports at the government level.Along with decentralization of the accounting func-tion, there may also be a case for the delegation ofasset management. The government’s investment ofcapital assets in agencies is usually significant. Toimprove the performance of capital assets, it couldbe argued, an agency should have the freedom to selland buy those assets insofar as doing so would notincrease the government’s total investment in theagency. The rationale is that it is difficult to hold agen-cies accountable for their performance if they do nothave the delegated authority to meet the performancegoals. With this delegated authority over assets, in-ternal control and audit would also need to be de-centralized (albeit with central oversight). Similarly,audit and finance regulations would most likely needamending. Agencies would provide new sets of finan-cial statements under the accrual accounting system.The impact on the audit procedure needs to be care-fully examined in the implementation process, notleast because successful implementation of accrualaccounting depends heavily on the willingness of theexternal auditor to support the move. Significant revi-sions are likely to be required in financial regulationsto provide for accrual accounting policies and proce-dures, departmental reporting requirements and for-mats, and management of departmental assets andbank accounts. Last, but far from insignificant, is theneed to upgrade computer and human resource ca-pacities. Accrual accounting is more complicated thancash accounting and requires a heavier investment incomputerization and skilled labor.

The change to public sector personnel policies isperhaps the most profound and revolutionary. At thesame time that the above changes create demand forskilled labor, the agencies will be shifting to a mode ofoperating that has more in common with the privatesector, which means that there will be greater oppor-tunities for labor mobility between the public and pri-vate sectors—especially for workers with managerial

©International Monetary Fund. Not for Redistribution

Page 73: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

and accounting skills. In fact, the new reforms not onlypresent such possibilities, but also provide incentivesto exploit them. For example, if agency heads are to beheld much more tightly accountable for the perfor-mance of their departments, they will become muchmore interested in the quality of the staff assigned toassist them in this task. They are likely to want themajor say in key appointments, and they may alsowant a freer hand in dismissing nonperforming staff.This could lead, and in some OECD countries has led,to fundamental changes in public sector employmentarrangements.

Preconditions for the Move to Accrual Accounting

The potential benefits of moving to an accrual-basedsystem make it easy to appreciate why advocates ofPEM best practices tend to include accrual accountingsystems as an essential element.104 However, this doesnot imply that they advocate its universal and imme-diate application. Accrual accounting is difficult todivorce from other elements of wider and more funda-mental budget system reforms, which is why, despiteits obvious advantages, it is likely not to be readilyadopted by governments.

For a large number of developing and transitionalcountries, a move to accrual accounting is of dubiousrelevance. Until basic compliance with the rule of lawis established and some basic controls are in place toallow government spending to be adjusted to ensuresome reasonable level of fiscal stability, the move froma centralized, traditional budget management model toa more decentralized model has inherent dangers andis perhaps to be discouraged. For countries with gov-ernance problems, the complexity of accrual ac-counting can be readily employed to hide and distort.Accrual accounting lends itself to creative accountingeven more than cash-based systems. In many coun-tries, the administrative capacity may be insufficient toundertake the major reforms implied by this move orto apply the required skills to make them work. Giventhe higher level of skills required, the move to accrualaccounting should not be encouraged for countries thatfind cash accounting difficult.

Even more advanced middle-income countries mayfind daunting the costs and challenges of introducingaccrual accounting. Much preparatory work is re-quired to launch the new accounting system. Informa-tion bases need to be built up, including the purchaseand pilot testing of new computer systems. This takes

64

time, money, and a great amount of training for thosepreparing and using the new information. Completeand consistent accounting standards and practices needto be formulated. Existing assets must be identified andvalued. In addition to the administrative challenge, itcan be difficult to foster an understanding of the fullimplications of the new performance regime that usu-ally accompanies such a move. For example, althoughgovernment financial reports in an accrual regime maylook similar to private sector financial reports, theiranalysis is different and this difference must be ac-commodated in the decision-making process. Doingthis requires determining the most meaningful methodof displaying the data in financial statements and defin-ing the required reporting requirements.105 In sum, theinvestment required to introduce an accrual system andreform the budget management system to gain the fullbenefits from this new system may be too onerous formany middle-income countries.

Apart from the resource costs of introducing accrualsystems, there are significant costs for their mainte-nance. The introduction of accrual accounting in thepublic sector is relatively recent, and the balance ofbenefits to costs is still subject to debate. It is perhapsironic that this initiative to promote efficiency has notitself been subjected to such critical analysis.106 Whilethe benefits of accrual-based accounts and reports areyet to be clearly demonstrated in practice, the costs ofsuch reforms are clear and significant, and many arerecurrent rather than being one-time. They include theincreased costs of employing and training significantlymore qualified accountants107 to serve both as prepar-ers and as auditors of government accounts and to useand interpret the accounting information. Training isalso often required for members of the legislature, thegovernment, and the MoF. The organization, careerstructure, and training of auditing staff may need to besignificantly strengthened. At the same time, there maybe the need for comprehensive training of managers invarious government agencies in the interpretation anduse of accrual-based accounts.108

It is all too easy to overstate the case for accrual ac-counting by underestimating the substantial costs in-

104The central role of accounting in the modernization of publicsector management is discussed critically in Lapsley (1999). Seealso Guthrie (1998).

105The GFSM 2001 (IMF, 2001a) prescribes a consistent frame-work for this purpose.

106For example, in the United Kingdom, which took the decisionto move to accruals in 1995, there has been no comprehensive cost-benefit study (Chow and Humphrey, 2003). As Likierman (2000)admits, “A full analysis will only be possible once the new systemshave been working for a number of years” (p. 253). Concurrently,there is some evidence that accrual accounting has not producedthe gains originally expected of it (Guthrie and Humphrey, 1999;Stanton and Stanton, 1998).

107Wynne (2004) estimates that the number of qualified accoun-tants working across the U.K. central government increased four-fold, from almost 600 in 1989 to 2,200 in 2003 (p. 13).

108As recently recommended by the European Federation of Ac-countants (2003).

©International Monetary Fund. Not for Redistribution

Page 74: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

A Five-Stage Transition to Accrual Accounting

65

volved in its introduction and maintenance. Withoutadoption of the full-fledged performance budgeting sys-tem, based on a decentralized modern agency model ofbudget management, the main benefits of accrual ac-counting are likely to be at the agency or departmen-tal level, not at an aggregate, government-wide level(Box 32). In the absence of a full accrual-based sys-tem, other changes in cash budgeting and accountingsystems—albeit not so comprehensive—can be sub-stituted at the agency level to send the correct signalsto managers.

One example is the introduction of a capital charge.A typical weakness of traditional, cash-based budgetsystems is poor asset management. The introduction ofa capital charge would foster the consideration of thetotal investment in assets and the recognition that suchinvestment has an opportunity cost which can be cap-tured in management decision making. Capital charg-ing would thus help internalize the full costs of thedelivery of goods and services and quantify the wideropportunity cost of the capital held by agencies. Thiscould generate a number of efficiency benefits: facili-tate the assessment of competitiveness of public deliv-ery; enhance transparency in costing of public outputs;provide a basis for introducing user payments; andprovide incentives for agencies to identify and disposeof underutilized or surplus assets.109 Alternative mod-els for such a capital charge are shown in Box 33. An-other possibility for introducing more realistic costingin agency outputs is the imputation of various over-heads, such as workers’ compensation and pensioncosts (as recently introduced in the United States fed-

eral government).110 Thus, within a cash system it ispossible to make departments pay for common ser-vices (e.g., accommodation, transportation) throughinternal charging and/or through the commercializa-tion of the service provider.

Finally, accrual accounting supports good PEM butdoes not replace it.111 Essentially, moving from cash-based to accrual-based accounting requires a changein the recognition of a transaction. In the cash system,a transaction is recognized when cash is received orpaid, and the financial result for a period is measuredas the difference between cash received and cash dis-bursed. Under the accrual basis, the transaction addsto or subtracts from the government’s net worth, andso assets or liabilities are recognized when obligationsare exchanged. As indicated, in terms of PEM, whichviews the spending process as a series of stages—appropriation, commitment, delivery verified, pay-ment order prepared, and payment order executed—anaccrual system records transactions near the verifi-cation stage and a cash system records them whenthe payment order is executed. However, a goodmanager will seek to monitor all stages of the spend-ing process—indeed, good expenditure control startswith good commitment control. The internal account-ing of transactions should therefore be able to recordall stages of the spending process, not just the verifi-cation (or recognition of the obligation to pay) or thepayment stage. Introducing accrual accounting maymake this task easier but will not replace it.

A Five-Stage Transition to Accrual Accounting

The experience of OECD countries indicates thatimplementation of accrual accounting is not easy; ittakes time and requires sustained political commit-ment. Moreover, it is possible to introduce many el-ements of performance-based budget managementon a cash basis. Not surprisingly, therefore, it cannotbe considered a top priority for most countries.When should countries attempt to move to accrualaccounting? Accepting that the nature and speed of

Box 32. Usefulness of Accrual Accounting forGovernment Operating Units

• Facilitates assessment of program performance byshowing the full cost of programs.

• Facilitates assessment of financial position by showingall resources and obligations.

• Enhances the accountability of managers for their per-formance.

• Acts as a spur to better management performance dueto increased transparency.

• Provides a wider range of information needed for de-cision making.

• Enables more effective use to be made of a given levelof resources.

• Provides a more effective basis for decisions aboutsuch matters as user charging, identifying savings op-tions to finance high-priority objectives, and work-place bargaining.

109In Australia, some states have had years of experience operat-ing capital charging regimes. See the discussion in Robinson(1998). See also Heald and Dowdall (1999).

110As noted, there is a wider question about how national pay-as-you-go pension schemes are treated under accrual accounting. Pre-sent practice requires treating future public service pension costs asa liability and thus recognizing the emerging expense but not the fu-ture national pension costs. GFSM 2001 recognizes only liabilitiesfor government employer pension schemes and not for social secu-rity, with contingent liabilities included only as a memorandumitem.

111As succinctly put by the OECD (2002a): “Accruals is not a‘magic bullet’ for improving the performance of the public sector.It is simply a tool for getting better information about the true costof government. It needs to be used effectively and in tandem witha number of other management reforms in order to achieve the de-sired improvement in decision-making in government.” (p. 11)

©International Monetary Fund. Not for Redistribution

Page 75: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

66

Stage Two: Integrate Operating Accounts andFinancial Asset and Liability Accounts to Move to Modified Accrual

Integrating asset and liability accounts and operatingaccounts is a significant step forward. It has the advan-tage of ensuring that all transactions are treated in aconsistent, self-balancing framework. For example, thepractice, say, of financing the recapitalization of banksby issuing government debt to build up a bank’s assets,without showing a corresponding deficit above the line,would be fully reflected in any form of accrual ac-counts. An acceptable second-best solution may be topropose an explicit standard for reporting such trans-actions, such as a memorandum item in the operatingaccounts of cash-based systems.

An essential element of this adjustment of cash ac-counts is to include payables and receivables:

• Accounts payable: This allows for the recording of liabilities that have not resulted in the payment ofcash in the current accounting period. It should in-clude goods delivered but not paid for and agreementsto pay subsidies and grants to the private sector.

Box 33. Design of the Capital Use Charge

Net Assets

Charge is determined on total assets controlled by agenciesless liabilities under their control.

Pros

• Uses basic information available in accrual accountingand applies consistent valuation principles.

• Identifies total cost of capital.• Allows fair comparison between public and private

providers.• Gives agencies an appreciation of their cost of capital.

Cons

• Relies heavily on the valuation base, which is usually notwell developed in cash accounting regimes.

Capital Flow

Charge is levied on funds committed to new capital expen-diture, net of asset sales.

Pros

• Is easy to implement.• Can be linked with appraisal of capital projects.

Cons

• Only covers flows of new capital expenditure and cannotadequately accommodate management of capital stock.

• Cannot provide managers with information on full capitalcost of outputs.

Debt Assignment

Debt is assigned to agencies based on the proportion ofassets they control, and charge is applied to agencies’proportion of the debt.

Pros

• Provides incentives to reduce debt.

Cons

• May underestimate capital use if agencies’ asset bases arenot debt-funded.

• Does not capture new capital funded from nondebtsources.

Debt Equity

Debt and equity are assigned to agencies and agencies arecharged interest on their proportion of debt; in addition, acapital charge is made to the residual government equityafter deducting the debt allocated to them.

Pros

• Encourages focus on all assets and provides for full cost-ing of outputs.

• Is more in line with private sector practice.• Identifies the separate debt and equity elements of the

asset base.

Cons

• Is more sophisticated and requires substantial managementskills and complementary information systems.

the progression is likely to differ among countries,and that the transitional steps may be unique to eachgovernment, it is possible to delineate five broad stagesin a possible strategy for the reform of accountingsystems.

Stage One: Get Cash Accounting to Work Well

This will typically involve a two-pronged approach.First, purge the system of common abuses, and second,supplement the cash accounts with adjustments to im-prove fiscal reporting.

As suggested, some benefits of accrual accountingcan be derived by adjusting the cash basis to include ac-crual data. For example, accrual accounting may betterreflect emerging liabilities such as unfunded public ser-vice pensions, but multiyear cash-based expenditureplans equally highlight the problem. Similarly, mostcash systems are weak in providing information onpayment arrears. A logical step is to add systems thatcan generate information on commitments undertaken,bills payable, and bills due for payment.

©International Monetary Fund. Not for Redistribution

Page 76: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

A Five-Stage Transition to Accrual Accounting

67

• Accounts receivable: This allows for the recordingof revenue earned by the government that has notresulted in the receipt of cash, although it is suffi-ciently close to cash to be reasonably secured. Itshould include taxation and nontaxation revenuesbeing processed, including credit sales of goodsand services.

Stage two should be regarded as a reasonable targetfor most developing countries, because any movementto incorporate accrual concepts results in more usefulinformation for fiscal policymakers. A system basedon cash recording and reporting but with an integratedset of operating and financial asset and liability ac-counts would achieve the operational advantages ofintegration, while not involving the additional com-plexities of maintaining accounts on an accrual basis.Box 34 indicates some of the steps involved in movingto this stage.

Stage Three: Introduce More Elements ofAccrual Recording and Move to a PartialAccrual Presentation in ex post Reporting

Additional elements of accrual accounting that couldbe recognized in this stage include the following:

• Provisions for employee entitlements, such as pen-sions (linked to years of service and leave), as theemployee earns them: Such entitlements havebeen generally shown to be a significant hiddencost of government which result in large unfundeddemands on budget resources in future years.Recognition of the buildup in such demandsthrough provisioning provides budget managerswith useful early warnings of possible future prob-lems in the cash funding of these entitlements andenables corrective action to be implemented.

• Prepayments received by government: These re-ceipts can range from deposits on the sale of assetsto installment payments on the provision of gov-ernment goods and services. Such receipts can beused to inflate the fiscal result for the current ac-counting period and consequently understate thefiscal result for future periods. However, becauseconditions have not been met for their recognitionas government revenue, they should not be treatedas revenues, but shown as financial transactions af-fecting assets and accounts receivable.

• Interest payable: Interest on debt can be a sig-nificant drain on budget resources, and simplyrecording interest as it is paid may not provide ad-equate information on future trends in interestpayments and whether they will place acceptabledemands on budget finances. This is particularlythe case with zero coupon bonds. This informa-tion would complement expenditure control andthe level of funds held within the commercialagency banks.

At the end of this stage of the move to accrual ac-counting, ex post reporting of the budget would in-clude a partial balance sheet with selected financialassets and liabilities, and the adjusted cash flow oper-ating statement would include some items on an ac-crual basis.

Stage Four: Recognize Nonfinancial Assets—Final Stage for Accrual Accounting

This transition from recognizing only financial assets torecognizing both financial and nonfinancial assets greatlycomplicates the accounting process. This requires thatconsistent valuation practices be applied to all govern-ment nonfinancial assets—many of which are not easilysubject to a market-related assessment of value. Once thistask has been completed, depreciation can be charged asan expense for each accounting period, thereby providinga better indication of the full costs of government opera-tions. This stage has not proved easy, even for the mostadvanced OECD countries.

Only at this stage would full accrual ex post report-ing be introduced to include:

• operating statement of performance showing howrevenues and expenses explain the movement inthe net stock of assets;

Box 34. Steps Involved in Moving to Modified Accruals

1. Adopt a classification structure that facilitates therecording of revenues, expenses, assets, liabilities,and cash flows (see IMF, 2001a).

2. Ensure that the general ledger is based on a double-entry system.

3. Explore the best option for recording and reportingselected assets and liabilities.

4. Generate and agree on trial balances.5. Establish a process for the reconciliation of assets

and liabilities in the general ledger with subsidiaryrecords, such as accounts receivable and payable andfixed assets.

6. Similarly, reconcile accounts with independent third-party information where available (e.g., ledger bal-ances with bank statements).

7. Publish statements of contingent liabilities and out-standing commitments as part of budget documen-tation.

8. Establish and train an asset valuation unit to developappropriate valuation methods and value all govern-ment financial assets.

9. Develop a statement of government financial assets(initially at historic cost, unless market valuation hasbeen established), including investments in all para-statals and liabilities.

©International Monetary Fund. Not for Redistribution

Page 77: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

• balance sheet of financial position for the begin-ning and end of the accounting period; and

• cash flow statement showing cash flows embodiedin assets, liabilities, revenues, and expenses clearlydistinguishing between operations, investmentactivities like loans, and advances and the financ-ing of cash flows through the issue of governmentsecurities.

Stage Five: Move from Accrual Accounting toAccrual Accounting and Budgeting112

Accrual budgeting should be differentiated from ac-crual accounting. Only New Zealand, Australia, andIceland have so far moved to a system where the bud-get is targeted, appropriated, and reported on an ac-cruals basis. Others may report on an accruals basis expost, but they budget (and also report ex post) on a cashbasis. Logic may suggest that this final stage is re-quired for introducing a performance-oriented system,but there has been a lot more resistance to this than tothe move to accrual accounting. Undoubtedly, a majorreason is the difficulty of implementing this final stage,notably the additional information required to preparean accrual-based budget that shows projected cashflow (as existing budgets); projected revenues, ex-penses, and operating result in the operating statement;and projected assets, liabilities, and equity in the state-ment of financial position. The additional informationrequired includes:

• information included in the cash budgets currentlyprepared, namely, movements in cash and cashequivalents, cash being spent on purchase of as-sets and received for sale of assets, and estimatedfinancing transactions;

• estimated movements in inventories, receivables,payables, employee entitlements, and other liabil-ities; and

• details of asset depreciation policies.A second reason for the lack of enthusiasm for ac-

crual budgeting, especially on the part of legislators,is the suspicion that the move carries the risk of loos-ening fiscal discipline. Cash gives a very clear match-ing of the political decision to spend and the actualamounts spent, but the methodological complexity ofaccruals is thought to undermine the ability to main-tain this basic check on the executive.113

A third reason is the general lack of experience withaccrual budgeting. Accrual budgeting has been nei-ther well defined nor widely discussed in the general

68

literature, and there is not adequate country experi-ence to form a basis for guidelines in its use.

Preliminary Guidance for Accrual Accounting

The experience of countries with accrual account-ing to date may be insufficient to create firm guide-lines, but it does indicate some pointers on the way toestablishing such guidelines.

Accrual Budgeting Can Take Place at Different Levels

First, it is possible to envisage accrual budgetingbeing applied at different levels of budget decisionmaking, addressing three different objectives:

• Macroeconomic stabilization: At the aggregatelevel, fiscal targets and/or rules could be specifiedin terms of accrual concepts such as net worth, theoperating balance, net financial worth, and netlending/borrowing.

• Resource allocation: Accrual concepts—particu-larly expense concepts—could be used for makingdecisions about resource allocation between agen-cies, specifically, in prioritizing agency budget bidsand refining agency estimates, but perhaps also indetermining legislative appropriations to agencies.

• Internal resource use within agencies: As the basisof operationalizing a performance-based budgetmanagement system within an agency, accrualexpense concepts could be incorporated into theproduct costing systems used for deciding resourceallocation between different activities within indi-vidual agencies.

These three uses of accrual budgeting are, to a con-siderable degree, separable. That is, it would be pos-sible to adopt an accrual budgeting approach at anaggregate level to address stabilization issues (speci-fying fiscal targets and/or rules) while having budgetappropriations to agencies on a largely cash, or mod-ified cash, basis. The decision to introduce accrualbudgeting, therefore, cannot be made without consid-ering its use. Instead, each country has to determinewhich uses of accrual budgeting make sense given itsspecific circumstances.

Accrual Budgeting Can Be on a Partial Basis

Just as for accrual accounting, budgeting can be ona partial or full accruals basis.

• Macroeconomic stabilization: For this purpose,aggregate accrual budgeting on a partial accrualsbasis would involve either or both (i) using netlending/borrowing as one’s budget balance mea-sure for fiscal policy purposes instead of the cash

112This section has benefited considerably based on the commentsof Marc Robinson of the IMF’s FAD.

113Blondal (2003) has made the point that “it is noteworthy thatthe legislatures in those countries that have adopted accrual bud-geting generally have a relatively weak role in the budget process”(p. 44).

©International Monetary Fund. Not for Redistribution

Page 78: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Preliminary Guidance for Accrual Accounting

69

(or modified cash) budget balance, and (ii) settingtargets/limits for net financial worth (instead of thenarrower cash net debt measure). By contrast, ag-gregate accrual budgeting on a full accrual basiswould set targets for the operating balance and/ornet worth (quite possibly in addition to, rather thanas substitutes for, partial accrual variables such asnet financial worth).

• Resource allocation: Resource allocation on a par-tial accruals basis would mean that agencies wouldbid for “current” budgets that would include notonly cash expenses, but also those noncash ex-penses that create financial liabilities (obligationsto make future payments, such as accruing leaveand pension entitlements of civil servants). Itmight (or might not) also involve placing agencyappropriations on the same basis. By contrast, bud-getary resource allocation on a full accruals basiswould involve taking the further step of includingdepreciation as a component of agency’s currentbudgets (as in Australia and New Zealand).114

• Internal resource use within agencies: In the partialaccruals approach to the internal allocation of costsof programs and their activities, agencies would beexpected to include cash and most noncash ex-penses. Similarly, a full accruals performance bud-geting system might also include depreciationwhen estimating program or output costs.

There Is a Logical Sequence for IntroducingAccrual Budgeting

There is very limited experience on which to basegood practice for accrual budgeting and even less onwhich to delineate the appropriate sequencing of accrualbudgeting for emerging economies. International experi-ence to date suggests the following tentative conclusions:

• Most of the benefits of accrual budgetary resourceallocation and accrual-based performance budget-ing can be realized by placing them on a partial ac-cruals basis. The advantages are less clear, evenfor advanced OECD countries, of moving theseforms of budgeting to a full accruals basis.

• For macro stabilization objectives, accrual bud-geting could be generally applied at the aggregatelevel, based upon partial accrual concepts. Cer-tainly, an accruals budget produces economic in-dicators that are of great importance for fiscalanalysis. For example, it is much better for fiscalsustainability purposes to target net financial worththan to target net (let alone gross) debt, because net

financial worth takes into account highly relevantforms of quasi-debt.115 This suggests that, at thetime when countries are ready to move to partialaccrual accounting (stage three), they should per-haps also formulate at least some of their fiscaltargets/rules in partial accrual terms.

• Aggregate accrual budgeting on a full accrualsbasis—in the sense of setting targets and/or rulesfor the net worth and operating balance—is a use-ful means of bringing a longer-run perspective tobudgeting, particularly in the management of ag-gregate public investment. However, this is mostrelevant to more advanced countries with assuredfiscal sustainability (in particular, no present orprospective debt problems), and therefore couldwait until after stage four, when accrual account-ing is more firmly entrenched and accrual con-cepts are also more acceptable to the legislature.

• With respect to using accrual budgeting for re-source allocation, a case could be made to movequickly after stage three to place agency budgetbids and estimates on a partial accruals basis. Thisis because partial accrual accounting introducesnoncash expenses which are variable costs andwhich are therefore as relevant to decision makingas cash expenses. However, to make this move, forconsistency it would be necessary also to place ap-propriations on a partial accruals basis. Similarly,the case could be made for partial accrual budget-ing in an agency’s internal resource allocation de-cisions. Again, providing an incentive system formanagers to productively employ accrual-baseddecisions would also imply making appropriationson a partial accruals basis. However, until a fullperformance budgeting framework is in place toenforce accountability, this move to partial accrualappropriations may prove dangerous.

• The speed of moving to accrual budgeting shouldbe linked to wider budget management reforms.Appropriating even on a partial accruals basis as-sumes a well-developed performance budgetingsystem. It means that agencies are given up frontthe funds to cover such accruing liabilities as de-preciation and then are expected to reserve thosefunds to meet the relevant payments when they falldue.116 Unless performance management is well

114The drawback of including depreciation in current budget ap-propriations has been net capital appropriations, where the capital ap-propriation to agencies is expressed as gross capital expenditureminus depreciation—and with the depreciation component of the ex-penses appropriation being also available to fund capital expenditure.

115As well as relevant financial assets such as the share portfoliosof civil service pension funds, where such funding arrangementsexist.

116Agencies would receive a cash appropriation that would reflecttheir full operating costs. They would be expected to replenish theircurrent assets from accumulated depreciation, and would have theauthority to do so. But if managers are given freedom to improveperformance, what is to stop the cash allowances for depreciationto be spent elsewhere than on replacing future capital assets? In thisway, the MoF and the legislature may lose control over the acqui-sition of capital assets.

©International Monetary Fund. Not for Redistribution

Page 79: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

established, a more prudent option is for suchfunds to be held back by the MoF or other appro-priate central bodies (e.g., civil service pensionfund). But if the latter approach is followed, thenplacing appropriations on a partial accruals basiswould remain something of a fiction—the realitywould be that the portion of the appropriations thatagencies were actually given would continue to bedetermined on a cash or modified cash basis. Thissuggests that moving appropriations even to a par-tial accruals basis for use in resource allocation andfor internal agency budgeting is less pressing andperhaps should wait until accrual-based account-ing and financial reporting are well established.

The GFSM 2001 Standard for Fiscal Reporting

With the publication of GFSM 2001 (IMF, 2001a), theIMF has “raised the bar” in terms of the internationalstandards of reporting statistics for general govern-ment. The manual reorients the IMF’s prescribed fis-cal reporting system to a format consistent with anaccruals basis, harmonizing the GFS with other statis-tical systems—specifically, the 1993 revised Systemof National Accounts (SNA) and Balance of Payments(BOP) manuals. However, because GFSM 2001 is in-tended to serve a different purpose than SNA, the waysin which the data are recommended to be reported andmost of the balancing items are different from the SNAand also different from the previous cash-based GFSanalytical structure. GFSM 2001 continues to focus onthe general government, but it also recognizes the use-fulness of expanding the coverage of fiscal data to thepublic sector. Box 35 summarizes the key features ofthe GFSM 2001.

The GFSM 2001 presentation is similar in many re-spects to conventional accounting, having an operat-ing (profit/loss) account, balance sheets, and cash flowstatement. However, data is presented in a form suit-able for time series analysis, and all data flows andbalance sheet items are linked within an articulated setof accounts. GFSM 2001 also distinguishes clearlybetween economic flows that can be influenced bygovernment (“transactions”), from those that occur in-dependently of government decisions (“other eco-nomic flows”). For this reason, GFSM 2001 is mucheasier to use than normal accounting data for most dataanalysis purposes.

Countries converting to the GFSM 2001 reportingrequirements could follow the transition path previ-ously mapped out for the change from cash to accrualaccounting. Countries unable to meet the previousGFS standards should perhaps not attempt the conver-sion, but countries that are able to meet the previousGFS standards are in a position to begin the conversion

70

to the new GFSM 2001. Moreover, many of thesecountries have a substantial amount of accrual data intheir national accounting standards which can be usedto move GFS reporting closer to a full accrual basis.For example, many Latin American countries have apartial accruals system, with transactions recorded atthe time that the exchange of obligations occurs andwith generation of accounts payables and receivable.In addition, government agencies in many countriesare often required to prepare balance sheets coveringthe assets and liabilities that they control, with depre-ciation of such assets also recorded. For most coun-tries, however, the accounting for assets does not

Box 35. Key Features of GFSM 2001

• The revised GFS is strictly harmonized with the 1993revised System of National Accounts (SNA). There-fore it is on a full accruals basis, covering transactionsand other economic flows and providing complete bal-ance sheets.

• Since GFSM 2001 serves a different purpose than theSNA, the ways in which data are reported, and most ofthe balancing items, differ from the SNA. They arealso very different from the previous cash GFS analyt-ical structure.

• The essential structure of GFSM 2001 is:

Opening balance sheet + Transactions +Other economic flows = Closing balance sheet.

• The balance sheet lists all government assets and liabil-ities, differentiated by type of asset/liability, identifyingthe accumulated resources available to government forthe provision of services and the distribution of thoseresources by asset type.

• Liabilities measure the constraints placed on govern-ment policies as a result of past events. The change be-tween opening and closing balance sheets indicateswhether the government improved or worsened its po-sition as a result of economic events during the period.

• The classification of the type of asset/liability distin-guishes financial and nonfinancial assets as major cat-egories, allowing separate investment (in real assets)and financing accounts to be provided.

• A fundamental distinction is made between trans-actions that affect the net worth of the government—revenues and expenses—and those that do not. Asubaccount for revenues and expenses (i.e., an oper-ating account) provides an economic breakdown ofthese flows as well as (for expenses) a functionalclassification.

• Other economic flows are all economic events affect-ing assets and liabilities that are not transactions (e.g.,valuation changes); changes to assets from damage orloss due to natural disasters, wars, asset depletion ordiscovery; and debt write-off.

Source: IMF (2001a).

©International Monetary Fund. Not for Redistribution

Page 80: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The GFSM 2001 Standard for Fiscal Reporting

71

include common assets such as major infrastructure as-sets (roads, bridges, water supply), and for these coun-tries, it would be opportune to adopt a “cash plusaccruals” strategy based on the availability of suchaccrual-based data. For some countries, a move in thisdirection would be an important step in strengtheningtheir cash-based systems. As noted, the conversionpath to GFSM 2001 mirrors the path for making thetransition from cash to accrual accounting.

Stage One: Restructure Existing Cash Data

Without the need for additional data, the cash trans-actions are restructured. Those relating to the opera-tions of government (tax and other current revenue,wages and salaries, purchases of goods and services,interest and other current payments) are included in acash operating accounting.117 Purchases of nonfinan-cial assets are classified in an investment account, andchanges in financial assets and liabilities would beshown in a separate financing account.

Recognition rules would be simplified. The rules forrevenue recognition would stay very similar to thosein the cash system, although tax revenues would berecognized on an accrual basis. In terms of expensecategories, category No. 22 (“use of goods and ser-vices”) would not be used except to distinguish currentexpenditures from longer-lasting material assets ac-quisitions (Box 36). The latter assets would be subjectto a 100 percent first-year depreciation rule to achievea better correlation between budget “expenditure”under a cash basis and accrual-based “expense.”118

Category No. 23 (“consumption of fixed capital”), aperiodic depreciation of fixed assets, would not beadopted at this stage.

There are four advantages to implementing this firststage. The first is the ease at which it can be done. Sec-ond, the generation of a cash operating account and aseparate investment account distinguishes the gov-ernment’s operating activity from its investment pro-grams, which cause considerable problems in a purecash-based GFS system. Third, although the cash op-erating statement excludes depreciation and may havesignificant timing problems, it provides a more usefulmeasure of government operations than the previouscash-based GFS aggregates. Fourth, all changes in

financial assets and liabilities are included in the fi-nancing account, eliminating distortions resulting from,for example, privatization proceeds.

Stage Two: Use Partial Accrual Data

Some elements of the statement of economic flowsshould be accommodated. Although other economicflows are typically not recorded in cash systems, thesecan be derived by subtracting transactions from thetotal of changes between the opening and closing bal-ance sheets. While admittedly less useful than accrualrecording of these flows, at least this gives a measureof the impact of revaluations and other changes in vol-ume of government assets and liabilities which areimplicit in the opening and closing balance sheets.

Most of these other economic flows involve reval-uation of fixed and financial assets and liabilities atmarket rates. This can be substantially diluted duringthe early transition when these flows could begranted recognition only in some (illustrative) cases:for example, write-off of assets (if on the books) dueto irretrievable losses; revaluation of financial as-sets/liabilities due to exchange rate fluctuations;recognition of windfall gains/losses of assets due tonatural disasters or unforeseen accretion to wealththrough discovery of new mineral assets; and perma-nent transfer of functions to a corporation or quasi-corporation without remuneration assets (i.e., byreclassification).119

There should be during this stage a progressiverecognition of existing financial and nonfinancial as-sets. Consolidated government balance sheets cover-ing a substantial proportion of real assets, as well asfinancial assets and liabilities, could also be provided.A possible transition path is as follows: First priorityis given to financial assets including on-lending andthen to easily quantifiable assets (government build-ings and land). More difficult assets such as infra-structure assets would follow, but these can be avoidedfor an indefinite period during transition. Only in stagethree would it be necessary to attempt to include min-eral resources and other hard-to-define assets such asforest wealth, and finally (if at all), heritage assets.

This stage should also bring the replacement of cashtransactions in the operating account by the corre-sponding accrual transactions. All liabilities should, ofcourse, be recognized as soon as they arise. Even con-tingent liabilities (such as orders placed for supply) canbe accommodated in accounts by opening accountheads to capture accounts receivable in revenue andaccounts payable in expense. Depreciation costs couldalso be included as expenses for those assets for whichthey are available.

117There are not many differences between the previous GFSclassifications for expenditures and revenues and the expense andrevenue classifications in GFSM 2001, although there could be con-fusion because an accrual-based “expense” is conceptually differentfrom a cash-based “expenditure.” Certainly, when planning a newchart of accounts and designing future information systems, it maybe expedient to begin by adopting the GFSM 2001 classifications.

118Expenses are the consumption or loss of future economic bene-fits resulting in the reduction of assets or increases in liabilities. Theycontribute to an accrual fiscal indicator. Expenditures are above-the-line cash payments that contribute to a cash fiscal indicator.

119Other specific examples are discussed in section 4.49 of GFSM2001 (IMF, 2001a).

©International Monetary Fund. Not for Redistribution

Page 81: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

72

will be required for the full implementation of GFSM2001.

Stage Three: A Progressive Move to Full Accrual Data

At this stage, full introduction of accrual accountingfor government cannot be avoided. GFSM 2001 rec-ognizes that implementation of the fully integratedGFSM 2001 system will take time and will progressaccording to the circumstances of each country. In par-ticular, it highlights the expectation that, ultimately,

At completion of this stage, it should be possible toswitch over to GFSM 2001 classification structuresand a GFSM 2001–compatible chart of accounts andto generate new fiscal reports without adopting full ac-crual accounting. Although the information availablewill not allow the full implementation of the GFSM2001 framework, the changes implemented during thisstage will provide enhanced analytical information.Even the use of partial accrual data will help get themaximum value from existing accounting systems,and undertaking this work will help identify thoseareas for which a modification of accounting standards

Box 36. Stages in Accommodating GFSM 2001 Reporting

StageDescription of Government Transactions

1

Revenue11 Taxes12 Social contributions13 Grants14 Other revenue

Expense21 Compensation of employees22 Use of goods and services23 Consumption of fixed capital24 Interest25 Subsidies26 Grants27 Social benefits28 Other expenseNet/Gross Operating Balance

Transactions in Nonfinancial AssetsNet Acquisition of Nonfinancial Assets311 Fixed assets312 Change in inventories313 Valuables314 Nonproduced assetsNet Lending/Borrowing

Transactions in Financial Assets and Liabilities32 Net acquisition of financial assets33 Net incurrence of liabilities

Change in Net Worth Resultingfrom Other Economic Flows41 & 51 Nonfinancial assets42 & 52 Financial assets43 & 53 Liabilities

Full accommodations Partial accommodations

2 3

©International Monetary Fund. Not for Redistribution

Page 82: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Concluding Remarks

73

countries will have to revise their accounting systemsto an accrual basis, as well as adopt the GFSM 2001classification systems.

During the first two stages, as indicated, accrual ac-counting is not required. GFSM 2001 does admit that,because GFS is a statistical reporting system, it candiffer in important respects from the underlying ac-counting system of a country from which GFS sta-tistics are derived.120 Certainly, it is possible for acountry to operate an accrual accounting system withnational accounting standards and coding structuresthat differ from the GFSM 2001 classification struc-ture (for example, Australia). Alternatively, a countrycould accept the GFSM 2001 accrual-based systembut with very significant dilutions of the accrual con-cepts.121 The central point is that the migration of cashto accruals increasingly provides more useful infor-mation for fiscal policy formulation, execution, andmonitoring. For example, accruals allow GFS analyt-ical balances, such as net operating balance and netlending/borrowing, to be derived. In this way, puttingeven modified data in the GFSM 2001 frameworkwould aid fiscal policy and analysis. At the same time,it is difficult to see how the full GFSM 2001 statisti-cal reports could be derived from an accounting sys-tem that is not on an accruals basis. Moreover, togenerate GFSM 2001 in-year reports for fiscal man-agement purposes from current rather than historicaldata requires the integration of GFSM 2001 classifi-cations into the government chart of accounts.

The implications are clear: if countries are to fullymeet the new international fiscal reporting standards,they cannot avoid moving to an accrual-based gov-ernment accounting system. The challenge then, asoutlined here, is to ensure that the correct precondi-tions exist and that an orderly and properly sequencedtransition path is formulated and followed.

Concluding Remarks

This section examined the essential role of accrualaccounting as an underpinning for introducing budgetsystem reforms and emphasized a number of points.First, the adoption of accrual accounting should beseen as an integral part of wider budget system reform.Accounting serves rather than leads budget system re-form, and therefore, a country’s budget system modelshould determine the government’s accounting needs.

The converse of this argument also holds: to be effec-tive, and to derive maximum benefits from accrual ac-counting, other features of the budget managementsystem must be in place. In any case, moving to an ac-crual accounting system typically necessitates parallelchanges elsewhere in a country’s PEM system. Indeed,implementation of an accrual-based system for gov-ernment accounting, given the costs involved, is per-haps only worthwhile in the context of an overalltransformation of the management of the governmentsector. Accordingly, it is essential to appreciate the fullramifications of implementing accrual accounting.

Second, budget systems are not created overnight.They evolve by passing through distinct stages—fromcompliance to performance-oriented systems—andthis transformation is accompanied by an accommo-dating evolution in accounting requirements. One candetect this parallel development of accounting re-quirements in the move from cash-based accountingto modified cash/modified accruals, and then to fullaccruals. Therefore, it is instructive to examine thetypical path by which accounting systems evolve, sothat countries contemplating the move can learnfrom the experiences of other countries and better plantheir transition process.

Third, as budget systems have developed, they haveadopted more comprehensive budget managementobjectives which have put increased demands on ac-counting systems. This development should be seenas one of accretion—adding to accounting require-ments rather than substituting one accounting systemfor another. There has been a temptation to debate therelative virtues of cash versus accrual accounting, yetthe two systems are not substitutes. Because accrualaccounting allows the generation of cash accounts, thedebate should focus on why there is a need for the in-creased information derived from accrual accounting.Particularly because the introduction of accrual ac-counting is hardly a costless exercise, it is importantto recognize that there may be less onerous interme-diate solutions or alternative ways of gaining some ofthe benefits afforded by accrual accounting.

Fourth, it must not be forgotten that accounting sys-tems perform two central functions. The first is the needto account for the financial position of the governmentat a particular point in time, determined by legal re-quirements and directed to ensuring accountability forpublic funds. The second is the need for a day-to-dayrecording of transactions as a basis for the informationsystems that support management decision making.Fulfilling the requirements of budget managers usuallynecessitates more information than preparing the legalfinancial statements, and on an ongoing basis. Regard-less of whether the accounting system is cash or accru-als based, there is usually a need for budget managersto record each stage of the spending process and tobe aware of other “off-balance-sheet” items such as

120See IMF (2001a, p. 5, para. 1.32).121Examples include depreciating new buildings 100 percent in

the first year, to reflect the entire “capital cost” as an expenditure inthe same fiscal year (as in Canada) or valuing heritage assets by atoken value. Other countries eliminate the difference between “ex-pense” and “expenditure” by assuming immediate consumption ofstocks on purchase and other means.

©International Monetary Fund. Not for Redistribution

Page 83: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VI MOVING FROM CASH TO ACCRUAL ACCOUNTING

contingent liabilities. From this viewpoint, accrual ac-counting systems are more comprehensive.

Finally, improved fiscal reporting requirements,such as those implied by the new GFSM 2001 stan-dards, cannot be divorced from a parallel process ofreform to the accounting system that underlies thegeneration of the basic data. From a statistical re-porting standpoint, the accrual-based GFSM 2001reporting requirements can be viewed as distinct

74

from accounting requirements. However, ensuring thequality and comprehensiveness of the GFSM 2001 re-ports will ultimately necessitate a move to accrual ac-counting. In particular, if reports are to be generatedwithin the budget year, and if these reports are to beuseful for budget management as distinct from purelyfor statistical purposes, then full compliance withGFSM 2001 will require a supporting move to accrualaccounting.

©International Monetary Fund. Not for Redistribution

Page 84: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII The Institutional Framework for Budget System Reform

75

The reforms of OECD budget systems have givenrise to fundamental institutional changes in the way

governments operate. Underlying all of these changeshas been the perceived need to move away from highlycentralized budget systems and associated control mech-anisms and toward more decentralized arrangements,which have come in many guises. This section providesa simplified typology as a framework to review the ad-vantages and disadvantages of the most popular of thevarious institutional approaches. From this review, somelessons for emerging economies are distilled. First, themove to institutionalize decentralized decision makingshould take place in stages, and some arrangements thatare more risky than others should be avoided. Second, itis necessary to put critical enabling conditions in place andnot to neglect the wider control and accountability envi-ronment in which the new arrangements must operate.Third, all solutions involve considerable investment inhuman resources to develop skills not only in the newlycreated decentralized units but also within supervisingministries, and especially in the MoF.

The Impetus for Performance Budgeting

Recent budget management reforms emphasize theneed to create an institutional environment that will ensureperformance—one that lets managers manage and alsomakes them manage. This stress on performance hasbeen triggered by two mutually reinforcing trends withinthe OECD membership: a downsizing of governmentoperations, combined with attempts to make these oper-ations more efficient and effective. The first trend arosefrom a basic questioning of the rationale for governmentintervention. Failure of the market mechanism creates aprima facie case for government intervention, and thiswas pursued rigorously by most governments after theSecond World War. Experience has since shown thatthere is no guarantee that government can do any better.The new public sector economics grew out of an increasedawareness both of government failure (partly throughinformation asymmetries) and of the incentives for bu-reaucracies to pursue their own interests (which may notentirely conform to the public good), which led to a ques-

tioning of the technical and allocative inefficiency ofgovernment operations. The result was a move to down-size government operations by outsourcing or transferringthem to the private sector.

The second trend arose from growing budgetary pres-sures, which spurred reforms aimed at making govern-ment interventions more effective—or more performanceoriented. These initiatives sought to tackle the usualcauses of government failure by “re-inventing” govern-ment through innovative institutional arrangements tobetter align incentives within government and thereby tobetter meet declared government objectives. The waythis has been implemented has been influenced by de-velopments in institutional economics, and in particular,principal/agent principles. In contrast to more traditionalapproaches to government service delivery, a key charac-teristic of these innovations is that the focus is on seekingto define objectives and to specify expected performancein terms of outputs (or in some cases outcomes) ratherthan inputs.

A result of these two mutually reinforcing trends hasbeen a shift in the predominant budget management modelused in the majority of OECD countries—away from tra-ditional, hierarchical, centralized decision making andtoward an accountability framework emphasizing decen-tralized decision making. This fundamental shift in theway that government operates can not be accomplishedwithout complementary institutional changes.

Different Approaches to Less-Centralized Budget Management

The case for greater decentralization in budget man-agement is summed up in Box 37. An important objectiveof decentralization is to introduce managerial flexibilityin order to give budget institutions the necessary freedomto devise the best ways of organizing their resources andto thereby generate productivity gains as a way of copingwith budget pressures. Another objective is to providemore effective service by being more responsive to clientneeds. One channel to accomplish this is to improve thedivision of labor in government by separating policy andimplementation functions. This move does not come with-out a price. Decentralized decision making raises issues

©International Monetary Fund. Not for Redistribution

Page 85: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

about how to ensure transparency and accountability,especially given the difficulties of defining, measuring,and monitoring performance.122 Several concerns are alsoraised about the capacity of central agencies to adapt totheir new role and the considerable transaction costs involved in establishing and maintaining the new decen-tralized control mechanisms. The way various govern-ments have addressed these issues has determined theinstitutional arrangements employed to operationalizethe move to decentralization.

To emphasize their main features, these new institu-tional arrangements can be broadly grouped according tohow they represent a progressive move away from cen-tralized government decision making.123 The latter can becharacterized as organized through the classic, generic,vertically integrated ministry or department of the ex-ecutive, which is under direct hierarchical control of a

76

minister or president. The degree to which a given de-centralized solution moves away from this model revolvesaround a number of key characteristics: the power of theminister over the ministry’s or department’s day-to-dayoperations; the scope of discretion of the chief executive;the determination of the organization’s budget; the staffingarrangements of the organization; and its legal status. Forexpository convenience, the institutional arrangementscan be ordered by degree of decentralization, here char-acterized as the five “Ds”—five progressively moredecentralized management models—deconcentration,decentralization, delegation, devolution, and divestment.Of course, within the OECD there are so many institu-tional hybrids that any all-encompassing categorizationis nearly impossible. This typology therefore only repre-sents the stylized characteristics of each approach; in re-ality, many overlapping arrangements exist.124

Step One: Deconcentration

As a first step away from centralized decision making,deconcentration represents no reallocation of adminis-trative authority from the center to the operating agencies.Centralized decision making is maintained, but local man-agers are given some flexibility in administering centrallydetermined policies that are difficult to implement cen-trally. Unit managers, however, are not managers in thetrue sense but are instead administrators of centrally de-termined decisions. In terms of the budget, the unit is stillintegrated into the ministry’s budget, so that the ministrydetermines the unit’s budget and has the right to reallocateresources among units. Under most traditional budgetsystems, there is a need to deconcentrate operations foreffective service delivery.

Step Two: Decentralization

A further step away from centralized decision making,characterized here as decentralization, represents a fullreallocation of implementation from the central level tothe line agencies, again with no major reallocation ofdecision-making authority. In this arrangement, agencymanagers remain administrators, but they are given greaterfreedom to administer central policies to fit day-to-daycircumstances. In terms of the budget, the central ministryretains hegemony and maintains the right to reallocateresources among units, and the agency’s budget is subjectto the usual annual budget review process. The agencyhas no separate status or legal identity and is predomi-nantly funded from taxes, but the head of the agency isgiven responsibility for day-to-day management. To reapmore of the potential benefits from this arrangement, cen-tralized controls on managers’ use of inputs are often re-laxed. This can take many forms: carrying over funds from

122For a discussion of the significant measurement problems facedby performance-oriented budget management, see Section IV.

123It is more convenient to group institutional arrangements alongthe spectrum of the movement away from centralized decision mak-ing because, internationally, it is not easy to distinguish central orcore government institutions and there is no generally acceptedclassification of the vast variety of public institutions outside thecentral or core level.

Box 37. The Pros and Cons of Devolved Budget Management

Pros

• Reduces transaction costs associated with layers of hierarchical controls.

• Increases efficiency and innovation.• Makes management of services more responsive to

customers.• Allows more effective partnerships between different

levels of government.• Enables central ministries to concentrate on policy.• Provides greater transparency in all aspects, opera-

tions, appointments of boards and chief executives;parliamentary reporting.

• Promotes clarity between responsibility and authority.

Cons

• Increases transaction costs in designing and imple-menting contract management.

• Raises greater accountability issues, due to lack ofstrategic management by activity, clear definition ofoutcomes, and transfer of greater authority to officials.

• Increases the risk of patronage and corruption (moralhazard).

• Limits the capacity of central ministries to specify per-formance, monitor, and analyze data.

• Carries risk that “civil service ethic” will be lost andthat unified career structure will be undermined.

124This represents a typology of vertical decentralization, whichhas parallels and can be applied on a horizontal basis, in the rela-tionships between different levels of government.

©International Monetary Fund. Not for Redistribution

Page 86: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Five Ds in Practice

77

one year to next (within predetermined limits); convertinga portion of funds for running costs into capital and viceversa; retaining within the unit some or all of the savingsrealized; building a capacity to generate and retain income;and exercising some flexibility in staff management.

Step Three: Delegation

This arrangement represents a significant step beyonddecentralization. Agencies, although legally still a partof the ministry or central government, are given greaterautonomy and independence in decision making. Centralsupervision is at arm’s length, with significant manage-ment responsibility resting with an intervening supervisoryboard or single person authority (although the ministrygenerally retains ultimate power over the board or author-ity). A quasi-contractual form of budgeting is put in place,with targets set by the reporting ministry in consultationwith the agency head. This type of institutional arrange-ment perhaps epitomizes the new performance budgetingapproach. The reporting ministry has no power to movefunds between agencies without prior approval of eitherthe MoF if amounts are small, or the legislature if amountsare large. For the most part, the agency can be consideredto have its own budget. However, this budget is part of ageneral budget law, and so the agency reports to the min-ister and is held at least nominally accountable by himfor the use of budget funds. Typically, the chief execu-tive will be nominated by the minister and will be a civilservant. The chief executive, once appointed, is givencomplete freedom in decision making but, in return, hasfixed performance targets. In meeting these targets, thechief executive usually is given flexibility in fixing gradesand pay, recruitment, and promotion. Staffing rules varybetween a full civil service system and more differentiatedcontrols. In many OECD countries, such agencies havebeen in the vanguard of performance budgeting reforms,being the first asked to move to output/outcome budgetingand accrual accounting.

Step Four: Devolution

This represents a more advanced move toward reallo-cating decision making from the central level to the im-plementing unit: legally separating the agency by giving itits own legal personality (partial or full). Such devolvedagencies typically have greater freedom in policymaking,often with clear restrictions on the ability of central minis-tries to intervene in decision making. In many countries,these devolved agencies have boards to set strategy andto develop policies for implementing the strategy—sometimes it is only an advisory board, but more usuallyit is a management board, and sometimes it is a governingboard. The minister usually participates in these boards,and in some cases directs them. Most typically, however,the minister has only indirect control. These devolvedagencies are often subject to different management andfinancial rules from traditional, vertically integrated min-

istries, and their staffs are often outside the civil service.In terms of the budget, these agencies are given direct fi-nancial support from the ministry in the form of a transfer,and/or are financed through fees and charges or by theircommercial operations. In many cases, they have thepower to borrow, usually under the scrutiny of the report-ing ministry or MoF. Typically, when they rely signifi-cantly on budget support, this is in the form of a contractualrelationship with the reporting ministry. Although prac-tice varies among countries, the minister’s role would beto inform the agency of the government policy, maybe tooffer advice, and certainly to monitor performance.

Step Five: Divestment

Divestment can be viewed as representing perhaps themost extreme approach to decentralizing decision-makingauthority—but in this case, it would mean decentralizingoperations outside the government sector. Such commer-cialization can take different forms, and a useful distinc-tion can be drawn with respect to the degree to whichassets change ownership between the government andprivate sectors. For example, when the government con-tracts out for services, few physical assets change hands.This can be contrasted with the complete transfer of own-ership of physical assets from the public to the privatesector that occurs with privatization. In the case of con-tracting, with few assets exchanged, the client withingovernment is able to maintain some control over activ-ities, monitor performance, impose financial penalties,and replace the contractor if dissatisfied. At the same time,there is a recognition that the contractor is carrying outwhat are still primarily public functions. In the case of pri-vatization, with extensive asset transfer, control is givenover to the markets, perhaps within some general regu-latory framework for that market, but with recognitionthat the new owner is not primarily carrying out a publicfunction. Bodies created for this purpose can be regardedas quasi-corporations which usually operate under privatelaw with their staff employed under general labor laws.They are mostly financed by their own sales; they have thepower to borrow and lend; and their budgets are separatefrom those of the ministries. Concessions and other public-private partnerships fall into this divestment category.

The Five Ds in Practice

The first two Ds—deconcentration and decentraliza-tion—reflect institutional arrangements associated withtraditional, vertically functioning ministries. The secondtwo Ds—delegation and devolution—are offshoots of thenew performance budgeting approach which has led tothe creation of a wide range of public bodies that are partof the government but have been given varying degreesof managerial autonomy, or independence. Moving fromdelegation to devolution can be viewed as simply a pro-gression in managerial or operational autonomy.

©International Monetary Fund. Not for Redistribution

Page 87: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

Delegated and devolved organizations have a longerhistory in some OECD countries, but the latter three de-centralization trends—delegation, devolution, and divestment—noticeably accelerated in the 1980s and1990s. The form and the name of the resulting organiza-tions vary widely among countries, and they represent animportant share of government spending.125 They gener-ally share the objective of providing greater managementflexibility, but the level of this flexibility varies greatlybetween countries. These bodies function outside the

78

usual vertical ministerial controls, typically under quasi-contractual or full contractual relationships with the min-istries, especially those associated with service delivery.126

Again, the degree of autonomy tends to vary betweencountries, as indicated in Box 38, as do the functions theycarry out. Generally, the departmental agencies and publiclaw administrations are engaged in the delivery of non-commercial services to citizens, support services to otherstate bodies, and some regulatory and quasi-judicial func-tions. The private law bodies, included here for complete-ness, are most often judged to operate in the commercialsector by SNA standards and often include governmentinstitutions on their way to privatization. In this study,the two former groups—departmental units managed atarm’s length from their supervising ministry, and publiclaw administrations—are referred to as agencies.

125Partial data available suggest their share in public expenditure andtheir share of civil servants employed can be over 50 percent, and insome countries above 75 percent. For example, the United Kingdom has131 executive agencies that employ over 75 percent of the civil service;in Spain, 51 percent of the budget is spent by government-related en-tities; the New Zealand Crown entities employ around 80 percent ofstate-sector employees; in Germany, 22 percent of public employeeswork in federal agencies; in the Netherlands in 2002, 30 percent of thecivil service worked in agencies, a percentage which continues to in-crease steadily. (OECD, 2003, p. 8)

Box 38. Different Types of Autonomous Public Bodies

Departmental Public Law Private Law Agencies Administrations Bodies

Legal Status

Governance

Staff

Funding

Examples

Source: OECD (2002b, pp. 13–19).

Part of ministries with no separateidentity; function under public law

Chief executive, directly appointedby minister, has operational control

Employed under general civil ser-vice rules

Allocations from state budget

Germany: Direct federal administrationNetherlands: AgenciesNew Zealand: Semi-autonomousbodiesSpain: Autonomous organismsUnited Kingdom: ExecutiveagenciesUnited States: Performance-based organizations

Partially or completely separatefrom ministries; function underpublic law

Usually have a governing oradvisory board, with ministerexerting indirect control

Generally subject to rules for civilservants, but staff rules vary evenincluding general employment rules

Usually financed by tax revenuesbut often allowed to carry oversurpluses

France: Administrative publicestablishments; professional publicestablishments; autonomousadministrative authoritiesGermany: Indirect publicadministrationsNetherlands: Public Law ZBOs(independent administrativeagencies);New Zealand: CrownEntitiesSpain: OAs (organismos auto-nomos), administrations, andselected regulatory bodiesUnited Kingdom: Executivenondepartmental public bodies

Not companies, but with fullseparate legal identity; functionunder private law

Governing Board, with ministerexerting indirect control

Usually under generalemployment law

Have budget separated fromministry, mostly financed bysales revenue; can carry forwardsurpluses, borrow, and lend

France: Industrial andcommercial publicestablishmentsGermany: Private lawadministrations and chargedadministrationsNetherlands: Private Law ZBOsUnited Kingdom: Somenondepartmental public bodies

126This excludes state-owned enterprises, bodies resulting fromadministrative decentralization, and constitutional bodies (such ascourts, audit bodies, and the like).

©International Monetary Fund. Not for Redistribution

Page 88: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Five Ds in Practice

79

In many countries that have pursued these institutionalreforms, it subsequently became clear that managing ona decentralized basis, of whatever form, creates specificaccountability and control issues. As emerging economiesundertake parallel reforms, it is important that these issuesbe fully recognized. Many relate to two areas: how to es-tablish accountability at arm’s length, focusing on the roleof contracting within government; and the choice anddesign for decentralized institutions to implement thisapproach, focusing on the role of the “agency.” These twocomplementary aspects of the new institutional frame-work are critical for ensuring the success of wider perfor-mance budgeting reforms.

Lessons from the OECD Experience:Delegation by Contract?

Performance contracting has emerged as a key tool forincreasing the efficiency of service delivery for govern-ment activities where market failures weigh against mar-ket provision. It has proved to be a useful mechanism togenerate desired behaviors in a devolved managementstructure, when day-to-day management is in the hands ofautonomous rather than centrally controlled units. The aimis to move to a system that continuously monitors man-agement’s attainment of objectives and makes managersaccountable for results. Performance “contracts”127 can bea way to define responsibilities and expectations betweenparties to achieve mutually agreed results (OECD, 1999a,p. 7). By doing so, they can strengthen both central andline units’ effectiveness by clarifying their respective roles.The center no longer manages directly but sets objectives,assists autonomous units, evaluates results, and facilitatesin redefining objectives.

Contracts tend to have three main elements. First, theylink budget negotiations more closely to units’ objectivesand performance measures, and in doing so they assist indeveloping a framework through which performance isgiven greater emphasis. Second, they focus the work ofcentral units on strategic priorities and away from inter-vention in line units’ day-to-day operations, clarifying thedivision of labor between the different levels. Third, theydevelop appropriate measurement and reporting arrange-ments to ensure full accountability. Although the perfor-mance contracting approach is often characterized asseparating policy and delivery, typically in practice thisis not a strict separation. While separation is sometimesneeded for managerial autonomy of units, the center mustalso recognize it is these units that are most aware ofpractical issues and impacts, and hence may deserve asay in policy.

In form, contracts vary widely, reflecting the fact thatdifferent approaches to performance contracting areappropriate for different circumstances. The OECD hasidentified seven broad types of performance contract(Box 39), stressing that the specific design of the con-tracting arrangement will depend on a variety of factors,such as the nature of the transactions covered, specificnational features of the country’s legal and administrativesystems, and the broader governance arrangements withinwhich the contract would function.

From this wide variety of experience, some key issuescan be identified that affect the effectiveness of perfor-mance contracting:

• Transaction costs: The costs of performance contractscan be high, at least initially. These include transac-tion and compliance costs associated with negotiatingand monitoring contracts, assessing and managingrisk, and enforcing contract provisions. There is alsoa considerable investment in human resources. Per-formance contracting demands new skills and waysof operating that may impose considerable costs onan organization. Similarly, when there is a need toshare data, the required IT systems may impose othersignificant costs.

• Defining performance: Many of the fundamentalproblems identified in the literature revolve aroundthe difficulty of defining performance in a contract,involving issues discussed in Section IV. Shouldperformance be defined in terms of outputs or theoutcome of these outputs? Should performance bedefined as a static concept or evolve over time? Forexample, after an initial period in which priority isgiven to reducing costs and improving productivity,performance may need to be adapted to broader ob-jectives such as effectiveness. For many OECD coun-tries, defining objectives has proved a difficult task.A common problem has been to set too many objec-tives and to cover all sectors of a unit’s activity, butfail to specify priorities. The difficulty of measuringperformance often means that any dimension of per-formance that can be measured may automaticallybecome an objective.128 There is a marked and well-documented tendency to define procedures ratherthan simply stating objectives and instruments ofcontrol/evaluation.

• Measuring performance: The other side of the prob-lem is obtaining the information on which to assessperformance, however defined in the contract. Skep-tics of this approach have compared governmentcontracting with the stricter view of contracting pre-vailing in the private sector. They point to funda-

127Note these agreements tend to be “relational” rather thanstrictly legal, as in the private sector. They are enforceable and arecharacterized by less formal dispute-resolution mechanisms. Thisis discussed more fully in Petrie (2002, pp. 120ff).

128Poor specification of performance can also lead to “gaming” inthe setting of performance targets or misrepresentation of perfor-mance statistics (Laking, 2002, p. 8).

©International Monetary Fund. Not for Redistribution

Page 89: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Box 39. Seven Broad Types of Performance Contract

Framework agreements cover strategies and priorities fora department and agency between a minister and chief ex-ecutive. The agreement provides the chief executive withautonomy in managing the organization in exchange for acommitment to meet specific strategic goals. Examples in-clude “framework documents” for Next Step Agencies in theUnited Kingdom and “letters of allocation” in Norway.

Budget contracts and resource agreements cover agree-ment over budget levels between the central budget officeor MoF and the chief executive of a department or agency.They provide aggregate budget authority and flexibility formanaging resources in exchange for agreed performancetargets and a method for monitoring performance. An ex-ample is Danish Contract Agencies which originally hadmultiyear budget guarantees.

Organizational performance agreements between a min-ister and chief executive break down overall strategic goalsinto program elements, setting specific and often detailedoperational, process, and output targets in exchange for in-creased operational autonomy in achieving targets. Ex-amples include Danish Contract Agencies, French TaxAdministration, U.S. Performance-Based Organizations.

Chief executive performance agreements made betweenministers and chief executives (often to complement orga-

Source: OECD (1999a, pp. 10ff).

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

mental problems in operating a contract with provenmeasurability problems, given the ambiguity of defin-ing outputs in the public sector (not to mention thequality of these outputs) and of determining “prices”for such outputs. Their strongest criticisms are directedto the challenges of enforcing contracts internallywithin government (Robinson, 2000)—how effec-tive can sanctions be against an agency that fails tomeet the requirements of its contract? An additionalproblem is the availability of information, particularlyif virtually all the information on the effectiveness ofoutputs comes from the service provider. Certainly,there is agreement across countries that setting targetsand indicators is difficult and requires continuous re-assessment—and is usually the most problematicaspect of performance contracting (OECD, 1999a,p. 29). The process of setting targets and indicatorshinges on the degree of specificity in the contract.Determining this requires balancing the center’s needto ensure accountability for policy outputs and out-comes, which pushes toward specificity, against theprovider’s need for flexibility in delivering servicesefficiently and effectively at reduced cost, whichpushes toward generality. Ultimately, the contractmust settle this trade-off, ensuring transparency and

80

“reasonableness” in accountability but at the sametime allowing both parties access to information nec-essary to effectively conduct their business. It is im-portant also to ensure that the relationship betweencontracting parties remains cooperative rather thanadversarial and to build a dispute-resolution mecha-nism into the contract.

• Wider budgeting environment: Some basic level ofmanagement expertise is required to make the contractapproach effective. Previous studies have stressed thelinks between the level of public management sys-tems and successful implementation of performancecontracting (OECD, 1999a, p. 17). In successful coun-tries, the existence of accrual accounting, contractingout, and other management tools that allow programsand services to be unbundled have offered a startingpoint for performance contracting. A performancecontracting regime is not a substitute for overall per-formance management, but is merely one element ofa performance management framework for budgetresources.

The verdict on contracting in government is that, whileit takes a very bad contract to be less efficient than cen-tralized regulation, creating a good contracting regime isextremely difficult.129 Overall, despite the difficulties in

nizational performance agreements) or between senior man-agement and staff at various levels. Examples include suchagreements in Australia, New Zealand, Norway, and theUnited Kingdom.

Funder-provider agreements focus on clarifying respon-sibilities by separating the role of the funder and theprovider of the services. An example is that New Zealandministers and chief executives negotiate agreements for thepurchase and supply of specified outputs, detailing factorssuch as timing, volume, cost, and quality. Similar pur-chaser-provider agreements are found in Australia.

Intergovernmental performance contracts and partner-ship agreements are often linked to devolution of programsor funding from national to subnational level, providingstate and local governments with funding in exchange forproviding specified levels and quality of service. These arefound in Canada, France, Germany, Norway, Spain, Swe-den, and Switzerland.

Customer service agreements, statements of service stan-dards provided by a program or service to its clients thatspecify the quality and level of services to be expected andin some cases, avenues of redress and compensation whereservices fail to meet standards. These are used in Belgium,Denmark, France, Italy, United Kingdom, and UnitedStates.

©International Monetary Fund. Not for Redistribution

Page 90: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Five Ds in Practice

81

OECD countries, the experience has been generally pos-itive (Box 40). It has been the basis for the introductionof much of the current “commercialization” within OECDgovernments (Box 41). A not inconsiderable part of thisendorsement arises from the value added by the con-tracting process itself, including the definition of roles,specification of the elements of the contract, and the effortto design mechanisms to measure and monitor whetherthe participants are fulfilling the contract (both during thecontract period and at its end). In this way, contractingcan be viewed as a learning and strategic planning exercisethat is important for performance-oriented management.

Lessons from the OECD Experience: Devolved Agencies as a Panacea?

Since the 1980s there has been an explosion of interestin devolved agencies in many countries.130 The drivingforce for most OECD governments has been the need toreduce spending, on the one hand, while on the other hand,strengthening the effectiveness of the remaining spendingby improving the quality and responsiveness of public ser-vice delivery. The agency model promises the advantageof separating implementation tasks from policymakingand of making them more achievable by better definingresponsibilities. Performance targets can be specifiedmore clearly, leading in turn to higher efficiency, econ-omy in the use of resources, closer customer relations, andhence more responsiveness to customer needs—that is, ahigher quality of service. To these “traditional” argumentshave been added several new, perhaps less justifiable,ones.131 However, what is clear is that decentralized insti-

tutional arrangements are an integral part of OECD bud-get practices.

For countries contemplating the move to more devolvedinstitutional arrangements, there are certain key lessonsfrom OECD experience.

There Is a Need for “Institutional Clarity”

An important message in a major study by the OECD(2002b) on country experiences with devolved agenciesis the need to clarify their institutional status, and hencetheir accountability arrangements.132 Countries should beclear on the criteria for creating devolved bodies. Histor-ically, individual agencies have been established in re-sponse to specific needs, with their powers and attributesdesigned to respond to those needs as perceived at themoment of their creation. Only later, as part of the reformprocess, have OECD countries made an effort retrospec-tively to divide agencies into groups and to create commonmanagement rules for the agencies in each group (OECD,2001, p. 27). One criterion often used, following theNew Zealand model, is to split an organization into sep-arate purchaser and provider agencies. However, this mayinvolve costs that become apparent only with time. Forexample, one cost may be the loss of institutional capacityto learn from experience, and another may be a growingdisconnect between the policy and service levels.133 Agood deal of caution is also appropriate when determin-ing the legal status, mandate, and powers to be accordedto a particular agency because an identical design modelcan produce vastly different consequences when appliedto agencies with different types of tasks. One key factorappears to be scale: if the task requires that an agency bevery large, it is likely to succumb to the same bureau-cratic tendencies as any other organization of similar sizeand yet, on the other hand, it may prove powerful enough

129Matheson, in OECD (1997b, p. 176).130However, such devolved units have a long history, for exam-

ple, dating back to the sixteenth century in Sweden and the 1870sin Germany.

131One specialized reason is legitimizing policy and ensuring im-partiality through politically independent organizations, for example,central banks and regulatory and investigatory institutions. How-ever, as Laking (2002) points out, there are many other reasons thatare less justified: “Agencies may also be created to protect a func-tion for legislative interference, pay off political allies or createpower bases for specific functions or to capture public assets or re-sources for private interests” (p. 5).

132Gill (2002) points out how some governments have createdproblems for themselves by setting up many such organizations op-erating under different rules. He points to New Zealand, a pioneerin the field, which created a wide range of new agencies out of gov-ernment departments in the 1980s and 1990s, which in turn led thepresent Labor administration to introduce legislation to bring someorder and consistency into their governance arrangements.

133An early criticism of the performance contracting system inNew Zealand was the lack of incentives in purchase and perfor-mance agreements for chief executives to coordinate their policiesand programs, to operate across organizational boundaries, and tofocus on longer-term issues of corporate capacity beyond the termsof the contract (OECD, 1999, p. 20).

Box 40. Pros and Cons of Contracting

Pros

• Helps lower costs and improve the quality of service.• Encourages greater innovation and responsiveness.• Allows a sharper focus on “core” government func-

tions.• Enhances accountability by introducing more rigorous

performance monitoring and, usually, establishingmechanisms for redress when loss and damage occurs.

Cons

• Raises the costs of monitoring and supervision.• Is difficult to control for quality.

©International Monetary Fund. Not for Redistribution

Page 91: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

to resist the control efforts of the supervising ministry(OECD, 2001, p. 24).

A strong inference from the OECD experience is that theprocess of creating devolved agencies is more effectivewhen there is a clear policy on the degree of autonomyto be granted as a basis for determining their institutionalstatus and accountability arrangements. If this has notbeen done, there may be a need to classify existing insti-tutions to clarify their accountability arrangements.

There Is a Need to Create a Framework for theManagement of Devolved Agencies

If agencies are to increase efficiency in service delivery,the correct management framework must be put in place.There are dangers in granting independence to agencieswithout clarifying their responsibilities and their account-ability. For those governments beginning such reforms,there are inevitably problems in defining an agency’s de-gree of operational independence and the role of the min-istry or supervisory organization. This challenge has been

82

addressed differently by different countries so that thereis no ready model to apply.134 Even when there is strongdesire to make line units autonomous, the relationshipwith the center inevitably remains hierarchical becausethe center must protect the public interest and must retainthe power to amend the terms of any contract it has with theagency. The limits of this asymmetry in power have beendifficult to define, but they determine the design of themechanism by which the center will endeavor to controlat a distance.135 The extent of the center’s discretion overthe agency needs to be made explicit.

Ultimately, a central question to be answered is theagency’s degree of autonomy to propose its objectives.On the whole, the trend has been to increase this auton-omy. Some OECD countries have come to recognize thevalue not only of allowing agencies latitude in choosingtheir modes of operation (subcontracting, partnerships,and others), but even of allowing them to adjust prioritiesaccording to their specific situations, provided they canjustify the adjustment. Some go so far as to allow agen-cies to define their objectives and also translate them intotargets, simply because of their superior knowledge ofproblems on the ground.136 It is important, however, thatsuch flexibility is not granted by default through the poorcapacity of the center to provide policy guidance andoversee results.137 The center must ensure adequate mon-itoring and must clearly define accountability, so that thecontracting parties know who is accountable for what andin what form, and what the reporting requirements are.Some accountability features of current OECD practicesrelated to devolved units are summarized in Box 42.

There Is a Need to Strengthen Reporting andAccountability Mechanisms

In a devolved management setting, in return for theirautonomy, agencies should have to report back on theiractivities, results, and performance. In exchange for flex-ibility in the use of inputs, they must report more system-

Box 41. OECD Practices: Degree ofCommercialization in Budget Management

Percentageof OECDCountries

1. Systems in place to charge a price for goods/services by one government organization to another:a. Used a lot 20b. Used moderately 32c. Used rarely 16

2. Government organizations allowed to collect user charges:a. Generally 27b. Only for specific purposes 70

3. Government organizations allowed to keep user charges:a. Generally 24b. No general rule, determined

case-by-case 484. Capital charge imposed on government

organizations for using capital assets:a. Generally applied 17b. Specific to certain sectors 22

5. Managers of government organizations can keep savings from efficiency gains to make other expenditures:a. Without restrictions 15b. With restrictions 33

6. Use of performance agreements specifying expected outputs 42

Source: OECD (2003, Tables 4.3 and 5.4.a).

134The United Kingdom’s Next Steps model has been widely in-fluential. Even though each agency has operational independence,it should be clearly responsible to its minister or parent department.However, many contemporary agencies in other countries clearlyfall short of this. (Schick, 2002)

135The center must learn to move from “rowing to steering.” AsSchick (2002) points out, improperly applied, this prescription maylead to a government that does neither rowing nor steering well (p. 35).Delegating management freedom always carries the risk that theagency will not be fully compliant with more general public interests.

136Within some OECD countries, it is possible to detect a reevalua-tion of the risks entailed in this approach, and in some countries such asthe Netherlands, there has been an attempt to redefine agencies’ auton-omy to bring them under closer control of ministries.

137This risk increases with the number of agencies. How is the centerto ensure that agencies work together to meet important public objec-tives? or to ensure there is no overlap in responsibilities? or that im-portant areas are not missed? The United Kingdom’s recent initiativefor “joined-up government” attempts to address this problem.

©International Monetary Fund. Not for Redistribution

Page 92: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Five Ds in Practice

83

atically on their outputs and outcomes. This is the majorreorientation at the core of performance budgeting reforms.As noted, most countries have started to implement output-and some outcome-oriented reporting through activity-basedcosting, and some have put in place multiyear agreementsand monitoring mechanisms. Again, country experienceis varied, and there remain many open questions. Forexample, should performance be gauged by outputsonly, or by outcomes? Some OECD countries such as theUnited Kingdom and New Zealand stress the measure-ment of outputs given that the autonomous unit is notgenerally perceived as responsible for outcomes (whichis seen to be a matter for the ministry). Others such asAustralia measure outputs but emphasize the importanceof knowing outcomes. Because agencies are involved inproducing both outcomes and outputs, they have to de-sign impact indicators and measure results. Most OECDcountries concentrate on measuring outputs, or if theymeasure the impact of public policies, they do so sepa-rately from the measurement of outputs (OECD, 2003,p. 83). Whatever the particular country solution, what is

clear is that it takes several years of adjustment for largeagencies to move from input-focused to more output- oroutcome-focused management. For many OECD coun-tries, this is a continuing process.

There Is a Need for New Skills

The time required to adjust to the new devolved man-agement framework largely reflects the time required toacquire new skills. Of course, the type of work that au-tonomous units carry out, and the skills they need to apply,will differ greatly depending on the degree of autonomythey are accorded. The necessary managerial skills canusually only be developed if the unit has the means, eitherfrom its own resources or from outside, to gain a thoroughunderstanding of its customers, to develop a strategicvision of its priorities, and to undertake a detailed analy-sis of its environment. Small units rarely have such skills.For this reason, perhaps not surprisingly, a devolution ofbudget management tends to go hand in hand with a de-volution of human resource management, or even wagenegotiations, to allow agencies to acquire the proper skillmix. However, this again brings managerial staff intoareas for which they may not be trained.138

Similarly, the center may also have to gear up to meetthe challenges of the new environment. Some have focusedon the changing role of the budget office, but the specificcontribution of other central units must also be redefined.In some cases, the central units may not have undertakenadequate policy analysis or development, and have insteadlimited their role to being regulators.139 Central units areobliged to define their objectives more clearly and to givethemselves the means to achieve these objectives. Thisinvolves creating a forecasting capability, improving theirunderstanding of technological issues, investigating cus-tomers’ overall expectations, and setting up units to carrythrough and monitor reforms. These units must provideadvice to line units; collate and evaluate approaches; drawup methodologies, guidelines, and rules for action; andmonitor autonomous units’ performance.

Devolution also poses new challenges for parliamen-tary control. Parliamentary committees often find it dif-ficult to keep track of the different bodies created and thefinancial and management rules that apply to them, andthey need to develop the capacity to analyze the new databeing reported.

Box 42. OECD Practices: PerformanceManagement in Government Agencies

Percentageof OECD Countries

1. Use of performance management in agencies:a. Arm’s length agencies 59b. Public law agencies 68c. Mixed agencies 67

2. Combination of output and outcome targets included in budget process:a. Arm’s length agencies 54b. Public law agencies 64c. Mixed agencies 70

3. Use of yearly contracts on programs and activities:a. Arm’s length agencies 75b. Public law agencies 58c. Mixed agencies 50

4. Regular performance reports submitted to the parent ministry:a. Arm’s length agencies 33b. Public law agencies 41c. Mixed agencies 30

5. Financial targets written in a formal contract arrangement with the parent ministry:a. Arm’s length agencies 41b. Public law agencies 41c. Mixed agencies 67

Source: OECD (2003, Table 7.2.c).

138The quality of human resources may be a key factor in the successor failure of reform (OECD, 2003, p. 76). Employment contracts maybe required to encourage managerial talent. In the United Kingdom,agency directors are appointed on fixed-term renewable (three- or five-year) contracts—that is, for the duration of the framework contractgoverning the agency. Incentive schemes may also be required for staffto accept and integrate change.

139This occurred, for example, in France. See OECD (2001, pp. 66ff).

©International Monetary Fund. Not for Redistribution

Page 93: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

There Is a Need to Nest Devolution in a WiderReform Framework

In a number of countries, particularly at the politicallevel, there has been a notable reluctance to embracedevolution. In some countries, this is because compli-ance issues are judged to be more important than perfor-mance issues. In other countries, provision of servicesby the public sector is regarded as superior to provisionby the private sector. In other countries, a lack of stabilityin the resource base and the need for in-year adjustmentsto budgetary funding create managerial problems.140 Instill others, such changes are adversely perceived as giv-ing too much power to unelected officials at the expenseof elected representatives. They may also be perceivedas opening up too many possibilities for improper or in-appropriate behavior, especially in countries with notedgovernance problems. For example, agencies must notmake borrowing decisions that create unsustainable con-tingent liabilities for the government, increase public debtbeyond the limits envisioned in the government’s fiscalplan, or exceed the borrowing authority approved by par-liament (OECD, 2001, p. 32). This caution to nest devo-lution in a broader framework of reform is even morerelevant to countries that have created agencies only toavoid fundamental political or institutional reforms suchas civil service reform, or to avoid hard policy choices suchas downsizing or deregulation. Agencies should not beviewed as a “magic bullet.” The full benefits of devolvedbudget management can only be reaped in a nurturing andsupportive reform environment.141

In developing and transitional economies, devolvedagencies have been proposed as a way of countering poorgovernance, insulating some important functions (such asregulatory activities), or making aid more effective. How-ever, experience with this approach is unconvincing. Forexample, independent revenue authorities (IRAs), whichhave often been created on such grounds, may experiencesome initial success but, because they cannot be com-pletely protected from the wider governance environment,generally become subject to political and bureaucraticinfluences. In Peru, the state revenue authority (SUNAT)

84

was protected initially, but eventually was undermined.Similar fates befell IRAs in Ghana and Uganda (Laking,2002, p. 6). Even when governance is not the main issue,the experience of many developing countries highlightsthe importance of establishing supporting preconditionsto enable the agency concept to succeed. The difficultiesexperienced by Tanzania, one of the most reform-mindedAfrican countries, illustrates some of the problems en-countered by agencies in less than fully supporting envi-ronments.142 As part of its public sector reform programlaunched in 1991, Tanzania passed legislation in 1997establishing “executive agencies,” modeled on the UnitedKingdom’s Next Steps agencies. Although the executiveagency program is still in its infancy, major challengeshave emerged in its implementation (Box 43).

Lessons from the OECD Experience with Divestment

There are many options for divestment, short of outrightprivatization. All seek to commercialize public services,making service provision more responsive to customersby giving them a purchaser role (Box 44). The exact ap-proach should be determined according to the market en-vironment. When it is decided that services can beprovided in a competitive market, schemes relying onvouchers or similar instruments can be employed to giveconsumers a direct purchaser role in selecting their ser-vice provider. When it is inappropriate or impracticableto give consumers a direct purchaser role, a popular ap-proach is competitive tendering and contracting that allowsthe in-house public sector service provider to competewith outside contractors by submitting a tender. Somegovernments have also experimented with user chargesfor internal administrative services. Contracting out, oroutsourcing, signifies competitive tendering only amongoutside contractors. This is most successful where theservice is easily defined, there is little specific capital re-quired for its delivery, and there is an existing, compet-itive private market.143 When market conditions requirethat the service be provided through monopolistic supplyarrangements, potential market failure dictates the need togrant a monopoly in a particular area. In such cases, con-cessions and public-private partnerships are increasinglypopular. What follows is a discussion of lessons from theOECD experience with various divestment options. Be-cause some are more relevant from a public expendituremanagement perspective, this section concentrates oncompetitive tendering and concessions.

140The need for safety valves is recognized even in OECD countries.For example, the use of medium-term budgetary strategies is importantin avoiding cuts. Some countries have a budgetary reserve to meet un-expected problems, others respect contractual commitments (Aus-tralia) or guarantee resources for a specific period (Denmark, Ireland),so that cuts can only apply to areas not covered by agency contracts.At the same time, moving too far to protect agencies from cuts imposedon others could be regarded as working against the flexibility they aretrying to promote.

141As so eloquently put by Schick (1996), “It is naive to assume thatagents who opportunistically pursue self-interest when they are formallycontrolled by principals will be more compliant when they are organi-zationally independent. It is equally naive to assume that departmentswhich gave inadequate attention to service performance when operationswere run by their own administrative subdivisions will be more attentivewhen services are hived off to independent entities.” (p. 31)

142As Scott (2001a) observes: “Public management reform cre-ates winners and losers and the losers are those that benefit fromthe old system and they are usually in very powerful positions”(p.12). The conclusion is that technical reform cannot substitute forpolitical reform.

143Examples are computers, transport, cleaning services, buildingmaintenance, minor capital works, printing stationary, and officeaccommodation, among others (Laking, 1996, p. 14).

©International Monetary Fund. Not for Redistribution

Page 94: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Five Ds in Practice

85

There Is a Need to Avoid Three Common Pitfallsthat Can Limit the Cost Savings of ContractingOut and Outsourcing

Contracting out and outsourcing—and, more generally,the commercialization of public service provision—in-volve opening up to competition a set of activities thatwere previously immune from it. Box 44 outlines variousstrategies for commercializing public services. Contract-ing out is a widely used mechanism for reform of publicsector service provision, which involves introducing com-petition for the market through competitive tendering.144

Other things equal, the lowest price tendered wins theright to supply for the duration of the contract term. In

this way, the government is able to secure the provisionof services at the lowest possible cost. The government isexpected to retain a fair measure of control over the activ-ities concerned, monitoring performance, imposing finan-cial penalties, and replacing the contractor in cases ofoutright performance failure.

Empirical evidence suggests that contracting can havesignificant impact in reducing the cost of providing awhole range of publicly funded services. Hodge (2000),when reviewing the international experience of contract-ing out of public services, estimated cost savings aver-aging 20 percent for a whole range of services, mainlyderived from the cost savings gained from exploiting flex-ibility in the workforce. A key finding is that cost reduc-tions are attained whether the contracts are won by publicsector organizations (contracting in) or private sector or-ganizations (contracting out) (Domberger and Rimmer,1994). This indicates that competition is the driving forcebehind savings rather than the means of delivery. How-

Box 43. Tanzania’s Experience in Creating Executive Agencies

In 1997, Tanzania passed the Executive Agencies Act.There are four notable features of the program:

Decentralized management

Agencies were deliberately constituted to be distancedfrom often politicized central departments. While the min-ister is theoretically responsible for the agency and techni-cally establishes its advisory board, it is the chief secretary(the senior civil servant) who approves the agency’s man-date, and it is the ministry permanent secretary (PS) whonegotiates the performance agreement with the chief ex-ecutive (CE). CEs report to their PSs, not to the ministers;in turn, the PSs report to the chief secretary, who interactswith the president. In this way, the decentralized agencymanagement system has been introduced in a highly cen-tralized power structure with weak ministerial responsibil-ity and ultimately weak governance of executive agencies.

Performance contracting

The CE is recruited into a principal/agent relationship, isappointed through open competition, and operates under aperformance agreement. Performance is specified in a frame-work document, setting out objectives, roles, authority, andperformance standards. However, performance measure-ment has been left largely to the agencies, and PSs often ap-pear unsure as to their role as “strategic managers,” leavingeverything to the agencies. In turn, the CEs feel let down bytheir parent ministries. There is little performance monitor-ing in place in the ministries, and this is left to the advisoryboards, which receive half-yearly reports. The adoption of

accrual accounting systems by the agencies has compoundedreporting problems to the ministries because the latter lackcapacity to interpret accrual accounts (most ministries usecash-based systems). This is aggravated by a weak externalaudit environment.

Output focus

The Executive Agencies Act defined agencies’ mandateas providing the best service to their customers and being re-sponsive to their needs. Pilot agencies were identified withpotential to operate on a cost-recovery basis. Agencies havehad much less success in operating on a commercial basiswhen their customers are other government departments.Government departments had no culture of paying for ser-vices, nor were they able to pay obligations because of theirown uncertain budgetary positions due to the uncertainty ofthe release of their budgeted funds. This uncertainty and un-predictability also affected the operations of those agenciesdependent on government subventions through their parentministries. The expectation that internal markets will work,especially for units providing a public good and dependenton public sector clients, proved unrealistic.

Human resource capacity

Performance management has foundered on staffing is-sues. An objective of public sector reform is to downsize,and this left agencies with substantially reduced personnel.At the same time, there has been the need to train staff innew ways of working. This has resulted in a heavy burdenon middle managers, who have little time for strategic andbusiness planning. Most agencies have yet to develop per-formance measures or benchmarks for evaluating outputs.

144The distinctive feature of contracting out is the element of exante competition—competition for the market as opposed to com-petition in it. The market in this case is defined by the contract spec-ification, and the bidding procedure resembles an auction.

Source: Caulfield (2003).

©International Monetary Fund. Not for Redistribution

Page 95: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

ever, there are some drawbacks to commercialization, andthe potential savings should be adjusted to accommodatethese problems:

• The problem of transaction costs: In specific circum-stances, the costs of contracting in the marketplacecan be high enough to offset the benefits. Even forOECD countries with adequate human resources andexpertise, the contracting process (developing spec-ifications, preparing contracts, organizing the contractprocess, etc.) represents a sizable overhead. Moreover,the inherent incompleteness of contracts (see the nextitem) complicates and adds to the costs of contractmonitoring. When this occurs, in-house or integratedproduction may be a better option.

• The problem of making comprehensive contracts:Contracts are inherently incomplete. It is impossibleto write a contract that incorporates every contingency,or every dimension of the service. In particular, be-cause certain aspects of quality are noncontractible,it is difficult to establish that a private contractor isfailing to provide the level of service specified in thecontract. As a consequence, the contractor’s incentiveto reduce costs tends to override the incentive to main-tain or improve service quality—referred to as the“quality-shading hypothesis.”

• The problem of ensuring that the benefits to societyare adequate and sustainable: It is important that thegains from contracting are passed on to the commu-nity and are not absorbed by the managing organi-zation. Experience shows the very real danger thatbenefits will accrue internally, instead of externally.Moreover, in countries with poor general governance,the move to contracting out poses a very real risk of

86

promoting corruption.145 It is also important that costreductions are sustainable over time and do not dis-sipate. In this regard, the contracting party must beaware of all considerations embodied in the tenderer’slow price. For example, private vendors may have alarge enough customer base to enable them to lowerthe price to gain the contract, and then to make up thedifference by attracting additional customers, espe-cially if they have many lines of activity. Moreover,private firms generally can take a longer-term view,and so they may be prepared initially to cut the priceas a loss leader, making up initial losses through sub-sequent business (Davenport, 1996).

Box 45 offers some general guidelines for contractingout and outsourcing.

There Is a Need to Carefully Design Projects andto Put in Place the Correct Control Mechanismsin Order to Realize the Full Potential ofConcessions and Public-Private Partnerships

Concessions and public-private partnerships (PPPs) takemany forms: leases, afterimages (a type of lease widelyused in France), B-O-T contracts, divestitures with revo-cable licenses to operate, and more. They are generallyused for publicly provided services with natural monopolycharacteristics—that is, when least-cost production re-quires that there be only a single service provider at any

Box 44. Commercialization of Government Services

Governments have adopted different strategies to intro-duce competition as a means to reduce operating costs andimprove service delivery.

Retention with internal reengineering

The government continues to provide a service, but mod-ifies its approach along market lines to improve delivery, toreduce operating costs, or to do both.

Example: A department reorganized; the in-house bidderwins a competitive tender; users are charged for internalservices.

Commercialization through outsourcing

The government outsources a service to another provider,retaining responsibility for that service.

Example: A private sector vendor wins a competitivetender.

Commercialization through partnering

The government enters a partnership with anotherprovider to provide a service through shared delivery and/orresponsibility.

Example: A build-operate-transfer (B-O-T) agreementis reached with a private consortium to develop infra-structure.

Commercialization with partial divestment

The government transfers a service to another provider,retaining ultimate responsibility for that service.

Example: A concession is granted to a private company.

Commercialization with full divestment

The government transfers a service to another provider,divesting itself of any responsibility for that service.

145Corruption, of course, is not confined to only those countries.A clear and direct connection between political party campaigncontributions and subsequent contracting out decisions has oftencome to light in the United States.

©International Monetary Fund. Not for Redistribution

Page 96: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Five Ds in Practice

87

one time.146 Public infrastructure projects and servicesmost often fall into this category: water, power, gas, rail-ways, and roads. Governments often grant exclusive rightsto the concessionaire, but in many cases this may not bedesirable—permitting entry by new competitors helpsensure that direct competition will take place whereverpossible and can pressure the incumbent to maintain goodperformance.

Certainly in emerging economies, and also in OECDcountries, schemes for private participation in infrastruc-ture have expanded dramatically in the last decade. SuchPPPs have at times succeeded in transferring significantcommercial risk to the private sector, mobilizing privatefinancing for public services, and improving the efficiencyand availability of such services. At the same time, mostgovernments have taken care to ensure a high level ofaccountability for the use of public funds. However, inthe developing world, the results have generally been dis-appointing. In many developing countries, PPPs have beenadopted to enable an expansion in government serviceswithout showing the full cost on the budget. In other coun-tries, incentives for efficiency and innovation have been

weak, accountability for performance has been absent,and opportunities for leveraging scarce public resourcesthrough private financing have been limited (Smith, 2001,p. 91). This experience stresses that the full potential ofPPPs can only be realized when projects are well designedand when the correct control mechanisms are put in place.Unfortunately, there are no standard models or blueprints,and approaches need to be adapted to the characteristicsof the service and the environment in which they will bedelivered. However, some of the requirements for suc-cessful concession contracting have emerged:

• Outputs must be clearly defined: Performance is de-fined as rigorously as possible, including specifyingthe recipient target group and defining an identifiablelevel of quality. Correctly identifying indicators forthese outputs is critical; misspecified or incompleteindicators can lead to counterproductive or biasedbehavior by the service providers. In order to pre-serve the flexibility of the concessionaire’s operationalarrangements, performance should focus on the endresults to be achieved rather than on the means to beused. Performance targets also can be designed toallow for renegotiations under specific preestablishedprocedures.

• Design issues must be fully addressed: The allocationof risks between the involved parties is at the core ofthe concession design.147 A careful analysis is nec-

146The capacity of ex ante competition to yield efficient outcomes de-spite structural conditions of natural monopoly was persuasively arguedby Demsetz (1968). In a similar way, Baumol (1982) and Baumol,Panzar, and Willig (1982) introduced the concept of a “contestable”market which, while not competitive in the sense of having several sup-pliers, can nevertheless generate competitive outcomes in terms of priceand output. Contestability occurs when the sole supplier does not have apermanent hold on the market and could be displaced by a more efficientproducer, charging lower prices.

Box 45. Guidelines for Contracting In and Outsourcing Government Services

These are key factors for realizing the potential benefitsof contracting in and out:

• Contracting in and outsourcing should be integrated withthe overall corporate strategy of the organization. Itshould not be a mechanistic exercise but an opportunity toreevaluate and reengineer internal work practices.

• It is of primary importance to involve affected staff andminimize any period of uncertainty.

• Competitive supplier markets are required to achieve thefull benefits of contracting out. These should be fosteredby contracting out practices, where government can playa role in developing markets for relevant services.

• Effective contract management requires a new set of skills,and so recruitment and staff training need to be geared toproviding these. To deal adequately with the contractor, or-ganizations need to maintain their knowledge of the mar-ket and their technical knowledge of the activity.

• All alternatives, including continued in-house provision,must be comprehensively evaluated. All costs and outcomes/outputs should be considered, including com-parative quality and any risks assessed.

• In-house bids should be treated the same as outside bids,with care taken to incorporate all cost items faced by privatecontractors as well as potential improvements in the workprocess. If successful, the work should be awarded to the in-house bidder on the basis of a formal document obligingstaff to meet the terms of the bid.

• Service requirements should be specified in terms of out-comes or outputs, not inputs, and should be specified asfully as possible to include appropriate service qualitymeasures.

• Contracting out does not diminish the responsibility of theorganization for performance, and so there should be pro-cedures established to regularly and formally monitor theperformance of the contractor. This process should reflectthe mutual interest in developing a cooperative rather thanadversarial relationship.

147While theoretical principles are well known—risks should beborne by the party best able to control, manage, or hedge against them—their application in practice often raises numerous difficulties.

Source: OECD (1997b).

©International Monetary Fund. Not for Redistribution

Page 97: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VII THE INSTITUTIONAL FRAMEWORK FOR BUDGET SYSTEM REFORM

essary to distinguish between costs that are truly ex-ogenous to the operator and those that are not. Onlyexogenous costs should be passed on to other partiessuch as consumers, suppliers, or the conceding au-thority. The choice of how long the monopoly will lastis part of the process of ensuring that the provider hasan adequate incentive for efficiency. It is generallyagreed that competitive bidding for time-bound con-cessions can provide a useful discipline over suppliers.

• The form, level, and structure of payment must be de-cided: This is crucial for determining the incentivesfor suppliers and the possibility of mobilizing privatefinancing. Here there are a number of difficult issuesto resolve. How tightly should payment be linked toperformance? Linking pay to performance indicatorsprovides stronger incentives, but since quality is moredifficult to measure, this approach increases the riskthat suppliers will engage in undesirable behavior(for example, cutting costs by reducing quality). Howto deal with trade-offs between quality and quantity?This often hinges on the degree of choice offered toconsumers of the service based on their preferencesand ability to pay. Should cross-subsidization be al-lowed? It is generally agreed that cross-subsidies areto be avoided—being distortionary, anticompetitive,and nontransparent. Other alternatives may be prefer-able, such as direct financing subsidies from the bud-get, or subsidies through special funds. There are alsoopen questions with no definitive answers. How heavyshould be the reliance on user fees when there aremerit good implications? (Smith, 2001, p. 3) Shouldthe service provider receive some up-front payment,or should the provider be paid only after a satisfac-tory delivery of services has been verified? All suchissues must be resolved on a case-by-case basis.

• Mechanisms to regulate the concession must be estab-lished: A central issue is how to ensure that serviceproviders are in a position to deliver in response toincentives, at arm’s length from regulators and thefunding source. Some body must be charged with thistask. Should a single entity be in charge, or shouldsome functions be delegated or contracted out? Estab-lishing specialized, cross-sectoral regulatory bodiesthat are independent of the government has oftenproved advantageous. “Competitive aware” mecha-nisms are also effective. However, when it is moreappropriate to negotiate procedures, then there shouldbe some built-in safeguards like benchmarks or a pro-cess for other providers to better the proposed terms.Not surprisingly, there is always a need for some reg-ulatory discretion, but at the same time, its successfulexercise requires technical capacity.

• There is a need for specialized skills: The negotiationand supervision of concessions is equally if not moredemanding for government organizations. Conces-sions generate high transaction costs, and the weakerthe institutional capacity, the higher these costs are

88

likely to be. On average they can be as much as 5 to10 percent of total project costs (Klein, So, and Shin,1996). Governments must organize themselves tomanage the process of designing and awarding con-cessions, and then to formulate laws and regulationsthat affect the operation of those concessions. Kerfand others (1999, p. 9) warn that inefficient organi-zation can result in substantial costs to government,developers, and consumers. They stress the need forgovernments to retain qualified and experienced ex-perts who are able to provide sound advice on a rangeof issues. While detailed technical expertise can becontracted out, governments must have staff with rel-evant expertise to hire and oversee the consultants andto incorporate the lessons of experience for futureconcessions.

Concluding Remarks

It is necessary to view this discussion of institutionalarrangements in the wider context of how the move tomore decentralized budget management decision makingis operationalized, and hence how the government oper-ates. These institutional changes are a means to an end,and not an end in their own right. Countries consideringsuch institutional changes should recognize this at theoutset and be prepared to commit to undertaking the morefundamental changes to their budget systems implied inthis process. Obviously, high-level policy commitment isrequired if for no other reason than institutional change—of any kind—takes time to implement.

OECD countries have generally adopted a phased ap-proach to introducing such changes. Some have adopteda policy of gradual voluntary change, deciding to let op-erating units move to a new management culture at theirown pace, albeit with some technical support and guidancefrom the center. Others have piloted the new systems inmore progressive units, with more active encouragementby central organizations. Once reforms have taken hold,and have passed some critical threshold, they may thendecide to make reforms compulsory or at least to promotethem more vigorously against an agreed timetable, solid-ifying the new arrangements in new legislation. However,given the variety of country experiences, answering basicquestions on the sequencing of such reforms is difficult.Where should reforms start? What is a reasonable time-table for their completion? Should changes be required ofall units, or should the reform process be selective? Thereis no single optimal answer, and the pace of reform hasto be country specific.

With this significant qualification, OECD experiencedoes suggest some necessary elements for successful in-stitutional reform.

• A staged approach is less risky. Institutions shouldbe progressively prepared to move from traditional,input-focused budget management to output-focused

©International Monetary Fund. Not for Redistribution

Page 98: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Concluding Remarks

89

and then to outcomes-focused management systems—that is, to move progressively along the continuumof the five Ds, from deconcentration to devolution,and in some cases to divestment.

• Preparing institutions to progress in this way usuallyrequires that their management and control systemsbe upgraded. These enhanced systems then need tobe tested by a staged process of granting themgreater and greater autonomy and flexibility in man-agement decision making.

• The process involves a progressive change in the in-ternal culture within government, and this must bereflected in a heavy investment in human resources.Experience indicates that technical innovations meantto improve performance must be reinforced by a par-allel process for putting in place mechanisms to im-

prove managers’ accountability to users and, moregenerally, to taxpayers.

• Following from this, a point will be reached when fur-ther budget management reform will require a changein the legal status of the staff and units concerned.

• Finally, and most important, it is necessary to commitresources to steer and manage the reform process. Themere fact that institutional change takes time impliesthat there is a high likelihood of it being derailed. Aphased approach implies that new principles and pro-cesses may not apply to the entire government sector,and this means that managerial imbalances can arise,human resource problems can develop, and reform canstall before a critical threshold is reached. Success orfailure will depend on high-level commitment andskill in managing the reform process.

©International Monetary Fund. Not for Redistribution

Page 99: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

This section examines the implementation of bud-get system reform, using the experience of coun-

tries that have attempted such reform to identify theprincipal elements of a successful change-managementstrategy and to highlight some of the lessons to belearned. Particular attention is paid to overcoming theconstraint of limited managerial capacity, which canlimit a country’s ability both to operate the new bud-get system and to engineer the move from one bud-get system to another.

What Can Be Learned from the OECD Experience?

It is perhaps not surprising that the fundamentalchanges in budget management that have been intro-duced in the more institutionally advanced membercountries of the OECD have caught the attention ofmany middle-income and emerging market countries,which now seek to adopt similar reforms. As a con-sequence, the Fiscal Affairs Department of the In-ternational Monetary Fund has increasingly faceddemands for technical assistance arising from thesesources. The department’s experience has revealedthat, although it is not too difficult to design budgetreform measures, nor to specify detailed implemen-tation plans, the lack of management skills is often animportant constraint. This constraint has two dimen-sions. The first is the substantial management capacityrequired to operate the new type of budget managementmodel. The second, and arguably more important, is thechange-management skills required to introduce newsystems, to sustain reform efforts, to follow throughon implementation, and to adapt to contingencies andchanges in the external environment. There are severalimportant lessons for middle-income countries thatlook to emulate the OECD budget reforms.

Budget Reform Is an Evolutionary and Not a Discrete Process

Budget reform is an evolving process, with budgetsystems moving through distinct stages. A modern

budget system should be able to meet three main re-quirements: first, to ensure control over expendituresso that they are consistent with the budget law; sec-ond, to stabilize the economy by timely and efficientadjustment mechanisms for the fiscal aggregates; andthird, to promote allocative and technical efficiency inservice delivery through procedures that provide in-centives for greater productivity. It is possible to char-acterize the evolution of budget systems as a gradualmovement through distinct stages defined by progres-sively assuming, and placing different emphasis on,these three requirements.

There Is a General Recognition of the Need for Budget System Reform

Middle-income countries typically find themselvesfacing common problems arising from the heritageof traditional budget management systems. Specif-ically, past attempts to ensure that budgetary pro-cesses would deliver a satisfactory aggregate fiscaloutcome in support of macroeconomic stabilizationhave often involved imposing a number of limita-tions on budget managers. Even when care was takento incorporate some degree of budget flexibility,typically the budget systems have been left with acomplex set of restrictions. These are increasinglyrecognized as diminishing the allocative and oper-ational efficiency of budget execution, normally byleaving budget managers to operate with only lim-ited responsibility for results. Accordingly, there isgrowing acceptance that the next stage of reform isto provide greater inducement for managers to focuson possible improvements in allocative and opera-tional efficiency at the program delivery level. How-ever, before this can be done, there must be a strongenough consensus—a critical mass of impetus for reform—to enable a shift in the budget managementmodel. In the OECD countries, this consensus wasreached from two separate directions: the increas-ingly obvious limitations of traditional compliance-oriented budget systems, and increased realizationthat stabilization and efficiency objectives were notnecessarily in competition.

90

VIII Managing the Reform Process

©International Monetary Fund. Not for Redistribution

Page 100: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

The Budget Reform Strategy Is a Major Part ofthe Government’s Overall Policy

Once the need for reform is recognized and ac-cepted, a common approach is to make it part of thegovernment’s fiscal strategy and a central element ofgovernment policy. The reform initiative is “owned”and supported by all ministries, not just the MoF orthe budget office. In this way, it forms a critical ele-ment of management strategy for all budget institu-tions. All public sector managers assume responsibilityfor its implementation, rather than leaving it to thosemanaging the agencies’ budgets or accounting sys-tems. This high-level commitment facilitates the re-quired changes in administrative procedures andmakes central agencies more willing to devolve bud-get management.

There Must Be Adequate Levels of Fiscal Control

Before advancing down the road of a more flexible,more decentralized budget management model, allcountries must establish adequate levels of control toensure compliance and stabilization objectives. Thethree prongs of reform—providing flexibility to bud-get managers, achieving greater certainty about budgetfunding, and increasing pressure to perform—need tobe pursued in parallel to ensure the overall success of the reform process. Allowing budget managersgreater freedom without putting in place an account-ability structure to ensure improvements in perfor-mance, as described in Section IV, could increaserather than decrease budget inefficiencies. At thesame time, providing a greater degree of resource cer-tainty may involve a trade-off against the need to cutspending for stabilization purposes. It is, therefore, es-sential to proceed down these three tracks in a highlycoordinated manner.

There Is a Need to “Engineer” Budget Reforms

Reforms must be engineered—a reform plan formu-lated, an implementation strategy agreed, and imple-mentation of the reform program managed to achievethe objectives and sustain the reform initiative. Thereform program has to be “sold” to the main stake-holders in the budget system. Perhaps more important,a reform team must be identified and empowered tocarry out the reform. It is this aspect of the reformprocess—the exercise in change management—thatis central to this section. Change management addressesthe question of how to engineer the shift from onebudget management model to the other—literally, in-troducing a shift within government from a compli-ance to a performance culture.

The Overall Approach to Reform

91

The Overall Approach to Reform

As suggested in previous sections, in many emerg-ing economies, reservations about the budget reformprocess relate less to the specific changes proposedand more to concerns about how the reform will beimplemented. It is possible to identify the sources ofsome of these concerns.

A Big Bang or an Incremental Approach?

A conservative approach is recommended for pur-suing changes to budget procedures that will improvethe trade-off between the objectives of stabilizationand efficiency in resource use. Specifically, the reformstrategy should be based on relatively small, sequen-tial changes rather than on concentrated reforms thatare part of a single, high-profile package. This is rec-ommended for two main reasons:

• The incremental approach to reform addressesthe “chicken or egg” problem of building finan-cial management skills within spending agencies.Highly centralized budgeting discourages the development of financial and allocative skills inagencies. An evolutionary approach to budgetreform allows agencies to expand their financialmanagement skills as greater management re-sponsibilities devolve to them. This reduces the risk,which may arise under a more radical approach toreform, that a control vacuum will emerge becausespending agencies are unready to abruptly takeover responsibility for allocative decisions previ-ously determined at the central level.

• The incremental approach minimizes other risksinvolved in reform. The pace at which additionalresponsibility is devolved can be matched againstthe pace at which capacity builds in agencies. Theincremental approach, therefore, minimizes therisk that increased flexibility to spending agenciesmight contribute to budget overruns should rev-enue conditions deteriorate.

Unfortunately, a cautious, step-by-step approach toreform may make the process appear to lack coher-ence. When pursued separately, important legal andinstitutional initiatives, programs, and projects maynot appear to fit together into an overall, coordinatedreform plan. The incremental approach makes it diffi-cult to specify a clear timetable for the reform processand to place emphasis on the sequencing of the indi-vidual components of the reform. This raises ques-tions about the role of the agents of change—and mostspecifically, the central budget office.

Top-Down or Bottom-Up Reform?

The OECD-type budget reforms require a differentbudget management model. The central budget office

©International Monetary Fund. Not for Redistribution

Page 101: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

149Scott (1996) analyzes PEM reforms in New Zealand, emphasiz-ing some important elements of this change-management process.

VIII MANAGING THE REFORM PROCESS

92

must take measures to enhance performance withinbudget institutions, which is rather different from itstraditional role. In traditional budget systems, thefunction of the central budget office is to act moreas a policeman, ensuring compliance with the budget,monitoring budget progress, and intervening whenthings go off track. The new budget management modelrequires a fundamental institutional shift in public ex-penditure management, first, by reorienting the centralbudget office’s role, and second, by devolving fiscalmanagement decision making. This in turn implies asustained effort for training and adequate time to de-velop a new culture, as indicated by Malaysia’s expe-rience in introducing its Modified Budget System, oneof the more successful reform efforts (Box 46).

Not only does the central budget office have to in-crease the capacity of government agencies to assumean enhanced management function, but, at the sametime, it has to change its own work practices and realignits objectives. It must still ensure that the resource allo-cations are consistent with overall fiscal targets and thegovernment’s program objectives. However, ratherthan dictating how these are to be attained, as in the oldsystem, the budget office must now establish guidelinesand procedures for reviewing policy changes, main-taining baselines and databases for measuring the bud-getary impact of changes, and conducting an ongoingevaluation of programs and measurement of perfor-mance.148 Moreover, the budget office is expected toprovide central guidance to the reform process, pro-ducing handbooks and advising on best managementpractices. In attempting to do so, it is likely to initiallymeet its own capacity constraints.

The problems are compounded by the fact that the re-formed budget management model is decidedly moredecentralized. Attempting to install a bottom-up sys-tem from the top down may involve a fundamentalcontradiction: the greater the success in decentraliz-ing decision making, the weaker the central budgetoffice’s leverage over agencies. With decentraliza-tion, and the resultant managerial freedom, the centerloses effective control over the budget, which will mostlikely work against efficiency in resource allocation.The budget office must adapt its work practices andbuild the information systems required for this newenvironment—and may, of course, encounter internalresistance to undertaking such fundamental changes.

Therefore, in terms of engineering these reforms,there may be conflicts inherent in a top-down approachthat directs managerial improvement from the center.For example, although it may be necessary, it may notbe appropriate for the central budget office to leadthe way and to increase its capacity to assist othersin implementing the reforms. If it tries to do this too

aggressively, it may risk smothering managerial ini-tiative and encouraging the old compliance mental-ity. It is also unclear whether a greater devolution ofdecision making is in the central budget office’s inter-est, both in the narrow view of attaining its complianceand stabilization objectives, or more generally, in re-ducing its overall influence over the budget process.There are costs as well as benefits to moving awayfrom traditional budgeting methods, and the principalagents of change may not necessarily see a net gain tothemselves in advancing this process. For all thesereasons, it is important to buttress budget system re-forms with wider governance reforms—as indeedhas been the case with pioneering OECD countries—stressing transparency and reinforcing the legal frame-work of budget management.

Elements of a Reform Strategy

The fact that OECD countries made this change inan evolutionary rather than a discrete manner and thattheir reform efforts are continuing brings out the im-portance of the change process. The implementationstrategy and the mechanics of implementation shouldbe focused on managing the reform process as awhole, rather than concentrating on individual reformelements.149

Change-management methodology emphasizes theneed to address crucial components for successful in-stitutional change that should be followed in a logicalsequence (Kotter, 1996). One of the most widely knownprescriptions for successful change management isshown in Box 47. Although this literature is derivedfrom the private sector and the experiences of trans-forming large corporations, there are obvious paral-lels with attempts to reform a government’s budgetmanagement system.

The experience of government modernization pro-grams in emerging market countries indicates that thefirst two steps in Box 47, getting agreement from topdecision makers that the budget system needs to bemade more flexible, may not be such a problem.That is, there already exist a sense of urgency andpeer group pressure for reform. Similarly, creating acommon vision about the organizational and proce-dural changes required is not so difficult: the vision isto adopt OECD budget systems. However, the realconstraint is the human element in implementingsuch institutional change (steps four through seven).Generally, not enough attention is paid to the agentsof change—identifying them, offering them incentivesto undertake reforms, and removing the constraintsthey face in sustaining these reforms. Accordingly,

148See Schick (2001).

©International Monetary Fund. Not for Redistribution

Page 102: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Box 47. The Steps for Creating Major Institutional Change

1. Recognize a sense of urgency to make a change.2. Develop a powerful enough coalition to support the

change.3. Identify a champion of change who creates and com-

municates a common vision.4. Empower and give resources to others to implement

this vision.5. Plan for early successes (“easy wins”) for positive

demonstration effects.6. Build on these early successes to sustain the

momentum for reform.7. Institutionalize the changes so that they become a

part of the organization’s culture.

Source: Adapted from Kotter (1995).

successful budget system reform requires engineeringthese three steps:

• Identify agents of change. Who are the championsof reform? Who is going to recognize the need forreform, design the reform, and monitor and imple-ment the reform?

• Assess and ensure the adequacy of the means attheir disposal. By what processes and with whatmeans will reform be carried out? What is thecapacity to carry forward reform—is this capacityadequate or does it require administrative restruc-turing, changes in procedures, or new skills?

• Restructure the incentives to undertake and sus-tain reform. In what ways will reforms benefitthe individual rather than the system as a whole?Will individuals be compensated for the costs ofthe effort involved? Once reform incentives areestablished, how are they to be maintained andprotected?

Elements of a Reform Strategy

93

Box 46. The Malaysian Modified Budget System

The existing planning, programming, and budgetingsystem (PPBS) system, modeled on the U.S. system, wasreplaced by the Modified Budgeting System (MBS) in 1990to place greater emphasis on outputs and outcomes. TheMBS was piloted first in three agencies and, based on thelessons learned, it was phased in over a period of five yearsfor the operating budget only. This was slightly modified in2001, when a two-year budgeting system was adopted,along with an important annual review of policies.

Key Elements of the Reform

• Based on a longer-term strategic plan: Policies aregeared to Vision 2020, which envisions Malaysia as afully developed country by that year. Policies are catego-rized as either existing policies, new policies, or nonre-curring (“one-off”) policies. Operational plans of agenciesdifferentiate short-term from longer-term strategies toachieve desired policy objectives. For the budget, theseare reviewed centrally to determine whether they will re-alistically meet policy objectives, and after five years thereis a more fundamental review of the programs.

• Improved financial compliance: Macro and agency fis-cal discipline are strengthened throughout the system,with improved financial planning and control systems thatuse information technology. This is supported by severalother reforms: strengthened policy development and co-ordination, organizational restructuring, and improvedmanagement of human resources.

• Enhanced quality of management: MoF from the starttightened financial discipline internally. In return, man-agers were given more autonomy, with flexibility tomove funds between programs and activities. They were

also given more stability in their resource base, especiallyafter adoption of the two-year budget framework. Allbudgetary requests are linked directly with policy, andeach ministry is given an expenditure target for existingpolicies. New policies can be created as a consequence ofexpanding existing policy, and the agency can requestextra funding for a new policy or a one-off policy.

• Improved efficiency of government operations: To de-termine impact, agencies are required to draw up perfor-mance indicators and a Program Agreement to indicatehow the use of resources will result in outputs and im-pacts, as indicated by the agreed performance measures.The Program Agreement emphasizes identifying theagencies’ clients.

Key Challenges

There were delays, even with reform efforts carefullyplanned, managed, and phased in over time.• This effort was centrally managed, with two divisions in

the MoF in charge of training and implementation. Theircoordination, at least initially, posed problems.

• It took time to convince the civil service that the MBSwas workable and to overcome the difficulties of categorizing outputs/impacts and developing costingsystems.

• Training was a large component of the reform effort, butthe scale of this effort meant it was difficult for all second-and third-tier organizations to fully implement the MBS.

• It took time to get managers to manage. This change inculture within agencies was difficult to engineer, and therewas a lingering preference to let the MoF make decisionsas in the past.

©International Monetary Fund. Not for Redistribution

Page 103: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VIII MANAGING THE REFORM PROCESS

94

The remainder of this section proposes some answersto these questions.

The Agents of Change

In a traditional, highly centralized budget system,there is a presumption that changes are derived topdown, from the MoF and, specifically, from the budgetoffice. This is not necessarily the case under a devolvedbudget management model. At least initially, any re-laxation of centralized controls may meet some resis-tance from the central budget office, which insteadshould be the key agent of change. This resistance canarise from a number of sources.

First, there is the fear of the unknown. Why adoptdifferent approaches when the present ones have notbroken down and appear to be delivering at least thebottom-line fiscal result? The budget office may notbe convinced of the need for a new budget manage-ment model which moves away from compliance andstabilization objectives and begins to emphasize effi-ciency and effectiveness in resource use. Obviously,budget officers will be more easily convinced if theyalready have in place a management framework thatensures basic compliance with the law and allows suf-ficient certainty in the fiscal outcome to minimizeshort-run disruptions to spending plans that may arisefrom the need to ensure fiscal stability.

Second, this conservatism is often reinforced by aconcern that relaxing central controls and giving bud-get program managers greater personal freedom willresult in more waste and corruption. This concernarises because waste is often invisible in traditionalbudget systems, which focus more on the correct useof inputs but not on the possible waste embodied in theresulting outputs. Under traditional systems, control isover total inputs and not their composition, and thisleads to a concern that, with the limitations of existinginformation systems, allowing freedom in the compo-sition of inputs will subvert this macro control. It isalso important to recognize that, under any system,where line agency accounting and internal controlsare rudimentary and poor, corruption may increase. Itfollows that there should be a major program to up-grade agencies’ financial management skills to somebasic level to offer the budget office assurances abouttheir fiscal responsibility.150

Third, there is often a natural reluctance to give upthe power that goes with centralized control. Changeis not only threatening, but it may also mean a loss ofpecuniary or other benefits that derive from the author-ity to release detailed line items to spending agencies.As a consequence, the central budget office, whichshould be the leader of reform, may become its chiefimpediment. The question becomes how best to re-

structure incentives to move the MoF’s budget officetoward a new role. The answer is to use consistentpressure from across the government and to continueto redefine the MoF’s new role in the context of areform of the whole budget system.

Fourth, the MoF generally has limited capacityfor change management. In private sector compa-nies, change management receives much top-levelattention and tends to be heavily funded. In the pub-lic sector, it typically suffers from benign neglect.Often there is a need to inject new capacity or to buildup existing capacity for change within government.Indeed, with such considerations in mind, the leadersof reform to a decentralized budget management modelmay be more likely to be found in the line agencies orlower levels of government than in the MoF.

Moreover, the management constraint does notmerely hinder the successful implementation of thereform process. The new budget model being intro-duced implies that public officials will cease beingadministrators and will assume a management roleakin to that in the private sector.151 Unfortunately, acentralized, compliance-oriented budget managementsystem does not foster the creation of such manage-ment capacity in either the MoF or the line agencies.These officials typically have spent most of theirworking lives in a compliance-oriented environmentwith a “command and control” management culture.Their reflex responses typically are to manipulate de-tailed external control systems to protect their pro-grams from the cash-rationing operations of the MoF.Line agency officials often see their role as adminis-trators, consisting primarily of distributing limitedcash to keep basic services functioning. This role doesnot recognize the value of good agency financial man-agement, nor foster the acceptance of increased man-agerial responsibility.

How best to build management capacity in a compliance-oriented traditional budget system? Thereis no easy answer to this question. However, experi-ence suggests that a “big bang” approach is difficult.An interesting case is Bolivia, where in 1988 thegovernment developed a program to recruit a limitednumber of middle-level managers and to place themin a few key posts in each agency in order to enhanceperformance. A total of 653 key posts were origi-nally planned but, by June 1993, only 69 had beenfilled. The strategy of sprinkling a small amount ofmanagerial talent across a large number of agenciesevidently had little impact, and a new approach wastried. From 1992, the government adopted a moresuccessful, broader civil service reform with a longertime frame, with about 2,500 targeted posts to be in-

150A point emphasized in Section V.

151For a discussion of the importance of the management capac-ity and ethic in this budget management model, see Schick (1998,pp. 130ff).

©International Monetary Fund. Not for Redistribution

Page 104: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

corporated into a newly formed Civil Service Program.The new strategy was to achieve a critical managerialmass within selected central administration agenciesand to progressively expand the number of agencies inthe program.

The experience of Bolivia, as well as of othermiddle-income countries, suggests that the solutionto the management constraint involves tackling twomain problems. First, building this capacity will taketime and must be done in stages. Second, this phasedapproach should be an integral part of the reform’s de-sign. One strategy for minimizing the risk of misuse orabuse of managerial authority, while recognizing thatagency capacity building takes time, is the adoption ofresource agreements. The approach used in Thailand,summarized in Box 48 and discussed more fully below,makes agencies’ access to greater managerial auton-omy conditional on improved financial management.These systems, and the managerial capacity to oper-ate them, if functioning properly, will be important as-surances of the quality of data and the reliability ofcontrols in the new devolved budget system.

The Means to Manage the Reform Process

The efforts of many middle-income countries andeven more developing countries to improve budgetsystems have been stalled prematurely and overtakenby changing external events. This is a result in part offailures in the management of the change process andin part of a lack of adequate resources to carry out thereforms. There are some general lessons to be learnedabout change-management strategies, all subject to thequalification of exceptions.

• Emphasize objectives. The cultural shift from acompliance-oriented to a performance-oriented

budget management model essentially involves ashift from focusing on inputs and how they areemployed to focusing on outputs and how theyfulfill the original budget objectives. This is a moveaway from traditional administrative proceduresand toward a more modern management orien-tation, focused on meeting clear objectives. Notsurprisingly then, a key to making a successfultransition from one budget management model tothe other is to focus on objectives. Those involvedin the pathbreaking public sector reform processesin New Zealand and Australia have stressed thisaspect of reform.152

• Keep participants focused on objectives. Publi-cizing the objectives is one way to create incen-tives and put pressure on participants to meettheir objectives, as well as a way to hold them ac-countable for their results. The United Kingdom’s“Citizen’s Charter” publicly committed public en-tities to meet specific performance standards andidentified the means of redress when they failed todo so. This approach has given way to the use ofPublic Sector Agreement targets to publicize agen-cies’ objectives and hold them accountable (seeBox 20). In the United States, the National Perfor-mance Review during 1992–2000 placed heavyemphasis on measuring and publishing results.

• Evaluate performance as a tool for enforcing per-formance. There must be feedback mechanismsto continuously improve the means by which agen-cies pursue their objectives. At the same time, thesemechanisms must be cost-effective, avoid exces-sive transaction costs, and carry minimal risks thattheir application will produce biased, inaccurate,or inconclusive results.

• Reward good performance and sanction poor per-formance. The reform leaders must ensure a clearlink between performance and rewards. Establish-ing clear performance accountability involves, first,a threshold level of basic financial and personnelmanagement systems that can be used to report onperformance, and second, a performance-orientedincentive framework that links rewards to perfor-mance, as discussed in Section IV. Many middle-income countries are handicapped in followingthe above strategy, however, because traditionalbudget systems put little emphasis on performance,and managers in line agencies, therefore, are notused to subjecting new policy proposals to criticalanalysis. Nor are they accustomed to managing ex-penditures to achieve a particular outcome. Thereis a need, therefore, in many countries to develop

Box 48. Thailand: Line Agencies Reform Contract

Main elements of the approach:• Identify reform-oriented line agencies.• Offer them a contract or resource agreement.• Reduce external control by Bureau of the Budget, in

exchange for demonstrated improvements in agencyfinancial management.

• The agency must demonstrate the achievement of“hurdle standards” in seven areas:– Budget planning– Output costing– Procurement management– Budget/funds control– Financial and performance reporting– Asset management– Internal audit

Source: Dixon (2002).

152For example, the 1989 New Zealand Public Finance Act, whichintroduced major reforms in fiscal management, developed a set ofprinciples of financial management that began with the clarificationof strategic and operational objectives.

Elements of a Reform Strategy

95

©International Monetary Fund. Not for Redistribution

Page 105: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VIII MANAGING THE REFORM PROCESS

96

such evaluation skills in agency management andalso to establish information systems based on indi-cators of output and performance that support suchactivities.

• Upgrade basic management tools to provide ad-equate, relevant, and timely information. Someelements of this are described in Section V, suchas the need to strengthen internal controls andinternal audit, to apply information technologyin upgrading information systems, and to expandcapacity for cost accounting. In particular, there isa critical need to upgrade the accounting system,as explained in Section VI.

• A top-down reform process may rapidly producediminishing returns. As noted, any attempt to in-stall what is essentially a bottom-up system fromthe top down involves a fundamental contradictionin that success in decentralizing decision makingwill simultaneously weaken the budget office’sleverage over agencies. The budget office musttherefore expeditiously adapt its work practicesand build up the information systems required forthis new environment—and must overcome anyinternal resistance to making such fundamentalchanges.153

Restructuring Incentives to Support Reform

The reform program cannot ignore the interests ofstakeholders, particularly the key players such as thecentral budget office and the targeted agencies. If thebudget office cannot see that the reform process willsafeguard its basic controls over the budgetary process,it is unlikely to be a strong proponent of reform. If thesystems being created under the reform process do notprovide agency-specific management tools, it is un-likely that those systems will yield the desired results.There may also be a need to enlist the support of indi-vidual agency staff by involving them in the designand implementation, providing performance-linkedinstitutional rewards, and introducing job transitionassistance for retrenchment or retraining.

Some middle-income countries have used perfor-mance agreements between the central budget officeand agencies to address some of these issues (Box 49).The process is managed centrally, typically by the bud-get office, and is described in a rather stylized fashionas a twofold approach of applying pressure from aboveand increasing capacity from below.

Applying Pressure from Above

• The MoF encourages agencies to undertake in-depth evaluations of activities that might war-

rant a reconfiguration of resourcing to improveperformance. These evaluations could be con-tracted out and could serve as an input to thepreparation of an MTBF. The MoF might alsoinstitute an “efficiency dividend” policy, wherebyagencies are required to find savings in theirbaseline budgets of an agreed percentage eachyear (Box 50). Over time, as improved accountingfor assets allows, a capital charge could be intro-duced to ensure better asset management at theagency level.

• Much of this effort is predicated on the MoF beingable to obtain better information on agency per-formance. However, by itself, performance in-formation provided to central agencies is a ratherineffective mechanism for ensuring that agenciesactually use their new financial freedoms to en-hance program outcomes rather than bureaucraticor personal goals. The center may set the newframework for budget management and put inplace mechanisms to tighten control if agencymanagement fails, but success ultimately dependson the capacity of agency management to adapt tonew outside pressures. And this capacity comesfrom within an agency itself, rather than from thecentral level, through effective leadership of theagency, clear frameworks for operation and re-sponsibility, effective management informationsystems, and proactive central coordinating divi-sions in the areas of agency accounting and bud-geting. Hence the second approach:153This is discussed more fully in Schick (2001).

Box 49. Formal Performance Agreementsbetween Central Authorities and Agencies

in Latin America and the Caribbean1

Benefits derived by agencies:

• Easier to hire key human resources• Enhanced managerial autonomy• Reliable cash flow relative to budget• Performance-linked institutional rewards• Technical assistance to improve management processes

Accountability requirements:

• Minimum reporting standards in financial and person-nel management systems

• A financial restructuring plan or institutional strength-ening agreement with detailed implementation plan anddeliverables

• Agreed performance targets and associated reportingrequirements

1Based on programs in Bolivia, Ecuador, Jamaica, andNicaragua; see Reid (1998).

©International Monetary Fund. Not for Redistribution

Page 106: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Increasing Capacity from Below

The central budget office, as the agency responsiblefor budget management, adopts a proactive role in thedevelopment of good management in the spendingagencies to which it provides budget funding. This isa key element in the budget reform model based ondevolution of financial freedoms and a more certainoperating environment. This takes the budget office

out of its traditional detailed control mode and towarda more managerial approach. Typically, some of themain elements in this approach are:

• The budget office highlights and “sells” improvedmanagement in spending agencies as a precondi-tion of the budget reform process.

• The budget office develops competencies in theevaluation of internal resource allocation processesand structures in spending agencies, focusing

Box 50. The Experiences of Australia and Sweden with Using an Efficiency Dividend

A number of countries, including Sweden and Australia,have introduced a charge against the baseline budget fore-cast for agencies. This charge, often called an efficiencyor productivity dividend, aims to provide agencies withthe incentive to continually seek new or more efficientmeans of undertaking ongoing government business. Italso allows the government to redirect a portion of the effi-ciency gains to higher-priority activities. The charge shouldbe small to allow managers room for maneuver. Even so, asthe experiences of Australia and Sweden indicate, it is dif-ficult to sustain these dividends politically for a prolongedperiod. Because it represents an across-the-board cut oninputs, it implies old-fashioned incrementalism (or in thiscase decrementalism) in budgeting and ideally should beused in conjunction with more performance-oriented effi-ciency improvements that focus more on outputs and out-comes, as was the case in Australia and Sweden.

Australia

The Australian efficiency dividend was introduced in1987 on administrative budgets, applied to agencies’ totalrunning costs. It was a uniform percentage cut applied eachyear and taken from portfolios before their budgets were re-leased. It was applied irrespective of agency size or demon-strated efficiency. The rationale was to capture a proportionof productivity gains made by operational managers and re-turn funds back to the government to be directed to priorityareas. Initially, the rate was set at a flat 1.25 percent perannum in the 1987 budget, and in 1990 it was agreed toretain this rate for a further three years. Subsequently, aParliamentary Committee Report recommended its reten-tion but suggested that the government consider both re-ducing the rate and increasing the number of exemptionsfor small agencies. The rate was reduced to 1 percent as ofJuly 1994 and by mid-1994, the number of exemptedagencies had grown to 46. One reason the efficiency div-idend could be relaxed was that from the late 1980s otherefficiency and performance contracts were being used notonly to promote efficiencies in input costs but also to pro-mote improvements in levels of outputs. By the early1990s, almost all agencies had various forms of resourceagreements in place. It was found that, although savingswere substantial, at least initially, total running costs didnot decrease as new policies were approved; rather an ele-

Elements of a Reform Strategy

97

Sources: Wanna, Kelly, and Forster (2001); Brunsson (1995).

ment of “gaming” was introduced, with funds being recy-cled rather than saved. The nature of the federal budget wassuch that the dividend was only applied to administrativecosts and many areas therefore escaped the discipline, in-cluding programs funded by the central government but ad-ministered by nonfederal agencies.

Sweden

In Sweden, across-the-board cuts of the appropriationsof all government agencies were introduced in the late1970s as a means of taking into account productivity in-creases in the public sector. At that time, the increase inproductivity in the private sector was estimated at 4 per-cent per annum and it was thought the government coulddo at least half as well. The 2 percent standard, or “thecheese-slicer” was introduced, subjecting all agencies to a2 percent cut in their appropriations after allowance wasmade for cost increases in salaries and expenditure. Ini-tially, some exemptions, full or partial, were allowed ongrounds of proven efficiency or in cases where serviceswould be reduced to unacceptable levels. It became politi-cally unpopular, and many questioned its ability to spur effi-ciency, given that it was applied across-the-board and basedon inputs with no regard to outputs. The rate became subjectto political negotiation. Exemptions were made for agencieswith particular problems and from the mid-1980s 2 percentcould be replaced with savings of 5–6 percent dispersed overa three-year period. The standard was also subject to tech-nical problems of interpretation. Before the cuts were ap-plied, all agencies were compensated for salary increasesand average inflation over the past two years so that theytended to be undercompensated in times of rising inflationand vice versa. Ex post changes were made after the cuts—agencies could be compensated for new assignments or havetheir appropriations further reduced for decreasing demandfor services. Agencies could overdraw their appropria-tions for reasons considered “legitimate” by the govern-ment, e.g., to compensate for salary increases not previouslytaken into account. As a result, the cheese-slicer lackedtransparency which was reinforced by a lack of interest inex post evaluations. After 1990 it was shelved because ofthe economic crisis. Thereafter, more emphasis was placedon the central element of budget reform, cutting back throughthe review of programs—the “cake-slice approach”—designed to cut considerable pieces of the governmentsector programmatically.

©International Monetary Fund. Not for Redistribution

Page 107: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

particularly on the ability of a spending agency toidentify cost-ineffective aspects of its activitiesand to reassign resources away from such areas.

• The budget office develops a checklist of the attri-butes of management systems in spending agenciesthat are conducive to allocative flexibility.

• The budget office then works with spending agen-cies in the application of these competencies toagency operations, with the ultimate goal that theagency will both undertake internal reallocationand identify savings options in response to exter-nal budget management requirements.

• The budget office also initiates a review of perfor-mance indicators for each major department. Thisshould begin with the enumeration of departmen-tal activities and the objectives being pursued byeach. Performance indicators already available indepartments should then be matched to these ob-jectives, and gaps in the existing range of perfor-mance indicators should be identified. Assistanceshould then be sought from sectoral experts to de-fine and collect new performance indicators to fillthese gaps, before the complete system of indica-tors is institutionalized.

Providing an Enabling Environment

Unlike a policy change, an institutional reform, likethat involved in budget system reform, typically has nopolitical constituency. No one in the general popula-tion is pressing for this reform, and even stakeholdersmay be difficult to mobilize because there may be a per-ception of powerlessness in bringing about enhancedexpenditure benefits or improved efficiency in servicedelivery. One key to sustaining reforms revolves aroundhow to generate the pressure for reform in society asa whole. From country experiences, what are the ele-ments of such an enabling environment?

An Adequate Level of Fiscal Stability

Fiscal crisis can help highlight the need for reform,or it can deflect attention away from it. OECD countriescarried out a major fiscal consolidation concurrentlywith introducing major fiscal reforms. The pressurefor budget reform largely arose from rapid growth inthe public sector in the 1970s and early 1980s, whichraised questions as to whether this was affordable andwhether it was providing value for money. The growthin public spending had been so sharp that there weredoubts as to its sustainability, and there was a recog-nized need to find reductions in some sectors to satisfyemerging demands. There were general concerns thattaxation levels were too high and that the public sectorwas crowding out the private sector, and these addedto this pressure for change. The consequence was a

VIII MANAGING THE REFORM PROCESS

98

substantial fiscal consolidation, which reduced budgetdeficits and stabilized debt levels, alongside the intro-duction of fundamental budget system reforms.154

In the OECD countries, then, fiscal crisis may haveprovided an opportunity to introduce fundamentalreforms, making changes more easily accepted andmore quickly implemented. However, in transitioneconomies the experience has often been that, whenthe degree of fiscal stress was severe, the initial re-action of public officials was to muddle through, adopt-ing short-term crisis management solutions that oftenwere hostile to longer-term reform. Undoubtedly, thedegree of crisis is important—too much can be dis-ruptive and counterproductive to reform.155

Some qualifications are perhaps in order. First, per-haps more important than the degree of fiscal stress isthe capacity of the budget system to adapt to it. If reli-able mechanisms are in place, then fiscal stress is prob-ably a secondary factor. Second, if fiscal stress is notso destabilizing as to allow some political and ad-ministrative stability, then its negative impact can beexpected to be much diminished. In this scenario, therewill be continuity in management to focus on solutionsto the immediate problem and also to put in place moresubstantive changes to address the root of the problem.

Third, much depends on the reform strategy adopted.An approach that lacks focus and does not set clear pri-orities is most liable to fall prey to short-run crisis man-agement. At the same time, an approach that is tooambitious and costly is also doomed. To be effective,performance-oriented public sector modernization in-volves immediate up-front costs and in all likelihoodrequires a substantial infusion of scarce managerialtalent. Recognizing this, reform should be designed ina way that phases in the cost increases over a longenough period of time that the government can meet theincreased fiscal costs, there is sufficient time to reap thebenefits of improved efficiency, and the reform effortdoes not jeopardize the overall macro fiscal policystance. A gradualist approach that covers agencies se-rially over a sustained period is also required by the factthat the professional human resources are in short sup-ply; such an approach enables learning from mistakes.Moreover, the serial approach, by allowing reform tostart with the most promising agencies, enhances thechances of demonstrating visible results early on, andthereby of providing momentum for reform.

Strengthened Overall Accountability

For reforms to be sustainable, it may be necessaryto strengthen the overall accountability framework.Section IV emphasizes that the core of the new perfor-

154See OECD (1996).155For a discussion of the case of Russia, see Diamond (2002a).

©International Monetary Fund. Not for Redistribution

Page 108: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Providing an Enabling Environment

99

mance budgeting management model is a reformulatedaccountability system. The focus is accountabilitywithin the executive branch and the need for the ac-countability of budget managers to be clearly definedand changed. Whereas in the past they were held ac-countable for the correct use of inputs, now they areto be held accountable for the results of using thoseinputs. However, this is only one aspect of the over-all accountability for the use of public funds. There isalso a need for the executive to give an account to thelegislature about how it is meeting its responsibilitiesas a whole, and this should be enforced by an inde-pendent external audit body that reports at least annu-ally to the executive. Often these latter dimensionsof accountability also need to be strengthened. Theaccountability of the executive often needs to be en-hanced by prescribing fuller reporting ex ante and expost of fiscal policy strategies and intentions, as wellas of the financial and nonfinancial outcomes. Theexternal audit office often needs to be better re-sourced, to be made more clearly independent fromthe executive, and to reorient its work to better servethe needs of the legislature. In the new performancebudgeting model, this requires it to widen the scopeof its audits beyond narrow financial compliance togive more emphasis to value-for-money audits. Tostrengthen accountability in these ways requiresgreater transparency within government.

Increased Emphasis on Fiscal Transparency

Transparency in government operations is increas-ingly regarded as an important precondition for goodgovernance and sustainable economic growth—but,for the emerging economies, it is also an essential aspectof sustaining confidence in government and, throughthis, building support for the democratic system andeconomic advancement. The need for the latter hasbecome particularly acute with the exposure of fiscalpolicies to international financial markets arising fromincreased market financing of deficits. Such exposurehas required many governments to modify past poli-cies and so has assisted in achieving fiscal disciplineand improving resource allocation.156 Transparency isalso an essential ingredient in establishing the newbudget management model.

Promoting greater transparency is a means to bolsterreform. Restructuring incentives by concentrating onthe stakeholders within the budget system is unlikelyto be enough—there is also the need for measures toactivate clients in support of reform. Clients need tobe empowered; quantitative performance indicatorsmust be supplemented with indicators of quality; andclient feedback must be strengthened and made moretransparent.

Empowering Clients

One channel for accomplishing this is widely pub-licizing performance standards expected from partic-ular public agencies and spelling out the specific stepsthe public can take to force agencies to meet thesestandards. The U.K. Citizen’s Charter attempted to dothis at an aggregate level, but such initiatives are pos-sible at a program level.

Supplementing Quantitative PerformanceIndicators with Indicators of Quality

The importance of performance indicators has longbeen recognized in reforms associated with perfor-mance budgeting, as evidenced by the ever-expandingliterature on the subject. Also important is improvingresource allocation for government by improving per-formance data and performance measurement. How-ever, one can detect a new trend to obtain differenttypes of performance data. Typically, performance datahas been directed to two questions: how well servicesare delivered (efficiency), and whether planned objec-tives are being met (effectiveness). However, increas-ingly important is a third type of question: whethercustomers are satisfied with the results (quality). Inthe United States, the Government Performance andResults Act (GPRA) was enacted in August 1993 toimprove the public’s confidence in government byholding agencies accountable for program results.Promoting a focus on results, improving the quality ofservices, and measuring customer/citizen satisfactionwith government services were important elementsof the act. The goal was to improve the delivery ofservices not only by requiring that managers establishplans for meeting their goals/objectives but also byproviding feedback to managers regarding actual pro-gram results and quality (Box 51).157

Strengthening Consumer Feedback

Nowhere is this push to improve the efficiency andeffectiveness of government service provision more

156“Global commercial liberalization and the free flow of capitalare exerting new pressures on systems of public governance. Recentexperience shows starkly that the quality of public institutions andthe trust in which they are held by economic players can have verydemonstrable effects on the behavior of these markets. Public sec-tor governance systems that induce loss of market trust impose costsnot only directly on their domestic economies, but more generallyas they reduce global growth rates below potential.” (Schiavo-Campo and Tommasi, 1999, p. 343). 157See Clinton and Gore (1997).

©International Monetary Fund. Not for Redistribution

Page 109: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VIII MANAGING THE REFORM PROCESS

100

noticeable than in the rapid adoption of IT in govern-ment, mirroring developments in the private sector.The spread of e-government has been increasinglydocumented.158 In the United States, there has been arapid expansion in constituent relationship manage-ment technology (CRM, also known as customer rela-tionship management)—a class of software designedto provide government agencies with the ability tomanage their constituent relationships by providingstakeholders the means to contact their governmentvia a variety of channels. Recorded contact informa-tion can be logged, analyzed, and then effectively em-ployed in various business processes. The widespreaduse of e-government means that the analysis of con-sumer feedback data can be a powerful input to agencydecision making. In this way, government perfor-mance can be monitored for the quality of differenttypes of contacts, showing where the government isdoing well and where improvements are required. Con-stituent data can be analyzed so that managers cansegment their client groups in order to deliver morespecifically tailored programs—and thereby improvethe quality of service provision.

The Need for a Compatible Legal andRegulatory Framework

A regulatory framework is required that establishesthe broad parameters for ensuring adequate budgetarycontrols and granting appropriate managerial auton-omy. The latter should be conditional on the agencymanagement providing adequate accountability forthose public resources under its stewardship.

At a minimum, this usually involves streamliningexisting budget procedures and associated financialregulations. In centralized, compliance-oriented sys-tems, there is usually a layered system of controls thathas accumulated over the years. A set of controls isintroduced to remedy some abuse. These controls arelater circumvented, leading to the institution of addi-tional controls to plug the gaps. The additional layersof controls further encourage managers to find ways tocircumvent them to get things done, leading to a viciouscycle. There is a need to rid the system of these redun-dant and counterproductive controls.

However, many countries have gone further andhave considered it necessary to set the new regulatoryframework in a wider context of a framework fiscallaw. In this they have often followed the model of NewZealand, a pioneer in budget reform. New Zealand laiddown principles of responsible fiscal managementas a framework for defining necessary fiscal proce-

158For a commentary on some developments in this rapidly de-veloping field, see the commentary and references contained inOECD (2001).

Box 51. The U.S. Government Performance and Results Act, 1993

The Act is an amendment to the Budget and AccountingAct of 1921, covering all agencies of the federal govern-ment (only excluding legislative and judicial branches andthe Central Intelligence Agency).

Important Elements of the Act

• To improve citizen confidence in government by holdingagencies accountable for program results.

• To encourage reform through a series of pilot projectsfocused on setting goals, measuring performance, andreporting progress against goals.

• To focus on results, improving the quality of services, andmeasuring customer/citizen satisfaction with governmentservices.

• To enhance decision making by disseminating informationrelated to efficiency/effectiveness of federal programs.

GPRA Requirements

• Establish a system of interrelated plans and reports that aredesigned to provide the basis to link resources and results.

• Each agency must produce a five-year strategic plan, whichmust be revised at least every three years.

• Agencies have to develop annual performance plans, es-tablishing specific program goals, identifying the resourcesrequired to meet those goals, and linking this with thestrategic plan.

• Agencies have to provide an annual performance reportreviewing its success in achieving the previous year’sperformance goals, and from this perspective explain anydeviations.

• Pilot projects are introduced to allow some agencies moreflexibility and to waive certain administrative requirementsfor their annual performance plan, but managers are heldaccountable for achieving higher performance.

• Introduction of a two-year pilot project of performancebudgeting in a few agencies, providing Congress withinformation on the direct relationship between proposedprogram spending and expected program results.

Evaluation

Since its enactment, the GPRA has been continuously an-alyzed and reviewed by stakeholders and independent bod-ies. It has been found that some agencies have madesubstantial progress and others have found it difficult tomeet its basic requirements. Major problems encounteredwere the ability to generate reliable data, problems in de-veloping measures for programs/activities spanning morethan one agency, and difficulty in making the transition toperformance budgeting. Notwithstanding these difficulties,basic support for implementation of the GPRA continues.

©International Monetary Fund. Not for Redistribution

Page 110: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

dures, fiscal targets, and fiscal reporting. As withNew Zealand’s 1994 Fiscal Responsibility Act, mostof the implementing legislation sought to enforce im-proved accountability and greater transparency in fiscalmanagement, and in some cases to assuage governanceconcerns. Another potentially important aspect ofadopting the new fiscal framework is its demonstrationeffect. In many countries there have been repeatedefforts at budget system reform or major reform initia-tives in the budget sphere. There is a natural tendency,therefore, for bureaucrats who have seen reforms comeand go to be cynical and to adopt a “wait and see”strategy. Adopting a new, wide, sweeping regulatoryframework thus has the added advantage of over-coming the bureaucracy’s “business as usual” attitude.The basic idea is to create rules and procedures thatimpose costs on governments and bureaucrats fordeviating from fiscal responsibility.

While New Zealand provided a model for this legalframework, other OECD countries have also followedthis path. The need for greater transparency in fiscalmanagement was an overarching objective. Not onlyare governments required to specify and adopt agreedfiscal targets, but they are required to explain the poli-cies underlying their attainment and, when deviatingfrom this policy commitment, they are required to ex-plain not only its rationale but also planned futurecorrective action. The Australian Charter of BudgetHonesty and the United Kingdom’s Code for FiscalStability have adopted this approach, albeit in differentforms. Notable among emerging economies, Brazil’sFiscal Responsibility Law has been quite successfuland influential. However, there have been many dis-tinct variations in the laws introduced.159 For example,the influential New Zealand law put more emphasison procedures,160 but Brazil’s is in many ways muchmore detailed and quite comprehensive in its cover-age, with well-defined sanctions and much more em-phasis on numeric fiscal rules to guide fiscal policy.

In practice, the experience of emerging countriesin introducing fiscal responsibility laws (FRLs) hasbeen quite mixed and suggests some points worthconsidering when introducing such legislation. First,the need for the law will depend on a country’s legaltradition.161 Much depends on the relative powers ofthe legislative and executive branches. Where theexecutive branch is powerful, budget reforms can beintroduced by regulation or by amendment without

introducing new legislation. Hence in countries fol-lowing the Westminster model, fundamental budgetreforms have been introduced and once proven areendorsed by changes in the law. Where the legislativebranch is more powerful, such as in many continentalEuropean countries, fundamental changes are impos-sible without changes in the law. The important pointis that for many emerging economies, a new legalframework should not be considered a precondition forundertaking reform. However, when the basic budgetframework law is considered to be deficient and whenthere is a need for a clear demonstration effect and apolitical declaration in support of reform, it may proveuseful to introduce a new budget system law beforeconsidering a full FRL. The new Public Finance Man-agement and Control Act of Turkey reflects such con-siderations and incorporates many elements of a FRL(Box 52).

Second, it should be noted that in the New Zealandcase, legislation followed rather than led budget systemreforms. In many ways, the incentive was to fill gapsin existing legislation to cover new procedural and fis-cal management reforms that had already been intro-duced. Therefore, in some cases it might be preferableto consolidate such reforms into a more comprehensivebudget system law rather than to introduce speciallegi\slation. New Zealand is in fact following this pathby repealing the Fiscal Responsibility Act and foldingkey features of that act into a revised version of themore comprehensive 1989 Public Finance Act to cre-ate a Public Finance, State Sector Management Act(Cullen and Mallard, 2003).

Third, merely adopting new legislation will not auto-matically bring reform. Typically, an FRL requires astrengthened PEM system to meet the new reportingrequirements and enhanced controls over fiscal mis-management. Most legislation of this nature providesfor a transition period to allow these systems to bestrengthened to meet the expanded accountability re-quirements. It would be imprudent to adopt new legalrequirements without putting these improved PEMsystems in place. From this perspective, the FRLs area double-edged sword. They have the advantage ofpolitical visibility and hence serve the objective of in-spiring confidence in a country’s fiscal management,especially in international markets. At the same time,failure to adhere to the law is equally visible and hencecan work to destroy confidence in fiscal management.To minimize the risks, it is important to anchor the newbudget system reforms on a firm PEM platform (theelements of which are outlined in Sections V and VI).

The Key Lessons for Budget Reform

Here are the key lessons to be learned from the FAD’sexperience in providing technical assistance for intro-

159See, for example, the summary of various fiscal responsibilitylaws contained in Baldacci and Corbacho (2004).

160For example, the New Zealand act sets out five principles forresponsible fiscal management, including a requirement to reducedebt to a “prudent level” and maintain it at that level. The law doesnot define the prudent level, which is left for the government to set.

161See discussion and description of various traditions in Lienertand Jung (2005).

The Key Lessons for Budget Reform

101

©International Monetary Fund. Not for Redistribution

Page 111: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

VIII MANAGING THE REFORM PROCESS

102

ducing and sustaining the move from one budget modelto another.

• First and foremost, do not underestimate the levelof management skills required. Who is going tomanage the reform process? Who is going to man-age the new budget system being introduced?Clearly addressing these questions can make a key difference in emerging countries wherethere is no great depth of managerial expertise ingovernment.

• Properly sequence the reforms. Management ca-pacities must be strengthened as a prerequisiteto devolving management autonomy. The de-volved budget management model must rest onsolid management foundations, especially at theagency level.

• Begin modestly and do not be too ambitious. Indesign, a gradualist rather than a “big bang” ap-proach is usually most appropriate. In implemen-tation, a serial approach is preferred over an attemptat blanket coverage. Rather than seeking to useadvanced technologies, making do with what isavailable and familiar may pay the highest divi-dends, at least initially. The relevant lesson frompublic sector reforms that have sought to installfully integrated FMISs (financial managementinformation systems) is that such complex sys-tems are often beyond the immediate capacities

of those who must use them and the process ofcomputerizing administrative procedures ends updeflecting resources away from more fundamen-tal managerial reform.

• Identify the right management teams. Managementteams must inevitably be allowed some discretion.The practical limits of holding managers fully ac-countable for all dimensions of their agency’s per-formance, however, imply that there is some riskthat this discretion will be misused. It is importantthat the change-management teams share commonobjectives that are fully congruent with those of thegovernment. The more fully the government cantrust the management teams, the more likely the re-form is to succeed. It may seem daunting to findmanagerial leadership with common vision, tech-nical competence, authority to spearhead the re-form, and a full commitment to the process. But thefact is such leadership has been found in manycountries, which offers living proof that in budgetreform the human factor cannot be neglected.

• Assess the change-management capacity, and de-sign technical assistance accordingly. It may benecessary to develop a checklist that will enablesome assessment to be made of this capacity andhence the likelihood of success in delivering TA.The key dimensions and relevant considerationsfor this risk assessment are reflected in Box 53.

Box 52. Turkey’s Public Financial Management and Control Law

This landmark law, passed in December 2003, re-orients Turkey’s very traditional line-item budgets to amore performance-oriented basis, including by the fol-lowing measures:• The law introduces the basics of strategic management

and performance-based budgeting with the need to definestrategic goals/missions and measurable objectives; pub-lic administrations are required to base their budgets onprograms and projects related to their strategic plans. (Ar-ticle 9)

• The State Planning Office has been charged with intro-ducing this approach and determining the strategic plan-ning calendar. Public administrations prepare theirbudgets on a performance basis and in accordance withtheir strategic plans. Ministers are required to inform thepublic about the goals, objectives, strategies, and annualperformance plans in the first month of the fiscal year.

• Strategic plans are set in a medium-term program and mustbe consistent with the medium-term fiscal strategy preparedby the MoF, which includes targeted deficit and borrowingrequirements. (Article 16)

• There is a requirement for public administrations to mea-sure performance according to predetermined indicatorsand to employ these indicators in monitoring and evaluat-ing this process. (Article 9)

• The performance indicators are set in a consultativeprocess by the MoF, the State Planning Office, and rele-vant public administrations.

• The MoF is authorized to review and check the consis-tency of performance indicators and strategic plans.

• Performance audits are to be carried out on the basis ofthese indicators, and the mandate of the Court of Accounts(the supreme audit institution) is extended from financialaudits to include value-for-money audits.

• One of the fundamental principles of the public financeframework is that it should be administered to ensure theaccountability of public officials. (Article 5)

• Head administrators are to be accountable for perfor-mance. (Articles 10–11) They are required to prepare re-ports to explain their activities in line with the strategicplans and performance programs, according to deter-mined performance indicators. Reasons for deviationsare to be explained. (Article 41)

• Based on these reports, the MoF presents a General Ac-countability Report to the Court of Accounts and makes itpublic. The Court of Accounts submits to the National As-sembly its evaluation report on the General Accountabil-ity Report, accompanied by its opinions. This is discussedby the National Assembly along with the GovernmentBudget Bill for the next fiscal round. (Article 42)

©International Monetary Fund. Not for Redistribution

Page 112: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Concluding RemarksAt first glance, the reform agenda described in this

volume may appear quite daunting. However, eachof the steps described—establishing a strong PFMbase, organizing the reform process, investing inhuman resources, adopting a clearly phased strategy,and putting in place an enabling legal and institu-tional environment—has individual merits and payoffsin its own right. Moreover, other countries havetaken these steps on the road to reforming their bud-get systems, and so there is encouragement as wellas lessons to be derived from their experiences. Themere fact that performance budgeting reforms havebeen continued and extended in the countries that firstadopted them should be a spur for those just settingout on this reform path.

The prescription for reform described here should notbe considered unique or cast in stone. While this reform

agenda has been distilled from the experience of theOECD countries, the fact is that all the more advancedcountries started from different initial conditions andfound their own solutions to the challenges inherent inreform. Over time, they modified and adapted these re-forms to meet their own particular circumstances andspecific political and economic environments.

It is also worth remembering that the last word inbudget system reform has not been written, and mostlikely never will be. Even the most advanced perfor-mance budgeting systems continue to be subjected tonew reform initiatives and experimentation. From thisperspective, middle-income and emerging countriesthat are pursuing budget reform should recognize thereis no perfect system that can be taken “off the shelf”and applied without adaptation. What appears to haveworked in one environment may not work in theirs.Instead, embarking on reform should be viewed as alearning process for all participants, one that can beexpected to continue for many years before bringingits full results.

At the same time, although the pay-off from reformmay not be immediate, there should be no doubt thatthere are significant advantages to pursuing such re-forms and significant risks to countries that fail to doso. Globalization has had important repercussionsfor market-based economies—for their private sec-tors but for their public sectors as well. The disciplineof the global market limits the scope for private sec-tor inefficiency but at the same time also offers incen-tives to increase efficiency. Globalization imposessimilar discipline on governments, not only directlythrough the channel of increased market financing ofpublic sector deficits and public sector investments,but also indirectly and no less forcibly. An interna-tionally competitive private sector requires an inter-nationally competitive public sector that can ensurebasic infrastructure, security, public health, an edu-cated workforce, and basic utilities and transportation.

Performance budgeting is an attempt to ensure greaterefficiency in the provision of public services. By sodoing, it not only contributes to the sustainability offiscal policies but also removes constraints on privatesector competitiveness. Experience shows that theultimate results can be profound: though performancebudgeting may commence narrowly, with a focus sim-ply on promoting improved service delivery, it hasoften ended up fostering a fundamental review of theappropriate role of government and changing theground rules for government operations.

Concluding Remarks

103

Box 53. Checklist for Assessing the Risks inDelivering TA for Budget Reform

Are the agents of change well identified?• How deep in the administration is the recognition of

the need for change?• Is this recognition at the government level, the level of

the minister of finance, or lower?• How stable/established is the reform team?

Do the reformers have an adequate base of support?• Is there an adequate level of fiscal stability to ensure

success and avoid a diversion of energies?• Is there sufficient political and administrative

stability?• Can the up-front costs of reform be borne by the gov-

ernment in the short term?• Does the present system have basic levels of fiscal con-

trol and financial management to support reform?

Does the overall environment provide incentives tosupport reform?• What is the general level of managerial capacity to

implement reform?• Is the regulatory framework adequate for reform, or

does it need to be changed?• What is the general perception of the level of gover-

nance in the country?• How empowered are consumers to demand better per-

formance from government agencies?

©International Monetary Fund. Not for Redistribution

Page 113: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

104

References

Abrahams, M.D., and M.N. Reavely, 1998, “Activity-BasedCosting: Illustrations from the State of Iowa,” Govern-ment Finance Review, Vol. 14, No. 2, pp. 15–20.

Accounting Standards Board (ASB), 2003, “Statement ofPrinciples for Financial Reporting: Proposed Interpreta-tion for Public Benefit Entities” (London: ASB). Avail-able at: http://frc.org.uk/asb/publications/documents.cfm?cat=3

Allen, R.I.G., 1996, “Management Control in Modern Gov-ernment Administration: An Introduction,” in Manage-ment Control in Modern Government Administration:Some Comparative Practices, SIGMA Paper No. 4(Paris: Organization for Economic Cooperation andDevelopment).

———, and Daniel Tommasi, eds., 2001, Managing PublicExpenditure: A Reference Book for Transition Coun-tries (Paris: Organization for Economic Cooperationand Development, SIGMA Program).

Allen R., S. Schiavo-Campo, and T.C. Garrity, 2004, As-sessing and Reforming Public Management: A NewApproach (Washington: World Bank).

Anderson, B., 2004, “Performance Budgeting and Per-formance Information in the United States,” paperpresented to International Seminar on PerformanceBudgeting, Brasilia, Brazil, March (unpublished).

Athukorala, S. Lakshman, and Barry Reid, 2003, AccrualBudgeting and Accounting in Government and ItsRelevance for Developing Countries (Manila: AsianDevelopment Bank).

Auerbach, Alan J., Jagadeesh Gokhale, and Laurence J.Kotlikoff, 1991, “Generational Accounts: A Meaning-ful Alternative to Deficit Accounting,” in Tax Policyand the Economy, Vol. 5 (Cambridge, Massachusetts:MIT Press), pp. 55–110.

Auerbach, A.J., L.J., Kotlikoff, and W. Leibfritz, 1999,Generational Accounting around the World (Chicago:Chicago University Press).

Australian Department of Finance and Administration, 1999,Commonwealth Accrual Budgeting Guidelines (Canberra).

Australian National Audit Office, 1996, Performance In-formation Reviews (Canberra: Department of Finance).

Axelrod, D., 1995, Budgeting for Modern Government(New York: St. Martin’s Press, 2nd ed.).

Babunakis, M., 1976, Budgets: An Analytical and Proce-dural Handbook for Government and Nonprofit Orga-nizations (Westport, Connecticut: Greenwood Press).

Baldacci, E., and A. Corbacho, 2004, “Recent Experimentswith Fiscal Responsibility Laws around the World—What Lessons Can Be Learned?” FAD Draft GuidanceNote (Washington: International Monetary Fund).

Barrett, K., and R. Greene, 2000, “Truth in Measurement,”Governing (July), p. 86.

Baumol. William J., 1982, “Contestable Markets: An Up-rising in the Theory of Industry Structure,” AmericanEconomic Review, Vol. 72, No. 1, pp. 1–15.

———, J.C. Panzar, and R.D. Willig, 1982, ContestableMarkets and the Theory of Industry Structure (NewYork: Harcourt Brace Jovanovich).

Bellamy, S., and R. Kluvers, 1995, “Program Budgeting inAustralian Local Government: A Study of Implemen-tation and Outcomes,” Financial Accountability andManagement, Vol. 11 (February), pp. 39–56.

Blondal, J.R., 2003, “Accrual Accounting and Budgeting:Key Issues and Recent Developments,” OECD Journalon Budgeting, Vol. 3., No.1, pp. 51–71.

Boston, J., ed., 1995, The State under Contract (Wellington:Bridget Williams Books, Ltd.).

Boyne, G., and J. Law, 1991, “Accountability and LocalAuthority Annual Reports: The Case of Welsh DistrictCouncils,” Financial Accountability and Management,Vol. 7, No. 4, pp. 179–94.

Brace, P., R. Elkin, D.D. Robinson, and H.E. Steinberg,1980, “Reporting of Service Efforts and Accomplish-ments,” FASB Research Report R09 (Norwalk, Con-necticut: Financial Accounting Standards Board).

Brook, P.J., and M. Petries, 2001, “Output-Based Aid:Precedents, Promises, and Challenges,” in Contractingfor Public Services: Output-Based Aid and Its Appli-cations, ed. by P.J. Brook and S.M. Smith (Washing-ton: World Bank), pp. 3–11.

Brook, P.J., and S.M. Smith, eds., 2001, Contracting forPublic Services: Output-Based Aid and Its Applications(Washington: World Bank).

Broom, C.A., 1995, “Performance-Based GovernmentModels: Building a Track Record,” Public Budgetingand Finance, Vol. 15 (Winter), pp. 3–17.

Brumby, J., 1999, “Budgeting Reforms in OECD MemberCountries,” in Managing Government Expenditure, ed.by S. Schiavo-Campo and D. Tommasi (Manila: AsianDevelopment Bank), pp. 349–62.

———, 2005, “Budget Management Reform in China,” inEngaging the New World: Strategic Economic Studies inthe Knowledge Economy, Employment, Health Care andFiscal Governance, ed. by B. Grewal and M. Kumnick(Melbourne: Melbourne University Press, Melbourne).

Brunsson, K., 1995, “Puzzle Pictures: Swedish BudgetaryProcesses in Principle and Practice,” Financial Account-ability and Management, Vol. 11, No. 2, pp. 111–25.

Burkhead, J., 1956, Government Budgeting (New York:John Wiley and Sons).

©International Monetary Fund. Not for Redistribution

Page 114: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

References

105

Carlin, T., and J. Guthrie, 2000, “A Review of Australian andNew Zealand Experiences with Accrual Output-BasedBudgeting,” paper presented to a conference of theInternational Public Management Network, MacquarieGraduate School of Management, Sydney, Australia,March 4–6.

Carter, N., R. Klein, and P. Day, 1992, How OrganizationsMeasure Success (London: Routledge).

Caulfield, Janice, 2003, “New Public Management in aDeveloping Country: Creating Executive Agencies inTanzania,” in Unbundled Government: A CriticalAnalysis of the Global Trend to Agencies, Quangosand Contractualisation,” ed. by Christopher Pollittand Colin Talbot (London: Routledge).

Chan, J.L., 1998, “The Bases of Accounting for Budgetingand Financial Reporting,” in Handbook of GovernmentBudgeting, ed. by R.T. Meyers (San Francisco: Jossey-Bass), pp. 357–80.

———, 2003, “Government Accounting: An Assessment ofTheory, Purposes and Standards,” Public Money andManagement, Vol. 23, No. 1, pp. 13–20.

Chow, D., and C. Humphrey, 2003, “Whole of GovernmentAccounting: An Aide to Performance, Managementand Accountability?” 9th Biennial CIGAR (Compara-tive International Government Accounting Research)Conference, Bodo, Norway, June 13–14.

Clinton, Bill, and Al Gore, 1997, The Blair House Papers:National Performance Review (Washington: Govern-ment Printing Office).

Cooper, R., and R.S. Kaplan, 1999, Design of Cost Man-agement Systems (Upper Saddle River, New Jersey:Prentice Hall, 2nd ed.).

Cullen M., and T. Mallard, 2003, Pre-Introduction: Parlia-mentary Briefing on Public Finance (State Sector Man-agement) Bill, August (Wellington).

Davenport, L.W., 1996, “Internal Service Fund Functions:Should They Be Required to Compete with PrivateVendors?” Government Finance Review (October),pp. 11–13.

Davis, G., B. Sullivan, and A. Yeatman, eds., 1997, TheNew Contractualism? (South Melbourne: Macmillan).

Dean, P.N., 1989, Government Budgeting in DevelopingCountries (London: Routledge).

Demsetz, Harold, 1968, “The Cost of Transacting,” Quar-terly Journal of Economics, Vol. 82, pp. 33–53.

Diamond, Jack, 1990, “Measuring Efficiency in Govern-ment: Techniques and Experience,” Chapter 9 in Gov-ernment Financial Management: Issues and CountryStudies, ed. by A. Premchand (Washington: Interna-tional Monetary Fund), pp. 142–66.

———, 2001, “Performance Budgeting: Managing the ReformProcess,” paper presented at UN Workshop on FinancialManagement and Accountability, Rome, November.

———, 2002a, “The Micro Basis of Budget System Re-form: The Case of Transitional Economies” (unpub-lished; Washington: International Monetary Fund).

———, 2002b, “The Role of Internal Audit in GovernmentFinancial Management: An International Perspective,”IMF Working Paper 02/94 (Washington: InternationalMonetary Fund).

———, 2002c, “Performance Budgeting: Is Accrual Ac-counting Required?” IMF Working Paper 02/240(Washington: International Monetary Fund).

———, 2002d, “The New Russian Budget System: A CriticalAssessment and Future Reform,” IMF Working Paper02/21 (Washington: International Monetary Fund).

———, 2002e, “Budget System Reform in TransitionalEconomies: The Experience of Russia,” IMF Work-ing Paper 02/22 (Washington: International MonetaryFund).

———, 2002f, “The Strategy of Budget System Reform inEmerging Countries,” Public Finance and Manage-ment, Vol. 2, No. 3.

———, 2003, From Program to Performance Budgeting: TheChallenge for Emerging Countries, IMF Working Paper03/169 (Washington: International Monetary Fund).

———, 2004, “The Role of Internal Audit in GovernmentFinancial Management,” in Accounting and Account-ability in Emerging and Transitional Economies 6,ed. by Trevor Hopper and Zahirul Hoque (New York:Elsevier), pp. 55–80.

———, and P. Khemani, 2005, “Introducing FinancialManagement Information Systems in DevelopingCountries,” IMF Working Paper 05/196 (Washington:International Monetary Fund).

———, and B.H. Potter, 2000, Setting Up Treasuries in theBaltics, Russia, and Other Countries of the FormerSoviet Union, IMF Occasional Paper No. 198 (Wash-ington: International Monetary Fund).

Dixon, G., 2002, “Thailand’s Hurdle Approach to BudgetReform,” Poverty Reduction and Economic Manage-ment (PREM) Note 73 (Washington: World Bank).

Domberger, S., 1998, The Contracting Organization: AStrategic Guide to Outsourcing (New York: OxfordUniversity Press).

———, and P. Jensen, 1997, “Contracting Out by the Pub-lic Sector: Theory, Evidence, Prospects,” Oxford Re-view of Economic Policy, Vol. 13, pp. 67–78.

———, and S. Rimmer, 1994, “Competitive Tendering andContracting Out in the Public Sector: A Survey,” Inter-national Journal of Economics and Business, Vol. 1,No. 3, pp. 439–53.

Estela, M., 2000, “Strengthening the Integrity of a Tax Col-lection Agency: The Case of SUNAT in Peru,” paperprepared for a World Bank–Inter-American Develop-ment Bank seminar on “Radical Solutions for Fight-ing Corruption in the Public Sector,” November 2–3,Washington, D.C.

European Commission, 1996, “Public Procurement in theEuropean Union: Exploring the Way Forward,” GreenPaper COM(96) 583 (Brussels).

———, 2004, “Public Finances in the European MonetaryUnion 2004” (Brussels).

European Federation of Accountants, 2003, “The Adoptionof Accrual Accounting and Budgeting by Governments(Central, Federal, Regional and Local),” FEE Discus-sion Paper (Brussels). See: www.fee.be/.

Fantone, Denise M., 2004, “Performance-Based Budgetingin the U.S.: A Congressional Perspective,” paper pre-sented to Working Party of Senior Budget Officials(SBO), Organization for Economic Cooperation andDevelopment, Paris, April 1–2.

France, Ministère de l’Economie, des Finances et de l’Industrie, 2001, Public Finance Reform, No. 1 (Paris).

Garamfalvi, L., and W. Allan, 1994, “Value for MoneyAuditing, Evaluation, and Public Expenditure Man-

©International Monetary Fund. Not for Redistribution

Page 115: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

REFERENCES

agement: Experience in Selected OECD Countries andLessons for Italy” (unpublished; Washington: Inter-national Monetary Fund, Fiscal Affairs Department).

Gearhart, J., 1999, “Activity-Based Management and Perfor-mance Measurement Systems,” Government FinanceReview, pp. 13–16.

Gill, D., 2002, “Signposting the Zoo—From Agencifica-tion to a More Principled Choice of Government Or-ganisational Form,” OECD Journal on Budgeting,Vol. 2, No. 1, pp. 27–79.

Glyn, D., B. Sullivan, and A. Yeatman, eds., 1997, The NewContractualism (South Melbourne: MacMillan).

Gray, A., W.I. Jenkins, and B. Segsworth, 1993, Budgeting,Auditing and Evaluation: Functions and Integration inSeven Governments (New Brunswick, New Jersey:Transaction Publishers).

Grindle, M.S., 1997, “Divergent Cultures? When Public Or-ganizations Perform Well in Developing Countries,”World Development, Vol. 25, No. 4, pp. 481–95.

Groszyk, W., 2001, “Outcome-Focused Management in theUnited States,” OECD Journal on Budgeting, Vol.1,No. 4, pp. 129–50.

Guthrie, J., 1998, “Application of Accrual Accounting in thePublic Sector: Rhetoric or Reality?” Financial Ac-countability and Management, Vol. 14, No. 1, pp. 1–19.

———, O. Olson, and C. Humphrey, 1999, “Debating De-velopments in New Public Financial Management: TheLimits of Global Theorising and Some New Ways For-ward,” Financial Accountability and Management,Vol. 15, No. 3/4, pp. 209–28.

Harper, M., J. Arroyo, T. Bhattacharya, and T. Bulman, 2000,Public Services through Private Enterprise: Micro-Privatisation for Improved Delivery (London: Inter-mediate Technology Publications).

Hatry, H.P., 1999, Performance Measurement (Washington:Urban Institute Press).

Heald, D., and A. Dowdall, 1999, “Capital Charging as aVFM Tool in Public Services,” Financial Account-ability and Management, Vol. 15, No. 3/4, pp. 209–28.

Henderson, Simon, 2004, “The Challenges of MeasuringPerformance,” paper presented at “Performance Infor-mation in the Budget Process,” OECD/PUMA SeniorBudget Officials (SBO) meeting, Organization for Eco-nomic Cooperation and Development, Paris, April 1–2.

Henke, E.O., 1992, Introduction to Nonprofit Organization Ac-counting (Cincinnati, Ohio: South-Western Publishing Co.)

H.M. Treasury, 2004, The U.K. Government’s Public ServiceAgreement Framework, Better Public Services Team(London).

Hill, Alex, 2004, “The U.K. Government’s Public ServiceAgreement Framework” (unpublished; London: H.M.Treasury, Better Public Services Team).

Hodge, G.A., 2000, Privatization: An International Reviewof Performance (Boulder, Colorado: Westview Press).

Hood, C., 1991, “A Public Management for All Seasons?”Public Administration, Vol. 69, No. 1, pp. 3–19.

———, 1995, “The New Public Management in the 1980s:Variations on a Theme,” Accounting, Organization,and Society, Vol. 20, No. 2/3, pp. 93–109.

Hyndman, N.S., and R. Anderson, 1995, “The Use of Perfor-mance Information in External Reporting: An EmpiricalStudy of U.K. Executive Agencies,” Financial Account-ability and Management, Vol. 11, No. 1, pp. 1–17.

106

Institute of Internal Auditors (IIA), 1999, Internal AuditStandards (Altamonte Springs, Florida: IIA).

———, 2001, “Consulting Implementation Standards,”Standards for the Professional Practice of InternalAuditing (Altamonte Springs, Florida: IIA, InternalAuditing Standards Board).

International Federation of Accountants (IFAC), 1998,Guidelines for Governmental Financial Reporting(New York: IFAC, Public Sector Committee).

———, 2003a, Cash Basis International Public Sector Ac-counting Standards: Financial Reporting Under theCash Basis of Accounting (New York: IFAC).

———, 2003b, Transition to the Accrual Basis of Accounting:Guidance for Governments and Government Entities,Study 14 (New York: IFAC, 2nd ed.).

International Monetary Fund, 2001a, Government FinanceStatistics Manual 2001 (Washington).

———, 2001b, Manual on Fiscal Transparency (Washington).———, 2001c, World Economic Outlook, October (Wash-

ington).International Organization of Supreme Audit Institutions,

1995, “Auditing Standards,” adopted by AuditingStandards Committee at the XVth INTOSAI Congress,Cairo (Stockholm: INTOSAI). Available at: http://asc.rigsrevisionen.dk/composite-14.htm

Jackson, P., and B. Palmer, 1989, First Steps in MeasuringPerformance in the Public Sector (London: Public Fi-nance Foundation).

Jones, L.R., and J.L. McCaffery, 1997, “Implementing theChief Financial Offices Act and the Government Per-formance and Results Act in the Federal Government,”Public Budgeting & Finance, Vol. 13, No. 1, pp. 35–55.

Joyce, P.G., 1993, “Using Performance Measures for Fed-eral Budgeting: Proposals and Prospects,” PublicBudgeting & Finance, Vol. 13, No. 4, pp. 1–15.

Kaplan, R.S., and D.P. Norton, 1996, The Balanced Score-card: Translating Strategy into Action (Boston: HarvardBusiness School Press).

Kaul, M., 1997, “New Public Administration, The Manage-ment Innovations in Government,” Public Administra-tion and Development, Vol. 17, No. 1, pp. 13–26.

Kerf, M., R.D. Gary, T. Irwin, C. Levesque, R. Taylor, andM. Klein, 1999, “Concessions for Infrastructure: AGuide to Their Design and Award,” World Bank Tech-nical Paper 399 (Washington: World Bank).

Klein, M., So, J., and Shin, B., 1996, “Transaction Costs inPrivate Infrastructure Projects—Are They Too High?Viewpoint, No. 95 (Washington: World Bank, PrivateSector Development Department).

Kopits, G., 2001, “Fiscal Rules: Useful Policy Frameworkor Unnecessary Ornament,” IMF Working Paper 01/145(Washington: International Monetary Fund).

———, and J.D. Craig, 1998, Transparency in GovernmentOperations, IMF Occasional Paper No. 158 (Washing-ton: International Monetary Fund).

———, and S. Symansky, 1998, Fiscal Policy Rules, IMFOccasional Paper No. 162 (Washington: InternationalMonetary Fund).

Kotter, John P., 1995, “Leading Change: Why TransformationEfforts Fail,” Harvard Business Review, Vol. 73, No. 2,pp. 59–67.

———, 1996, Leading Change (Boston: Harvard BusinessSchool Press).

©International Monetary Fund. Not for Redistribution

Page 116: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

References

107

———, and J.L. Heskett, 1992, Corporate Culture and Per-formance (New York: Free Press).

Kristensen, J.K., 2002, “Overview of Results Focused Man-agement and Budgeting in OECD Member Countries,”paper presented at an expert meeting held by the PublicManagement Committee at Organization for EconomicCooperation and Development, Paris, February 11–12.

———, W. Groszyk, and B. Bühler, 2002, “Outcome-FocusedManagement and Budgeting,” OECD Journal on Budget-ing, Vol. 1, No. 4, pp. 1–29.

Laking, R., 1996, “Public Management in the OECD: SomeQuestions for Developing Countries,” paper presentedto the World Bank, Washington, D.C., May (unpub-lished; Victoria: University of Wellington).

———, 2001, “Governance of the Wider State Sector: Prin-ciples for Control and Accountability of Delegated andDevolved Public Bodies,” paper presented at the Orga-nization for Economic Cooperation and Development,Public Management Service (PUMA) Conference on“Devolving and Delegating Power to More Autono-mous Public Bodies and Controlling Them: The Gov-ernance of Public Agencies and Authorities,” Bratislava,November. Available at: http://www.oecd.org/dataoecd/8/19/2730680.pdf

———, 2005, “Agencies: Their Benefits and Risks,” OECDJournal on Budgeting, Vol. 4, No. 4, pp. 7–25.

Lapsley, I., 1999, “Accounting and the New Public Man-agement: Instruments of Substantive Efficiency or aRationalising Modernity?” Financial Accountabilityand Management, Vol. 15, No. 3/4, pp. 201–207.

Larbi, G.A., 1998, “Institutional Constraints and CapacityIssues in Decentralizing Management in Public Ser-vices: The Case of Health in Ghana,” Journal of Inter-national Development, Vol. 10, No. 3, pp. 377–86.

Larsson, K., and J.S. Madsen, 1999, “Protecting the FinancialInterests of the State and of the European Union,” Pub-lic Management Forum, Vol. 5, No. 6 (Paris: OECD),pp. 4–17.

Leibenstein, H., 1966, “Allocative Efficiency vs. X-Efficiency,” American Economic Review, Vol. 56(June), pp. 392–415.

Lienert, Ian, and Moo-Kyung Jung, 2005, “The Legal Frame-work for Budget Systems, and International compari-son,” OECD Journal on Budgeting, Special Issue,Vol. 4, No. 3.

Likierman, A., 2000, “Changes to Managerial Decision-Taking in UK Central Government,” ManagementAccounting Research, Vol. 11, No. 2, pp. 253–61.

Maholland, L., and P. Muetz, 2002, “A Balanced ScorecardApproach to Performance Measurement,” GovernmentFinance Review (April), pp. 12–15.

Martinez-Vazquez, J., and R. McNab, 2000, “The Tax Re-form Experiment in Transitional Countries,” NationalTax Journal, Vol. 53, No. 2 (June), pp. 273–98.

Mellor, T., 1996, “Why Governments Should Produce BetterBalance Sheets,” Australian Journal of Public Admin-istration, Vol. 55, No. 1, pp. 78–81.

Mikesell, J., 2003, Fiscal Administration: Analysis and Ap-plications for the Public Sector (Belmont, California:Thomson/Wadsworth, 6th ed.).

Organization for Economic Cooperation and Development(OECD), 1994, “Performance Management in Govern-ment: Performance Measurement and Results-Oriented

Management,” Public Management Service (PUMA)Occasional Paper No. 3 (Paris).

———, 1995, Budgeting for Results: Perspectives on PublicExpenditure Management (Paris).

———, 1996, “Managing Structural Deficit Reduction,”Public Management Service (PUMA) Occasional PaperNo. 11 (Paris).

———, 1997a, “Benchmarking, Evaluation and StrategicManagement in the Public Sector,” paper presented atthe 1996 Meeting of the Performance ManagementNetwork of the OECD’s Public Management Service(PUMA) (Paris).

———, 1997b, “Best Practice Guidelines for Contractingout Government Services,” Public Management Ser-vice (PUMA) Occasional Paper No. 20 (Paris).

———, 1999a, Performance Contracting: Lessons fromPerformance Contracting Case Studies (Paris).

———, 1999b, “Managing Accountability in Intergovern-mental Relationships,” Public Management Service,PUMA/RD (99) 4 (Paris).

———, 2000, The OECD Outputs Manual, PUMA/SBO(2000)7 (Paris).

———, 2001, “Financial Management and Control of PublicAgencies,” SIGMA Paper No. 32 (Paris).

———, 2002a, “Accrual Accounting and Budgeting: KeyIssues and Recent Developments, Twenty-Third AnnualMeetings of OECD Senior Budget Officials, Washington,D.C., June 3–4,” Public Management Service, PUMA/SBO(2002)10 (Paris).

———, 2002b, Distributed Public Governance: Agencies,Authorities, and Other Government Bodies (Paris).

———, 2002c, Governing for Performance in the PublicSector, OECD/Germany High-Level Symposium,March 13–14, Berlin (Paris). Country reports availableat: http://www.oecd.org/document/43/0,2340,en_2649_37457_1813355_1_1_1_37457,00.html

———, 2003, OECD/World Bank Budget Practices andProcedures Survey (Paris).

———, 2004, “Public Sector Modernisation: Governing forPerformance,” Policy Brief (Paris). Available at: http://www.oecd.org/dataoecd/52/44/33873341.pdf

Osborne, D., and T. Gaebler, 1992, Reinventing Government(Reading, Massachusetts: Addison-Wesley).

Osborne, S.P., T. Bovaird, S. Martin, M. Tricker, and P. Waterston, 1995, “Performance Management andAccountability in Complex Public Programmes,” Fi-nancial Accountability and Management, Vol. 11, No.1, pp. 19–37.

Pendlebury, M.R., R. Jones, and Y. Karbhari, 1994, “Developments in the Accountability and Financial Reporting Practices of Executive Agencies,” Finan-cial Accountability and Management, Vol. 10, No. 1,pp. 33–46.

Perrin, B., 1998, “Effective Use and Misuse of PerformanceMeasurement,” American Journal of Evaluation, Vol. 19,No. 3, pp. 367–79.

Petrie, M., 2002, A Framework for Public Sector Perfor-mance Contracting (Paris: Organization for EconomicCooperation and Development).

Pollitt, C., 1986, “ Beyond the Managerial Model: The Casefor Broadening Performance Assessment in Govern-ment with the Public Service,” Financial Accountabil-ity and Management, Vol. 2, pp. 155–70.

©International Monetary Fund. Not for Redistribution

Page 117: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

REFERENCES

———, 1993, Managerialism and the Public Services(Cambridge, Massachusetts: Basil Blackwell, 2nd ed.).

———, K. Bathgate, J. Caulfield, A. Smullen, and C. Tal-bot, 2001, “Agency Fever? Analysis of an InternationalPolicy Fashion,” Journal of Comparative PolicyAnalysis, Vol. 3, No. 3, pp. 271–90.

Premchand, A., 1969, Performance Budgeting (Bombay:Academic Books).

———, 1977, “Restructuring Budgetary Systems in Devel-oping Countries: Relevance of PPB Systems,” IMFWorking Paper (Washington: International MonetaryFund).

———, and J. Burkhead, eds., 1984, Comparative Interna-tional Budgeting and Finance (New Brunswick, NewJersey: Transaction Books).

Public Expenditure and Financial Accountability (PEFA)Program, 2005, Public Financial Management Perfor-mance (Washington: PEFA Secretariat, World Bank).

Reid, G.J., 1999, “Performance-Oriented Public SectorModernization in Developing Countries,” in Researchin Public Administration, Vol. 5, ed. by J.C.N. Raad-schelders and J.L. Perry (Greenwich, Connecticut: JAIPress).

Robinson, M., 1998, “Capital Charges and Capital Expen-diture Decisions in Core Government,” Journal ofPublic Budgeting, Accounting and Financial Manage-ment, Vol. 10, No. 3, pp. 354–74.

———, 2000, “Contract Budgeting,” Public Administra-tion, Vol. 78, No. 1, pp. 75–90.

Schiavo-Campo, S., and D. Tommasi, 1999, ManagingGovernment Expenditure (Manila: Asian DevelopmentBank).

Schick, Allen, 1966, “The Road to PPB: The Stages of Bud-get Reform,” Public Administration Review, Vol. 26,pp. 243–58. Also published in Perspectives on Bud-geting, ed. by A. Schick (Washington: ASPA, 1980).

———, 1971, Budget Innovation in the States (Washington:Brookings Institution).

———, 1996, The Spirit of Reform: Managing the NewZealand State Sector in a Time of Change (Wellington:State Services Commission).

———, 1998, “Why Most Developing Countries ShouldNot Try New Zealand’s Reforms,” World Bank Re-search Observer, Vol. 13, No. 1, pp. 123–31.

———, 2001, “The Changing Role of the Central BudgetOffice,” OECD Journal on Budgeting, Vol. 1, No. 1,pp. 9–26.

———, 2002, “Agencies in Search of Principles,” in Dis-tributed Public Governance: Agencies, Authorities,and Other Government Bodies (Paris: Organization forEconomic Cooperation and Development), pp. 33–52.

Schultze, C., 1968, The Politics and Economics of PublicSpending (Washington: Brookings Institution).

———, 1996, Government Reform in New Zealand, IMFOccasional Paper No. 140 (Washington: InternationalMonetary Fund).

———, 2001a, “Managing Governments for Better Perfor-mance and Results: Some Lessons of Experience,”Workshop on Financial Management and Account-ability, Rome, November 28–30.

———, 2001b, Public Management in New Zealand: Lessonsand Challenges (Wellington: New Zealand BusinessRoundtable). Available at: http://www.nzbr.org.

108

nz/documents/publications/publications-2001/public_management.pdf

Scott, C.D., 1996, “The Nature of Public Policy,” in NewZealand Politics, ed. by R. Miller (Auckland: OxfordUniversity Press), pp. 247–55.

Shand, D., 1997, “International Developments in PublicSector Financial Management,” paper presented toAustralasian Officers’ Conference, Perth, Australia,November 19–21.

Sharp, T., 2001, “Linking Performance Measurement toBudgeting,” Government Finance Review, Vol. 17,Part 2, pp. 15–18.

Smith, W., 2001, “Designing Output-Based Aid Schemes: AChecklist,” in Contracting for Public Services: Output-Based Aid and Its Applications, ed. by P.J. Brook andS.M. Smith (Washington: World Bank), pp. 91–117.

Stanton, P., and J. Stanton, 1998, “The Questionable Eco-nomics of Governmental Accounting,” Accounting,Vol. 11, No. 2, pp. 191–203.

Strauch, R., 2000, Managerial Flexibility and Fiscal Perfor-mance (Paris: Organization for Economic Cooperationand Development).

Tanzi, V., and H.H. Zee, 2000, “Tax Policy for EmergingMarkets: Developing Countries,” IMF Working Paper00/35 (Washington: International Monetary Fund).

Thompson, F., 1998, “Cost Measurement and Analysis,”Chapter 15 in Handbook of Government Budgeting,ed. by R.T. Meyers (San Francisco: Jossey-Bass).

United Nations, 1965, A Manual for Programme and Per-formance Budgeting (New York).

United States Commission on Organization of the Execu-tive Branch of the Government, 1949, Budgeting andAccounting: A Report to Congress by the Commissionon Organization of the Executive Branch of the Gov-ernment (Washington: Government Printing Office).

United States, 1949, Task Force Report, Fiscal, Budgetingand Accounting Activities (Washington: GovernmentPrinting Office).

U.S. Congressional Budget Office (CBO), 1993, UsingPerformance Measures in the Federal Budget Process(Washington).

U.S. Federal Accounting Standards Advisory Board(FASAB), 1993, Statement of Federal Financial Ac-counting Concepts 1, Objectives of Federal FinancialReporting (Washington).

U.S. General Accountability Office, 2004, “PerformanceBudgeting: Current Developments and Future Prospects,”GAO-04-174, April (Washington).

U.S. Office of Management and Budget (OMB), 1993,“Statement of Federal Financial Accounting Con-cepts 1: Objectives of Federal Financial Reporting”(Washington).

Wanna, J., J. Kelly, and J. Forster, 2001, Managing PublicExpenditure in Australia (St. Leonards, NSW: Allen &Unwin).

Waterson, A., 1965, Development Planning: Lessons of Ex-perience (Baltimore: John Hopkins).

Wholey, J.S., 1999, “Quality Control: Assessing the Accu-racy and Usefulness of Performance MeasurementSystems” in Performance Measurement, ed. by H.P.Hatry (Washington: Urban Institute Press), pp. 217–39.

Wildavsky, A, 1969, “Planning-Programming-Budgeting:Rescuing Policy Analysis from PPBS,” Public Admin-istration Review, Vol. 29.

©International Monetary Fund. Not for Redistribution

Page 118: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

References

109

———, 1979, The Politics of the Budgetary Process(Boston: Little, Brown).

Williams, W.A., 2002, “Trusting the Numbers: The Powerof Data Verification,” Government Finance Review,Vol. 18, No. 2 (April), pp. 18–21.

Willoughby, K., and J.E. Melkers, 2001, “Assessing the Im-pact of Performance Budgeting: A Survey of American

States,” Government Finance Review, Vol. 17, No. 2(April), pp. 25–30.

Wynne, A., 2004, Is the Move to Accrual Based Accountinga Real Priority for Public Sector Accounting? (Lon-don: Association of Chartered Certified Accountants,London). Available at: http://www.accaglobal.com/doc/publicsector/ps_doc_009.pdf

©International Monetary Fund. Not for Redistribution

Page 119: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

OCCASIONAL PAPERS

Recent Occasional Papers of the International Monetary Fund

245. Budget System Reform in Emerging Economies: The Challenges and the Reform Agenda, by JackDiamond. 2006.

244. Monetary Policy Implementation at Different Stages of Market Development, by a staff team led by BernardJ. Laurens. 2005.

243. Central America: Global Integration and Regional Cooperation, edited by Markus Rodlauer and AlfredSchipke. 2005.

242. Turkey at the Crossroads: From Crisis Resolution to EU Accession, by a staff team led by Reza Moghadam.2005.

241. The Design of IMF-Supported Programs, by Atish Ghosh, Charis Christofides, Jun Kim, Laura Papi, UmaRamakrishnan, Alun Thomas, and Juan Zalduendo. 2005.

240. Debt-Related Vulnerabilities and Financial Crises: An Application of the Balance Sheet Approach toEmerging Market Countries, by Christoph Rosenberg, Ioannis Halikias, Brett House, Christian Keller,Jens Nystedt, Alexander Pitt, and Brad Setser. 2005.

239. GEM: A New International Macroeconomic Model, by Tamim Bayoumi, with assistance from DouglasLaxton, Hamid Faruqee, Benjamin Hunt, Philippe Karam, Jaewoo Lee, Alessandro Rebucci, and IvanTchakarov. 2004.

238. Stabilization and Reforms in Latin America: A Macroeconomic Perspective on the Experience Since theEarly 1990s, by Anoop Singh, Agnès Belaisch, Charles Collyns, Paula De Masi, Reva Krieger, GuyMeredith, and Robert Rennhack. 2005.

237. Sovereign Debt Structure for Crisis Prevention, by Eduardo Borensztein, Marcos Chamon, Olivier Jeanne,Paolo Mauro, and Jeromin Zettelmeyer. 2004.

236. Lessons from the Crisis in Argentina, by Christina Daseking, Atish R. Ghosh, Alun Thomas, and TimothyLane. 2004.

235. A New Look at Exchange Rate Volatility and Trade Flows, by Peter B. Clark, Natalia Tamirisa, and Shang-Jin Wei, with Azim Sadikov and Li Zeng. 2004.

234. Adopting the Euro in Central Europe: Challenges of the Next Step in European Integration, by Susan M.Schadler, Paulo F. Drummond, Louis Kuijs, Zuzana Murgasova, and Rachel N. van Elkan. 2004.

233. Germany’s Three-Pillar Banking System: Cross-Country Perspectives in Europe, by Allan Brunner, JörgDecressin, Daniel Hardy, and Beata Kudela. 2004.

232. China’s Growth and Integration into the World Economy: Prospects and Challenges, edited by EswarPrasad. 2004.

231. Chile: Policies and Institutions Underpinning Stability and Growth, by Eliot Kalter, Steven Phillips, MarcoA. Espinosa-Vega, Rodolfo Luzio, Mauricio Villafuerte, and Manmohan Singh. 2004.

230. Financial Stability in Dollarized Countries, by Anne-Marie Gulde, David Hoelscher, Alain Ize, DavidMarston, and Gianni De Nicoló. 2004.

229. Evolution and Performance of Exchange Rate Regimes, by Kenneth S. Rogoff, Aasim M. Husain, AshokaMody, Robin Brooks, and Nienke Oomes. 2004.

228. Capital Markets and Financial Intermediation in The Baltics, by Alfred Schipke, Christian Beddies, SusanM. George, and Niamh Sheridan. 2004.

227. U.S. Fiscal Policies and Priorities for Long-Run Sustainability, edited by Martin Mühleisen and Christo-pher Towe 2004.

226. Hong Kong SAR: Meeting the Challenges of Integration with the Mainland, edited by Eswar Prasad, withcontributions from Jorge Chan-Lau, Dora Iakova, William Lee, Hong Liang, Ida Liu, Papa N’Diaye, andTao Wang. 2004.

225. Rules-Based Fiscal Policy in France, Germany, Italy, and Spain, by Teresa Dában, Enrica Detragiache,Gabriel di Bella, Gian Maria Milesi-Ferretti, and Steven Symansky. 2003.

224. Managing Systemic Banking Crises, by a staff team led by David S. Hoelscher and Marc Quintyn. 2003.

110

©International Monetary Fund. Not for Redistribution

Page 120: in Emerging Economies - IMF eLibrary...Budget system reform in emerging economies : the challenges and the reform agenda / by Jack Diamond — Washington, D.C. : International Monetary

Occasional Papers

223. Monetary Union Among Member Countries of the Gulf Cooperation Council, by a staff team led by UgoFasano. 2003.

222. Informal Funds Transfer Systems: An Analysis of the Informal Hawala System, by Mohammed El Qorchi,Samuel Munzele Maimbo, and John F. Wilson. 2003.

221. Deflation: Determinants, Risks, and Policy Options, by Manmohan S. Kumar. 2003.

220. Effects of Financial Globalization on Developing Countries: Some Empirical Evidence, by Eswar S. Prasad,Kenneth Rogoff, Shang-Jin Wei, and Ayhan Kose. 2003.

219. Economic Policy in a Highly Dollarized Economy: The Case of Cambodia, by Mario de Zamaroczy andSopanha Sa. 2003.

218. Fiscal Vulnerability and Financial Crises in Emerging Market Economies, by Richard Hemming, MichaelKell, and Axel Schimmelpfennig. 2003.

217. Managing Financial Crises: Recent Experience and Lessons for Latin America, edited by Charles Collynsand G. Russell Kincaid. 2003.

216. Is the PRGF Living Up to Expectations?—An Assessment of Program Design, by Sanjeev Gupta, MarkPlant, Benedict Clements, Thomas Dorsey, Emanuele Baldacci, Gabriela Inchauste, Shamsuddin Tareq,and Nita Thacker. 2002.

215. Improving Large Taxpayers’ Compliance: A Review of Country Experience, by Katherine Baer. 2002.

214. Advanced Country Experiences with Capital Account Liberalization, by Age Bakker and Bryan Chapple.2002.

213. The Baltic Countries: Medium-Term Fiscal Issues Related to EU and NATO Accession, by JohannesMueller, Christian Beddies, Robert Burgess, Vitali Kramarenko, and Joannes Mongardini. 2002.

212. Financial Soundness Indicators: Analytical Aspects and Country Practices, by V. Sundararajan, CharlesEnoch, Armida San José, Paul Hilbers, Russell Krueger, Marina Moretti, and Graham Slack. 2002.

211. Capital Account Liberalization and Financial Sector Stability, by a staff team led by Shogo Ishii and KarlHabermeier. 2002.

210. IMF-Supported Programs in Capital Account Crises, by Atish Ghosh, Timothy Lane, Marianne Schulze-Ghattas, Ales̆ Bulír̆, Javier Hamann, and Alex Mourmouras. 2002.

209. Methodology for Current Account and Exchange Rate Assessments, by Peter Isard, Hamid Faruqee, G. Russell Kincaid, and Martin Fetherston. 2001.

208. Yemen in the 1990s: From Unification to Economic Reform, by Klaus Enders, Sherwyn Williams, NadaChoueiri, Yuri Sobolev, and Jan Walliser. 2001.

207. Malaysia: From Crisis to Recovery, by Kanitta Meesook, Il Houng Lee, Olin Liu, Yougesh Khatri, NataliaTamirisa, Michael Moore, and Mark H. Krysl. 2001.

206. The Dominican Republic: Stabilization, Structural Reform, and Economic Growth, by a staff team led byPhilip Young comprising Alessandro Giustiniani, Werner C. Keller, and Randa E. Sab and others. 2001.

205. Stabilization and Savings Funds for Nonrenewable Resources, by Jeffrey Davis, Rolando Ossowski, JamesDaniel, and Steven Barnett. 2001.

204. Monetary Union in West Africa (ECOWAS): Is It Desirable and How Could It Be Achieved? by Paul Massonand Catherine Pattillo. 2001.

203. Modern Banking and OTC Derivatives Markets: The Transformation of Global Finance and Its Implicationsfor Systemic Risk, by Garry J. Schinasi, R. Sean Craig, Burkhard Drees, and Charles Kramer. 2000.

202. Adopting Inflation Targeting: Practical Issues for Emerging Market Countries, by Andrea Schaechter, MarkR. Stone, and Mark Zelmer. 2000.

201. Developments and Challenges in the Caribbean Region, by Samuel Itam, Simon Cueva, Erik Lundback,Janet Stotsky, and Stephen Tokarick. 2000.

Note: For information on the titles and availability of Occasional Papers not listed, please consult the IMF’s Publications Catalog or contact IMFPublication Services.

111

©International Monetary Fund. Not for Redistribution